SOURCE: "Jacques Derrida: Language against Itself," in Deconstruction: Theory and Practice, 1982. Reprint by Routledge, 1988, pp. 18-41.

[Norris is an English critic and educator who has authored numerous studies on Derrida and deconstruction. In the following excerpt, he offers a detailed summary of Derrida's theories on language, philosophy, and writing.]

The texts of Jacques Derrida defy classification according to any of the clear-cut boundaries that define modern academic discourse. They belong to 'philosophy' in so far as they raise certain familiar questions about thought, language, identity and other longstanding themes of philosophical debate. Moreover, they raise those questions through a form of critical dialogue with previous texts, many of which (from Plato to Husserl and Heidegger) are normally assigned to the history of philosophic thought. Derrida's professional training was as a student of philosophy (at the École Normale Supérieure in Paris, where he now teaches), and his writings demand of the reader a considerable knowledge of the subject. Yet Derrida's texts are like nothing else in modern philosophy, and indeed represent a challenge to the whole tradition and self-understanding of that discipline.

One way of describing this challenge is to say that Derrida refuses to grant philosophy the kind of privileged status it has always claimed as the sovereign dispenser of reason. Derrida confronts this claim to power on its own chosen ground. He argues that philosophers have been able to impose their various systems of thought only by ignoring, or suppressing, the disruptive effects of language. His aim is always to draw out these effects by a critical reading which fastens on, and skilfully unpicks, the elements of metaphor and other figurative devices at work in the texts of philosophy. Deconstruction in this, its most rigorous form acts as a constant reminder of the ways in which language deflects or complicates the philosopher's project. Above all, deconstruction works to undo the idea—according to Derrida, the ruling illusion of Western metaphysics—that reason can somehow dispense with language and arrive at a pure, self-authenticating truth or method. Though philosophy strives to efface its textual or 'written' character, the signs of that struggle are there to be read in its blind-spots of metaphor and other rhetorical strategies.

In this sense Derrida's writings seem more akin to literary criticism than philosophy. They rest on the assumption that modes of rhetorical analysis, hitherto applied mainly to literary texts, are in fact indispensable for reading any kind of discourse, philosophy included. Literature is no longer seen as a kind of poor relation to philosophy contenting itself with mere 'imaginary' themes and forgoing any claim to philosophic dignity and truth. This attitude has, of course, a long prehistory in Western tradition. It was Plato who expelled the poets from his ideal republic, who set up reason as a guard against the false beguilements of rhetoric, and who called forth a series of critical 'defences' and 'apologies' which runs right through from Sir Philip Sidney to I. A. Richards and the American New Critics. The lines of defence have been variously drawn up, according to whether the critic sees himself as contesting philosophy on its own argumentative ground, or as operating outside its reach on a different—though equally privileged—ground.

In the latter camp it is F. R. Leavis who has most forcefully asserted the critic's right to dissociate his habits of thought from the logical checks and procedures demanded of philosophic discourse. Criticism on Leavis's term is a matter of communicating deep-laid intuitive responses, which analysis can point to and persuasively enact, but which it can by no means explain or theorize about. Philosophy is kept at arm's length by treating literary language as a medium of 'lived' or 'felt' experience, a region where the critic's 'mature' responses are his only reliable guide and where there is no support to be had from abstract methodology. Hence Leavis's insistence on the virtues of 'practical' criticism (or close reading), allied to such moral imperatives as 'relevance', 'maturity' and an 'open reverence before life'. The effect of this programme is to draw a firm line of demarcation between literary language and the problems of philosophy. Leavis rejects the idea that criticism need concern itself with epistemological problems, or rhetorical modes of working, implicit in literary texts. His ideal critic works within a discipline defined by qualities of responsiveness and intuitive tact, rather than subtlety of philosophic grasp.

Such was the tenor of Leavis's famous 'reply' to Rene Wellek [in 'Literary Criticism and Philosophy,' Scrutiny VI], who had asked (in an otherwise appreciative essay) why Leavis should not provide a more coherent or worked-out rationale for his critical judgements. To do so would amount to a betrayal, it seemed, of the different but equally disciplined activity required of the literary critic. That activity was justified in so far as it preserved the life-giving wholeness of critical response from the deadening weight of abstract theory.

Leavis represents the most rooted and uncompromising form of resistance to philosophy on the part of literary criticism. The American New Critics, with their penchant for rhetorical system and method, tended to strike a somewhat more ambiguous stance…. [Allen Tate has written] despairingly of criticism as a middle-ground activity torn between the warring poles of imagination and philosophic reason. Typically, the New Critics managed to contain these tensions by devising a rhetoric of figure and paradox which closed the poem off within its own formal limits. Poetry (and fiction, so far as they dealt with it) took on a kind of self-authenticating status, confirmed by the various dogmas of critical method. Conceptual problems—like that of relating poetic 'form' to communicable 'meaning'—were neatly side-stepped by being treated as if they were somehow constitutive of poetry's uniquely complex mode of existence. Paradox and irony, which Tate saw as bearing (to some extent at least) on the critic's own predicament, were generally regarded by the New Criticism as objectively 'there' in the poem's structure of meaning.

