De Man's notion of phenomenality is compared with the idea of material vision attributed to him in the recent reception of his work, and with ideas of mental 'seeing' or the impossibility thereof in the work of Elaine Scarry and Timothy Bahti. These various critical constructions of literary phenomenality reinstate transcendental models of mind for divergent ends. De Man's material vision, in contrast, may be understood on the model of Kant's use of hypotyposis, as a figure for the analogy to which we rightly resort when dealing with speculative propositions about cognition. This essay appears in _Legacies of Paul de Man_, a volume of _Romantic Circles Praxis Series_, prepared exclusively for Romantic Circles (http://www.rc.umd.edu/), University of Maryland.

1. What do we see in reading? It might seem that "see" is a murky word, one whose conflation of sensory perception with cognition makes it a poor lens for the inspection of either. This suggestion, common in the last twenty years' work on lyric, takes its cue from Paul de Man's emphasis on the discontinuity of phenomenal and cognitive processes. In "Phenomenality and Materiality in Kant," de Man finds that in the Third Critique Kant needs "a phenomenalized, empirically manifest principle of cognition on whose existence the possibility of . . . an articulation [between conceptual and empirical realms of discourse] depends,"¹ but instead registers "a deep, perhaps fatal, break or discontinuity" between "language as a performative as well as a cognitive system" and "the powers of transcendental philosophy" (AI 79). This discontinuity "becomes apparent in the text" (AI 79) as what de Man calls "a material vision," "purely material, devoid of any reflexive or intellectual complication" (AI 82, 83). De Man goes on to propose that the "equivalence . . . in the order of language" of Kant's material vision is "the prosaic materiality of the letter" (AI 90, 89). For Kant's purposes, then, "a material vision" is the very opposite of the "phenomenalized, empirically manifest principle of cognition" that the aesthetic was supposed to provide; material vision is the Dantean hell that de Man writes for Kant as a parody of his desire for "phenomenal cognition." "No degree of obfuscation or ideology," de Man writes, "can transform this materiality into the phenomenal cognition of aesthetic judgment" (AI 90). In tracing the possible impacts of de Man's remarks on material vision, then, we might begin by remembering that for de Man, the foundering of Kant's transcendental system in material vision is a failure of redemption, the failure of materiality to "transform." As such, material vision is the manifestation in de Man's late work of his lifelong analysis of renunciation.² The anticlimactic sinewave of which de Man's notion of materiality is part is tonal evidence for how de Man reception might interpret it: the message of the larger narrative is that we are (only) what we are, that the world is what it is—which is not to say that we know what it is.

2. De Man distinguishes a linguistic function from "perception" in another way in "Resistance to Theory." Here de Man stipulates that "linguistic terminology"—the terminology of "literary theory"—"considers reference as a function of language and not necessarily as an intuition. Intuition implies perception, consciousness, experience, and leads at once into the world of logic and of understanding with all its correlatives, among which aesthetics occupies a prominent place."³ De Man is doing a lot, probably too much, in this single sentence. He disarticulates linguistic functioning from all other faculties conceived as a network. Reference is not only "not necessarily . . . an intuition" or perception, but not necessarily part of the "world of logic and of understanding" either, if that world is construed as a complete set of "correlatives." In this context the discontinuity between seeing and reading would be typical of the always possibly contingent relations between kinds of mental acts.

3. The contrast between kinds of mental acts, however, is not as stark as it might at first seem. The objects of seeing are perceptual and intuitive; the products of linguistic functions are "not necessarily" so. By de Man's own logic, his carefulness is not a direction to purge the word "seeing" from literary theory.⁴ The aesthetic concept of the phenomenalization of thought is not the only possible meaning of the word "seeing"; still less is it what seeing actually is.The word "seeing," in all its ambiguity, is split between perceptual and cognitive, literal and figurative, meanings, and only our own interpretation can unify and hence aestheticize it. In itself, I'll suggest, it represents what we know—and don't know—of perception and cognition more accurately than "perception" and "cognition" do, while the attempt to use "seeing" narrowly, over and against "reading," tends to entangle itself in aesthetics. Not that there is, after all, a passageway between perception and cognition. Rather, in coming to the place where such a passageway is needed and missing, we find ourselves in a difficulty for which "seeing" can be a rather honest figure, one that does not necessarily resemble aesthetic ideology's appropriation of it.⁵

