13. Jonathan Culler¹

[1 The key for Culler citations is: OD: On Deconstruction (1982); PO: Structuralist Poetics (1975); and PS: The Pursuit of Signs (1981).]

        Jonathan Culler is best known for mediating literary theories of other scholars, especially Europeans. His major objective is to encourage a general theoretical reflection on the enterprise of literary studies. Certainly, a stong trend in this direction has appeared in America, and Culler's work may have been a material contributor. Whereas Structuralist Poetics (completed in 1973) tried to “introduce” "critical and theoretical writings" to “an English and American audience that had little interest in continental criticism," On Deconstruction (1982) aspired to “intervene in a lively and continuing debate” already well under way (OD 7). Culler came to devote more attention to American critics, albeit most of those he mentions had been working in theory for some time and he is often sceptical about their efforts.

        Culler's own theoretical position is a complex issue. Despite his apparent empathy with particular theorists, he functions less as a disciple or advocate than as a weathervane for the climate in contemporary critical theory. I will therefore concentrate on how his positions have been formulated and modified over the years. Since he does not necessarily endorse what he reports, his own commitments have to be inferred from his selection and evaluation of theories. His seeming inconsistency in moving his focus from one mode of theorizing to another can be read as a consistency on the higher level from which he surveys the “apparently incompatible activities" within "the field of criticism” (OD 17).

        As his career reveals, he welcomes in later trends a more consequential treatment of problems he was contemplating all along. The reservations and critiques in his earlier work eventually led to more radical consequences. In poststructuralist theories he saw a means to bring into the theoretical centre the problems he had found structuralism marginalizing as dissonant or disruptive. He was thereby able to attenuate his uneasiness regarding his sources. If he had seemed to treat structuralism as if he were preparing a vaccine from an originally pathogenic substance, he now seems to advertise deconstruction as a basically sound or benign medicine some critics mistake for a malady.

         “Structuralism" is a vague general heading for a generation of theories.² [2. Culler's survey has some major gaps, notably his omission of Marxist ideas (Note 17). In exchange, he covers at least briefly some work often classed under other headings, such as “formalism” (Propp, Sklovskij, Eikhenbaum), and, with stronger reservations, “New Criticism” (Empson, Brooks, Crane).] Although “to call oneself a structuralist" may have been a “polemical gesture,” many "disciplines” “have long been concerned with structure”; one might even “describe all theoretically oriented critics as structuralists” (PO 3; OD 19). But this usage would be unduly broad. The “diverse projects of structuralists” are “unified” by their use of “linguistics” as a “methodological model” for the “investigation of a text's relation to particular structures and processes, be they linguistic, psychoanalytic, metaphysical, logical, sociological, or rhetorical" (PO 4; OD 21). “Languages and structures, rather than authorial self or consciousness, become the major source of explanation” (OD 21).

        During the ascendancy of descriptive linguistics, researchers developed “procedures of segmentation and classification” and looked for "abstract units of structure” and "functional distinctive features which determine class membership” (PO 206, 10). Great attention was paid to “oppositions” and "differences," under the Saussurian assumption that these alone “make meaning possible”; "in the linguistic system, “there are only differences with no positive terms" (PO 10, 245; OD 28).³ [3. This thesis is routinely invoked (as here) without considering Saussure's immediate qualification: “But the statement that everything in language is negative is true only if the signified and the signifier are considered separately; when we consider the sign in its totality, we have something that is positive” (1966 [1916]:121). On the residual “logocentrism” of Culler's argument, see Fletcher (1984).] This assumption worked well enough in “phonological analysis,” “for many structuralists” “the model of linguistics itself” and "based on a reduction of the sound continuum to distinctive features, each of which “involves a choice between two terms of an opposition" (PO 14; cf. PO 93). Similar success was attained in analysing the sound structure of poetry, notably by Jakobson (cf. PO 65ff). Culler too keeps using word sounds as the star evidence for the centrality of binary oppositions (PO 10-14; PS 28; OD 96, 224).

        But both linguistics and poetics had trouble applying the same conception to other levels of structure. The assumption that “elements of a text acquire meaning as a result of oppositions into which the various areas of experience have been organized” raised the prospect of using the methods of linguistics to analyse “culinary, gustatory, olfactory, astronomical, acoustic, zoological, sociological,” and “cosmological" “codes” (PO 52).⁴ [4. This proliferation of codes was particularly pronounced in the work of Lévi-Strauss and Barthes (cf. PO 32-54, 227).] Yet as Culler demurred, “binarism” “permits one to classify anything"; “binary oppositions can be used to order the most heterogeneous elements” and "can be very misleading" when deployed to “present factitious organization” (PO 15f). If all language levels are broken down this way, "linguistic categories are so numerous and flexible that one can use them to find evidence for practically any form of organization” (PO 62).

        The drift in linguistics from descriptive to generative encouraged a new look at critical theory. just as “semiotics"⁵ [5. Culler feels “it would not be wrong to suggest that structuralism and semiology are identical," but "such shifts in terminology are of little moment" except for "history” (PO 6). He does stipulate that "semiology" is the more inclusive term. ] should “make explicit the implicit knowledge which enables people within a given society to understand one another's behaviour," a "theory of literary discourse” should “make explicit what is implicitly known by all those" “concerned with literature" and “interested in poetics" (PO 32, 118, 258). “Just as the speaker of a language has assimilated a complex grammar which enables him to read a series of sounds or letters as a sentence with a meaning, so the reader of literature" “has semiotic conventions which enable him to read series of sentences as poems or novels endowed with shape and meaning” (PO viii). Culler uses notions from generative theorizing to depict the work of Ruwet, Barthes, Todorov, and Kristeva (cf. PO 25, 118, 215, 218), although they were not following Chomsky's version of “generative grammar” very directly (cf. PO 7).

        Culler declines to see an “opposition" between “structural linguistics and generative grammar (PO 27). “At the level of generality which concerns those looking to linguistics for models,” “Chomsky's work can be taken as an explicit statement of the programme implicit in linguistics” “but not hitherto adequately or coherently expressed” (PO 7, 27). For Culler, all “grammars must be generative”; “they have simply not been explicitly” so (PO 24). He quotes Chomsky's (1965: 9) own technical definition of the term “generate" (“assign a structural description to”), but routinely uses it in the everyday sense of “produce." Things that get “generated" here include: "episodes," “forms of the text," a “novel," an “interpretation,” “metaphors,” "writing,” and “paradox” (PO 109, 146; PS 63, 193; OD 90, 201). The “generating" agents include: a “system,” “formal devices," "formal procedures," a "reordering of codes,” an “interpretation," "neatness,” "metonymy," and even "nature” (PO 109, 146, 107; PS 63; OD 90, 201; PS 193, 162). Most of these uses, though hardly compatible with the technical definition, are characteristic of much literary theorizing in that period.⁶ . [6. Especially confused from a technical standpoint is the locution “generate structural descriptions" (i.e., assign them to themselves) (PO 218). To be sure, Culler's sources have the same metaphorizing tendency, as when Barthes (1966: 57) writes of “works” “generating” "variations of meaning” (PO 118).]

         All in all, Culler distrusts the importation of linguistic models into the literary theories of structuralism. He decries the naive “assumption” that when an “analysis” or “interpretation" proceeds from the "methods" and "metalanguage” of "linguistics,” the results must be “correct” and have "interest and value” (PO 73, 218). Since "the linguists" they "read did not devote much time to discussion of the conditions which a linguistic analysis must meet," the "structuralists" believed that "if a metalanguage seemed logically coherent" and "if its categories” resulted from "systematic inquiry,” “then no further justification was required" (PO 206f; cf. PO 49). The “terms" of "linguistics," being “already linked by a theory," were thought to endow any study with "ready made” “coherence" (PO 102). Culler objects: “the value" of "conclusions and interpretations" is “totally independent of the linguistic model” being applied” (PO 109). At worst, a "linguistic argument" may be "pure obfuscation” (PO 108). “Specious arguments,” “spurious rigor,” "confusion," and "failure" are diagnosed in the projects of Jakobson, Barthes, and Greimas (PO 34f, 37f, 61, 72, 76f, 84f).