Hence the circularity and self-sufficient character of New Critical rhetoric. It kept philosophy at bay, not, like Leavis, by flatly denying its relevance, but by translating its questions into a language of irreducibly aesthetic paradox and tension. As critics came to interrogate this rhetoric of closure, so it became more evident that the problems had merely been repressed or displaced, and that criticism had yet to discover its relation to the modes and exigencies of 'philosophic' discourse. It was at this point in the history of American criticism and its discontents that Derrida's influence came as such a liberating force. His work provided a whole new set of powerful strategies which placed the literary critic, not simply on a footing with the philosopher, but in a complex relationship (or rivalry) with him, whereby philosophic claims were open to rhetorical questioning or deconstruction. Paul de Man has described this process of thought in which 'literature turns out to be the main topic of philosophy and the model of the kind of truth to which it aspires' [Allegories of Reading: Figural Language in Rousseau, Nietzsche, Rilke, and Proust, 1979]. Once alerted to the rhetorical nature of philosophic arguments, the critic is in a strong position to reverse the age-old prejudice against literature as a debased or merely deceptive form of language. It now becomes possible to argue—indeed, impossible to deny—that literary texts are less deluded than the discourse of philosophy, precisely because they implicitly acknowledge and exploit their own rhetorical status. Philosophy comes to seem, in de Man's work, 'an endless reflection on its own destruction at the hands of literature'.

Derrida's attentions are therefore divided between 'literary' and 'philosophical' texts, a distinction which in practice he constantly breaks down and shows to be based on a deep but untenable prejudice. His readings of Mallarmé, Valéry, Genet and Sollers are every bit as rigorous as his essays on philosophers like Hegel and Husserl. Literary texts are not fenced off inside some specialized realm of figurative licence where rational commentary fears to tread. Unlike the New Critics, Derrida has no desire to establish a rigid demarcation of zones between literary language and critical discourse. On the contrary, he sets out to show that certain kinds of paradox are produced across all the varieties of discourse by a motivating impulse which runs so deep in Western thought that it respects none of the conventional boundaries. Criticism, philosophy, linguistics, anthropology, the whole modern gamut of 'human sciences'—all are at some point subjected to Derrida's relentless critique. This is the most important point to grasp about deconstruction. There is no language so vigilant or self-aware that it can effectively escape the conditions placed upon thought by its own prehistory and ruling metaphysic.

The passage 'beyond formalism' was broached in various ways. Some critics (like Geoffrey Hartman) have adopted a wayward and teasingly indirect style, while others—notably Paul de Man—have attempted to think through the paradoxes of New Critical method. De Man's essays in Blindness and Insight (1971) were a powerful application of Derridean ideas to the rhetoric of modern poetics. To read the New Critics with an eye to their founding metaphors is to discover, in de Man's terminology, a 'blindness' inseparable from their moments of greatest 'insight'. Their formalist notion of the poem as 'verbal icon'—a timeless, self-possessed structure of meaning—is shown to deconstruct its own claim through unrecognized twists of implication. Their obsession with 'organic' form was undermined by those very 'ambiguities' and 'tensions' which they sought out in order to praise, and so contain, them. 'This unitarian criticism', as de Man puts it, 'finally becomes a criticism of ambiguity, an ironic reflection on the absence of the unity it had postulated'. 'Form' itself turns out to be more an operative fiction, a product of the interpreter's rage for order, than anything vested in the literary work itself. The organicist metaphors of New Critical parlance result from what de Man calls the 'dialectical interplay' set up between text and interpreter. 'Because such patient and delicate attention was paid to the reading of forms, the critic pragmatically entered into the hermeneutic circle of interpretation, mistaking it for the organic circularity of natural processes'.

Deconstruction draws no line between the kind of close reading appropriate to a 'literary' text and the strategies required to draw out the subtler implications of critical language. Since all forms of writing run up against perplexities of meaning and intent, there is no longer any question of a privileged status for literature and a secondary, self-effacing role for the language of criticism. De Man fully accepts the Derridean principle that 'writing', with its own dialectic of blindness and insight, precedes all the categories that conventional wisdom has tried to impose on it.

This amounts to a downright refusal of the system of priorities which has traditionally governed the relation between 'critical' and 'creative' language. That distinction rested on the idea that literary texts embodied an authentic or self-possessed plenitude of meaning which criticism could only hint at by its roundabout strategies of reading. For Derrida, this is yet another sign of the rooted Western prejudice which tries to reduce writing—or the 'free play' of language—to a stable meaning equated with the character of speech. In spoken language (so the implication runs), meaning is 'present' to the speaker through an act of inward self-surveillance which ensures a perfect, intuitive 'fit' between intention and utterance. Literary texts have been accorded the status of a self-authenticated meaning and truth, a privilege deriving (in Derrida's view) from the deep mistrust of textuality which pervades Western attitudes to language. This mystique of origins and presence can best be challenged by annulling the imaginary boundaries of discourse, the various territorial imperatives which mark off 'literature' from 'criticism', or 'philosophy' from everything that stands outside its traditional domain.

This redistribution of discourse implies some very drastic shifts in our habits of reading. For one thing, it means that critical texts must be read in a radically different way, not so much for their interpretative 'insights' as for the symptoms of 'blindness' which mark their conceptual limits. De Man puts the case most succinctly:

Since they are not scientific, critical texts have to be read with the same awareness of ambivalence that is brought to the study of non-critical literary texts, and since the rhetoric of their discourse depends on categorical statements, the discrepancy between meaning and assertion is a constitutive part of their logic.