4. If we take de Man's readings of Kant seriously, then after our discovery of aesthetic ideology we are called upon to go beyond transcendental philosophy's elements. Beyond, and not simply back to empiricism: de Man leaves us with a materialism "more radical" than that of empiricism (AI 121) in that it is found even at the heart of form. Although it is compatible with an unredemptive formalism, however, this materialism should not be made to serve retranscendentalizing operations. Frances Ferguson criticizes de Man for putting the mind in the position of always needing to start over, and compares him to Adorno in this respect.⁶ She is right, and that's one thing for which we may value de Man: the position to which de Man returns us (from which, he reveals, we have never yet actually budged) is the de facto condition of epistemology, and a cause for neither celebration nor disappointment. From Blindness and Insight on de Man presents this reality of our epistemological circumstances steadily and urbanely, without ascending into hysteria or theology. I worry that the current direction of the reception of de Man's ideas about materiality, in contrast, invents a new, philosophically reactionary transcendentalism that erodes de Man's legacy, or at least the attitude of resilient skepticism that is part of what I would like de Man's legacy to be.

   I.

5. For the editors introducing Material Events: Paul de Man and the Afterlife of Theory (2001), the constellation of ideas around "materiality" inspires the project of an alternative to the Third Critique, one that seems not to give up on transcendentalism but to reformulate it. They see opportunities in the a priori. "Whatever inscription designates, it conjures sheer anteriority," they suggest; "it does not deliver us to any immediacy of reference, [or] to any historical narrative that presumes to encode such, but to mnemonic programs that appear to precede and legislate these" (viii).⁷ De Man's work is called "the portal for a wide-ranging interrogation into how 'the event' operates in history, and what intervention in the order of inscription entails" (xii-xiii). "Intervention" is depicted as the storming and occupation of the imperial palace of faculties:

   By way of de Man's late work on "materiality" a project emerges that relates less to a "seventies" venture in theory than to still future and proactive investigations of and interventions in the hypertextual relay systems and programs out of which the "human" (and nonhuman) appears constituted, temporalization produced and managed, the "sensorium" altered, the virtuality of the present and the technicity of inscription brought to a point of passage or crossing. (xiii)

   Editors Cohen, Cohen, and Miller remodel the infrastructure of the Third Critique. De Man appears as an "engineer" who approaches "the archive, the prerecordings out of which experience is projected and semantic economies policed" (ix). CCM cite the career of Benjamin as another example of arche-engineering on the production lot of phenomenality, "where this trajectory finds an ultimate articulation as a radical (re)programming of the (historial) archive out of which the sensorium' would be alternatively produced" (x). Still, "experience is projected," and "the sensorium' . . . produced," out of the engineer's workplace. Engineering has a polemical meaning in the history of literary theory; this engineer seems to be the short-circuiter of structuralism's empiricist base. If Levi-Strauss is the anti-engineer of inductive reason, assembling significance from observable surfaces, in Material Events the engineer enters the studio at night and with a few keystrokes changes what is projected on the screen—the blockbuster known as SensoriumRedux. That the archive, and the source code, are external, and that there is no way to infer archive from screen, does not dilute the strictly transcendental nature of the fantasy.

6. Is de Man's work a "portal" to this project?⁸ There is fascination with the a priori in de Man, but there is irony toward it, too. Arguably, his work on "reference prior to designating the referent" (RT 8) and the "autonomous potential of language" (RT 10) show such an ambivalent fascination. In "Kant and Schiller" de Man speaks of "the historicity a priori" of the textual event (AI 133); the ever-expanding motion of the unfolding textual event is pursued to a logical extreme in "Phenomenality and Materiality in Kant" (AI 77-79). Part of de Man's point, however, is that this is not a wave that can be caught. His vision of "the saturation of the tropological field as language frees itself of its constraints" (AI 79) adapts the problem of plenitude in Neoplatonic metaphysics that he considers in his early work: the infinite generativity that would seem to be required of an infinitely powerful being ends in the existence of everything—a blow to unity and value whose damage, according to Lovejoy, the "great chain of being" is inadequate to repair.⁹ In his early "Criticism and the Theme of Faust," de Man condenses quotations from Lovejoy to conclude that "rationality, when conceived as complete, including all arbitrariness, becomes itself a kind of irrationality....The world of concrete existence...is no impartial transcript of the realm of essence....It is, in short, a contingent world.'"¹⁰ De Man's narrative here discovers materiality at the end of power as the anticlimax of an investigation into the a priori.

