Copyright © 1986 by Michael Macrone.

Naturalizing Culture

In the so-called “human sciences,” the objects of perception are the phenomena of life (man in his biological essence), labor (man in his social essence), and language (man in his cultural essence). But there are no eternally constant objects corresponding to the words “life,” “labor,” and “language.” What these terms meant in the different epochs of history of consciousness from the sixteenth to the twentieth century changes constantly, and changes moreover in conformity to trans­for­mations that occur on a metalinguistic level of apperception, a level on which different modes of discourse generate different categories for the constitution of the elements and relationships presumed to inhabit the “human” world.

— Hayden White[1]

The status of literary studies, within an institution and within our culture, rests on its claim to certain, stable, delimited objects of investigation and investment. The fields of textual study not only claim such objects, but are constituted by them, are indebted to their cultural and ontological privileges.

How do our objects distinguish themselves, hold themselves together and announce their status, their being? Are they scientific objects, cultural objects, unified objects, mystical objects, conscious objects; are they alive? What type of typology governs the project of literary criticism?

The answers depend on place and time. Answers that once had satisfied changed when the objects of “natural science” and of “the science of man” were split into opposing types. Answers changed again in the age of relativity, structuralism, the “new physics” and the interdisciplinary ideal. We may celebrate a New Age of the reinvestment of phenomena with multiple dimensions.

The natural and human sciences are said to be discovering objects in each other’s fields, are learning to speak each other’s languages. And to the extent that disciplinary boundaries are eroding, there must be a repositioning of the objects of knowlege. Is the scientific object becoming more humanistic, or is the humanistic object becoming more scientific, or both? Has the story of nature versus culture come to an end? Have we agreed to the postulate that all objects are the product of culture?

At least, this is half the story. At the moment that the natural sciences were developing a relativistic notion of the epistemological object as a product of epistemological practice,[2] certain sciences of man were committing their pursuits to a notion of cultural phenomena as subject to a “scientific” epistemology.

The latter development, the naturalization of culture, is what interests me here. I shall take the work of Michael Riffaterre as a test case for the classification of cultural artifacts—literary “monuments”—as both empirically coherent and rigorously classifiable. Most of what follows is a pretext for an analysis of Riffaterre’s reading of William Blake’s “The Sick Rose,” a reading he presents as itself a test case for his general claims.

I : The Frame of Reference.

It has become a central concern of recent literary criticism, under pressure from “post-structuralist” problematics, to examine and question the ways textual boundaries are constituted, the way texts are conceived as objects of study.[3]

At first glance, defining the limits of a text seems to pose little difficulty. To take a simple example, one could say that a conventional sonnet comprises fourteen verses of iambic pentameter, and further that these verses rhyme within a certain scheme and deploy a certain kind of thesis along structured lines of development. In short, the text is bounded and motivated by its well-defined form; this form implies a dynamic structure and thereby conditions a certain kind of interpretation.

One may question interpretation of a given sonnet, but the integral assumptions of what the text is are usually taken as “given.”

There is a wrinkle, however, in the event of textual variants. In adjudicating the intention of the author, the critic faced with variants must consider what in practice is a multiple text, a text which is actually several texts, whether or not one of these texts has been or may be chosen as “the” text.

Consider next a sonnet (with or without variants) that is an element of a sonnet sequence. The critic usually considers the poem’s position within the sequence as somehow significant. As an element in a chain, the sonnet may signify unto itself yet cannot help reference to its context, the sequence. It is only the text that it is by virtue of its diacritical position in a larger chain, a chain which becomes the smallest unit of essential meaning. The textual boundary is extensional.

It is my claim that there is no ideal situation in which any constitution of the text is not subject to such dissolutions and extentions. One is always confronted with the problem of where a text begins and where it ends. And every practicing critic must in some way address—even if to dismiss—a nervousness about her or his assumptions of textual constitution.