This argument cuts both ways when it comes to defining the critic's position vis-à-vis the literary text. Clearly it denies him the kind of methodical or disciplined approach which has been the recurrent dream of a certain critical tradition. On the other hand it offers a way beyond the rigid separation of roles which would cast him as a mere attendant upon the sovereign word of the text. What it loses in methodical self-assurance, criticism stands to regain in rhetorical interest on its own account. A similar reversal of priorities occurs in the deconstructive reading of 'literary' texts. There is no longer the sense of a primal authority attaching to the literary work and requiring that criticism keep its respectful distance. The autonomy of the text is actively invaded by a new and insubordinate style of commentary which puts in question all the traditional attributes of literary meaning. But at the same time this questioning raises literature to a point of rhetorical complexity and interest where its moments of 'blindness' are often more acutely revealing than anything in the discourse of philosophy.

Such has been the effect of Derrida's writing on a deeply entrenched conservative tradition—that of American New Criticism—which had already started to question its own ideology. What might have carried on as a series of skirmishing tactics (or virtuoso exercises in Hartman's manner) was galvanized by Derrida into something far more radical and deeply unsettling. We can now look more closely at the major texts in which Derrida sets forth the terms and implications of deconstructive reading. Rather than take his books one by one, I shall fasten upon certain crucial themes and argumentative strategies, acting as far as possible on Derrida's reiterated warning that his texts are not a store of ready-made 'concepts' but an activity resistant to any such reductive ploy.

If there is a single theme which draws together the otherwise disparate field of 'structuralist' thought, it is the principle—first enounced by Saussure—that language is a differential network of meaning. There is no self-evident or one-to-one link between 'signifier' and 'signified', the word as (spoken or written) vehicle and the concept it serves to evoke. Both are caught up in a play of distinctive features where differences of sound and sense are the only markers of meaning. Thus, at the simplest phonetic level, bat and cat are distinguished (and meaning is generated) by the switching of initial consonants. The same is true of bag and big, with their inter-substitution of vowels. Language is in this sense diacritical, or dependent on a structured economy of differences which allows a relatively small range of linguistic elements to signify a vast repertoire of negotiable meanings.

Saussure went on from this cardinal insight to construct what has become a dominant working programme for modern linguistics. His proposals broke with traditional thinking in two main respects. He argued, first, that linguistics could be placed on a scientific basis only by adopting a 'synchronic' approach, one that treated language as a network of structural relations existing at a given point in time. Such a discipline would have to renounce—or provisionally suspend—the 'diachronic' methods of historical research and speculation which had dominated nineteenth-century linguistics. Second, Saussure found it necessary to make a firm distinction between the isolated speech-act or utterance (parole) and the general system of articulate relationships from which it derived (la langue). This system, he reasoned, had to underlie and pre-exist any possible sequence of speech, since meaning could be produced only in accordance with the organizing ground-rules of language.

Structuralism, in all its manifold forms and applications, developed in the wake of Saussure's founding programme for modern linguistics. This is not the place for a detailed account of that development, which the reader will find expounded in Terence Hawkes's Structuralism and Semiotics (1977). Briefly, structuralism took over from Saussure the idea that all cultural systems—not only language—could be studied from a 'synchronic' viewpoint which would bring out their various related levels of signifying activity.

The precise status of linguistics in regard to this newfound enterprise was a topic of considerable debate. Saussure had argued that language was but one of many codes, and that linguistics should therefore not expect to retain its methodological pre-eminence. With the advent of a fully fledged semiotics, or science of signs, language would assume its proper, participant place in the social life of signs in general. Paradoxically, it was Roland Barthes—the most versatile of structuralist thinkers—who originally wanted to reverse this perspective and reinstate linguistics as the master-science of semiology. Barthes was quick to exploit the possibilities of structuralist method across a diverse field of cultural codes, from literary texts to cookery, fashion and photography. Yet in his Elements of Semiology (1967) Barthes is to be found expressing the conviction that 'the moment we go on to systems where the sociological significance is more than superficial, we are once more confronted with language'. And this, he explains, because 'we are, much more than in former times … a civilization of the written word'.

Of course, this text belongs to an early stage of Barthes's development, a phase he was later to criticize precisely for its overdependence on concepts of metalinguistic or 'scientific' knowledge…. [He] eventually travelled toward deconstructing such concepts through a textual activity aware of its own shifting and provisional status. But the kind of linguistic analogy that Barthes once deployed is representative of structuralism at a certain definite point in its development. It was at this point that Derrida intervened, with the object of wrenching structuralism away from what he saw as its residual attachment to a Western metaphysics of meaning and presence. In particular, he questioned the role of linguistics in dictating the methodological priorities of structuralist thought. Derrida's critique of Saussure, in his essay 'Linguistics and Grammatology' [in Of Grammatology], is therefore a crucial point of encounter for the deconstructive enterprise.

