Paul de Man’s Proverbs of Hell

Geoffrey Hartman

The death of Paul de Man at the age of 64 deprives us of a literary critic whose influence, already immense in the United States and on the Continent, was beginning to be received in England. This influence is not linked to a large body of published work. De Man’s career started late. His studies in philosophy at the University of Brussels were interrupted by the war; after the war, he emigrated to America, taught at Bard, participated in Harvard’s Society of Fellows, took his PhD only in 1959 (his thesis on Mallarmé and Yeats still awaits full publication), and served as a teacher at Cornell, Johns Hopkins and the University of Zurich before settling at Yale in 1970. And although his earliest essays appeared in French during the 1950s (especially in Critique), they were not well-known until the Minnesota edition of Blindness and Insight (1983) incorporated some of them. Blindness and Insight was his first collection, published in 1971; a second major book, Allegories of Reading, appeared in 1980.[*]

Anyone who has read even a single essay of de Man’s can gauge the quality of his mind. Many of his early pieces circulated as if they were dangerous to the academy, and assured him a samizdat reputation. His was an analytical temper that preferred essay to book, and each essay left its mark. Students went to whatever university he was at. That he became a controversial and widely-known scholar only in the last years of his life was in harmony with his bearing. His courtesy was absolute, but so was his refusal to accommodate either the text or his thoughts. The teaching was superb because his intellect was quick and penetrated always beyond the canonical aspects of the philosophical or literary works he took up. The tragedy of his passing is made more acute by the fact that he was, even in his own eyes, just starting to say what he wished to say. His past work he considered prolegomenal to a study of Hegel almost complete at the time he died. In his final years he made his mind increasingly severe; it was hard, in fact, to get him to release earlier essays. (Columbia and Minnesota will, however, bring out some of the uncollected pieces.) The pressure on both text and reader was heightened by a prose that stripped all pathos and uplift from its subject. Yet even as prose it often approached a strange vertigo in its reversing intellectual movement, as if already caught up in that sober revel characterising the ‘absolute spirit’ in Hegel’s Phenomenology.

To gain an estimate of de Man one must first acknowledge the antagonism his work aroused. This adverse reaction had many sources. Some of it was based on vulgar error, and some was intelligent. (De Man himself points out that a piece called ‘The Resistance to Theory’, in the Spring 1982 issue of Yale French Studies, had been commissioned and then rejected by the Modern Language Association, a decision he accepted without rancour.) The most general charge by those unable or unwilling to read him was that his mode of exegesis, the intricate pressure he put on parts of a text, sinned against the direct or public meaning of the work as a whole, erecting rather than knocking down barriers between author and reader, and fascinating the seducible young by strength and ingenuity, rather than justness of mind.

This is an allegation made against intellectuals in almost every generation. They stand accused, in Dryden’s phrase, of injuring the page, of making it speak whatever they please. Books shaped by commonsensical values are turned into a foil for the idiosyncratic thinker: ‘The text inspires not them, but they the text inspire.’ That the ‘injury’ inflicted by de Man is a rigorous ‘deconstruction’ of the text, and, far from being subjective, is curiously impersonal, does not itself calm the polemics. For the critical wars of today seem to have replaced the religious wars of Dryden’s day, and intense speculation in literary matters is treated as a species of enthusiasm.

The modern polemical phase, it should be said, did not begin with de Man, Derrida, and the movement of deconstruction. The storm broke because of the surprising inroads of the New Criticism into the academy. We may consider the New Criticism rather tame, and appreciate it for introducing a tougher pedagogical stance, but traditional scholars feared that its exclusive emphasis on the specifically literary qualities of novel or poem would isolate these from the public. ‘We have been urged to investigate,’ Bonamy Dobrée wrote in The Broken Cistern, the Clark Lectures for 1953, ‘the recondite significance of imagery and symbol, of paradox and ambiguity, of irony and wit, and to embark on the treacherous oceans of the philosophy of language. New instruments have been thrust into our hands ... But haven’t we perhaps ... too exclusively pursued some ultimate, to the shouldering aside of what is most commonly valuable in poetry?” Overanalysed, the well-wrought urn becomes a broken cistern. Common humanity and even the survival of poetry are affected. ‘In paying as we do such attention to matters which only the specialist can be at home with, haven’t we, with great ceremony, brought poetry, not into the wide halls of judgment, but into the academic laboratory?’