7. Similarly, a strongly "critical"-sounding passage of "Resistance to Theory" defines "literary theory" by its movement from "meaning" and "value" to "the modalities of production and of reception of meaning and of value prior to their establishment" (RT 7). Thus semiotics is, at least in principle, a more theoretical discourse than interpretation. De Man's conclusion, though, is that for the very reason that the logical conclusion of semiotics (for example, in Greimas) is a transcendental system applicable "to the generation of all texts," it rejoins the aesthetic project (RT 14). Only "reading," rather than any properly critical—in linguistic terms, grammatical—investigation, is "a negative process in which the grammatical cognition is undone" (RT 17). Literary theory has no exemption from aestheticization by virtue of its theoretical or specifically literary nature: that literary theory shifts to "modalities of production and of reception of meaning and value" makes it theoretical, but this is not necessarily a compliment to literary theory. Rather, it shows that literary theory remains susceptible to aesthetic ideology. Passages like this one should qualify the claim that the enterprise of "proactive investigations" into the general operation or generation of textual events "emerges" from the late de Man with his imprimatur. Still less does de Man write about intervening in or reprogramming mnemonic structures and the sensorium. Not only does he not mention the possibility of intervention in mnemonic and sensory structures; it's hard to reconcile such a possibility with his assertion that the linguistic structure that, he finds, disarticulates the Third Critique, cannot be mobilized by any drive (AI 147).¹¹

8. The notion of the sensorium as film studio (in the head or in the world), replete with engineers, set designers, directors, and projectionists, is one of western metaphysics' favorite motifs—the fantasy production lot of the aesthetic project. In this case politically progressive, it is in any case philosophically regressive. I do not find this motif advocated in de Man, and it can be found in the most undeManian places—in the startlingly strange perceptual literalism of Elaine Scarry, for example. Scarry takes the position opposite to contemporary deconstruction on "seeing" and "reading": she believes that in reading we "see" mental images that surpass the vivacity of nonliterary imagining, images that "somehow . . . acquire the vivacity of perceptual objects" (5). (As she notes, many philosophers and cognitive psychologists don't think "mental images" exist, but some do, and she does [258-259n6].) The verbal arts "displac[e] the ordinary attributes of imagining," Scarry asserts in Dreaming by the Book,

   —its faintness, two-dimensionality, fleetingness, and dependence on volitional labor—with the vivacity, solidity, persistence, and givenness of the perceptible world . . . . this comes about because we are given procedures for reproducing the deep structure of perception, and because the procedures themselves have an instructional character that duplicates the "givenness" of perception. (38)

   In another passage from Dreaming By the Book, "imaginary vivacity comes about by reproducing the deep structure of perception . . . . what in perception comes to be imitated is not only the sensory outcome (the way something looks or sounds or feels beneath the hands) but the actual structure of production that gave rise to the perception; that is, the material conditions that made it look, sound, or feel the way it did" (9). In approximating the material conditions of perception, texts make readers feel as though they are having particular perceptions and not just material visions of letters. This feeling is what Scarry calls seeing a mental image.

9. In "Kant and Schiller," de Man shows how Schiller's sublime forms a symmetrical set of chiasmi that allows for safe travel in the universe. Scarry is extravagantly Schillerian in her attachment to imitation as play and in her equanimious, symmetrical transfers;¹² flowers (and other small diaphanous things) are her ultimate imaginable objects because they already resemble mental images—they are the entry points to her aesthetic Paradiso. Scarry even cites Schiller's Aesthetic Education as an authority on this point: Schiller, she notes, "places the flower in the space of passage between material and immaterial," contributing to the "explanation of the easily imagined as something that can enter the mind precisely because it is always already in a state of passage from the material to the dematerialized" (63).