I wish to keep these questions in mind: Given a reading, what sort of framework is presented as adequate to an enclosure and definition of the text, and how is that framework grounded outside of the boundaries thereby constituted for the text? Assuming such an object of interpretation, what is the effect of the mode of observation upon the object observed? Is it possible to secure a frame around the object, to define it, to “find” it, so that the procedures and authority to which the interpretation appeals do not themselves colonize the object?

Textual definition is, after all, an act of positing an inside and an outside, intrinsic and extrinsic elements, a self and an other, of the text. As we shall see, in telling the story of the victory of the self over the other, criticism always brings defeat to the victor. The other has conquered the self before the heroic narrative begins.

I shall call a “frame of reference” the criteria, the epistemology, put into effect to limit what is relevant to the reading of a text. It is a functional boundary constructed to introject (as “intrinsic”) and project (as “extrinsic”) data which might bear on reading—for example, biographical events, available topoi, generic convention, putative allusion, ambiguity, textual variants, and so on. The reader must judge what is “there in the text” and in so doing specify how the relevant or intrinsic materials at hand are organized in significant relations.

My intent in this essay is to highlight the problems of textual constitution and of the theoretical “scientificity” of this project. The reason Riffaterre’s analysis of “The Sick Rose” is particularly apt is that the poem has been typecast as a particu­larly difficult one to contain, to cordon off and interpret as a “thing itself.” My ultimate claim is that Riffaterre, who takes as axiomatic the absolute coherence, certifiability and autonomy of this and every poetic text, replicates a trope in the rhetoric of referentiality that he strenuously criticizes. In setting his frame around poetic texts, a frame defining what type of objects poems are, Riffaterre shows that his own frame cannot hold, and that whether or not a poem is a purely self-sufficient object, his criticism cannot help allowing what the poem is not to contaminate that which it is.

II : Contours. The Reader Referred to Appendices.

Criticism of “The Sick Rose”[4] is notoriously thin. There are two predominant reasons for this:

(1) Although seen as essential to a full constitution of the “Blake canon,” the lyric poems attributed to Blake are characterized as mere “rough sketches” or “warm ups” for the truly “significant” Prophetic Books. This “prophetocentrism,” which marks the lyrics as parasitic upon a serious, fully realized and present oeuvre dominated by the prophe­cies, allows the critic to discount the lyrics in the name of defining a symbolic system that may be given the coherent name “Blake.”

(2) “The Sick Rose” is seen to “resist criticism,” and has readily been appropriated from the context of Blake studies into the context of “critical theory.” The poem, thus thought to defy the standard practices and procedures of interpretation, becomes a circulating exemplum of a literary object that announces its ontology outside of interpretation. Negated, in one sense (frame), as an object of criticism, it constitutes itself in another sense (frame) as an object that sustains itself beyond framing. Thus it becomes a sort of bargaining chip in the large-scale transactions of literary studies; it is the centerpiece of several large claims for the ontology of the poetic object.

I must here refer the reader to appendix A, in which I supply a digest of prophetocentric readings. My demonstration there is sketchy yet at the same time tedious enough to be exiled from the body proper of this essay. I assume that the reader is, like me, anxious to move on to business. Yet it is still necessary for me to present here a digest of the digest, so that the contours of investigation, the tropes of my and other criticisms, are clear.

Digest of Appendix A

Within Blake studies, “The Sick Rose” has become the occasion for a return of the repressed. In reading the poem, critics have revealed the axiomatic assumptions that condition all claims to a putative “coherent symbolic system” said to code both the writing and the reading of all Blake texts. These repressed axioms are both the basis and the negation of an allegedly self-adequate frame of reference. They mark the frame at every point of application, yet usually in a less obvious manner than in the case of “The Sick Rose.”

S. FOSTER DAMON was the first critic to claim that Blake encoded his writing according to a fixed, closed and private symbolic system. A READING SHOULD decode the text from within. In the case of “The Sick Rose,” HIS READING ACTUALLY grounds itself in a public arena of literary convention.