If the New Criticism academised and so isolated art, then deconstruction, which suffers under the additional charge of turning everything into text (‘pantextualism’, ‘wall-to-wall textuality’), will obviously have an even harder time in being read by a wide audience. Certain pronouncements like Derrida’s ‘There is no hors texte’ have become notorious: they are taken out of context as statements about reality itself rather than about the difficulty of turning texts inside out. The assumption, now challenged, that texts have the interior/exterior structure of worldly facts – like gloves or houses, for instance – influences a model of interpretation that is all too common: we must find the core of the text, unveil something hidden, or harmonise outside features like figures of speech or verbal tricks with inside features presumed to be psychological.

De Man too, long before knowing of Derrida, began a critique of such models of reading. While still at Harvard, and teaching with Reuben Brower (author of the New Critical and devotedly pedagogical The Fields of Light), de Man insisted that the New Criticism take the full consequences of its emphasis on the literary qualities of a work. For in terms of a theory of literature, these deft critics recoiled from their own discovery that texts were defined pre-eminently by the fact of being texts – that is, made of language rather than ideas. They began to close off rebellious textual complexities (what John Crowe Ransom called, tongue in cheek, ‘irrelevant texture’) by a species of the very ‘heresy of paraphrase’ they had condemned. Having found that words were not rendered less ambiguous by being organised in a literary way – that the ambiguity, or, beyond it, the ambivalence, became more complex and discomfiting – a tendency arose to distinguish the literary from the linguistic in terms that relapsed into humanistic cant. De Man’s early essay, ‘The Dead-End of Formalist Criticism’, on Richards, Empson, Wheelright and Eliot, remains authoritative on this turn toward what he calls incarnational or salvational criticism. Dobrée, in short, wins out; and while close reading continues as a technique, ways are found to short-circuit the contradictions or divisions revealed by that technique.

‘It may be said,’ Empson wrote in Seven Types of Ambiguity, ‘that the contradictions must somehow form a larger unity if the final effect is to be satisfying. But the onus of reconciliation can be laid very heavily on the receiving end.’ The receiving end is, of course, the reader; and rather than developing a reader-response theory, with its subjectivist dangers, or facing clearly the problematics of closure in the realm of literary interpretation, the New Criticism described that ‘larger unity’ by means of naturalised religious or metaphysical concepts: the reconciliation of opposites, organic form, or (as in Northrop Frye) archetypal form.

The ‘insight’ of the New Criticism, then, was accompanied by a ‘blindness’ in direct relation to the intolerable force of that insight. Like a Blakean Giant Form those close readers shrunk back, at the level of theory, into a non-contradictory state far simpler than any text. De Man himself, in these early essays, uses not Blake but Hegel, Hölderlin and Heidegger. He reverts to such concepts as Hegel’s ‘unhappy consciousness’ to indicate that art could never heal the split between consciousness and being, or the division within being itself.

Yet Hegel and Co are by no means exempt from the ‘error’ of the New Critics, who feel compelled to separate out a work as literary, only to deny that specificity by talking about an overall coherence or ideal unity or happy incarnation. De Man’s essay on Heidegger, also written in the 1950s, shows that Hölderlin says exactly the opposite of what Heidegger makes him say. By treating Hölderlin as Holy Writ, by erecting him as the one Western (post-Greek) poet who ‘speaks Being’, Heidegger substantiates the same poetics of presence or of the undissociated sensibility which the New Critics also could not give up.

