For the past forty or fifty years, teachers of literature in American colleges and universities have acted upon a few simple assumptions, mainly derived from I.A. Richards’s early books Principles of Literary Criticism (1924) and Practical Criticism (1929). The first assumption is that in reading a poem you think of the words on the page as a transcription of a voice speaking; not necessarily the poet speaking in his own person, but a hypothetical person, speaking in imagined circumstances sufficiently indicated by what he says. The second assumption is that you are interpreting the poem, trying to understand the context, the speaker’s sense of it, and the cogency of that sense. The meaning of the poem is what the speaker means to say. The third assumption is that you read poems to imagine experiences you have not had, to exercise sympathy and judgment upon them, and to take part in richer communications. It follows that it is essential, in reading a poem as in taking part in a conversation, to judge the speaker’s tone correctly, because tone indicates his relation both to his own feeling and to the person or persons he is addressing. These assumptions, suitably elaborated, prescribe an orthodoxy of reading.

Take, for instance, Robert Frost’s poem “Acquainted with the Night,” which begins

I have been one acquainted with the night.
I have walked out in rain—and back in rain.
I have outwalked the furthest city light.

I have looked down the saddest city lane.
I have passed by the watchman on his beat
And dropped my eyes, unwilling to explain.

In an orthodox reading, you follow the speaker’s feeling from first word to last. In the first line, for instance, you think of the degree of assertiveness in “I,” the precise degree of knowledge claimed in “acquainted,” the relation between the apparent precision of “acquainted” and the vagueness of its object, “the night.” You gauge the tone of those repeated “I have” phrases. And so on. Reading a poem is like meeting its speaker.

These assumptions are defined and proposed in most of the textbooks that have established themselves in American courses in literature, whether survey-courses or more advanced classes in the criticism of literature. The most influential textbook is still Understanding Poetry (1938) by Cleanth Brooks and Robert Penn Warren, a book that begins with these sentences: “Wordsworth called the poet a man speaking to men,” and “Poetry is a form of speech, written or spoken.” The best motto for the orthodoxy of reading is “hearing with eyes,” as in Shakespeare’s Sonnet 23:

O learn to read what silent love hath writ,
To hear with eyes belongs to love’s fine wit.

The orthodoxy has been challenged from time to time, mostly by people who feel that Richards, Brooks, Warren, and the New Critics generally have ignored the historical understanding of literature and encouraged students to raise only the questions that can be answered by pointing to a few privileged terms, notably irony, ambiguity, tension, and complexity. And it has sometimes been remarked that certain poems, such as Pound’s Cantos, defeat the elocutionist and have to be read in some other way. But generally the orthodoxy has stood up pretty well. The vocal and acoustic character of poetry is widely accepted.

Within the past ten years or so, a new challenge has arisen. If it were to prevail, it would surround with anxiety and misgiving not only the reading of poems but the negotiation of every major theme in Western literature and philosophy. The most vigorous challenger is Jacques Derrida, and the name of the challenge is Deconstruction, a form of commentary that relies upon a diverse if selective reading of Nietzsche and Freud. Derrida reached Deconstruction mainly by an aggressive reading of Husserl and Heidegger, a discipline there is no reason to think all his followers have practiced. He does not claim that Deconstruction is in every respect a distinctive activity: how could it be, since it is mainly a commentary written in the margin of other philosophical and literary texts? And it shares some of the arguments put forward by Structuralism, even though it eventually accuses structuralists of many follies ascribed to traditional forms of thought.

What form would a deconstructive reading of Frost’s poem take? I have not seen one, and it might surprise me, but if it did not surprise me it would start by suffusing the “I” of the first line with doubt; questioning its neo-Cartesian assumption, and the blatant punctuality with which it implies a speaker. It would note that the printed words are given only as script, and that the reader is urged to convert them into acoustic signs: speech is supposed to be more fully present than print. The deconstructive critic would question the apparent assumption, in the first line, that someone—the speaker—exists, and has existed even before the “I” of his self-assertion; and that this “person” guarantees the authenticity of what he says by presiding over it as a controlling consciousness.