10. What does Scarry's apparatus have in common with Cohen, Cohen and Miller's? Surprisingly much, considering their contrary critical lineages and ideological aims. Scarry, too, pursues the post-Kantian project of securing the prosaic conditions, or "actual structures," of what she believes to be the sensory effects of reading. Moving from an investigation of textual effects to the conditions of possibility of the sensorium, her argument recalls eighteenth century arguments from the very existence of multiple senses to a supersense that supports them. For her, the fact that "verbal art, especially narrative, is almost bereft of any sensuous content" (5) only suggests an alternate, subterranean route to sensuous content—"miming the deep structure that brings the sensation about" (256n6). While she does not believe that we can alter the structure itself, she does believe that we can manipulate it for our own ends. Primary among the methods that Scarry claims enable such mimesis are the suppression of the sense of volition—the imaginer likes being given "direction" by someone, as this makes the content seem "given," therefore real—and a "sequence of coherent steps for constructing the image" (20). Scarry, too, is a programmer. Imagining under authorial direction is more than a way of being open to a mere illusion of sensation. Her version of mimesis is strong enough for virtual worldmaking: it is a repeatable method for stimulating in the body an image that responds to the content of a particular idea. If the methods are sleights of hand, they are sleights that provoke physiological responses and specifiable perceptions. For her, the fact that texts have their own sensory properties as ink shapes on paper does not interfere with their power to evoke images related to their semantic content. Rather, the sensory fact of their materiality rubs off on the mental images the words conjure, much as the ongoing sensory experiences of the dreaming body may contribute to the reality effects of dream content.¹³ Like Cohen, Cohen, and Miller, Scarry organizes her work through cinematic metaphors, down to two of her section titles, "Making Pictures" and "Moving Pictures." For the utterly nondeManian, indeed aesthetically ideological Scarry, then, language is again "material production" that elicits a "perceptual outcome" (20) in a "projective space" (14).

    II.

11. De Man, in contrast, reads Kant partly in order to give an account of a relatively straightforward attempt at aesthetics, one that does not conceal its limits. In "Kant and Schiller," de Man both points out how Schiller edits Kant's troubled Third Critique into a system that raises fewer questions, and admires Kant for his relative inability to paper over its problems. Just after having discussed Schiller's channeling of the "organic, sensory," "chaotic," and "concrete" elements of discourse into a symmetrical exchange with order and system, de Man writes:

    Here, the comparison to make with Kant is with Kant's statements about figuration, about what he calls hypotyposis, which is the difficulty [my italics—RT] of rendering, by means of sensory elements, purely intellectual concepts. And the particular necessity which philosophy has, to take its terminology not from purely intellectual concepts but from material, sensory elements, which it then uses metaphorically and frequently forgets that it does so . . . . At any rate, hypotyposis for Kant is certainly a problem for understanding, and a very difficult problem that again threatens philosophical discourse; whereas here it is offered by Schiller as a solution, again in the form of a chiasmus, for a similar opposition. (AI 153)

    What Schiller offers as a solution, Kant, to his credit, sees as "the difficulty," which he calls hypotyposis. The thinking of hypotyposis, then, could be a model for the realization of aesthetics' limits and the self-critical registration of this realization.

12. Literally, hypotyposis is a sketch, an outline, thus "form" with an emphasis on the emptiness, as Rodolphe Gasché observes (207). The rhetorical tradition uses this spatial figure of outline as an analogy for crisp verbal description. Hypotyposis is "clear explanation and almost visual presentation of events as if practically going on," writes Cicero (quoted in Gasché 207). The only thing that becomes clear in such definitions is that hypotyposis is as thin referentially as a term of art could be. The invocation of hypotyposis implies exactly nothing about how hypotyposis gets done. Nor do we even know exactly what its effects are, for hypotyposis is a figure whose effects are themselves described figuratively, with an "as if." In that way, it's close to what I mean by "seeing" and, as I'll explain later, what de Man means by "materiality." These are placeholders in language for something we do not know anything about, even whether it actually exists or not.