NORTHROP FRYE claimed that all of Blake’s poems resolve into “tonalities” which are regions or topoi of archetypal formulae. A READING SHOULD reconstitute the formulae by which Blake has encoded the text. In the case of “The Sick Rose,” HIS READING ACTUALLY claims that the poem escapes these tonalities and in fact all conceptual thought; it “speak[s] with the unanswerable authority of poetry itself.”

HAROLD BLOOM adopted Frye’s notion that the Songs frame themselves as “satires” of each other’s narrative limitations. A READING SHOULD recognize each poem in the sequence both as a critique of narrative assumptions elsewhere and as itself an inadequate narrative. In the case of “The Sick Rose,” HIS READING ACTUALLY presents the poem as a direct and self-adequate moral allegory.

ROBERT F. GLECKNER, the best of the readers of the lyrics, accepts the frame which locates each poem within a private systematization of archetypal referents. But he claims that the system developed organically and was never fixed, let alone at the time Blake wrote the Songs. A READING SHOULD trace the cumulative development of the system and refer each text to the texts that precede it. HIS READING ACTUALLY retrospectively fixes the system before it had any chance to constitute itself organically.

The point of these digests is that, among those interpreters who have expended much energy on Blake’s poetry, the problem of reference is seen as a problem of context. The residence of signification is a total context grounded in a systematic construction of “Blake’s symbolism.” This system is seen to be exemplified and fully realized in the Prophetic Books, and the task of criticism is seen to be a recuperation of the field of poetic objects into this total context. I have called this epistemology prophetocentric, and it operates to at once constitute and negate its objects, to comprehend by sublating them in a prophecy-centered referential network.

As readings of “The Sick Rose” demonstrate, this project reveals itself as incoherent at the moment that its methodical, soi-disant “closed” interpretive systems admit open-ended axioms that emerge as self-evident assumptions, as corroboration of a systematic interpretation, or as explicit ruptures in the framework at hand.

In a larger context, “The Sick Rose” has been validated precisely for its rupturing quality, what is taken to be its “resistance” to stable referentiality. The poem is used as an exemplum not of Blake’s general intention or method, but of an even more general epistemological theory. Taken as a poem resolutely “meaningless” in Housman’s sense (see appendix A, note 21), “The Sick Rose” is employed as an icon for meditation on the failure or inherent partiality of interpretation. In thwarting practice, the poem becomes, by catachresis (referential emptiness), the basis of theory. For Riffaterre, “The Sick Rose” illustrates powerfully what is merely the case with every literary object: what the text says is that efforts to make it referential are doomed, and that it is the object it is by virtue of its radical self-sufficiency.

Interpreters who have celebrated the poem’s resistance to referentiality (and thus to interpretation) have usually not done so purely as an exercise in self-negation. Such a celebration inevitably invokes claims for the “independent life” of the text, and therefore for its ontological privelege and for its “sublimity” (in the traditional master/slave dynamic of the sublime). A recurrent, if latent, theme is that “symbolic” poetry, as exemplified by “The Sick Rose,” transcends the nature/culture and object/representation polarities. I have provided a digest of a few notable examples of such readings in appendix B.

It is these sorts of romanticism that Michael Riffaterre claims to defuse at the moment of maintaining his own claims for the radical self-referentiality of “The Sick Rose,” and indeed of all literature. Riffaterre, and a literary critic with more concern for his or her own scientificity cannot be imagined, claims first of all that a literary text is a precisely defined object, and is such because of its formal self-sufficiency and freedom from referentiality. Riffaterre invokes the epistemology of the “natural sciences” by naturalizing literary texts. That the object of interpretation be well-defined ontologically is precisely the first requirement of any claim to a scientific discourse; in Riffaterre’s work, claims for the specificity of objects and their epistemological field are followed immediately by other axioms, definitions and rules, produced with the most “scientific” rhetoric of incontro­verti­bility.