The critic would then question the confidence with which Frost’s first line begins and ends; begins, in the assertive presentation of the “I” who speaks; and ends, with the equally assertive presentation of whatever experience “the night” is supposed to denote. He would ask himself whether the apparent slide from the dogmatic “I” to the vague, third-person-pronoun “one” is an evasion making possible the more extreme evasiveness of “the night,” a phrase as sonorous as it is obscure; or merely a decent confession of misgiving about the assertiveness of “I” in the first place. The critic would then go through the poem, diagnosing every example of blindness or naïveté in Frost’s relation to his language; the false confidence with which he proceeds; the uncritical assumption that by miming a voice he is verifying a personality.

Deconstruction is a style of accusation. Its main charges are these. One: Western thought, whether in literature or philosophy, has been grounded upon bogus axioms of being and presence; upon being, understood as presence. Meaning has been offered by analogy with presence; the presence of an object to one’s sight; the self-presence of subjectivity and consciousness, as in the speaking “I” of Frost’s poem.

Two: what Derrida calls “the metaphysics of presence” has been endorsed by the presentation of meaning as speech. He calls this prison house of language “logocentrism” or “phonocentrism,” and claims that it involves “the absolute proximity of voice and being, of voice and the meaning of being.” The meaning of Frost’s poem begins with the reader’s feeling that he is in the presence of its speaker. According to Derrida, the cardinal word in Western thought is logos, taken to mean the original uttered word, identified in some theologies with the audible presence of God. The entire philosophic tradition is infatuated, Derrida alleges, with the notion of a first moment, authentic and paternal.

Three: the axiom of being as presence is mere wish-fulfillment, and those who indulge themselves in it are inauthentic; it testifies not to presence but to absence, to a desire for presence; it keeps us always turned back toward an ostensibly first moment in which we were guaranteed authenticity by the speaking voice of the Father. Reading Frost’s poem by translating it into an individual voice, we are really consoling ourselves, trying to persuade ourselves that our entire lives are dialogues, continuous in some sense with the original moment in which God uttered Himself.

Four: the logos-prison has forced upon us delusions of self, personality, subjectivity, creativity, imaginative power, ego-psychology, and so forth. Frost’s poem is full of these delusions of personal grandeur: “And dropped my eyes, unwilling to explain.” Five: psychological terms (such as feeling, emotion, desire, motive, drive) are misleading, they encourage us to think of them as attributes of our suffering and masterful selves; they must be replaced by linguistic terms.

So the best answer to our infatuation with voices and presences is writing—writing stripped of all delusions. Writing in this sense has given up yearning for a lost father, it knows that it is an orphan, it is merely what it is and therefore it has vetoed nostalgia, accepted its separation from any origin and settled for that separation. Deconstruction seeks, in Geoffrey Hartman’s phrase, “the eclipse of voice by text.” Text in that phrase means language set not to work but to play among its internal possibilities, having been released from all the old claims of meaning as voice, personal presence, and dialogue.

What Deconstruction urges is not a new system of thought but skepticism toward all the old ways, which are construed as really only one way. It proposes a certain critical position.

Deconstruction and Criticism is offered as a manifesto of “one of the most stimulating movements in contemporary letters.” It is an odd book. The critics who have joined to produce it have only one property in common: they teach at Yale. The book has more to do with the rhetoric of power in American universities than with its ostensible subject: its evident purpose is to claim that Yale houses a major school of deconstructive criticism. But the truth is that only two of the five critics are deconstructors: Derrida and Paul de Man. J. Hillis Miller practices occasional conformity, he has written essays in the spirit and according to the letter of Deconstruction. But he seems to me still what he ever was, a phenomenologist, a critic of consciousness, his heart in the Geneva of Georges Poulet. Geoffrey Hartman is one of the most vigorous opponents of Deconstruction. His essay in the present book is a meditation on Wordsworth’s poem “A Little Onward”: a fine essay, indeed, but it has nothing to do with Deconstruction, and it could have been published with greater propriety in the Review of English Studies.