De Man’s lucid and unsentimental tracking of these moments of blindness has in it the obsession and chill of a good detective story. There is no plot in the ordinary sense, no pseudo-narrative or definitive peripety: yet he never loses focus. A central essay in Blindness and Insight defines the type of reading he is after. The reader ‘has to undo the explicit results of a vision that is able to move toward the light only because, being already blind, it does not have to fear the power of this light’. Such hermeneutic ‘Proverbs of Hell’ have poetic force despite their doctrine, which de Man humorously called his ‘scorched earth’ policy. Often, however, a slight ambiguity remains, either because the thought is so dense or because the style falters. What is the status of the ‘has to ... have to’? What necessity may be implied by the auxiliary power of that modest English verb? Other sentences come at us with a hyperbolic twist, yet strike home like a curveball right over the plate. ‘Not only does the critic say something that the work does not say, but he even says something that he himself does not mean to say. The semantics of interpretation have no epistemological consistency and can therefore not be scientific.’

The point made repeatedly, as if in danger of being blunted by the gentle reader, is that the relation between critic and text (including the critic’s own) is never that of a sophisticated revision replacing a naive insight. Text and commentary live each other’s life, die each other’s death. Each is a monumental project, and each is a mortal thesis astonishing in its vigour and fallibility. We cannot feel superior: like a Greek chorus, the critic is part of what he exhibits, compelled to say what he sees from a vantage-point that is neither above nor below, neither cosily inside nor on the envious periphery.

These brief remarks on de Man’s earliest concerns indicate lines of thought he will follow to the end. His focus is, first and foremost, on the problematics of reading. By examining the ‘rhetoric’ of critical readers he discovers a structure of revelation and recoil which he will name ‘misreading’ and say that it can be exposed but not scientifically corrected. ‘Deconstruction’ is not a bad label for this negative task, though de Man uses the term sparingly. The deconstructive method works with the language of each text rather than demystifying it from an authoritative, transcendant point of reference, backed up by a new terminology or metalanguage.

Closely associated with de Man’s view that we read to close off rather than to open texts is a critique of literary history and traditional concepts of periodisation (Modernity, Classicism, Romanticism). Such period terms are often unanalysed and contradictory, and serve mainly to privilege one kind of literature over another. So his essays on Rousseau, Wordsworth, Shelley and Hölderlin are not devoted primarily to rescuing the Romantics from disesteem, but to demonstrating that a later and favoured generation (Mallarmé, Rilke, Yeats, Valéry) has been read even more sloppily than the Romantics. In Defence of Reading was the title of the volume edited by Reuben Brower and Richard Poirier to which he contributed a remarkably balanced essay on ‘Symbolic Landscape in Wordsworth and Yeats’. Ten years later, in ‘The Rhetoric of Temporality’ (Blindness and Insight), his redefinition of allegory deconstructed a poetics of the symbol that had enmeshed – even duped – literary history from Goethe and Coleridge on.

As de Man’s career proceeds, it is fascinating that ‘reading’ moves increasingly into opposition to ‘history’ (i.e. history-writing). In the preface to Allegories of Reading he confesses that he had intended to write a history, but without success. ‘I had to shift from historical definition to the problematics of reading.’ The cause is never stated but seems clear enough in retrospect. At first it was a matter of making us aware how often interpretation is governed by a received version of literary history rather than by a reading of the work itself. In fact, ‘the more ambivalent the original utterance, the more uniform and universal the pattern of consistent error in the followers and commentators.’ Rousseau is the exemplary victim, and de Man says amusingly: ‘It is as if the conspiracy that Rousseau’s paranoia imagined during his lifetime came into being after his death, uniting friend and foe alike in a concerted effort to misrepresent his thought.’ This ‘concerted effort’ or ‘consistent error’ would not be possible without the insertion of the work into a history – which is really its insertion into a comforting principle of order.