The argument turns on Saussure's attitude to the relative priority of spoken as opposed to written language, a dualism Derrida locates at the heart of Western philosophic tradition. He cites a number of passages from Saussure in which writing is treated as a merely derivative or secondary form of linguistic notation, always dependent on the primary reality of speech and the sense of a speaker's 'presence' behind his words. Derrida finds a dislocating tension here, a problem that other structuralists (Barthes included) had been content to regard as a puzzling but unavoidable paradox. What are we to make of this privileged status for speech (parole) in a theory which is otherwise so heavily committed to the prior significance of language-as-system (langue)? Barthes presents the question most succinctly:

A language does not exist properly except in 'the speaking mass'; one cannot handle speech except by drawing on the language. But conversely, a language is possible only starting from speech; historically, speech phenomena always precede language phenomena (it is speech which makes language evolve), and genetically, a language is constituted in the individual through his learning from the environmental speech. [Elements of Semiology]

The relation of language and speech is thus 'dialectical'; it sets in train a process of thought which shuttles productively from one standpoint to the other.

Where Derrida differs with Barthes is in his refusal simply to accept this paradox as part of a larger, encompassing project (that of semiology) which would overcome such apparent contradictions. For Derrida, there is a fundamental blindness involved in the Saussurian text, a failure to think through the problems engendered by its own mode of discourse. What is repressed there, along with 'writing' in its common or restricted sense, is the idea of language as a signifying system which exceeds all the bounds of individual 'presence' and speech. Looking back over the passage from Barthes quoted above, one can see how 'speech' terminology prevails, even where the argument is ostensibly stating the rival claims of language-as-system. Thus Barthes (drawing on Saussure) refers metaphorically to 'the speaking mass' in a context which purportedly invokes the totality of language, but which appeals even so to actual speakers and their speech as the source of that totality. Barthes may state, as a matter of principle, that language is at once the 'product and the instrument' of speech, that their relationship is always 'dialectical' and not to be reduced to any clear-cut priority. In practice, however, his theorizing leans upon metaphors which implicitly privilege individual speech above the system of meaning that sustains it.

Derrida's line of attack is to pick out such loaded metaphors and show how they work to support a whole powerful structure of presuppositions. If Saussure was impelled, like others before him, to relegate writing to a suspect or secondary status, then the mechanisms of that repression are there in his text and open to a deconstructive reading. Thus Derrida sets out to demonstrate

1. that writing is systematically degraded in Saussurian linguistics;
2. that this strategy runs up against suppressed but visible contradictions;
3. that by following these contradictions through one is led beyond linguistics to a 'grammatology', or science of writing and textuality in general.

Derrida sees a whole metaphysics at work behind the privilege granted to speech in Saussure's methodology. Voice becomes a metaphor of truth and authenticity, a source of self-present 'living' speech as opposed to the secondary lifeless emanations of writing. In speaking one is able to experience (supposedly) an intimate link between sound and sense, an inward and immediate realization of meaning which yields itself up without reserve to perfect, transparent understanding. Writing, on the contrary, destroys this ideal of pure self-presence. It obtrudes an alien, depersonalized medium, a deceiving shadow which falls between intent and meaning, between utterance and understanding. It occupies a promiscuous public realm where authority is sacrificed to the vagaries and whims of textual 'dissemination'. Writing, in short, is a threat to the deeply traditional view that associates truth with self-presence and the 'natural' language wherein it finds expression.

Against this tradition Derrida argues what at first must seem an extraordinary case: that writing is in fact the precondition of language and must be conceived as prior to speech. This involves showing, to begin with, that the concept of writing cannot be reduced to its normal (i.e. graphic or inscriptional) sense. As Derrida deploys it, the term is closely related to that element of signifying difference which Saussure thought essential to the workings of language. Writing, for Derrida, is the 'free play' or element of undecidability within every system of communication. Its operations are precisely those which escape the self-consciousness of speech and its deluded sense of the mastery of concept over language. Writing is the endless displacement of meaning which both governs language and places it for ever beyond the reach of a stable, self-authenticating knowledge. In this sense, oral language already belongs to a 'generalized writing', the effects of which are everywhere disguised by the illusory 'metaphysics of presence'. Language is always inscribed in a network of relays and differential 'traces' which can never be grasped by the individual speaker. What Saussure calls the 'natural bond' between sound and sense—the guaranteed self-knowledge of speech—is in fact a delusion engendered by the age-old repression of a 'feared and subversive' writing. To question that bond is to venture into regions as yet uncharted, and requires a rigorous effort of conceptual desublimation or 'waking up'. Writing is that which exceeds—and has the power to dismantle—the whole traditional edifice of Western attitudes to thought and language.

The repression of writing lies deep in Saussure's proposed methodology. It shows in his refusal to consider any form of linguistic notation outside the phonetic-alphabetical script of Western culture. As opposed, that is, to the non-phonetic varieties which Derrida often discusses: hieroglyphs, algebraic notions, formalized languages of different kinds. This 'phonocentric' bias is closely allied, in Derrida's view, to the underlying structure of assumptions which links Saussure's project to Western metaphysics. So long as writing is treated as a more or less faithful transcription of the elements of speech, its effects can be safely contained within that massive tradition. As Derrida puts it:

The system of language associated with phonetic-alphabetic writing is that within which logocentric metaphysics, determining the sense of being as presence, has been produced. This logocentrism, this epoch of the full speech, has always placed in parenthesis, suspended, and suppressed for essential reasons, all free reflection on the origin and status of writing. [Of Grammatology]

There is a deep connection between the craving for selfpresence, as it affects the philosophy of language, and the 'phonocentrism' which prevents linguistic method from effectively broaching the question of writing. Both are components of a powerful metaphysic which works to confirm the 'natural' priority of speech.