III : Rifaterre’s Large Claims.

Quot ego tuās petītiōnēs, ita conjectās ut
vītārī nōn posse vidērentur, parvā quādam
dēclīnātiōne et, ut āiunt, corpore effugī!

Riffaterre’s reading of “The Sick Rose,” “The Self-sufficient Text,”[5] presents itself as a test case of his general theory. He begins: “My purpose is to analyze a poem using internal evidence only and to determine to what extent the literary text is self-sufficient.” He continues:

It seems to me that a proper reading entails no more than a knowledge of the language.… [In modern criticism, reference] to things has been replaced as a semantic analysis criterion by the notion of reference to the corpus of topoi, themes and motifs. Which is a devious way of relapsing into the referential fallacy. The assumption seems to be that their [sic] symbolism cannot be understood without understanding them first as metaphors for man. (p. 39)

For Riffaterre, former formalisms (at least the structuralist and archetypal versions) never went far enough; in fact, they never went anywhere, merely replicating the pre-formalist referential fallacy—replicating, that is, a “lapse.”

Riffaterre’s quarrel with “humanistic” structuralism has a long history, dating back at least to his attack on Roman Jakobson and Claude Lévi-Strauss, “Describing Poetic Structures: Two Approaches to Baudelaire’s les Chats” (1966).[6] But where that essay presented itself as a correction of psychologistic tendencies in structuralism (Riffaterre did endorse “comparative structuralism”), later essays, including “The Self-sufficient Text,” have taken on the tone of repudiation. I will soon claim, however, that the more Riffaterre denies structuralist methods, the more he is forced to replicate them.

I should qualify my position by saying that Riffaterre is not oblivious to a certain dissonance in his method, a method nonetheless celebrated for its pedagogical effectiveness, if not for its theoretical coherence.[7] And, on the other hand, I hardly mean to imply that there is or should be a unitary “Riffaterre” who, were all well with his claims and were they to neatly line up, would be in control of himself, his theory and his texts. To provide a fairer basis for appraising “The Self-sufficient Text,” I shall first present a brief digest of Riffaterre’s positions as stated elsewhere. I shall not, however, be able to follow subtle evolutions and adaptations, slight swerves in his position. I will leave gaps in tracking the gaps in a theory that is fragmented and self-devisive, but no more so than most. To ask for a static and unitary theory would only be to pretend that I could somewhere find it, that there is such a unitary text, even in the ideal.

In the seminal piece, “Explaining Literary Phenomena,”[8] Riffaterre sets forth his project against interpretation:

Since my approach will be a formal one, it would perhaps be of some use to remember that textual analysis has nothing to do with traditional normative stylistics or rhetoric.… It must also be emphasized that analysis does not attempt to form value judgements; it is not literary criticism. It records the existence of something which is then, and only then, appropriated by criticism .…

[Linguists] hope … to find a place for literary expressions within the framework of a general theory of signs. Because it is a generalization, this kind of investigation, which belongs to the realm of poetics, cannot account for one particular characteristic of the literary verbal message: its textual nature. This kind of message is known only through texts, which is to say, monuments .…

The text is always one of a kind, unique. And it seems to me that this uniqueness is the simplest definition of literariness that we can find.… The text works like a computer program designed to make us experience the unique. The uniqueness is what we call style. It has long been confused with the hypothetical individual termed the author; but, in point of fact, style is the text itself. (TP, pp. 1–2).

As we can see from this excerpt, Riffaterre proposes his own project as ontological or typological rather than evaluative. He initially makes very limited claims for the epistemological value of this project, except to set himself against a reigning practice of intepretation: “Explication of texts is really a machine for taming a work, for defusing it by reducing it to habits, to the reigning ideology, to familiar mythology, to something reassuring” (pp. 2–3).