13. In §59 of the Third Critique, "On Beauty as the Symbol of Morality," Kant glosses "hypotyposis" as "exhibition [Darstellung], subiectio ad adspectum." He also specifies what hypotyposes are not: "mere characterizations, i.e. designations of concepts by accompanying sensible signs [bloße Charakterismen, d.i. Bezeichnungen der Begriffe durch begleitende sinnliche Zeichen]" such as "words, or visible (algebraic or even mimetic) signs [Worte, oder sichtbare (algebraische, selbst mimische) Zeichen]." Hypotyposes have intuitive content: they are schemata or symbols. A schema, in turn, is what mediates the assimilation of intuitions by concepts of understanding. And what is a schema? It's a rule. So far, the analysis of hypotyposis simply gives exhibition in general, as opposed to description, but explains what counts as exhibition only circularly. —What kind of a rule is a schema? Well, it is, for example, the rule that effects must follow causes in time, or the rule that substance has to have duration. Schemata are the rules of the classical natural world: if concepts are linked to intuitions through schemata, they're linked through their common fitness for that world. (A schematic hypotyposis might be, for example, a Euclidean proof.) So we can now say about hypotyposes that when they are schematic, they are natural, that is, they are plunged in space and time and in the natural world. But hypotyposes—I return now to §59—can also be symbolic:

    Symbolic exhibition uses an analogy . . . in which judgment performs a double function: first it applies the concept to the object of a sensible intuition; and then it applies the mere rule by which it reflects on that intuition to an entirely different object, of which the formal object is only a symbol. Thus a monarchy ruled according to its own constitutional laws would be presented as an animate body, but a monarchy ruled by an individual absolute will would be presented as a mere machine (such as a hand mill); but in either case the presentation is only symbolic. For though there is no similarity between a despotic state and a hand mill, there certainly is one between the rules by which we reflect on the two and on how they operate. This function has not been analyzed much so far, even though it very much deserves fuller investigation; but this is not the place to pursue it. Our language is replete with such indirect exhibitions according to an analogy, where the expression does not contain the actual schema for the concept but contains merely a symbol for our reflection. Thus the words foundation (support, basis), to depend (to be held from above), to flow (instead of to follow) from something, substance (the support of accidents, as Locke puts it), and countless others are not schematic but symbolic hypotyposes; they express concepts not by means of a direct intuition but only according to an analogy with one, i.e., a transfer of our reflection on an object of intuition to an entirely different concept, to which perhaps no intuition can ever directly correspond.
    
    [einer Analogie . . . in welcher die Urtheilskraft ein doppeltes Geschäft verrichtet, erstlich den Begriff auf den Gegenstand einer sinnlichen Anschauung[,] und dann zweitens die bloße Regel der Reflexion über jene Anschauung auf einen ganz andern Gegenstand, von dem der erstere nur das Symbol ist, anzuwenden. So wird ein monarchischer Staat durch einen beseelten Körper, wenn er nach inneren Volksgesetzen, durch eine bloße Maschine aber, (wie etwa eine Handmühle) wenn er durch einen einzelnen absoluten Willen beherrscht wird, in beiden Fällen aber nur symbolisch vorgestellt. Denn, zwischen einem despotischen Staate und einer Handmühle ist zwar keine Unlichkeit, wohl aber zwischen der Regel über beide und ihre Causalität zu reflectiren. Dies Geschäft ist bis jetzt noch wenig auseinandergesetzt worden, so sehr es auch eine tiefere Untersuchung verdient; allein hier ist nicht der Ort sich dabei aufzuhalten. Unsere Sprache ist voll von dergleichen indirecten Darstellungen, nach einer Analogie, wodurch der Unsbruck nicht das eigentliche Schema für den Begriff, sondern bloß ein Symbol für die Reflexion enthält. So sind die Wörter Grund (Stütze, Basis), Abhängen (von oben gehalten werden), woraus fließen (statt folgen), Substanz (wie Locke sich ausdrückt: der Träger der Accidenzen) [,] und unzählige andere nicht schematische, sondern symbolische Hypotyposen[,] und Ausdrücke fürBegriffe nicht vermittelst einer directen Anschauung, sondern nur nach einer Analogie mit derselben, d.i. der Übertragung der Reflexion über einen Gegenstand der Anschauung auf einen ganz andern Begriff, dem vielleicht nie eine Anschauung direct correspondiren kann.]