Derrida shows that these assumptions, though consistent and mutually reinforcing at a certain level, lie open to disruption as soon as one substitutes 'writing' for 'speech' in the conceptual order that governs them. The effect is unsettling not only for linguistics but for every field of enquiry based on the idea of an immediate, intuitive access to meaning. Derrida traces the exclusion or degradation of writing as a gesture perpetually re-enacted in the texts of Western philosophy. It occurs wherever reason looks for a ground or authenticating method immune to the snares of textuality. If meaning could only attain to a state of self-sufficient intelligibility, language would no longer present any problem but serve as an obedient vehicle of thought. To pose the question of writing in its radical, Derridean form is thus to transgress—or 'violently' oppose—the conventional relation of language and thought.

Such is the deconstructive violence to which Derrida subjects the texts of Saussure and his structuralist successors. It is not a question, he repeats, of rejecting the entire Saussurian project or denying its historical significance. Rather it is a matter of driving that project to its ultimate conclusions and seeing where those conclusions work to challenge the project's conventional premisses. In Derrida's words,

It is when he is not expressly dealing with writing, when he feels he has closed the parentheses on that subject, that Saussure opens the field of a general grammatology … then one realizes that what was chased off limits, the wandering outcast of linguistics, has indeed never ceased to haunt language as its primary and most intimate possibility. Then something which was never spoken and which is nothing other than writing itself as the origin of language writes itself in Saussure's discourse. [Of Grammatology]

Saussure is thus not merely held up as one more exemplar of a blind and self-deceiving tradition. Derrida makes it clear that structuralism, whatever its conceptual limits, was a necessary stage on the way to deconstruction. Saussure set the terms for a development which passed beyond the grasp of his explicit programme but which could hardly have been formulated otherwise. By repressing the problem which his own theory of language all but brought into view, Saussure transcended the express limitations of that theory. The very concept of 'writing' was enlarged through this encounter into something primordial and far removed from its place in traditional usage.

The point will bear repeating: deconstruction is not simply a strategic reversal of categories which otherwise remain distinct and unaffected. It seeks to undo both a given order of priorities and the very system of conceptual opposition that makes that order possible. Thus Derrida is emphatically not trying to prove that 'writing' in its normal, restricted sense is somehow more basic than speech. On the contrary, he agrees with Saussure that linguistics had better not yield uncritically to the 'prestige' that written texts have traditionally enjoyed in Western culture. If the opposition speech/writing is not subjected to a full critique, it remains 'a blind prejudice', one which (in Derrida's phrase) 'is no doubt common to the accused and the prosecutor'. Deconstruction is better provided with texts, like Saussure's, which foreground the problematic status of writing precisely by adopting a traditional perspective. A repressed writing then reasserts itself most forcibly through the detours and twists of implication discovered in Saussure. It is the 'tension between gesture and statement' in such critical texts which 'liberates the future of a general grammatology'.

Deconstruction is therefore an activity of reading which remains closely tied to the texts it interrogates, and which can never set up independently as a self-enclosed system of operative concepts. Derrida maintains an extreme and exemplary scepticism when it comes to defining his own methodology. The deconstructive leverage supplied by a term like writing depends on its resistance to any kind of settled or definitive meaning. To call it a 'concept' is to fall straight away into the trap of imagining some worked-out scheme of hierarchical ideas in which 'writing' would occupy its own, privileged place. We have seen how structuralism proved itself amenable to such uses. The concept of structure is easily kidnapped by a tame methodology which treats it as a handy organizing theme and ignores its unsettling implications. Derrida perceives the same process at work in the structured economy of differential features which Saussure described as the precondition of language. Once the term is fixed within a given explanatory system, it becomes (like 'structure') usable in ways that deny or suppress its radical insights.

Hence Derrida's tactical recourse to a shifting battery of terms which cannot be reduced to any single, self-identical meaning. Differance is perhaps the most effective of these, since it sets up a disturbance at the level of the signifier (created by the anomalous spelling) which graphically resists such reduction. Its sense remains suspended between the two French verbs 'to differ' and 'to defer', both of which contribute to its textual force but neither of which can fully capture its meaning. Language depends on 'difference' since, as Saussure showed once and for all, it consists in the structure of distinctive oppositions which make up its basic economy. Where Derrida breaks new ground, and where the science of grammatology takes its cue, is in the extent to which 'differ' shades into 'defer'. This involves the idea that meaning is always deferred, perhaps to the point of an endless supplementarity, by the play of signification. Differance not only designates this theme but offers in its own unstable meaning a graphic example of the process at work.

Derrida deploys a whole rhetoric of similar terms as a means of preventing the conceptual closure—or reduction to an ultimate meaning—which might otherwise threaten his texts. Among them is the notion of 'supplement', itself bound up in a supplementary play of meaning which defies semantic reduction. To see how it is put to work we can turn to Derrida's essays on Rousseau and Lévi-Strauss [in Of Grammatology], where the theme is that of writing in the context of anthropology and the cultural 'sciences of man'.