          The question is then whether such problems might be resolved by some correction or modification, or whether they are intrinsic to the entire research program. Whereas Culler's earlier work favours the former conclusion, his later work favours the latter. In this he parallels several prominent critics who traded their previous "structuralist" positions for “post-structuralist” ones (cf. OD 25).⁷ [7. Culler compares Harari's (1971) "bibliography of structuralism” with the same editor's (1979) .anthology of “post-structuralist criticism” as grounds to include among the transfers Barthes, Deleuze, Donato, Foucault, Genetic, Girard, Marin, Riffaterre, and Serres, leaving Levi-Strauss and Todorov as perhaps "the only true structuralists” (OD 25). However, anthologies often pursue famous names rather than theoretically sound groupings.] Culler doesn't accentuate this shift, since it reflects his own, but claims that enterprises now deemed post-structuralist” "were manifestly under way” in structuralist writings" (OD 25). The "theories" and “arguments" of the Tel Quel group "against the notions of a literary system and literary competence” still .presuppose these notions" (PO 243). Lewis (1982: 8) contended that “the structuralist enterprise" had "an acute self-critical awareness from the start," and Kristeva (1969: 30) called for "'perpetual self-criticism"; but such avowals collide with structuralism's “narcissistic relation to its own rhetoric” (OD 25; PO 251; PS 35; OD 21).

        Culler also opines that "structuralists generally resemble post-structuralists more closely than many post-structuralists resemble each other" (OD 30). This effect might be predicted, since the later movement depended vitally on the earlier and yet greatly surpassed it in diversification. Culler thought it “extremely difficult to go beyond" “structuralism": "any attack on structuralist poetics” will "fail to provide a coherent alternative" (PO 243, 253). “While structuralism cannot escape from ideology and provide its own foundations,” "critiques" of it “cannot do so either and through their strategies of evasion lead to untenable positions" (PO 253). Even though Culler still says "deconstruction has not refuted structuralism” (PS x), he has clearly moved his own point of orientation -- first judging post-structuralism by the standards of structuralism, and then doing just the reverse.

      Meagre. results in critical theory had caused general confidence in linguistics to decline, and scholars turned their attention to philosophers whose works foreshadow or circumscribe the dilemmas a science of language would encounter: Vico, Rousseau, Nietzsche, Husserl, Heidegger, Cassirer, Merleau-Ponty, and so on. Derrida inherited and reformulated these dilemmas in striking ways. Culler was initially unconvinced, accusing Derrida of ignoring “crucial differences between the conventions of oral communication and those of literature” “in Western culture”; and of "losing the distinction which translates a fact of our culture. Communication does take place. Many instances of language are firmly situated in the circuit of communication” (PO 133). These objections might “arrest the play" of Derrida's “concepts” and rebuke his attempts “to replace a metaphysic of presence by a metaphysic of absence" and “to invert the relation between speech and writing so that writing engulfs speech.” Culler also upheld the “distinction between understanding and misunderstanding” without which “there would be little point to discussing and arguing about literary works" (PO 121).

        Later, Culler saluted the trend by undertaking a synopsis of "deconstruction," now extolled as “the leading source of energy and innovation in recent theory," and "bearing on" its "most important issues" (OD 12). His earlier objections are expressly countermanded (OD 68, 89-103, 175f, 178). Sobered perhaps by his own prior reaction, Culler cautions that "deconstruction is ambiguously or uncomfortably positioned and particularly open to attack and misunderstanding” (OD 150f). He wryly hopes his "misreading of Derrida may in some contexts pass as sufficient understanding” (OD 178). It certainly suffices to disconfirm the simplistic outcry that Derrida is “playing with words” or “championing a principle or rule that any word in a text has all the meanings ever recorded for it” (CD 146, 219).

        "To deconstruct a discourse is to show how it undermines the philosophy it asserts, or the hierarchical oppositions on which it relies, by identifying in the text the rhetorical operations that produce the supposed ground of argument, the key concept or premise" (CD 86). “The logocentrism of metaphysics” is attacked, along with “the orientation of philosophy toward an order of meaning -- thought, truth, reason, logic, the Word -- conceived as existing in itself, as foundation” (OD 92). “Deconstruction” “reveals” "the inability of any discourse to account for itself and the failure of performative and constative or doing and being to coincide” (CD 201).

         “Derrida" “pursues with the greatest possible rigour the structuralist principle that in the linguistic system there are only differences, without positive terms” (OD 28). We can imagine no “first structure" or “originary event” in “language,” because “we must assume prior organization” and "differentiation” (CD 96). “If in the linguistic system there are only differences, Derrida notes, every 'sign’ must "relate to another element which is not simply present," and must participate in an “infinite referral in which there are only traces” (CD 99). A discourse is thus replete with “traces of forms that one is not uttering,” so that an item "can function as a signifier only insofar as it consists of such traces” (OD 96). This scheme enables endless inversions and displacements by bringing out presupposed “traces” from within the texture of the discourse, or, in Barbara Johnson's (1978: 3) well-known metaphor mixing militarism with hairdressing, by "teasing out the warring forces of signification within the text" (PS ix; OD 213; cf. OD 199, 220).

        If the identity of any entity depends on its own opposite, then the latter is implicitly included as well as excluded, and the opposition is not stable. Instead of “promoting one term” of the “opposition” "at the expense of the other,” “the second term,” usually "treated as a negative, marginal, or supplementary version of the first,” is now found to be “the condition of possibility of the first" (OD 213). Since, according to Derrida, "the hierarchy of binary oppositions always reconstitutes itself," "a movement that asserts the primacy of the repressed term is strategically indispensable" (OD 173).

         Yet a certain dualism inhabits such moves. On the one hand, “one demonstrates that the opposition is a metaphysical and ideological imposition" by “bringing out its presuppositions and its role in a system of metaphysical values” and by "showing how it is undone in the texts that enunciate and rely on it” (OD 150). On the other hand, “one simultaneously maintains the opposition" by “employing it in one's argument” and by “restating it with a reversal of status and impact." This technique is demonstrated on a series of key oppositions: "essential” versus "contingent" (or “inessential"), “inside" versus “outside," "central" versus “marginal,” "present” versus “absent," "performative” versus "constative" (or "cognitive"), "literal” versus “figural," "literary" versus “philosophical," “signifier" versus "signified," and so on (OD 146, 140, 107, 196, 95, 147, 182, 148, 188).

        For Derrida, the "traditional philosophical opposition" is "not a peaceful coexistence" but "a violent hierarchy" in which "'one of the terms dominates the other," so that "to reverse the hierarchy" might be to undo a "repression," (OD 85). Still, to the extent that these oppositions are presupposed in many discussions of language and thought, such an undoing can be mistaken for a wilful or aimless act of original violence, rather than a strategically designed response of counter-violence. The "'general displacement of the system" due to "'intervening in the field of oppositions"“ can be misconstrued as "anarchism" (OD 86, 151).

        Hence, Culler feels impelled to make the "large claims" of deconstruction seem “more comprehensible” and less "excessive" or “irritating” (OD 107, 204, 133). He solicitously deflates the "rumour" that “treats deconstruction as an attempt to abolish all distinctions, leaving" "only a general undifferentiated textuality” (OD 149). He assures us we have “no reason to stop work on theory" when we realize that “the language of theory always leaves a residue” (OD 133). He rejects the “belief” in the “humanities” that "a theory which asserts the ultimate indeterminacy of meaning makes all effort pointless." This belief is merely a pretext to evade the prospect that “the notions of meaning, value, and authority promoted by our institutions are threatened" (OD 179). "The identification of the normal as a special case of the deviant helps one to question the institutional forces and practices" of “legitimation, validation, or authorization that produce differences among readings and enable one reading to expose another as a misreading." These "forces" are naturally disturbed when we inquire whether “all reading is misreading."

        Structuralism had been quite emphatic about its scientific character. "Linguistics can give literature the generative model which is the principle of all science" Barthes (1966: 58) promised (PO 128). "Theories" would be “testable" in terms of “reproducing” “attested facts” about “literary competence” and about "the “grammar” of literature" whereby people "convert linguistic sequences into literary structures and meanings" (PO 122, 114). Research should best be carried out via some “algorithm," some “mechanical” and “automatic procedure" that would guarantee correct, consistent results and eliminate “the subjective decisions of the analyst” (PO 76, 81f, 94f, 123, 259). In accord with "generative grammar," "rules" "stated" as “formal operations" were judged the proper format “to make the implicit explicit” (PO 122). The powerfully metaphorical nature of this theorizing, a factor Culler stresses (PO 96-109, 255), was not widely acknowledged.

        Later on, deconstruction placed scientific ambitions in a very different light. “Theory may well be condemned to a structural inconsistency" (OD 109). “Theories grounded" on "an ideal norm that subsists behind all appearances” -- and a grammar has this status -- "undo themselves, as the supposed foundation or ground proves to be the product of a differential system, or rather, of difference, differentiation, and deferral." These three terms are needed to render Derrida's “différance," “alluding” to “the undecidable, nonsynthetic alternation between the perspectives of structure and event” (OD 97). “A scrupulous theory" “can never lead to a synthesis," because it "must shift back and forth between these perspectives," each of which “shows the error of the other" (OD 96). For Culler, this reservation undermines the linguists’ division between “langue" (“the system of a language”) and "parole" ("speech acts") (OD 96; PO 6), though he doesn't draw the corollary that "competence" (a main concept in his early theorizing) gets blurred with “performance." His treatment of "speech acts” points up “the impossibility of controlling effects of signification or the force of discourse by a theory, whether it appeal to intentions of subjects or to codes and contexts” (OD 128). This conclusion is argued via the hardly contestable “boundlessness" of contextual possibilities."