    Kant's examples of the equally symbolic corporeal and machinic presentations of monarchy might be said to reveal that symbolic hypotyposis all too conveniently imports qualities from the presentable to the unpresentable through the lie of poetry. The question arises why, in that case, Kant thinks symbolic hypotyposis is hypotyposis at all. What it is about the symbolic that merits comparison to the schematic, and makes it a kind of intuitive presentation? A monarchy, an animate body, and a "handmill"—whatever that is—are unlike; a monarchy is not an entity whose properties are completely definable. While it's fair to say that a monarchy must be answerable to the laws of physics, it's also fair to say that we don't know exactly how, because "the monarchy" itself is an approximation—if not an opaque name, then one that is dark around the edges. If someone tells us that it is at least an organization of bodies, however, then when we imagine its workings, we do so by using our concepts of other organizations of bodies. We don't have to know or be able to know what a tyranny is in order to be able to understand if someone tells us (wrongly or rightly) that, like a pepper grinder, it is operated by a single will.¹⁴ The hypotyposis indicates the schema to which disparate concepts are connected as an otherwise unpresentable cause is to its effects. Kant's text even illustrates hypotyposis by hypotyposis: His verbal comparison between the comparison between tyrannies and pepper grinders, on one hand, and symbolic and schematic hypotyposes, on the other, shows by symbolic hypotyposis how symbols and schemata are both hypotyposes. The proof concerning demonstration is not demonstrated directly: discourse and method occur together, method being talked about in discourse.¹⁵ My point is not that therefore, all is discourse; rather, this is the end of the line, and whatever conclusion one draws will be partly a matter of attitude. In this place, repeatedly, the philosophy of language takes a leap of faith, and claims to plug into natural science: it bids to become a natural science of language, grounded in necessity, as in Descartes, Russell, early Wittgenstein, or Chomsky. This is the place where philosophers start saying, with regard to analogy, that it really seems.¹⁶

14. One's degree of faith in the reality of analogy is not so much an interpretation as a decision. Hypotyposis, or any construct that serves its placeholding function, is a black box. For de Man, I would suggest, such a construct is not a solution but a "difficulty of rendering." We do not know what is there, or whether anything is there that is not a mere artifact of the terms of the problem; we do not know what real seeming consists in, or whether it consists in anything. We can only agree with Kant that if it is somewhere, there is where it should go.

15. Now, Derrida argues in "Typewriter Ribbon" that "materiality" is, like hypotyposis, a word to put down when one can go no further. In "Phenomenality and Materiality in Kant," de Man casts his characterization of Kant's description of the ocean as a catachretic act of nomination: "the only word that comes to mind is that of a material vision" (AI 82). In "Kant and Schiller" de Man again struggles to nominate "something [properly unnameable—RT] which one could call a progression—though it shouldn't be—a movement, from cognition, from acts of knowledge, from states of cognition, to something [again—RT] which is no longer a cognition but which is to some extent an occurrence, which has the materiality of something that actually happens" (AI 132). Noting de Man's appeals "to what he himself says he calls text'" (TR 336) and "what is called materialism'" (AI 128, quoted in TR 350), Derrida proposes that "materiality" for de Man is whatever fills "the place of prosaic resistance" (TR 350), "a very useful generic name for all that resists appropriation"(TR 353), or, going even further, "the name, the artifactual nomination of an artifactual figure . . . . a sort of invention by de Man, one could say, almost a fiction produced in the movement of a strategy" (TR 352-353). In this case the materiality of language is de Man's X at the spot where aesthetics cannot be completed.