For Derrida, writing (in its extended sense) is at once the source of all cultural activity and the dangerous knowledge of its own constitution which culture must always repress. Writing takes on the subversive character of a 'debased, lateralized, displaced theme', yet one that exercises 'a permanent and obsessive pressure … a feared writing must be cancelled because it erases the presence of the selfsame (propre) within speech'. This passage occurs in the course of a chapter on Rousseau, whose Essay on the Origin of Languages is the starting point for one of Derrida's most brilliant meditations.

Rousseau thought of speech as the originary form and the healthiest, most 'natural' condition of language. Writing he regarded with curious distrust as a merely derivative and somehow debilitating mode of expression. This attitude of course falls square with Rousseau's philosophy of human nature, his conviction that mankind had degenerated from a state of natural grace into the bondage of politics and civilized existence. Language becomes an index of the degree to which nature is corrupted and divided against itself by the false sophistications of culture. What Derrida does, in a remarkable tour of argument, is to show that Rousseau contradicts himself at various points in his text, so that far from proving speech to be the origin of language, and writing a merely parasitic growth, his essay confirms the priority of writing and the illusory character of all such myths of origin.

Rousseau, for instance, treats of writing as the 'supplement' of spoken language, existing in a secondary relation to speech just as speech itself—by the same token—is at one remove from whatever it depicts. Such arguments have a long prehistory in Western thought. Like Plato's mystical doctrine of forms, the effect is to devalue the activities of art and writing by constant appeal to a pure metaphysics of presence, their distance from which condemns them to an endless play of deceitful imitation. For Derrida, the 'supplementarity' of writing is indeed the root of the matter, but not in the derogatory sense that Rousseau intended. Writing is the example par excellence of a supplement which enters into the heart of all intelligible discourse and comes to define its very nature and condition. Derrida shows that Rousseau's essay submits to this reversal even in the process of condemning the subversive influence he attributes to writing and its 'supplementary' character. A whole strange thematics of the supplement runs through the detail of Rousseau's argument like a guilty obsession and twists his implications against their avowed intent. That Rousseau cannot possibly mean what he says (or say what he means) at certain crucial junctures is the outcome of Derrida's perverse but utterly literal reading. Rousseau's text, like Saussure's, is subject to a violent wrenching from within, which prevents it from carrying through the logic of its own professed intention.

Music was one of the manifold interests which went toward the Rousseauist philosophy of culture, and Derrida has some fascinating pages relating Rousseau's ideas on the subject to the general theme of speech versus writing. The argument turns on Rousseau's preference for the vocal or melodic style, which he identified with the Italian music of his time, as against the harmonic or contrapuntal, which typified the supposed weakness and decadence of French tradition. As a matter of musical history this view is open to all kinds of scholarly question. Derrida, however, is not concerned so much with musicological fact as with the textual symptoms of doubt and duplicity which mark Rousseau's argument. The primacy of melody in music is held to follow from its closeness to song, which in turn represents the nearest approach to the passionate origins of speech itself. Harmony enters music by the same 'degenerate' process of supplementarity which marks off writing from speech. As music developed, melody (as Rousseau explains it) 'imperceptibly lost its former energy, and the calculus of intervals was substituted for nicety of inflection'.

Derrida fastens upon this and similar passages in Rousseau's text, and shows that what Rousseau is really describing is the condition, not of music in a phase of historical decline, but of any music which aspires beyond the stage of a primitive, inarticulate cry. Forgetfulness of origin may be the ruse by which harmony and writing manage to efface the primordial 'warmth' of a pure communion with nature. Yet Rousseau is forced obliquely to acknowledge (through the blind-spots and contradictions of his text) that music is strictly unthinkable without the supplement of harmony, or swerve from origin, which marks the possibility of its progress. Rousseau's 'embarrassment' is plainest when he attempts to define the originary nature of melody and song. If song is already, as Rousseau suggests in his Dictionary of Music, 'a kind of modification of the human voice', then how can he assign to it (Derrida asks) 'an absolutely characteristic (propre) modality'? The text unconsciously confesses what Rousseau is at such pains to deny: that thought is incapable of positing a pure, unadulterated origin for speech or song. Rousseau's argument, as Derrida describes it,

twists about in a sort of oblique effort to act as if degeneration were not prescribed in the genesis and as if evil supervened upon a good origin. As if song and speech, which have the same act and the same birth pangs, had not always already begun to separate themselves.

Rousseau's text cannot mean what it says, or literally say what it means. His intentions are skewed and distorted by the 'dangerous supplement' of writing as it approaches the theme of origin.

Derrida perceives such discrepancies at every turn of Rousseau's argument. Wherever the primacy of 'nature' (or speech) is opposed to the debasements of 'culture' (or writing), there comes into play an aberrant logic which inverts the opposition and cuts away the ground of its very meaning. Thus Rousseau's quest for the 'origin' of language turns out to presuppose an already articulate movement of production which must be cut off at source from any such originating presence. The supplement has to be inserted, Derrida writes, 'at the point where language begins to be articulated, is born, that is, from falling short of itself, when its accent or intonation, marking origin and passion within it, is effaced under that other mark of origin which is articulation'.