That kind of history, in short, is not history at all, at least not temporal like the text it wishes to order. It maintains the myth, for example, of a parental structure or dialectical relation joining later and earlier authors. There is a relation, of course, but it is not edifying; de Man shows how Baudelaire remains an enigmatic stranger despite the effort of commentators (including Mallarmé) to fix his position as ‘the father of modern poetry’. Periodisation of this kind encourages a simplified notion of enlightenment, even a quasitheological notion: it divides history into the blindness of a previous era (e.g. the Synagogue) and the lux et veritas of a new era (the Revealed Church). Each new movement becomes a light-carrier of this kind; no wonder students of literature are sceptical about their discipline. History-writing subserves the drive for ‘correct’ or ‘canonical’ readings. Such readings create a pseudo-history and deprive interpretation of temporal development. They are a substitute form of belief, the anti-historical inheritance of a literate community, whether it calls itself by a secular or by a religious name.

De Man’s failure to pass from the activity of reading to literary history has the effect of keeping the historical dimension free. It is not foreshortened by a master-theory. Yet by the time he publishes Allegories of Reading, the hope of converting this freedom into a positive yield (actual history-writing) has faded, because the question, ‘Is history-writing possible?’ turns out to be identical to ‘Is reading possible?’ De Man, as if the fires of consciousness have consumed too much, and are feeding on themselves, now doubts that reading can result in understanding. Perhaps that doubt was always there, but it articulates itself with new vigour. He sweeps away wordly-wise definitions of criticism that see it ‘constructing a form of intelligibility for our time’ (Roland Barthes), as well as others that speak the language of natural desire, of psychopolitical imposition. That imposition occurs, and has immense influence, but it occurs precisely to the extent that pseudo-history dominates our consciousness of being-in-time, our ‘rhetoric of temporality’.

We can glimpse this further stage as soon as de Man develops the notion of allegory, perhaps by way of Benjamin’s attempt to grasp its mortifying effect. Benjamin’s study of an antiquated genre, the 17th-century German tragic drama (Trauerspiel), revalues emblematic allegory. Even later times have never quite succeeded in reducing allegory to a picturesque prop. What is now displayed is its basic structure, independent of a historical period; and this is allegory’s disquieting insistence on pure anteriority (on a time before time), which casts a shadow not only on the authenticity but also on the intelligibility of later, secular time. No doctrinal conclusions are drawn from this by Benjamin or de Man, though Benjamin is attracted to the linguistic version of pure anteriority: the idea of a primal, Adamic language. De Man, for his part, responds not only to the temporal predicament revealed by allegory as a rhetorical figure but also to its aberrant valuation in poetics. Allegory comes to mean a language of transparent signs (dove = peace), prosaic and disembodied, rather than poetic and embodied: yet it is precisely the possibility of this sort of allegory, of such arbitrary and contingent equations (stabilised only by the conventional memory), which de Man emphasises in his end-phase, especially his work on Hegel. Benjamin’s pathos falls away entirely, as does that of Hegel (which had entered the soul of de Man’s earlier work): and the difference between reading and understanding is purged of tragic overtones.

‘Allegories of Reading’ evokes, then, two different meanings of allegory which come together and obscure in a symptomatic way the most public part of a book: its title. First, a sense which remains close to riddle and the unreadable, and suggests how precarious reading is when confronted by literature. Reading seems relatively impotent from this perspective, or else impositional. The literary work rejects our hypotheses, showing them up as partial.

The danger of this attractively humble view is that we may forget that riddles have trivial as well as sublime solutions and so begin to view literature as secular scripture. De Man’s book, moreover, is not humble; its démarche toward the point where Rousseau, Rilke, Proust and Nietzsche reveal the contingent and arbitrary character of their fictions – where the linguistic penetrates the higher ‘aesthetic’ or ‘literary’ realm – is as elegant and ruthless (some would say mechanical) as ever. ‘Allegories of Reading’ moves therefore toward a second meaning which emphasises less the theme of making sense than of producing, through reading, a very basic and very ordinary rhetorical structure. The reduction of literature to such a figure is as foundational as de Man ever gets. Allegory is deconstructive, not only vis-à-vis the symbol, but vis-à-vis all tropological devices. It is itself no very secure bottom, however, because even its negative movement toward a moment of truth or epistemological certainty is undone by something outside the system itself, which de Man names irony.