16. Various philosophical choices may then be made. The resort to algebra—"here is where something should go"—leads to a literature in which indexicals and names are the foundations of knowledge: a new metaphysics, potentially, or a deconstructive nominalism.¹⁷ It also opens the way to negative theology, a recentering of metaphysics on the void (the path of Lacan and Zizek). Cohen, Cohen, and Miller continue to treat materiality as part of a realm "out of which experience is projected." They take their difference from Kant to mean that experience would be projected differently, and in this sense, seem to go on as though the transcendental system could be made to work to new ends. De Man's conclusion at the end of "Phenomenality and Materiality in Kant" is that the loss of "the architectonic unity of the system . . . marks the undoing of the aesthetic as a valid category" and "undoes the very project of such a philosophy" (AI 89). A forward-thinking literary theory, I would suggest, would not read in this conclusion the possibility of something like a transcendental philosophy persisting but no longer being predictable or unified,¹⁸ or persisting in ruins whose debris could be taken up and used for other purposes, or turned against itself, not least because one cannot turn against itself a machinery whose very existence is dubious. Rather, because Kant leaves a question mark at the most important place in his system, the transcendental philosophy will not have been something so substantial as to have produced debris. It will have been a plan for a system that never got finished. One can keep trying to revise and finish it, but that makes sense only if you wish it worked. If you are relieved that it doesn't work, the issue is still open, but for the moment, it makes more sense to say: So much for that. That way lies a renewed empiricism, not eighteenth-century empiricism, but a radical empiricism strong enough to encompass formalism.¹⁹

17. The attraction of the Cohen-Cohen-Miller model is its speculative liberation theology:

    We will take the position for the sake of argument and because it is interesting to consider, that what remains unengaged in de Man's text addresses the possibility of intervention in the mnemonic, the programming of the "historial," and a treatment of "materiality" that compels a rethinking of technicity and the "sensorium" on the basis of inscription. Among other things it would be an approach, given the "materiality of inscription," to the notion of the "virtual" and toward a rendering virtual—and hence, toward alternative histories to those programmed by inherited regimes of definition and perception (viii).

    The desire of this position lies in its ambition to reach into the structures that produce history and the sensorium, thereby arriving at a means of generating histories and sensoria, potentially for all. Changing the past is a crucial revolutionary desire for which Benjamin is a very good keynote. I do not mean to derogate it in the slightest by suggesting that it has been given most serious expression in the mode of impossibility. Rendered possible, it is no longer the same desire, no longer revolutionary but totalitarian. The idea of possibly intervening in regimes of definition and perception is downright frightening (it also opens up the black comic possibility of a Charlie Kaufmanesque nightmare of botched interventions), and with good reason "remains unengaged in de Man's text." As far as I can tell, the Cohen-Cohen-Miller angle on materiality is the wrong end of materialist criticism: it retranscendentalizes materialism rather than understanding materialism as something that detranscendentalizes form.

    III.

18. In his thoroughly brilliant book Ends of the Lyric, Timothy Bahti draws a distinction between "seeing" and "reading" using the example of Shakespeare's Sonnet 43:

    When most I wink, then do my eyes best see,
    For all the day they view things unrespected,
    But when I sleep, in dreams they look on thee,
    And darkly bright, are bright in dark directed.
    Then thou, whose shadow shadows doth make bright
    How would thy shadow's form form happy show
    To the clear day with thy much clearer light,
    When to unseeing eyes thy shade shines so!
    How would, I say, mine eyes be blessèd made,
    By looking on thee in the living day,
    When in dead night thy fair imperfect shade
    Through heavy sleep on sightless eyes doth stay!
    All days are nights to see until I see thee,
    And nights bright days when dreams do show thee me.

    Bahti notes that the poem's chiasmi of repeating words can be noticed even by a reader who doesn't know English, as in line 4: "darkly bright . . . bright, are bright . . . bright in dark." "So far, this is only seeing, not reading," he remarks (33). Others depend on semantics, like line 1's "most I . . . eyes best" (34). We could say that the chiasmus can be filled in with intuitive or symbolic content—literal or figurative seeing—while the chiasmic form channels us toward understanding ("seeing" in a metaphoric sense), that "dark, bright, bright, dark," is like "most, I, eyes, best." But "reading," Bahti asserts, "necessitates the distinction of sense as meaning from sense as the sense-perception of vision, the precise distinction of actual letters seen and actual letters not seen (no longer seen) but read acting" (38). He continues, "reading would appear' [at the end of the poem] if it were something one could see, but instead one can only read the vanishing of sight" (39). Bahti figures reading as seeing plus interpreting minus seeing, so the end of the process recedes over the horizon, under the line at the end of the column/poem: sublime. Described as though it were given, the perceptual process of seeing is made to provide ballast for reading-in-abstraction-from-seeing, an "end" of lyric in a reading remaining to be seen. There is a double sense of "reading" here: as participle, reading is what we do as we see and interpret; as abstract noun, it is the never wholly attained end product of seeing, interpreting, and learning not to see.