'Accent', 'intonation' and 'passion' are bound up together as positive terms in Rousseau's philosophy of man and nature. They all belong to that ruling ideology of voice-aspresence which equates the primacy of speech with the virtues of an innocent, unclouded self-knowledge. Rousseau constructs an elaborate mythology based on the contrast between 'natural' languages which remain close to their sources in passionate utterance, and 'artificial' language where passion is overlaid by the rules and devices of convention. The former he associates with 'the South', with a culture largely indifferent to progress and reflecting in its language the gracefulness and innocence of origins. The latter is identified with those 'Northern' characteristics which, for Rousseau, signalize the decadence of progress in culture. Passion is overcome by reason, community life invaded by the forces of large-scale economic order. In language the polarity (according to Rousseau) is equally marked. In the passionate, mellifluous, vowel-based language of the South one encounters speech near the well-spring of its origin. The tongues of the North, by contrast, are marked by a harsh and heavily consonantal structure which makes them more efficient as communicative instruments but widens the rift between feeling and meaning, between instinct and expression.

For Derrida, this Rousseauist mythology is a classic instance of the reasoning that always comes up against its limits in trying to locate any origin (or 'natural' condition) for language. He shows how Rousseau associates the threat of writing with that process of 'articulation' by which language extends its communicative grasp and power. 'Progress' involves a displacement from origin and a virtual supersession of all those elements in speech—accent, melody, the marks of passion—which bound language to the speaking individual and community at large. To deconstruct this mythology of presence, Derrida has only to pursue that 'strange graphic of supplementarity' which weaves its way through Rousseau's text. What emerges is the fact that language, once it passes beyond the stage of a primitive cry, is 'always already' inhabited by writing, or by all those signs of an 'articulate' structure which Rousseau considered decadent. As with Saussure's linguistic methodology, so with Rousseau's historical speculation: speech in its imaginary plenitude of meaning is disrupted at source by the supplement of writing.

This is why Rousseau occupies such a central place in Of Grammatology and Derrida's writing generally. He represents a whole constellation of themes which, in one form or another, have dominated subsequent discourse on language and the 'sciences of man'. His texts are a constant, obsessive repetition of gestures which miss their rhetorical mark and display the insufficiency of language when it strives for an origin beyond all reach. The deadlocked prolixity of Rousseau's text is also a lesson to the modern philosopher or linguist:

Our language, even if we are pleased to speak it, has already substituted too many articulations for too many accents, it has lost life and warmth, it is already eaten by writing. Its accentuated features have been gnawed through by the consonants.

Speech itself is always shot through with the differences and traces of non-present meaning which constitute articulate language. To attempt to 'think the origin' in Rousseau's fashion is therefore to arrive at a paradox which cannot be resolved or surpassed: 'The question is of an originary supplement, if this absurd expression may be risked, totally unacceptable as it is within classical logic.' The supplement is that which both signifies the lack of a 'presence', or state of plenitude for ever beyond recall, and compensates for that lack by setting in motion its own economy of difference. It is nowhere present in language but everywhere presupposed by the existence of language as a pre-articulated system. Philosophies that take no account of its activity are thereby condemned (Derrida argues) to a ceaseless repetition of the paradoxes brought to light in his reading of Rousseau.

This critique is extended to the structuralist anthropology of Claude Lévi-Strauss, where Derrida finds the same issues raised in terms of nature versus culture. Lévi-Strauss was among the first to perceive that the insights of structural linguistics could be applied to other 'languages' or signifying systems in the effort to elucidate their underlying codes. This gave rise to what is perhaps the most impressive single achievement of structuralism in its broad-based interpretative mode. Lévi-Strauss rests his analyses of myth and ritual on the conviction that, behind all the surface varieties thrown up by the world's different cultures, there exist certain deep regularities and patterns which reveal themselves to structural investigation. It is a matter of looking beyond their manifest content to the structures of symbolic opposition and sequence that organize these various narratives. At a certain level of abstraction, he argues, it is possible to make out patterns of development and formal relations which cut right across all distinctions of culture and nationality. Myths can then be seen as a problem-solving exercise, adapted to context in various ways but always leading back to the great abiding issues of human existence—mainly the structures of law and taboo surrounding such institutions as marriage, the family, tribal identity, and so forth. The end point of such analysis may well be to discover, as Lévi-Strauss frequently does, a formula of algebraic power and simplicity to express the logic underlying a dispersed corpus of myths.

Derrida reads Lévi-Strauss as an heir to both Saussure's 'phonocentric' bias and Rousseau's nostalgic craving for origins and presence. The two lines of thought converge in what Derrida shows to be a subtle but weighted dialectic between 'nature' and 'culture'. The phonocentric basis of Lévi-Strauss's method derives, quite explicitly, from the structural linguistics of Saussure and Roman Jakobson. But along with this methodological commitment there is also, according to Derrida, a 'linguistic and metaphysical phonologism which raises speech above writing'. In effect, Lévi-Strauss is seen as performing for modern (structuralist) anthropology the same ambiguous service that Rousseau performed for the speculative science of his day. The nature/culture opposition can be shown to deconstruct itself even as Lévi-Strauss yields to the Rousseauistic dream of an innocent language and a tribal community untouched by the evils of civilization.