If a text, engendered by these complexities, ‘puts an insurmountable obstacle in the way of any reading or understanding’, it is only natural to wonder why the human intellect should produce so strange a result. This question, in turn, presupposes that we know what we are doing, especially in the area of language. But for de Man, as for Benjamin, truth is the death of intention; and a text is as much a machine as a modality of understanding. ‘The machine,’ writes de Man, ‘is like the grammar of the text when it is isolated from its rhetoric, the merely formal element without which no text can be generated.’ In practice, de Man discloses an asymmetry between grammar (or a metagrammar like Jakobson’s) which tends to be normative, and rhetoric, which is disjunctive and figurative. It is this gap between figures of speech and ‘grammatical’ modes of expression that allows vertiginous possibilities of interpretation.

What reading produces, then, is not so much a replica or restitution of ‘inner experience’, as something more akin to vertigo than to understanding. This sense of vertigo had been associated in Blindness and Insight with the discovery of temporality: the sense of time as a process without synthesis, repetitive, un-reconciling, leading to death rather than to a recognition of permanence, and therefore no more ‘authentic’ than the naiver state it demystifies. De Man’s eloquence concerning this vertigo is unsettling in a critic who renounces affect. Describing the process in Baudelaire, he writes: ‘It may start as a casual bit of play with a stray loose end of the fabric, but before long the entire texture of the self is unravelled and comes apart. The whole process happens at an unsettling speed. Irony possesses an inherent tendency to gain momentum and not to stop until it has run its full course; from the small and apparently innocuous exposure of a small self-deception it soon reaches the dimensions of the absolute.’ And he concludes with an unusually emphatic assertion. ‘Irony is unrelieved vertige, dizziness to the point of madness. Sanity can exist only because we are willing to function within the conventions of duplicity and dissimulation, just as social language dissimulates the inherent violence of the actual relationships between human beings’ (‘The Rhetoric of Temporality’).

Scientific language is subject to the very same oscillation, the very same instability, as we read in ‘Criticism and Crisis’, the first chapter of Blindness and Insight:

every change of the observed subject requires a subsequent change in the observer, and the oscillating process seems to be endless. Worse, as the oscillating gains in intensity and in truth, it becomes less and less clear who is in fact doing the observing and who is being observed ... In the case of a genuine analysis of the psyche, it means that it would no longer be clear who is analysing and who is being analysed ... The need to safeguard reason from what might become a dangerous vertige, a dizziness of the mind caught in an indefinite regression, prompts a return to a more rational methodology.

It would be cheap to infer from this a repressed existential pathos that floods back. Yet there is affective writing in de Man, and it punctuates without diverting an always complex, vigorously prosecuted argument. The analogy now is Science Fiction rather than the detective story. I am reminded of how Science Fiction moves against the grain of its plot, how it collapses its spectacular narrative into a few instances of disorientation, Wordsworth’s ‘Blank misgivings of a creature/Moving about in worlds not realised’. In Wordsworth, however, such moments are still described in the vocabulary of inner experience, of ‘recollection’.