19. The assumption is that texts and reading, like tyrannies, are not phenomenal entities and processes, while pages and retinal activities, like pepper grinders, are. We are so used to thinking metaphysically, whether in deconstructive or in humanistic subgroups—especially those of us who work on lyric—that it is hard to imagine any other way of approaching seeing and reading. But what if texts and tyrannies and retinal activities were on the same level—were alike empirical entities, not aesthetic ones, only subject to more or less complex inspection? If there were, it would become evident that inspection itself is a difficulty all the way down (or up).

20. Daniel Dennett, in a typically brisk fifteen-page treatment of a vast question, in this case an essay called "Seeing is Believing—or Is It?," asks how we know when we've seen something. What Dennett poses as "the nature of takings'" is none other than the question of seeing and reading—the question of whether and when "a state of the nervous system" is to count as a "perception" (340,337). How do we know we've seen something? We remember it, so we believe we've had a perception; or we took a photo, and believe that the photo is a picture of what we perceived. "One of the reasons people tend to see a contrast" between belief and perception, according to Dennett, "is that they tend to think of perceptual states as much richer in content than mere belief states" (341). Not always: Scarry, stressing mental images, thinks of ordinary perceptual states as richer than what Dennett calls belief states (memories, conjurings), but thinks of literary belief states as being as rich as ordinary perceptual states. Bahti, stressing texts, understandably thinks of interpretations, belief states, as richer than the perceptual states of seeing letters. For my argument, though, it doesn't matter which way the values go. What matters is the dualism and its tendency to obscure "what happens in the middle," which in Dennett's opinion is everything. "No sane participant . . . would claim that the product of perception was either literally a picture in the head or literally a sentence in the head. Both ways of talking are reckoned as metaphors" (342), he remarks. "We should be leery of metaphor," he goes on, "but is there any alternative at this point?" (343). In the end, Dennett argues, "the idea that we can identify perceptual—as opposed to conceptual—states by an evaluation of their contents turns out to be an illusion" (352).

21. In illustrating this point, Dennett calls upon the classical celebration of hypotyposis as a presentation modeled on vision: "After all"—he channels an interlocutor—"perceptions are like pictures, beliefs are like sentences, and a picture's worth a thousand words. But," he goes on, "these are spurious connotations. There is no upper bound on the richness of content of a proposition" (341). The allusion to the rhetorical tradition is not casual: much of Dennett's discussion constitutes a commentary on metaphoric transfer from the perspective of empirical studies of perception. For the eliminativist Dennett, who believes that only technological obstacles prevent consciousness from being analyzed into directly or indirectly observable material elements, analogies between perception and cognition are not effectively ideological. Rather, it is the concept "cognition" that makes it sound as though there were supernatural substances or forces immune in principle to even the most powerful and indirect observation. For Dennett cognitions are in principle observable, while under the current scientific understanding, perception is scarcely less enigmatic than cognition. A cognition is like a perception not because it's as plain as day but because it's as clear as mud. Thus it is neither illuminating nor mystifying to compare cognition to perception. Because there is no nonmetaphorical way of talking about what even a perception is, seeing—in the ambiguous, sense-conflating sense—after all better represents the state of knowledge regarding perception and cognition than a distinction between "perception" and "cognition" which can be made logically but cannot be grounded in any difference in content.

22. Perceptions are like pictures, beliefs are like sentences—and perceptions are like sentences, and beliefs are like pictures. Until we know more, a thousand words on the topic are not yet worth a dime. As a deconstructive materialist writing in the wake of Paul de Man, I would rather tell you that than "burden the system with extra machinery," as Dennett puts it, "—scene-painting machinery or script-writing machinery" (344). Either kind of machinery resubscribes to systems of mind that understate the complexity, not only of reading, but of the very idea of sensory perception, which nonetheless remains the only channel epistemology gets.