        In exchange for withdrawing as a conventional science, deconstruction might exert significant pressure on science and politics as institutions. Culler concedes that "such effects may be slow to work themselves out," but is optimistic that “the most abstract or recondite problems may have more disturbing consequences than immediate and intense political debates” (OD 157f). “Deconstructive “analyses may “have potentially radical institutional implications” if the “concern" with “the conditions and assumptions of discourse" empowers us to "engage the institutional structures governing our practices, competencies, performances" (OD 159, 156). "Self-reflexivity" and “inversions of hierarchical oppositions” “open possibilities of change" in "assumptions, institutions, and practices” (OD 154, 179).

        This outcome would go far beyond the earlier project of merely “making the implicit explicit." Derrida (1982) recommends that "deconstruction” "should seek a new investigation of responsibility” and "question the codes inherited from ethics and politics” (OD 156). He does not advocate "a methodological reform that should reassure the organization in place, nor a flourish of irresponsible” “destruction” sure to “leave everything as it is and to consolidate the most immobile forces within the university.” Yet an underground philosophical revolution seems tenuous to the degree that deconstruction "remains implicated in or attached to the system it criticizes and attempts to displace” (OD 151).

        Consider in this perspective how deconstruction encouraged a shift in linguistic theorizing. Formerly, structuralism had been beleaguered by its tendency to expound modest if not vacuous ideas as significant findings,  and to mistake an ostensibly more formal statement of a problem for a genuine-], solution. ⁸ [8 Typical cases are Greimas's (1966) precepts that “first and second person pronouns” “are replaced by “the speaker” and “the listener," and that “the story of a quest will have a subject and an object" like a “sentence” (PO 82); or Todorov's (1969: 28) proposal to "treat characters as proper names to which certain qualities are attached” in the same way as "properties" accrue to "the grammatical subject" of a sentence via "conjunction with a predicate" (PO 235).] Typically, the only radical aspect was the reductiveness, such as treating the literary work as if it had the same structure as a sentence (cf. PO 82f, 104). Post-structuralism was much more radical in carrying linguistic theorems to the point where their implications become disconcerting and disruptive.

        Alarmed by the abruptly radical uses of their staid science, protests are voiced that linguistics is being misunderstood, e.g. by Searle.⁹ [9 Searle(1977: 203) is compelled to diagnose Derrida's “penchant for saying things that are obviously false” because conventional linguists delight in rehearsing what is obviously true, for example, that a promise implies one's intent to carry it out. This vacuity is probably due to the habit in linguistics of viewing language as a system independent of contexts (cf. Beaugrande 1987b).] Such protests fail to appreciate that a conscious and intentional displacement has been deployed to signal that any science of language is potentially deconstructable because of what it excludes and represses, and that even its own fundamental theorems can be rigorously pressed to support this argument. Deconstruction never claimed that pioneer linguists like Saussure or Austin would approve its new readings of their work. Their writings have merely been enlisted as strategic spaces for a discourse driven by a quite different program.

         Deconstruction clinches the point by yielding apparent misreadings and paradoxes that must be read performatively, because our habituated mentality makes it so hard to read them constatively. The philosophical conditional obverts the objects and concepts customarily discussed in the scientific indicative. This move offers literary theory the chance to initiate a paradigmatic impulse of dissonance and demystification vis-à-vis the sciences, where, despite a similar pressure from philosophical metatheory, the traditional disregard of the performative still serves the interests of professional decorum.

        Some illustrative radicalisations may clarify my projection. A binary opposition derived simply by opposing the presence of an entity to its own absence remains vacuous or trivial until we question or invert the privilege of presence and uncover the metaphysical stakes involved. To assert that language is a system containing only differences is a noncommittal abstraction or an expedient dismissal of substances¹⁰ [10. “Identity of substance” is made the hallmark of “traditional studies which treat individual works as “organic wholes," as compared to structuralist criticism that finds a “homology of differences” (PO 97). One might however see substantialisrn in the objectifying of grammatical concepts by Barthes of Tedorov.] until we regard the activity of differentiating as an endless deferring or cross-referring within a system that is thus either centripetal or else without any centre and can yield only “ex-centric” meanings; equally disquieting is the consequence that "referentiality” is continually “postponed" (OD 251).

         To assert that meaning is composed of “semes" (minimal units of meaning) and that coherence is a classifying of them (cf. PO 77, 79, 87) merely makes the whole issue less tangible than ever, at least in lack of consistent, effective means for discovering semes; but the issue regains momentum when such units are set in motion to perform uncanny dis-seme-inations, e.g. via puns and etymologies.¹¹ [11. Rorty (1978: 146f) finds "puns" and "etymologies"“ among "the most shocking things about Derrida's work" (OD 144). The “pun” is able to “treat” “an “accidental” or “external relationship between signifiers” “as a conceptual” one (OD 91f; cf. OD 190ff, 240). “Etymology" can “put in question” the distinction "between the contingent and the essential” (OD 146). Both figures are beloved among the [then] Yale group.] The division between constative and performative, or between serious and non-serious speech acts seems tranquillising until the opposition gets blurry and the supposedly “parasitical” second term begins to absorb the first as its "special case” (cf. OD 112f, 116-125, 134; PS 223). Though these readings of linguistic theorems are “ex-centric" (in the sense just proposed), they are evidently not impossible, and we might gain more insight by being concerned less about their (constative) authorization than about their (performative) impact.

        Culler's exposition has to walk a fine line by justifying these radicalised propositions without reducing them to mere reformulations of familiar aporias. Culler incurs such a risk when he adduces the variability of reading according to time, circumstance, and reader (cf. OD 176) as grounds to see in Derrida's transformations of linguistic theorems only reasonable circumscriptions of everyday reading. Or, Derrida's notion of “open” "possibilities" of meaning is explained in terms of the "boundless" room for “other specifications of context” (OD 123f, 131) -- an account with no more force than the earlier structuralist prospect of “multiplying the codes" when “reading" "any particular stretch of discourse" (PO 52). The concept of the “unreadable," set down especially by de Man to designate “violent ambivalence," “the way” "the system of values in the text both urges choice and prevents that choice" (PO 106; OD 81; cf. OD 257ff, 276f), is occasionally flattened by relating it to “tedious,” “modern," or not “intelligible” (OD 259; PO 190). It would be quite contrary to Culler's intent if his "brisk common sense" (cf. OD 27)¹² [12. The phrase is Miller's (1976: 336). Culler returns the compliment by describing as "canny” Miller's presentation of “uncanny” criticism (OD 27). Compare Note 27.]  adjusted the precepts of deconstruction until they lean toward the self-evident and inconsequential. This result would leave the study and use of language and literature exactly as before, except that its customary blind spots would have been remapped with a new cartography, just as Gödel's proof has been absorbed without forcing a genuine realignment of the ambitions of formal logicians (cf. OD 133).

        Culler's solution is a precarious one, as "ambiguously" "positioned" and open to attack and misunderstanding" as the theory he expounds (cf. OD 150f). He combines the disruption of reversals and interventions with the reassurance that they do not have the effects they would in a traditional disputation. Deconstruction is portrayed as perpetually requiring and thus conserving its ideological counterpoints. For instance, the contention that “causality” is “produced” by the "tropological operation" of “metonymy or metalepsis" “does not lead to the conclusion that the principle of causality is illegitimate” (OD 86f). Or, the thesis that "'every signified is also in the position of a signifier"“ "does not mean that the notion of sign” “should be scrapped,” nor that “there are no reasons to link a signifier with one signified rather than with another”,  nor that a “text" is merely “a galaxy of signifiers"; “on the contrary, the distinction between what signifies and what is signified is essential to any thought whatever” (OD 188f). Or again, the sign's “possibility of endless replication" "does not propose indeterminacy of meaning in the usual sense: the impossibility of choosing one meaning over another," but "only the failure of signifieds to produce closure" (OD 189).

        What detractors of deconstruction cannot seem to understand is that "an opposition that is deconstructed is not destroyed or abandoned but reinscribed” (OD 133). Trained in traditional binary logic, their minds reject "the double procedure of systematically employing the concepts or premises one is undermining" (OD 87f). They feel cheated because they expect that once a concept is put in question, it should be thrown out in favour of some less questionable counterpart which deconstruction does not provide. And they resentfully construe the stance of such criticism as “skeptical detachment," rather than as the unwarrantable involvement” it intends to demonstrate (cf. OD 87f). In the course of this chapter, we shall retrace some aspects of this involvement in Culler's own theorizing.