What Harold Bloom is doing in this book, I have no idea; he is not a deconstructor. A rabbi, a prophet, he would never let himself be shamed out of the language of self, presence, and voice. His essay in Deconstruction and Criticism rehearses his famous theory of poetry, using for his text John Ashbery’s “Self-Portrait in a Convex Mirror.” As always, Bloom’s categories are psychological rather than linguistic, he does not claim for language anything more than an instrumental and referential function.

Derrida’s essay is an account of Maurice Blanchot’s fiction, L’arrêt de mort, concentrating on one motif, that of living and living on (vivre; survivre): it has much to say about ghosts, the narrator as one who lives on, corpses embalmed, and so on. I seem to have heard parts of it in Derrida’s lecture at New York University a few months ago, the most opaque and withholding lecture I have ever heard. So far as I can understand it (but that’s not far), the essay is trying to dislodge from the story a language that retains an old-fashioned affection for truth and being; to dislodge it by projecting it beyond itself (la sur-vérité) into an empty future. In his earlier books, Derrida tried to dislodge such language by consigning it to an empty past for which the key word has always been déjà. Hence, in the new essay, his emphasis falls on the motif of a stairway that figures in Blanchot’s fiction; there are pages upon pages of word-play on escalier as escalade of truth, “one truth about another, one truth on (top of) another, one above or below the other, each step more or less true than truth.” These pages are pretty hectic, but I read them as part of the deconstructive effort to show that truth and word never coincide.

Paul de Man’s essay is a study of Shelley’s “The Triumph of Life”—a study in deconstruction. Most readers would probably take the poem as a vision in which Shelley sees a neo-Roman triumphal procession, with chariots, charioteers, captives, and an unnamed Shape. Shelley then speaks to Rousseau, who gives him a guided vision of Voltaire, Frederick the Great, Czar Paul of Russia, Catherine the Great, Leopold the Second, Plato, Aristotle, and Alexander the Great. Rousseau gives Shelley a vision of what we may call Ideal Beauty. I cannot report on all the various interpretations of the poem, but those I have read are compatible at least with the common sense of the poem as involving visionary meetings, partings, voices, questions, and answers.

None of this interests Paul de Man. He merely uses the poem, drags it on the wheels of his own chariot, determined to prove that poetic figures are not to be ascribed to the imagination that apparently engaged with language to produce them but rather to the arbitrariness of Language itself. The “Shape all light” encountered by Rousseau in “The Triumph of Life” is “the figure for the figurality of all signification,” according to de Man. It follows, he claims, that “the figure is not naturally given or produced but that it is posited by an arbitrary act of language.” By this he means:

The positing power of language is both entirely arbitrary, in having a strength that cannot be reduced to necessity, and entirely inexorable in that there is no alternative to it. It stands beyond the polarities of chance and determination and can therefore not be part of a temporal sequence of events.

I don’t understand this. De Man, implacable in denying to the poet any active power, is evidently willing to ascribe an “act” to “language”; he apparently does this merely for the satisfaction of reporting that the “acts” of language are mechanical, arbitrary, and repetitive. He will do anything, this critic, rather than read poems in terms of poets, speakers, imaginations, and languages in some sense willing to be invoked. When he adverts to rhyming words (billow, willow, pillow) in a poem, he takes the rhyming as primary, the meanings as secondary. Such sequences, he maintains, are generated “by random and superficial properties of the signifier rather than by the constraints of meaning.” And he goes far out of his way to avoid the ordinary reader’s understanding that it is a mark of the poet’s skill to reconcile the constraints of meaning and the properties of the signifier; to keep his chariot going happily by driving both horses at the same speed and in the same direction.

De Man’s new book Allegories of Reading is a study of figural language in Rousseau, Nietzsche, Rilke, and Proust. What de Man means by allegory is not as clear as it should be. Normally we think of allegory when we read such works as The Divine Comedy and The Pilgrim’s Progress and sense (or know) that the narrative has another meaning or other meanings beyond the literal one. Angus Fletcher’s Allegory: The Theory of a Symbolic Mode (1964) says of allegorical works that “as they go along they are usually saying one thing in order to mean something beyond that one thing.” Allegory is a mode of writing in which one set of meanings sustains another, and the two are kept separate by being kept parallel to each other.