The return to Hegel in de Man’s last essays is again a revision of the model of inner experience traced through Hegel’s own development from the Phenomenology to the Encyclopedia. The vanity of human understanding is one error de Man is not tolerant of, and because he has no system of moral terms, he questions the function of certain terms in systems whose very pride of system makes him suspicious. ‘Understanding’ becomes one of these terms, as if it had made itself too agreeable and now suggested complicity rather than rigour. ‘Aesthetic’ is another such term, despite Kant’s careful assessment of its role vis-à-vis cognition and judgment. How we view the ‘aesthetic understanding’, therefore, is crucial to all further morality, in the personal or the public sphere. De Man’s reinterpretation of Hegel’s Aesthetics can give us a sense of his tenacious moral analysis of an unpromising philosophical text:

The infrastructures of language, such as grammar and tropes, account for the occurrence of the poetic superstructures, such as genres, as the devices needed for their oppression. The relentless drive of the dialectic, in the Aesthetics, reveals the essentially prosaic nature of art ... Hegel summarises his conception of the prosaic when he says: ‘It is in the slave that prose begins’ ... Hegel’s Aesthetics, an essentially prosaic discourse on art, is a discourse of the slave because it is a discourse of the figure rather than of genre, of trope rather than of representation. As a result, it is also politically legitimate and effective as the undoer of usurped authority.[†]

De Man’s reversion to Kojève’s sense of the centrality of the master/slave dialectic in Hegel and the intricate transfer of that dialectic to a very different philosophical work are less important than the reversal of the canonical – and perhaps Hegel’s own – reading of the Aesthetics. De Man chooses sections, he calls them ‘enslaved’, which have little or no aura, as if at that point the dialectic were an emetic that had purged the master-thinker and brought him into ‘the light of common day’. Perhaps because he himself is facing an end, there is something touching, even recklessly didactic about passages like this. Derrida’s commentary on Hegel appears almost baroque compared to de Man’s: the one is a juggler, the other a tight-rope walker; and while both display a saturnine brilliance, Derrida’s semioclastic revel, at least in Glas, is far less sober than de Man’s dizzy footwork only an inch above the text.

Much more might be said about the issue of mastery, from a pedagogical point of view. That so many students are attracted to deconstruction (and use it modishly against whatever went before) triggers a pedagogical anxiety. Deconstruction’s intellectual rather than identifiably ideological character differentiates it from other movements. To be more fair about that: by now we have historicised these movements or demystified their techniques – in short, found out their wishful and fallible ambitions. The academy might like to regard deconstruction as equally cliquish, another fashionable wave that will pass. Yet even in America the movement associated with de Man and Derrida is felt to challenge more than a particular set of institutionalised values such as nostalgia for a Common Reader or a Public Style or a Unified Sensibility. The spirit of criticism embodied by de Man seems to threaten the institutionalisation of criticism itself.

A rude scholar like Leavis could see the academy as the very centre and diffusion-point of his discipline, and a polite scholar like Dobrée could see the incorporation of criticism into the university also from its de-formative side. Neither questioned the outcome of whatever academic struggle had to ensue. Matters would be resolved, not entirely, not peaceably perhaps, but a new curriculum would evolve out of this conflict of interests. In America, however, those who oppose deconstruction as obscurantist, and those who espouse it as intellectually necessary, agree that it pushes against the limits of the academic. Lectures and books are being announced at an alarming rate that raise the question: is it teachable? Even more subversively, is academic teaching (now itself deconstructed) any better than a worldy or pseudo-religious mode of authority? Can a teacher’s relation to students free itself of seductiveness or imposition or – to sum it up in one massively bad word – pathos?

It may not be an accident, but the training procedures of psychoanalysis are undergoing the same kind of questioning, one so traumatically introduced by Jacques Lacan in the 1950s. Even in circles independent of Lacan the relation of analyst and patient, together with a notion of counter – transference (the resistance, and attraction, of physician to patient, and how it may already have distorted the very system of psychoanalysis, because it was not fully acknowledged by its first physician, Freud) – these have now come forward in a disturbing way, accompanied by extreme charges and sensational publicity. The Women’s Movement too has emphasised the ‘masculine, persuasive force’ in Freud’s remarkable assumption of authority, which he transferred to the successor institution. Thus deconstruction and psychoanalytic literary criticism often surprisingly form an alliance, despite the anti-psychologistic bent of de Man’s practice.