        Deconstructive theses seem helpful for literature, a discourse that also subverts oppositions, such as fact versus fiction, reality versus imagination, specific versus general, internal versus external, and so on. It thereby systematically undermines attempts to impose definitive closure or complete interpretation, however much it may invite readers to try. It is a “mode of writing distinguished by its quest for its own identity” (OD 182). "The essence of literature is to have no essence, to be protean, undefinable”; it “transcends any account of it and can include what is opposed to it” (OD 182f). “Deconstruction” may agree with Hirsch in not viewing “literature” as “a privileged, superior mode of discourse" with an “authoritative epistemological status” (VAL 210; CD 1830 and still promote explorations which fit literature better than his (cf. Ch. 7).

        Indeed, deconstruction might be too readily absorbed into the established corpus of critical practices in America, leaving them hardly altered. Culler would like to ward off that outcome by clarifying what is at stake. Since “Derrida” “has not dealt directly" with “the task of literary criticism, the methods for analysing literary language, or the nature of meaning in literature,” "the implications of deconstruction for literary study must be inferred" (OD 180). Culler's deliberated, conscientious effort to draw out those implications can be mapped from the development of his stated positions during his apparent “structuralist" and "post-structuralist” phases.
        In some measure, Culler's view of the nature of literature has remained constant, as the following mosaic of quotations should prove. He upholds “the centrality of literary structures to the organization of experience” (PS 215). "Literature takes as its subject all human experience,” which it "orders, interprets, and articulates,” while “commenting on the validity of various ways" to do so (OD 10; PS 35). In addition, it is “a continuous exploration and reflection on signification in all its forms," “an exploration of the creative, revelatory, and deceptive powers of language,” and a “powerful, elegant, self-conscious” “manifestation” of “sense-making" (PS 3 5, 217). Hence, it is “the most complex of sign systems" and "explores" "the limits of intelligibility” (PS 35; OD 11). It offers “the best occasion” to “watch the complexities" of “the signifying processes” of “order and meaning” “work freely” (PO 264; PS 35).

        Like many theorists (e.g., Wellek and Warren, Frye, or Bloom) Culler maintains that “poetry lies at the centre of the literary experience" and “most clearly asserts the specificity of literature" (PO 162). “For structuralists,” “poetry undermines the function of ordinary language” and, for Kristeva (1969), “includes by definition all possible varieties of signification” (PO 183, 247). Invoking Derrida, Culler says that “in literature,” “we have the least cause to arrest the play of differences by calling upon a determinate communicative intention to serve as the truth or origin of the sign,” an insight restating the traditional view that “a poem can mean many things” (PO 133).

        Reading literature is correspondingly seen as a recovery against odds, a “process” Culler terms both “naturalization” and “recuperation" – “one of the basic activities of the mind," subsuming “the common operations of reading" (PO 137f, 178; cf. PO 178-88, 225). To “interpret something is to bring it" within "a mode of discourse which a culture takes as natural" (PO 137). “Structuralists" “imply that rhetorical figures are instructions about how to naturalize the text by passing from one meaning to another-from the “deviant” to the integrated,” using "rule-governed steps” (PO 179f, though much more may be involved. ¹³ [13. A comparable conception can be seen in many models: Fish's reader recovering from wrong hypotheses, Riffaterre's “transcending mimesis,” Hirsch's applying “validation” procedures, Iser's “filling in gaps,” Holland's reverting to “fantasy content,” and so on (cf. OD 37, 64-73; PS 52, 8Off, 219). But in each case, the recuperation is described differently. Compare also Culler's idea that “fantasy” is “the unnatural that the reader accepts as other nature” (PS 61).

          Throughout, Culler notes how "literature announces its fictional and rhetorical nature” and shows “awareness of rhetorical structures and forces," and of “textuality” (OD 183; PS 226). Yet a change of emphasis occurs. “Structuralism" followed “rhetoric” in classifying tropes into tidy formal categories (PO 179ff). “Metaphorical interpretation” was viewed as a “coexistence" of “added” "semantic features” with the “old" ones they “contradict” (PO 86). "Metaphor and metonymy" were argued to be cleanly opposed, for instance, via neurological function by Jakohson or via historical sequence by Hayden White, with Culler (1976) following suit (PS 60, 192, 63ff, 216).¹⁴ [14 White (1973) has the “sequence” arranged as “metaphor, metonymy, synecdoche, and irony” (PS 65). Culler “suspects” this scheme is “not a movement of history” but “a narrative curriculum with its own propulsive forces.” Earlier, though, he confidently made his own categorical classifications, e.g.: “in novels most symbolic operations follow the models of metonymy and synecdoche” (PO 226). He becomes steadily more uneasy about the status of "literal” meaning (PS 61, 70, 197). ]

      The later Culler finds this outlook, including his own contribution, “a very dubious enterprise” of manufacturing "contrasts,” such as the "opposition" having "an ambitious and deluded Romanticism, committed to an organic theory of imagination" and to “a continuity between form and meaning" or between "subject and object," on one side; and "an ironic, self-conscious modernism” that “questions these assumptions” on the other (PS 64, 155).

        The post-structuralist tendency is rather to trace displacements from one class of tropes to another (cf. Ch. 14, 15). "The distinction” between “Metaphor, based on" an “essential similarity,” and “metonymy, based on a merely accidental or contingent connection” is “put in question” as a plan to “assert the responsibility” of “rhetoric" by “privileging” “cognitively respectable" “tropes" (PS 190, 194, 198, 191, 199). De Man for instance is said to “reverse” the “metaphorical privileging of metaphor by assimilating metaphors metonymically to metonymy” (PS 199; cf. ON 243ff). Similarly, the “opposition” is blurred whereby we might “maintain the priority of the literal over the figurative” (PS 206). “The figurative is the name we give to effects that exceed, deform, or deviate from the code,” but these “get codified” and “create opportunities for new turns” (PS 209). “In general,” .any attempt to ground trope or figure in truth always contains the possibility of reducing truth to trope" (PS 204). However, Culler still vows that "the very notion of rhetorical effect" "requires there to be a distinction between literal meaning and metaphoric meaning” (PS 41) -- again remaining involved in an opposition while undermining it.

        At one time, "semiotics” was a “metalinguistic enterprise" “to describe the evasive, ambiguous, paradoxical language of literature in a sober, unambiguous metalanguage” (PS xi). "In recent years, it has become clear" that, just as in “deconstructive” “analysis” the "discourse is shown to repeat the structures it is analysing," "critical and theoretical discourse shares many properties with the language it attempts to describe" (OD 139; PS xi). “Deconstruction” has “particularly" illustrated how "theories" have an "uncanny involvement" in their own “domains" and "how critics become engaged in a displaced re-enactment of a text's scenario” (PS xi). Moreover, “literary works" "contain” their own “metalinguistic commentary" which determines “to a considerable extent" “the authority of critics” metalinguistic position" (OD 199). Though it “is always at work," “the distinction between language and metalanguage" “evades precise formulation." Still, Culler does not envision erasing the boundary between literature and criticism, unlike Bloom or Hartman,¹⁵ [15 Culler is strangely tolerant of Bloom and Hartman, who have clouded the public view of deconstruction far more than Miller, whom Culler so insistently rebukes. Bloom's failing is made out to be ours: "we” "take what he says about a poem and its intertextual, tropological genesis as an interpretation" and "are affronted” that it “should be so extravagant" (PS 14). Culler writes me he has now corrected" his “tolerance” of Hartman; and in a lecture cited my book for having uncovered the craftily obscured but essential theological centre of Hartman’s enterprise.] He can't see why “the critic must deem himself a poet”; “on the contrary," critics should “continue the pursuit of signs, the attempt to grasp, master, formulate, define," to “capture in their prose evasive signifying structures” (PS xi, vii). This “pursuit" leads to no final end, since “a literary analysis is one that does not foreclose possibilities of structure and meaning in the name of the rules of some limited discursive practice" (OD 182). We may not even "accept” a “definitive commentary" inside “the work," “telling us where to stop”; we may go on and finally “stop when we feel we have reached" "the place of maximum force" (PO 229).

         Following Derrida's lead, Culler is more inclined to erase the boundary between literature and philosophy (cf. OD 147, 181; PS 223), though, like other deconstructive gestures, this one subverts itself.. "the distinctiveness of philosophy is” “maintained within the argument that seemed to obliterate distinctions by treating philosophy as literature” (OD 184). Indeed, “the distinction between literature and philosophy" remains “essential to deconstruction's power of intervention” (OD 149f). So the “unwarrantable involvement” (OD 88) persists again.