The common preference for symbol rather than allegory, which is found in Romantic aesthetic since the later years of the eighteenth century, is explained by this separation. Coleridge, for instance, in The Statesman’s Manual and elsewhere, downgrades allegory as being merely “a translation of abstract notions into a picture-language, which is itself nothing but an abstraction from objects of the senses.” Symbol, on the other hand, as Coleridge says, “always partakes of the reality which it renders intelligible; and while it enunciates the whole, abides itself as a living part in that unity of which it is the representative.” Unity of experience, as a deeply desired state, is the promise of the symbol. The symbol asserts that we may live fully in the world; we are not fated to be aliens there.

From the point of view of Deconstruction, this promise of unity—indeed the quest for unity—is yet another instance of the hated “metaphysic of presence.” De Man calls it “self-mystification.” In an effort to cancel the promise of unity as offered by the symbol and by the general theory of imagination, he prefers allegory:

Whereas the symbol postulates the possibility of an identity or identification, allegory designates primarily a distance in relation to its own origin, and, renouncing the nostalgia and the desire to coincide, it establishes its language in the void of this temporal distance. In so doing, it prevents the self from an illusory identification with the non-self, which is now fully, though painfully, recognized as a non-self.

Allegory, then, holds sign and meaning to be discontinuous, just as it supposes man and the world to be strictly separate and that every attempt to blur their separateness is fraudulent.

It follows that de Man must have the same objection to metaphor that he has to symbol and imagination:

By suggesting the potential identification of tenor and vehicle, the traditional metaphor stresses the possible recuperation of a stable meaning or set of meanings. It allows one to see language as a means towards a recovered presence that transcends language itself.

Metaphor is therefore compromised, it is a language of desire, another form of wish-fulfillment. De Man wants us to understand “the impossibility for the language of poetry to appropriate anything, be it as consciousness, as object, or as a synthesis of both.”

The essay on Rilke pursues the logic of de Man’s argument. Again he ascribes to “language” what ordinary readers would ascribe to Rilke. If a poem seems to denote a movement of feeling, it is merely “rhetorical agitation.” If poems seem to refer to something, the something is merely virtual, it is there only for the rhymes. Themes inhabit these poems “not because they are the expression of Rilke’s own lived experience (whether they are or not is irrelevant) but because their structure allows for the unfolding of his patterns of figuration.”

At one point de Man examines Rilke’s poem “Ich liebe dich, du sanftestes Gesetz.” It is a difficult poem; it is hard to divine what the “gentlest law” (“sanftestes Gesetz“) enforces, or what the speaker’s relation to God is. However, an ordinary reader would register the feeling of constraint in various forms in the first stanza, and the gradual movement to a feeling of release and freedom at the end. He would follow the speaker’s voice as the several invocations lead to the change of imperatives from object to subject in the last line. He would try to interpret the poem on the assumption that it is about something rather than about nothing. He would get no help from de Man, who reduces the entire poem to euphony. The meaning of such poems “is the conquest of the technical skills which they illustrate by their acoustic success.” How little “acoustic success” stands for to a deconstructor, it is unnecessary to remark.

De Man also examines Rilke’s “Am Rande der Nacht,” but only to refute the reader’s assumption that the poem has something to do with man and nature. The speaker says, among other things

Ich bin eine Saite,
über rauschende breite
Resonanzen gespannt.

(“I am a string strung over wide, roaring resonances.”) But strings and violins do not matter, according to de Man, such references are merely ostensible: the poems “are composed of entities, objects and subjects, who themselves behave like words, which ‘play’ at language according to the rules of rhetoric as one plays ball according to the rules of the game.” (I interrupt to remark that the ball does not play anything: the player plays the game with a ball according to rules which he knows and which the ball does not know. Left to itself, the ball would not budge.)