De Man was cool to such alliances. He placed himself stubbornly within a tradition he still labelled as philological, perhaps most amusingly in the Times Literary Supplement symposium on ‘Professing Literature’ (10 December 1982). There Harvard’s Reuben Brower finds himself in the company of Nietzsche and recent French theorists, even though his teaching of literature was founded on Richards’s ‘practical criticism’. The turn to theory, de Man writes, occurred as a return to philology, to an examination of the structure of language prior to the meaning it produces. Brower, from a purely teacherly and pragmatic standpoint, was able to show that ‘mere reading ... prior to any theory, is able to transform critical discourse in a manner that would appear deeply subversive to those who think of the teaching of literature as a substitute for the teaching of theology, ethics, psychology, or intellectual history.’

How, indeed, was de Man to situate himself? What ‘history’ could he align himself with, given his scruples about false continuities? He forgets to mention, for example, that Reuben Brower still believed there was a ‘key’ that would emerge from all textual bafflement to unlock a particular literary work. Or that philologists like Leo Spitzer tried to psyche out the ‘spiritual etymon’ of each writer. Or that Erich Auerbach valued historicism, and the variety of perspectives it introduced, as a Western heritage it would be tragic to lose and which had enabled him to construct the colourful narrative of Mimesis.

De Man, in the TLS symposium, was composing a little family romance, but one which had its point. He refers us to a tradition in which philosophy, including aesthetics, is an extension rather than an abrogation of philology. He invokes the shade of Brower as a humanist not deceived by fellow humanists who appropriate aesthetics as a mode of thought that circumvents both philological detail and philosophical precision. Above all, Brower serves to caution de Man’s followers about their infatuation with theory. The patience of the philologist averts the spectre of mastery, and its pedagogical anxieties.

De Man’s legacy is an intellectual style of remarkable purity. That purity, moreover, seems to me a moral quality as well as a methodological one. Undoing the illusion of mastery in others could lead to condescension, but whatever judgment takes place is always founded on a recognition of the necessity of error, that is of rhetoric; and no vigilance, literary or scientific, is able to achieve an ultimate correction. Yet if this obliges de Man to demonstrate one thing over and again, there is no temptation to disenchant anything except, possibly, his own style. I refer to a style marked by didactic fervour, whose undertow takes us into strange seas of thought, but it remains analytic and prosaic, with a minimum of semiotic play, and no mixing by montage of fiction and criticism. Though de Man’s critique of canonical reading has helped to inspire a flood of essays on Marxist and feminist lines, his own pages note, but do not indulge in, a vocabulary of crisis. He rethinks familiar terms from the trivium or takes his polarities from authors and their critics, always shifting slightly, diacritically, the value of exemplary words. He prefers to come to a position of non-understanding rather than manipulate the screen of received ideas. The effect is that of a new and precariously impersonal diction.

His favourite authors, often called the Romantics, create their illusions out of a permanently disruptive self-consciousness. Whatever systematic unravelling (deconstruction) occurs, or whatever vertigo threatens, de Man’s writing also continues to maintain itself, without system, afflatus, pseudo-history or borrowed machinery. By a hyperbolic and constitutive gesture of his own, the critic claims a co-ordinate integrity for his work.

Critical thought is not simply a response to or conversational extension of a prior text. It fascinates us, rather, by a labour of the negative so consistent that an intuitive remark of Benjamin’s is confirmed. ‘Unravelling an artfully wound skein with such certainty – is that not the joy of all productivity, at least in prose?’

[*] The 1983 Blindness and Insight, reviewed by Christopher Norris in LRB, Vol. 6, No 1, is available from Methuen at £7. 50. Allegories of Reading is published by Yale at £24. 50 and £7. 95.

[†] ‘Hegel on the Sublime’ in Displacement: Derrida and After, edited by Mark Krupnick, Indiana University Press (1983).