        For some scholars, deconstruction is an occasion to erase the boundary between literature and criticism in the opposite direction, namely, by making the literature function as its own criticism. For Hillis Miller, “the text already contains the operation of self-deconstruction, in which two contradictory principles or lines of argument confront one another”; and "this undecidability “is always thematized in the text itself in the form of metalinguistic statements" (PS 15). "Great works of literature,” Miller" (1975: 30f) "insists, “have anticipated explicitly any deconstruction the critic can achieve."

        Culler repudiates this “shift" "taking place" "when deconstruction comes to America" and enabling it “to succeed" here whereas "Marxism and structuralism have not” (PS 15f).¹⁶ [16  Culler opines that the “shift” was “subtly inaugurated" by de Man and “transformed into a central methodological principle" by Miller (PS 15), both of whom certainly were successful.] If "the text” “is about self-deconstruction so that a deconstructive reading is an interpretation of the text,” then “deconstruction" is “tamed” and “made into a version of interpretation” specialized for “particular” “privileged” “themes, such as "undecidability” and “the relationship between performative and constative.” Culler expounds deconstruction so as to deauthorize such an outcome, however much the latter may have enhanced the success of the trend in his country. Evidently, new criteria for success are needed to promote the version of deconstruction Culler favors.

        Although he singles out “Paul de Man and Barbara Johnson” as its "best practitioners,” Culler does not proclaim only certain versions of "deconstruction" to be legitimate (PS 16; OD 227f). “One is tempted to speak of an original practice of deconstruction in Derrida's writings and to set aside as derivative the imitations of his admirers, but in fact these repetitions, parodies,” or “distortions are what bring a method into being” (OD 120). The fact that Culler "would not wish Miller's interpretations to be taken as the very models for deconstructive reading does not mean they are to be excluded from deconstruction" (letter to me).

       Culler has been quite consistent in opposing interpretation, a stance that attracted him to both structuralism and deconstruction. Like Iser, he considers “the interpretation of individual texts” to be "an ancillary activity" “only tangentially related to the understanding of literature" (PO 118; PS 5). "The notion that the goal of analysis is to produce enriching elucidations of individual works is a deep presupposition of American criticism” and foments "resistance to the systematic projects of structuralism, Marxism, and psychoanalysis” (ON 221). Though he disclaims the intent to "condemn interpretation," Culler makes it the “atomistic” “enemy of poetics” and the culprit that "subsumes and neutralizes the most forceful and intelligent acts of revolt in American criticism”; its misdeeds include "emasculating" the “promising mode of investigation” in “New Criticism,” and “nullifying” the “insights” of the “potentially valuable formalism” in Fish's “theory” (PO 118; PS 9, 11, 7, 130f).

        We are exhorted to “loosen the grip” of “interpretation” on “critical discourse” and "consciousness” (PS 5; PO 119). Culler borrowed from structuralism the “generative” idea that “the task of linguistics is not to tell us what sentences mean," but “how they have the meanings” that “speakers give them,” as an argument why literary theory cannot be expected to tell us what individual works mean (PO 74, 31, 97; PS 218; OD 21). Later, he welcomed deconstruction as a model that might resist “assimilation” to "interpretation” by "precluding the possibility of interpretive conclusions” (ON 222). He is dismayed when "American criticism has found in deconstruction reasons to deem interpretation the supreme task of critical inquiry and thus to preserve" “continuity between” “New Criticism" and "the newer criticism” (ON 220). If (in line with Miller's contention) "deconstruction reveals the impossibility of any science of literature and discourse" and thereby “returns criticism to the task of interpretation," problems arise: "deconstructive readings" fail to "respect” the “distinctiveness" and "integrity of individual works," and focus instead on “issues," “structures of language, operations of rhetoric, and convolutions of thought" (OD 220f).

        Culler has all along claimed that the proper "task" should be “to construct a theory of literary discourse” so as to "account for the possibilities of interpretation" and the “conditions of signification" (PO 119; OD 20). This approach can make "the proliferation of interpretations” into “an object of knowledge” (PS 48). We might “explain on what basis a range of interpretations” for a work "could be produced: what conventions and interpretive procedures enable critics to draw the inferences and make the statements they do" (PS 76). Although we won't prove what the work "really means, " we can uncover the “restrictive conventions of reading" without which "interpretation would be impossible” (PS 76; PO 250).

         How far interpretation can be de-emphasized in practice is uncertain. A theory that did not deliver any interpretations as vital evidence of its own range and necessity might be misprized.. Barthes and Todorov, for instance, rose to prominence because their unconventional studies of Balzac's Sarrasine and Boccaccio's Decameron opened new interpretive possibilities. Culler performed a similar service on the works of Flaubert. Even within deconstruction, we are told that “hierarchical reversals are likely to be the most convincing when they emerge from critical readings of major texts” (ON 174).

        Conversely, the unconvincing nature of structuralist interpretations reinforced Culler's criticism of their theories. When Greimas failed to present his “procedures of extraction at work on an actual text,” Culler suspected (as I do) that the "project" “may be impossible, in principle as well as in practice” (PO 84; cf. PO 213, 234). Culler is probably most uneasy about theories whose inadequacies are concealed behind the brilliant interpretations performed by their proponents -- a tactic he notices in Fish and, more guardedly, in Riffaterre (cf. PS 127ff, 93ff).

        Culler's real argument must therefore be that interpretation should be not so much curtailed as assigned a fundamentally different status and function: “the work” is “interpreted" as “the vehicle of an implicit theory of language” (PO 98; cf. PO 103). “The reordering of codes generates a different sort of interpretation": “the ultimate meaning of episodes and formulations is what they tell us about literary discourse” (PS 63). "The combination of context-bound meaning and boundless context" "urges that we continue to interpret texts" and "elucidate the conditions of signification” (ON 133). Most of the sample critics in my survey would accept this formulation, though for diverse motives.

        Culler emphatically called for "an explanation” of "the striking fact" that “a work can have a variety of meanings but not just any meaning” (PO 122). “If each text had a single meaning,” then "this meaning” might be “inherent to it and depend on no general system” (PO 243). “The text” can have a “plurality of meanings” “because it does not itself contain a meaning, but involves the reader in the process of producing meaning.” Thus, "the fact that a variety of meanings and structures are possible is the strongest evidence we have of the complexity and importance of the practice of reading.” "Variations in interpretation are not an obstacle," but “the fact with which one starts” (PS 124). “In general, divergence of readings is more interesting than convergence" (PS 51). An “emphasis on the variability of reading" “makes it easier to raise political and ideological issues" -- a gesture whereby Culler grants “Marxism," to be “not an illegitimate distortion, but one species of production" (OD 38), though its impact on structuralism via such figures as Lévi-Strauss, Goldmann, and Althusser was had utterly ignored in his survey.¹⁷ [17.Apart from asides like this one, I find a belated mention of Lévi-Strauss's dependence on Marx (PS 25ff). Derrida's “investigation” is compared in passing to “Marxism" as a “systematic expanding analysis of the overt and covert relations between base and superstructure or institutions and thought" (CD 221). Culler apparently shares “American criticism's” “resistance” to “Marxism” (cf. OD 221), although the latter should be prized because it does not "easily become a method of interpretation” (PS 16).

        Culler deliberates how the variability of readings might be circumscribed. He attributes "intelligibility" in "poetics" to "operative conventions” (PO 123). The "experienced reader of literature" has “assimilated" an “interpersonal” "system" (PO 128). Hence, “certain expectations" and "ways of reading" "impose severe limitations on the set of acceptable, plausible readings” (PO 127). Culler is confident that "an account of literary competence” should emerge when a critic "notes his own interpretations and reactions to literary works” and “formulates a set of explicit rules" for those results (PO 128; cf. PS 78). Such confidence looks odd next to some other judgments: "considering what particular prior readers have achieved, we tend to conclude that they failed to understand what they were doing" and “were influenced by assumptions they did not control"; and “few of the many who write about literature" have “the arguments to defend their activity" (OD 80; PO vii).

        Nonetheless, Culler believes the "considered reactions" of published critics to be “more Than adequate as a point of departure for a semiotics of reading” (PS 53). “By consulting the interpretations which literary history records for any major work, one discovers a spectrum of interpretive possibilities of greater interest and diversity than a survey of undergraduates could provide.” A critic's own reading makes a good model because "an explicit formulation of one's own interpretive operations would have considerable general validity." "The processes of writing” about them would “accentuate everything that is public and generalisable in the reading process" (hardly a deconstructive vision of writing!). If “to read is to operate with the hypothesis of a reader" and “to interpret is to posit an experience of reading" (OD 67), then a critic's own activities are already an implicit theory waiting to be made explicit and public. Naive, non-professional readers, for whom many authors wrote, can be disregarded.