Derrida's arguments are largely based on a single brief excerpt—'The Writing Lesson'—from Lévi-Strauss's book Tristes Tropiques (1961). Here the anthropologist sets out to analyse the emergence of writing and its consequences among a tribe (the Nambikwara) whose transition to 'civilization' he describes with undisguised feelings of sadness and guilt. He records how the motives of political power ('hierarchization, the economic function … participation in a quasi-religious secret') manifested themselves in the earliest responses to written language. Lévi-Strauss gives expression, like Rousseau, to an eloquent longing for the lost primordial unity of speech-before-writing. He takes upon himself the burden of guilt produced by this encounter between civilization and the 'innocent' culture it ceaselessly exploits. For Lévi-Strauss, the themes of exploitation and writing go naturally together, as do those of writing and violence.

Derrida's answer is not to deny the inherent 'violence' of writing, nor yet to argue that it marks a stage of irreversible advance beyond the 'primitive' mentality. On the one hand he points out that the Nambikwara, on Lévi-Strauss's own evidence, were already subject to a tribal order marked 'with a spectacular violence'. Their social intrigues and rituals of power are in manifest contrast to the retrospective feelings of the anthropologist, who elsewhere presents an idealized picture of their playful and uncorrupted nature. Moreover, as Derrida argues, this suggests that writing is always already a part of social existence, and cannot be dated from the moment when the anthropologist, that guilty spectator, introduced its merely graphic conventions. In truth, there is no such pure 'authenticity' as Lévi-Strauss (like Rousseau) imagines to have been destroyed by the advent of writing in this narrow sense. 'Self-presence, transparent proximity in the face-to-face of countenances … this determination of authenticity is therefore classic … Rousseauistic but already the inheritor of Platonism'. From this point it is possible for Derrida to argue that the violence of writing is there at the outset of all social discourse; that in fact it marks 'the origin of morality as of immorality', the 'nonethical opening of ethics'.

Thus Derrida's critique of Lévi-Strauss follows much the same path as his deconstructive readings of Rousseau and Saussure. Once again it is a matter of taking a repressed or subjugated theme (that of writing), pursuing its various textual ramifications and showing how these subvert the very order that strives to hold them in check. Writing, for Lévi-Strauss, is an instrument of oppression, a means of colonizing the primitive mind by allowing it to exercise (within due limits) the powers of the oppressor. In Derrida's reading this theme of lost innocence is seen as a romantic illusion and a last, belated showing of the Rousseauist mystique of origins. 'Writing' in Lévi-Strauss's sense is a merely derivative activity which always supervenes upon a culture already 'written' through the forms of social existence. These include the codes of naming, rank, kinship and other such systematized constraints. Thus the violence described by Lévi-Strauss presupposes, 'as the space of its possibility, the violence of the arche-writing, the violence of difference, of classification, and of the system of appellations'.

This latter has to do with the function of names in Nambikwara society, their significance and mode of designation. Lévi-Strauss offers a casual anecdote about some children who took out their private animosities by each revealing the other's name in a round of mutual revenge. Since the Nambikwara, according to Lévi-Strauss, place strict prohibitions on the use of proper names, this episode becomes symbolic of the violence that intrudes upon preliterate cultures when their language gives way to promiscuous exchange (or writing). Derrida counters with evidence—again from Lévi-Strauss's own text—that these were not, in fact, 'proper names' in the sense the anecdote requires, but were already part of a 'system of appellation'—a social arrangement—which precludes the idea of personal possession. The term 'proper name' is itself improper, so the argument runs, because it carries an appeal to authentic, individuated selfhood. What is really involved is a system of classification, a designated name which belongs to the economy of socialized 'difference' and not to the private individual. In this instance, what is prohibited by the Nambikwara is not the breach of any personal rights but rather the utterance of 'what functions as the proper name':

The lifting of the interdict, the great game of the denunciation … does not consist in revealing proper names, but in tearing the veil hiding a classification … the inscription within a system of linguistico-social differences.

Derrida's strategies are most clearly on view in these pages devoted to Lévi-Strauss. The 'nature' which Rousseau identifies with a pure, unmediated speech, and Lévi-Strauss with the dawn of tribal awareness, betrays a nostalgic mystique of presence which ignores the self-alienating character of all social existence. Writing again becomes the pivotal term in an argument that extends its implications to the whole prehistory and founding institutions of society.

Moreover, the evidence pointing to this conclusion is there in the texts of Lévi-Strauss, as it was in the writings of Rousseau and Saussure. It is not some novel and ultrasophisticated 'method' of reading devised to keep criticism one jump ahead. Nor does it impinge from outside and above, like certain forms of Marxist criticism which treat 'the text' as a handy support for their own superior knowledge of its meaning or mode of production…. Indeed, one of the myths or metaphysical ruses Derrida often attacks is the notion that writing is somehow external to language, a threat from outside which must always be countered by the stabilizing presence of speech. Carried down through a long tradition, from Plato to Saussure, this idea is most visibly (and paradoxically) inscribed in the Rousseauistic leanings of Lévi-Strauss. Writing becomes an exteriorized agency of violence and corruption, constantly menacing the communal values so closely identified with speech. Derrida's aim to is to show that, on the contrary, writing emerges both within the very theme of speech and within the text which strives to realize and authenticate that theme. Deconstruction is in this sense the active accomplice of a repressed but already articulate writing. In Derrida's much-quoted phrase, 'Il n'y a pas de hors-texte' ('There is nothing outside the text').