De Man’s general argument is that ostensibly human motives are properly understood as functions of textual systems. The rhetorical mode of Rousseau’s Emile, for instance, “produces the opposition between nature and society as a textual necessity.” Vetoing the presence of the writer, de Man replaces him by an invented ghost which he inserts in a machine called grammar. Thinking of “the arbitrary power play of the signifier,” he allows his mind to be driven by the metaphor of language as machine; with odd consequences for his reading of Rousseau’s Confessions.

De Man concentrates on the episode in the second Book about Rousseau’s theft of the ribbon: he steals a ribbon, tells a lie about it, and puts the blame on the servant Marion. De Man insists on reading this episode as if he were watching the repetitions of a machine. He quotes the passage in the fourth Rêverie where Rousseau reconsiders the theft of the ribbon, and another occasion on which he told a lie. “It is certain,” Rousseau says, “that neither my judgment nor my will dictated my reply, but that it was the automatic result of my embarrassment:” “et qu’elle fut l’effet machinal de mon embarras.” De Man applies this effet machinal to the text itself, and to textuality in general.

But the matter is well enough explained by going back a few pages in the Rêverie to the passage where Rousseau says, “My heart followed these rules of conscience mechanically before my reason had adopted them, and the moral instinct alone made the application: “Mon coeur suivait machinalement ces règles avant que ma raison les eût adoptées, et l’instinct moral en fit seul l’application.” It is reasonable to translate machinalement as “mechanically” or “automatically,” but it would not be far wrong to say “instinctively” or “spontaneously.” The point is simply that Rousseau feels his conscience acted before the obligation of reason. Nothing in the passage in the Confessions justifies de Man’s fixation upon the notion of the machine in a relentless application to the property of grammar. Besides, grammar is formal but not mechanical or automatic, it limits choice but does not prevent choice. A sufficient proof is that many different styles inhabit the same language.

But de Man can’t bear to have Rousseau do what he appears to do—discuss his motives, discriminate between one force of feeling and another, account for something he has done in psychological terms. De Man wants to remove the entire discourse from Rousseau to grammar. When the question of guilt comes up, de Man says that “any guilt, including the guilty pleasure of writing the fourth Rêverie, can always be dismissed as the gratuitous product of a textual grammar or a radical fiction: there can never be enough guilt around to match the text-machine’s infinite power to excuse.” This is witty, but trivial, at best a Wildean flourish. It is always possible, de Man goes on, “to face up to any experience (to excuse any guilt), because the experience always exists simultaneously as fictional discourse and as empirical event and it is never possible to decide which one of the two possibilities is the right one.” If you’ve ever felt guilty, the decision has been painfully easy. To remain suspended between de Man’s two possibilities is a clear instance of self-mystification.

Readers of the Confessions think they hear someone talking about his own experience, accusing himself, justifying himself, and so forth. De Man can’t bear to hear that voice, because he doesn’t want to hear any voice: he wants to see a machine working without human intervention. His application of a linguistic model to all situations is pedantic. If it were enforced in practice as rigorously as he proposes, it would dismiss the questions commonly considered in morality, ethics, politics, and psychology, and treat them as purely linguistic functions. De Man draws back from that conclusion, not a moment too soon. Indeed, from time to time in Allegories of Reading he makes concessions which, if added together, would put in question the whole ideology of Deconstruction. I shall list them more prominently than they are listed in the book.

One: “the notion of a language entirely freed of referential constraints is properly inconceivable.” When you insist upon the fallacy of reference, the insistence is referential. As de Man concedes, “The deconstruction states the fallacy of reference in a necessarily referential mode.” The mode of Allegories of Reading, for instance, is referential. Two: de Man concedes, in his reading of Proust’s novel, that as soon as you refer to the narrator, even if you reduce him to the status of a pronoun, you allow yourself to infer the intent of the subject from the structure of the predicate. The personal concession is enforced by reference to voice, even if it is only to the active or passive voice in grammar.