        We should not be surprised that Culler has consistently tended to supplant the reality of reading with an idealization. For him, our proper concern should be “not what actual readers happen to do, but what an ideal reader must know implicitly in order to read and interpret works in ways which we consider acceptable" (PO 123f; cf. OD 34, 41, 79). This "theoretical construct” can be inferred "not" from “the immediate and spontaneous reactions of individual readers,” but from "the meanings they are willing to accept as plausible and justifiable when explained,” these being precisely “the meaning of a poem within the institution of literature” (PO 124; OD 35). This argument suggests why the “critic" “does not begin by taking surveys to discover the reactions of readers” (PO 50). But it risks delivering a fresh rationale for interpretation, namely to “explain” “meanings" for other “readers” to “accept.”

        Culler resists evaluating the fact that “texts have meaning for those who know how to read them” (PO 50) as an urgent empirical problem. Throughout his oeuvre runs the self-deconstructive dualism of referring an issue to readers while vigorously arguing against studying them directly. On one side, he avers that “the meaning of the work is what it shows the reader”; "poetics is essentially a theory of reading"; “a literary taxonomy must be grounded on a theory of reading”; "semantic description must provide a representation of the structuring activity of the reader"; “a theory of plot structure ought to provide a representation of readers’ abilities to identify plots”; and so forth (PO 130, 128, 120, 92, 205). He counsels that "one might start from data about the effects of poetic language and attempt to formulate hypotheses"; “to test whether the patterns isolated are in fact responsible," one may “alter the patterns to see whether they change the effects” (PO 68f; cf. PO 256). He feels “the conventions of poetry” are “easier to study” "as the operations performed by readers than as the institutional context taken for granted by authors," since “the meanings readers give” and “the effects they experience are much more open to observation” (PO 117). “Though the meaning of poems may not be reducible to the judgments of individuals," such "are the only evidence we have” (PO 50f).

        On the other side, he warns against “the dangers of an experimental or socio-psychological approach which would take too seriously the actual and doubtless idiosyncratic performance of individual readers" (PO 258). “Holland's well-intentioned empirical research” is decried as “miscarried" and “irrelevant" (PS 53). The idea from linguistics that "performance may not be a direct reflection of competence” is deemed a reason why it “would serve little purpose" "to take surveys of the behaviour of readers” (PO 123). Just as "the competence that the linguist investigates is not behaviour," the latter being “always at some distance" from the "rule," “claims about literary competence are not to be verified by surveys of readers” reactions which the analyst attempts to explain" (PO 10, 8, 125f, e.a.).

        We are therefore to take it on faith that “literary competence” conveniently “excludes any readings which seem wholly personal and idiosyncratic" (PO 128). Executing the same manoeuvre he notes in Fish, Culler says “we have little difficulty setting aside the idiosyncratic response whose causes are personal and anecdotal" (PS 125; OD 41). The “we" must be “the competent readers” who have “learned” “a series of techniques and procedures" and can decide which “readings” ought to be "placed outside the normal procedures of reading" (PS 125).¹⁸ [18 Culler agrees with Frye (AC 10f) that literary study leads to an “implicit” "mental process" just as "coherent and progressive" as the “explicit” one in “science” (PO 121). But note the vigorous complaints I cited Frye lodging against many critics. This group stands to profit most by asserting that “the notion" of "critical argument makes sense only if reading is not an idiosyncratic process" (PO 258).]

        His deconstructive standpoint thus offers Culler a fresh rationale against empiricism – whose methods frighten most American “literary scholars”,  theorists or not, in any case. “Theories of reading” are designated “stories of reading" that may be “dramatic” (Riffaterre's), "sad” (Stephen Booth's), “merry” (Holland's), and so on (PD 69; cf. OD 64-83). But due to “a gap or division within reading,” these “stories" typically “argue for a response that no one" "ever had” (OD 67; cf. Reichert, 1977: 87). So Culler need no longer judge their accuracy, as he had tried to before, but merely portray “the problematic situation to which stories of reading" (including his own earlier one) “have led us" (OD 83; cf. OD 79). He seizes the occasion to remark rnagisterially that “deconstruction” is “the culmination of recent work on reading" “because projects which began with something quite different in mind are brought up against the questions that deconstruction addresses” (OD 83). This conclusion again illustrates how a deficit left by structuralism, this time the lack of an empirical grounding, becomes a plus for post-structuralism.

        Though even in his early view Culler did not “believe” “in a single correct reading" "for each work” (PO 122), he did not renounce the mentality that divides right from wrong or “understanding” from “misunderstanding" (PO 121f). “To reject the notion of misunderstanding as a legislative imposition is to leave unexplained the common experience” of “grasping a mistake” and “the tacit knowledge" needed to do so (PO 121, 123). His more recent formulations better fit the thesis that "a semiotics of reading leaves entirely open the question of how much readers agree or disagree in their interpretations" (PS 50). Now, being right is more properly viewed as a performance. “A correct reading" is "imposed” by “cultural authority"; “truth is but a fantasy of the will to power" (Johnson, 1980: 14); "'a single"“ "meaning"“ would be "'theological," a "message of an "Author-God"“ (Barthes, 1977:146); “the concepts of criticism" “are a displaced theology"; and so on (PS 77; OD 178, 33; PS 160; cf. OD 187).

        From a deconstructive perspective, “a gap or division within reading” “always prevents there from being experiences that might simply be grasped and adduced as the truth of the text" (OD 67f). At most, “we have a stake” in “maintaining our belief” in such an “experience” and “seeing misunderstanding" as “an accident” or “deviation" (OD 68, 175). But “in fact the transformation or modification of meaning” in “misunderstanding is also at work” “in understanding” (OD 176). “A formulation more valid than its converse” would be “that understanding is a special case of misunderstanding," one “whose misses do not matter” or “have been missed” (OD 176, 178). “This account of misunderstanding is not, perhaps, a coherent and consistent position," but “it resists metaphysical idealizations and captures the temporal dynamic of our interpretive situation” (OD 178). This argument forms the background against which Hartman, Bloom, and de Man “treat literature and reading as a repeated historical error or deformation" (PS 13).

        Again balancing out a disruptive move with a reassuring one, Culler bases his modified “position” upon “the most familiar aspects" of “interpretive practice” (OD 176). The “acts of reading" at different times or by “different readers” are .not identical.” Moreover, "every reading" is “partial" because it must “select and organize” when faced with “the complexities of texts, the reversibility of tropes, and the extendibility of context.” "Interpreters" “can use the text to show that previous readings" are "rnisreadings,” and the same can later happen to their versions; in this sense, “the history of readings” reveals how some “misreadings” .may have been accepted as readings" (OD 176). “The historical perspective enables one to recognize the transience of any interpretation” (PS 13).

        Thus, the temporal variability and the provisional, utopian nature of understanding are deployed to subvert the “hierarchical opposition with institutional implications” wherein “misunderstanding” is a “complication or negation" that .might be eliminated" (OD 175ff). But the danger impends that this reassuring account may seem revisionistic. ¹⁹ [19. "Revisionism" is defined by Jameson (MF xv) as “the act of making a theory comfortable and palatable by leaving out whatever calls for praxis or change" and is unfit for “purely contemplative intellectual consumption.".] Or, it may leave the way open for a critic such as Bloom or Hartman to continue traditional projects within a more ornate and dramatized rhetoric, or to reinject theology in a more personalized and devious form.

       Problems were again converted to advantages when the status of the text was shifted. Structuralism believed that “in reading poems or novels, one does establish a hierarchy of semantic features” (PO 53). The function of “the text” was to “throw up semantic features and invite one to group and compose them,” “bridging the gap between the semantic features of words and the meanings of sentences or texts" (PO 236, 77). If the text is “assigned" "properties” "only with respect to a particular grammar," we "need not struggle" “to find some objective property of language which distinguishes the literary from the non-literary” (PO 113f, 128f). The decisive factor is “conventions" of "reading," both “general" (e.g., “impersonality, unity, and significance”) and "specific” (PO 178, 114). “Genres are not special varieties of language but sets of expectations which allow sentences of language to become signs of different kinds in a second-order literary system" (PO 129).

        The issue of what texts do to people and vice versa remained a thorny question. For Riffaterre, “the reader is “under strict guidance and control" "because of the complexity"“ of the text's "structures," "the multiple motivations of its words, and its “saturation by the semantic and formal features of its matrix" (PS 94). For Eco, the “open text outlines a “closed” project"; "you cannot use the text as you want but only as the text wants you to" (OD 70). Fish vacillated between "a reader who actively takes charge" and one "buffeted by fierce sentences" (OD 71).

       Deconstruction offered a resolution by making this indecision axiomatic. “Theories of reading demonstrate the impossibility of establishing well-grounded distinctions between what can be read and what is read, between text and reader" (OD 75). Culler will admit “no compromise formulation, with the reader partly in control, and the text partly in control,” but only a "juxtaposition of two absolute perspectives" (OD 73).

       The tension between text versus context, or intrinsic versus extrinsic, which troubled Wellek and Warren, is equally suspended. Influenced by critics like Barthes and Kristeva, Culler came to view the "text" as a metaphorical “space,” which he usually calls "intertextual” (less often “dialogical,” “discursive,” “ironic,” etc.) (PS 105f, 109, 118; PO 107, 184, 261; PS 116, 113; OD 33).²⁰ [20. “Spatial fictions” abound in Culler's theorizing, though he diagnoses them in Fish (PS 119; on Eco's “spatial” vision, see PS 201). The term “space" is also applied to “literature" at large, “metaphor," "the individual,” and (quoting Barthes) “the reader,” plus the more banal “typographic space” of a poem on a page (PS 7, 207; PO 230; OD 33; PO 184). One “discursive space” is said to have "sentences” as “constituents” (PS 117). Elsewhere, a parallel is drawn to a “gap in the mental process" (PO 184; cf. PO 77; OD 67), variously typified by conceptions of Brooks, James, Ingarden, and Iser (OD 36f).]

       “Poetics” could adopt “two limited approaches to intertextuality”: "to look at the specific presuppositions of a given text, the way in which it produces" “an intertextual space whose occupants may or may not correspond to other actual texts"; or else at "the conventions which underlie that discursive activity or space" (PS 118).

        “Infinite intertextuality" shows that “the autonomy of texts is a misleading notion” (PS 103). “The possibility of endless replication” that “is constitutive" of the “structure” of the “sign" makes the "text” a "'multidimensional space in which a variety of writings" "blend and clash," "a tissue of quotations drawn from innumerable centers of culture" (Barthes, 1977:146) (OD 188f, 32f).. The corollary that the “text" is "a machine with multiple reading heads for other texts," "a weaving" "produced only through the transformation of another text" (OD 139, 99) is reflected when Bloom dwells upon combat with a mighty predecessor, and Hartman upon a haunting by alien voices (Chs. 15, 16). If certain "stories of reading" still “reinstate the text as an agent” that “produces stimulating, unsettling, moving, and reflecting experiences," this strategy is just a way to create “more precise and dramatic narratives" and to “celebrate great works” (OD 82).²¹ [21 Barthes says “the artistic text” "works untiringly, not the artist or consumer” (OD 70). Culler has “sentences” doing such things as “leading us through a garden" and “revealing" “an orchard,” or “forming a drama of innocence" (PO 194, 202). He also attributes to Derrida the theses “the sign has a life of its own," and “the written word is an object in its own right” (PO 248, 133).

        As we might surmise from Culler's treatment of the text, the question of its organic unity” (PO 137) has caused him some inconsistencies. Within structuralism, he saw “totality” as "the end which governs” the “teleological process" of "understanding," a view he attributes in “various forms" to Jakobson, Greimas, Todorov, and Barthes (PO 171f). "Unity is produced not so much by the intrinsic features” of texts "as by the intent at totality of the interpretive process" controlled by “the expectations” of “readers” (PO 91). This “intent” "may be seen as the literary version of the Gestaltist law of Prägnanz: that the richest organization compatible with the data is to be preferred” (PO 174). “As Merleau-Ponty” (1964) “says,” "it is only in the light of hypotheses about the meaning of the whole that the meaning of parts can be defined” (PO 92). The “expectation” of “totality" is “often” "disappointed,” but is still “the source of the effects,” as when "modern poetry" “fails to realize, except momentarily, the continuity promised by formal patterns" (PO 172).²² [22. Such passages indicate that certain theorizing may apply only to special kinds of works, such as “open works" that “invite the reader to play a more fundamental role”; or “works" that are “violently explicit in their dealings” and make a “radical contribution to a theory of signs and signifiers,” or “modern” “fiction” with its “faceless protagonists” (OD 37; PS 36; PO 231). Such avant-garde critical theories as structuralism and deconstruction typically have affinities for modem art, however much they may generalize their claims for all art. As de Man says, “the affinities between structuralism and the nouveau roman are obvious” (BI 61). Alain Robbe-Grillet once told me Derrida "feels close to me” (“se sent proche de moi”).

        Culler himself used “unity" and “totality" as central constructs for discussing particular poems (PO 172-74; PS 68-76). He invoked “the importance" and “power of the convention of unity” and averred that “readers of poems” "feel an overwhelming compulsion to transform” their “heterogeneous experiences unto a unified vision” (PS 69ff). Since "we think of a successful literary form as a synthesis," “critical interpretation seeks a unified totality” (OD 252).

        On the other hand, Culler wanted to swerve away from "traditional," “organic,” or "New” critics by questioning whether “the task of criticism is to reveal thematic unity" (PO 119; cf. PO 67, 116; PS 3ff; OD 202-05). Culler now thinks it a “major point” of "agreement” that “literary works" are “not “organic wholes, " but “intertextual constructs” (PS 38). A "principle of structural semiotic analysis" was that “elements of a text do not have intrinsic meanings" but “derive their significance from oppositions" "related to other oppositions in a process of theoretically infinite semiosis” (PS 29).

         Deconstruction further blurs the boundaries between the “inside and outside” of the text, projecting it as “already riven by contradictions and indeterminacies inherent in the exercise of language" (OD 199; PS 43). "Distance, absence, misunderstanding, insincerity, and ambiguity" are postulated to be “features of writing" and of “speech" as well (OD 101, 103). "Derrida suggests" "the meaning of meaning” “is infinite implication," "the unchecked referral from signifier to signifier" (OD 133). We now face “the impossibility of ever mastering and making present the intertextuality of a particular text” (PS 118). Still, we are not left only with “interpretation which applies one text to another in order to produce new readings,” as practiced by “Bloom and others." We should also strive to be “acute analysts of intertextuality" and “engage with all the pragmatic presuppositions, the conventions of discourse, and the sedimentation of prior texts."

        Unity ultimately gets the usual deconstructive treatment that both undermines and preserves, keeping an unwarrantable involvement. It is "not easy to banish" “notions of organic unity"; even "critical writings" that “celebrate heterogeneity" continue to "rely" on them (OD 200, 220). Though "deconstructive readings show scant respect for wholeness and integrity,” "deconstruction leads not to a brave new world in which unity never figures, but to the identification of unity as a problematical figure” (ON 220, 200). "Interpreters are allowed to argue that a work lacks unity, but to ignore the question of unity is to flout the obligations of their task" (OD 220p. “Deconstructive criticism" “engages with the hierarchical oppositions on which the unifying understanding depends" and with the “elements” it "represses"; “the result is not a new unified reading or an alternative unity,” but a “limit perpetually transgressed" by “the significations of the text” (ON 256, 260) (cf. Brenkman, 1976).

        Culler scrutinizes Brooks” analysis of Donne's Canonization poem and picks these lines:

 Our legend […]  I will be fit for verse;

And if no peece of Chronicle wee prove,

We'll build in sonnets pretty roomes;

As well a well-wrought urn becomes

The greatest ashes, as halfe-acre tombes,

And by the hymnes all shall approve

Us Canonized for Love:

And thus invoke us.

Culler adds up “the legend describing the lovers, the verse representation of this legend, the celebratory portrayal of the lovers in the response of those who heard the legend,” and so on (ON 203). He concludes: “we have not so much a self-contained urn as a chain of discourses,” a “series of self-representations, invocations, and readings” “within the poem and outside it," coming to “no end” (CD 203, 205).

        This tendency to interpret a work as both self-referential and self-disseminating is standard among deconstructionists and their heirs; de Man, Bloom, and Hartman all do it with dizzying virtuosity (Chs. 14-16). Indeed, this result is attained in so many deconstructive readings that it becomes the new unifying conception within whose purview a critic can discover “the same thing” or a “comparable logic” or "the same structure" in the most diverse pairs of authors, such as Marx and Hugo, Marx and Kant, or Austin and Lacan (OD 260f). The unity within the text is supplanted by a unity between texts -- the repressed converse of those procedures, "put in question" by Culler, whereby “a difference within” gets “transformed” “into a difference between” (ON 215; cf. ON 68, 133).

        The dissolution of the subject, about which we will hear more (Chs. 14, 16, 17), rearranges the constellation of author, text, and reader. We already noticed the structuralist proclivity for “rejecting" “the notion of subject" and treating “language and structures, rather than authorial self or consciousness,” as “the major source of explanation” (PO 28; OD 21). In this spirit, Lévi-Strauss (1964: 20) declaimed that "'myths think in men, unbeknownst to them” ; and Foucault (1966: 15) ordained that "'man is a fold in our knowledge who will disappear in his present form" (PO 50, 28, 231; PS 32; OD 223), where one has to wonder who “our” can refer to (apparently not to women). Barthes first looked into “the obsessions of the writer as subject,” but soon was “no longer willing to make the individual subject the source of the structures he discovers in the works" (PO 98).