If you go this far, I don’t see how you can prevent yourself from going further and admitting the full regalia of personal presence. De Man sometimes does this, referring with ordinary naïveté to “the essential barrenness of the self and of the world.” Three: in any case, de Man is willing to allow the self to be recuperated, recovered from its dispersal in the world, if only “in the highly abstracted and generalized form of a deconstructive process of self-denial.” Self-denial is just as naïve as self-delight, self-regard, self-disclosure, or any other such gesture. In de Man’s terms, the self would be reinvented “at the far end of its most radical negation,” no longer the dupe of its desires.

Four: de Man admits, as Derrida also does, that “deconstructive readings can point out the unwarranted identifications achieved by substitution, but they are powerless to prevent their recurrence even in their own discourse.” Five: irony itself is not secure. Allegories of Reading ends by saying that irony, no longer a trope, is “the undoing of the deconstructive allegory of all tropological cognitions, the systematic undoing, in other words, of understanding.” No wonder Harold Bloom has referred to de Man’s “serene linguistic nihilism.”

The deconstructive reading of texts is practiced nowadays not only in seminars at Yale but in many other graduate schools. If you want a deconstructive reading of, say, Daniel Deronda, you can have one. But there are several problems. Deconstructive readings are written mostly for the sake of the theory they are supposed to endorse; the strings are attached before the essay is sent off to Diacritics or Glyph or Semiotexte. Further: nothing I have read in deconstructive criticism is likely to suppress the common desire of readers to believe that in reading a poem they are listening to someone saying something about something—whatever critical questioning they may bring to the act of listening.

No consideration of form, grammar, or structure is likely to divert an intelligent reader from the assumption that a poem consists of words spoken, hypothetically, by someone. Nor have I seen anything that seriously damages the Romantic theory of imagination as variously but on the whole consistently outlined from the late eighteenth-century German and English critics to Emerson, Stevens, Frost, and Harold Bloom; the theory by which the imagination is understood as the mind in the aspect of its freedom and creativity. Genius is only an extreme degree of creative power, a degree not available, indeed, to most of us but demonstrably active in some people; in Wordsworth, Beethoven, Goethe, and Baudelaire, for instance. I have not read anything in deconstructive criticism that would convince me that Wordsworth’s “Immortality Ode” is best read as an anonymous work of Language proceeding with mechanical and repetitive force. Deconstruction bears approximately the relation to reading that Robbe-Grillet’s theory bears to the writing and reading of fiction; it is cogent enough to induce an occasionally felt scruple, but not a determination to change one’s ways.

Why, then, is it having a success in American universities, or at least in some of them? It can’t be explained by its charm, or the pleasure it offers. What Derrida and de Man practice is merely an extreme form of skepticism: an old story, after all. Derrida’s version is delivered with a certain Byronic chic. He reminds me of the Surrealists in the Thirties who, as William Empson says in a poem, enjoyed “a nightmare handy as a bike.” De Man’s style is a sober thing, painstaking as well as painsgiving: he seems to be the Malcontent of the movement.

Still, its success is very curious. I have only two or three thoughts on the subject. I think Deconstruction appeals to the clerisy of graduate students, who like to feel themselves superior to the laity of common readers, liberated from their shared meanings; liberated, too, from the tedious requirement of meaning as such, the official obligation to suppose that words mean something finite rather than everything or nothing. Deconstruction allows them to think of themselves as forming a cell, the nearest thing the universities can offer in the form of an avant-garde. The wretched side of this is that Deconstruction encourages them to feel superior not only to undergraduates but to the authors they are reading. There is also a suggestion of scientific method in Deconstruction which appeals to graduate students who have begun to doubt that the Humanities constitute a discipline. And there is the politics of Deconstruction: like Structuralism, it is antibourgeois, and particularly hostile to the ego-psychology which a bourgeois ideology is supposed to offer its members as a consolation prize.

Whatever the case may be, Deconstruction seems to be an instance of “serious folly,” a phrase Shelley used in “The Triumph of Life” with something quite different in mind.

This Issue

June 12, 1980