        Similarly, "deconstruction" “concentrates on conceptual and figural implications rather than on authorial intentions" (OD 110). “A totalising notion of the self” is “irreparably subverted by aspects of language," such as “citation and allusion, whose interpretation can never be limited by an authorial project” (PS 166). “Intention” is “not something prior to the text that determines its meaning," but “always a textual construct,” "an artifice of reading," and “a way of dramatizing” one's own “claim about the subject's relation to language and textuality” (OD 217f). Therefore, “the attempt to reconstruct an author's intentions is only a particular, highly restricted case of rewriting" (OD 38)-a special imitative performance. “

        To divert attention from the author as source and the work as object," deconstruction  “brought to the fore" (PO 131). "the concepts of “lecture" and “écriture”. 23 [Écriture” is not well translated as “writing,” since it is much broader, subsuming the “set of institutional conventions within which the activity of writing can take place" (PO 134). It is in this broad sense that Derrida includes speech inside writing (cf. OD 9Off).} Just as "écriture"“ must be "'grasped"“ via "the institution (literature)"“ rather than "'the idiolect of the author"“ (Barthes 1971: 8), “intertextuality comes to take the place" of “intersubjectivity" (PO 135, 139). The thesis that “language" is "the privileged, exemplary case" to be viewed as “a system of rules” “which escape the subject” fits generative linguistics (cf. PO 28f).. “The notion of competence does not lead” "to a reinstatement of the individual subject as the source of meaning”; "the subject” here is “an abstract and interpersonal construct" “constituted by a series of conventions" (PO 2 5 8; cf. OD 111). “Emphasis falls on the reader” not as a person," but as “a function," “a place where codes" “are inscribed” (OD 33).

        A striking duplicity among structuralists is that their theorizing is countered by their own conspicuous, flamboyant practices as writers. They proclaim the “decentering" of “the subject" (Foucault, 1969: 22) (PO 29) while parading their own subjectivity; they disperse the “self” while self-centeredly forcing their audience not merely to contend with their imperious style, but to grant them license to disdain usual evidential or logical procedures. Even the sympathetic Culler admits their "self-indulgent love of paradox” and “bizarre interpretations” (OD 21). This duplicity passed on into post-structuralism, yet there it was no longer an anomaly, but just one more self-deconstructing technique for undermining on one level what is asserted on another, and for mismatching the performative with the constative.

        The thesis that the “self can no longer be identified with consciousness” (PO 28) has a Freudian cast. As befits his non-empirical stance, Culler remarks that “students of theory" might "read Freud without enquiring whether later psychological research may have disputed his formulations” (OD 9) (which it certainly did; see Fisher and Greenberg 1977). Like Frye, Fiedler, and Holland, Culler is prepared to assume that "the logic of dreams and fantasies proves central to an account of the forces at work in all our experience” (OD 160). Yet unlike those critics, he doesn't want to psychoanalyse the author. Any .comprehensive” "intention” of an author would be “divided” between “conscious and unconscious," and “the line between” them is “highly variable, impossible to identify, and supremely uninteresting” (OD 127; PO 118).

        Instead, Culler “claims” that “literature can illuminate and situate the problems addressed" in “psychoanalysis” by sharpening our “awareness of rhetorical structures and forces” (PS 226) -- a project of the Yale group for different reasons. He recommends an “investigation of Freud's writings as simultaneously an analysis of tropes and a tropological construct” (PS 217). “Freudian theory makes narrative the preferred mode of explanation” by “reconstructing a story" with a “decisive” “primal event" that is found to be “in fact a trope" (PS 178ff). "The motive turns out to be a motivation of signs" (OD 191). Freud admitted that "the figurative language peculiar to psychoanalysis" is needed before we can “become aware" of, let alone “describe, the processes in question” (OD 266). “The determining event in a neurosis never occurs," but is “constructed afterwards" by a “textual mechanism of the unconscious,” like a “reproduction without an original" (OD 163f). ²⁴ [24. The “Wolfmna” (in real life, Sergei Konstantinovich Pankejev), who intrigues other critics, e.g., Holland (PIP 157; 5RR 257) and Bleich (SC 8Off), is exhumed here as well (OD 163, 190f; PS 179). Compare Note 6 to Ch. 19. On the whole case see Freud (1973).] For instance, the "Oedipus complex" is “a structure of signification,” a “product of discursive forces" (PS 175). Oedipus decides his guilt on the force of meaning, the interweaving of prophesies, and the demands of narrative coherence” (PS 174). This procedure whereby the cause is seen as an interpretation elicited by the effect is repeatedly adduced by Culler as a Derridean reversal.²⁵ [25. Culler's demonstration with “the phenomenal order, pain…pin” “producing the causal sequence, “pin…pain” (OD 86) (a “mosquito” does the deed in PS 183) is not well chosen, the causality being far more elementary than most perceptions of the world. Fletcher (1984: 52, 55) suggests that in Culler's syllogism, “philosophical concepts of experience are asserted, and then denied by philosophical concepts of rhetoric,” thereby making “rhetoric function as a transcendental signified.”

        Freud's own “texts” have “considerable deconstructive” “force" and provide an excellent example" of how “an apparently specialized or perverse investigation may transform a whole domain" (OD 159f; cf. OD 169). Freud “deconstructs" the “hierarchical oppositions” privileging the “normal” over the “pathological," the “real” over the “imaginary," the “conscious” over the “unconscious,” and "life” over “death,” and “shows that in fact each first term" is “a special case of the second” (OD 160). Moreover, “Freud invokes a complex writing apparatus” “to represent the paradoxical situation in which memories become inscribed” “in the unconscious without ever having been perceived. " Derrida hails Freud's "'formidable"“ "theme" that "the present is not primal but rather reconstituted" (OD 164).

        In the main, Freudian theorizing serves here to construct speculative versions of communicative activity, such as associating “metaphor" with “the father" and “metonymy" with the “maternal” (OD 60).²⁶ [26 Fiedler breaks literary creation down like this: “the personal element” is “the Son, the conscious-communal the Father, and the unconscious-communal the Mother” (NT 321).] Or, the "parallelisms and repetitions commonly at work in literary compositions” are not accounted for though “unity of meaning," formerly one of Culler's guiding interpretive principles but now seen in the “interest” of “phallogocentrism” (OD 61); instead, the new account invokes the "repetition compulsion,” a “powerful mobile psychic force” (OD 261f). This idea ties in with “the transferential structure of reading, as deconstruction has come to analyze it," which "involves a compulsion to repeat independent of the psychology of individual critics, based on a curious complicity of reading and writing" (OD 272). At most, an “interpreter" might “control" and "master the effects of repetition by casting them into a story” (OD 264).

       This usage of Freudian theory is contrasted with the practice of a critic like Frederick Crews that "makes psychoanalysis a source of themes” to use for “interpretation” -- here too, in Culler's opinion, “restricting the impact of potentially valuable theoretical developments” emerging (as usual) among “French" scholars (PS 9f). Early Holland would presumably be open to the same complaint, and Bleich to a lesser degree. Hillis Miller (1976: 335F), in contrast  “does abstract Freud's ideas into a theoretical model by opposing "canny" criticism, with its "promise of a rational ordering of literary study" through "solid advances in scientific knowledge about language," against "uncanny"'criticism Whose "deepest penetration into the actual nature" of "language"“ occurs "when logic fails" (OD 23). ²⁷ [27. Freud (1919) expounded "the uncanny"“ as the "class of the frightening that leads back to the familiar" (CD 24; cf. OD 262ff). Such a definition warns us that “uncanny” criticism, however bizarre its procedures, is apt to produce familiar results. Still, Miller often goes the other way, starting with a familiar theme and pushing it toward the frightening.] Culler feels “dubious" about this “division" between the “two camps" (OD 24f), perhaps troubled by the memory of his own progress from the one to the other. And as we saw, he deplores making deconstruction into one more tool for interpretation (p. 240f).

        A more disturbing outlook on Freud arises when Culler's exposé of deconstruction turns to feminism as a special hierarchical inversion.²⁸ [28. Deconstruction is episodically enlisted for feminist purposes, such as postulating “the differential meaning in the single term “woman," or noting that, when “Freud posits for the woman an original bisexuality,” he undercuts his own concept of penis envy by making the woman “the general model of sexuality” and enabling a “rhetorical reversibility of masculine and feminine” (CD 64, 171) –[a point also made by Millett (SX 27Offi. For a far more elaborate merger of deconstru