Furthermore, the literary object is well defined, coherent, unique, and monumental, and may be identified with style. [These are axioms, and classic cases of petitio principii.] Yet style, thus conceived, is not something that a text may claim for itself within its own monumentality; style is a function of reading. “[The] problems [explicators have had with ambiguity and obscurity] all point to the role of one particular property of stylistic phenomena: we must be able to perceive them” (p.10). A recognition of the phenomenological character of style, and therefore of the textual object, necessitates a rejection of structuralist claims for “latency of structures, “ for an underlying organization within which polysemy could be reduced to univocality (p. 10).

At first glance, these claims also seem to imply that the notion of “uniqueness” is equivalent to a rejection of the reducibility of the text to any paraphrase or coherent structuration. That is, Riffaterre seems to claim that what defines the object is the full and repeated contact of its irreducible surface with a general reader, rather than a “hidden” meaning that may only be educed by the trained critic. There are two problems with this account:

(A) Who is the general reader? How can the analyst sift out, as the “literary phenomenon,” “the reader and all of the reader’s possible reactions to the text” (p. 3; Riffaterre’s emphasis)? What does this reader know, how does s/he perceive and react, how are we to define the field of perception that is the reader in general?

(B) Riffaterre banishes “latent structures” to the end of introducing what he calls the “matrix,” “hypograms” and “models” which generate the text. These objects, to simplify, function as intertexts of which the text itself is an expansion and a distortion, or, in the case of “models,” as rhetorical templates by which such expansion and distortion occur. What distinguishes these “invariant” objects, of which the text is an encoded series of variants, from “latent structures” is that their variants must “necessarily be perceived” as variants; “The decoding of the text is controlled so that the reader perceives a given segment of the verbal sequence not only as stylistically marked, but also a variant; it is perceived as the corollary of something else” (p. 11). Yet, in defining the text as a semiotic field of affective stylistics, as a phenomenological presence (“necessarily ... perceived”), Riffaterre reintroduces structural absence:

A variant is encoded in such a way that, first, it reveals that it is hiding something and, second, it indicates how we can find that something. In this sense, and only in this sense [?], can we speak of latency....

The first of these features, the one that points to the structural function of a stylistic fact … is the impossibility of decoding without referring to some external element (external, that is, to a word’s meaning, to a group’s stylistic effect, to the special code, and so on). In short, it is perceived as a distortion of the mimesis.

The second of these features, the one that shows the reader where to look for the structural relationship and how to interpret it … seems to lie in the nature of the mimesis distortion …. (p. 12)

What immediately emerges is a tension between (a) claims for a self-sufficient, unitary object that comes into being as a reading, and (b) an analysis that appeals to a stylistically grounded “hidden something.”

As de Man pointed out, the tension between these claims for semiotic phenomena on the one hand, and stylistic phenomena on the other, are never resolved.[9] And further, it is hard to imagine how Riffaterre’s “matrix” etc.—all that constitutes the hidden something—is not a “latent structure,” especially since it is in a later work described as functioning by virtue of its repression.[10] In Semiotics of Poetry (1978),[11] Riffaterre re-presents his claim that the text points us to its matrix and hypograms by its ungrammaticality, that is, its self-announcing anti-mimetic self-reference, and that local ungrammaticalities collate into collective signs indicating the matrix of the text and its mode of distortion.

The poem’s significance, both as a principle of unity and as the agent of semantic indirection, is produced by the detour the text makes as it runs the gauntlet of mimesis … with the aim of exhausting the paradigm of all possible variations on the matrix.… The text functions something like a neurosis: as the matrix is repressed, the displacement produces variants all through the text, just as suppressed symptoms break out somewhere else in the body. (p. 19; my underline)

The phenomenon is indeed intertextual, since the agent of its effect, the ungrammaticality, cannot be seen, let alone defined, without a comparison between the text and its generator, the hypogram.… [T]he destroyed model [leaves] enough of a trace imprinted upon the text to create a unit of significance. (p. 42)

Riffaterre’s use of the concepts of overdetermination, suppression, repres­sion, symptom and trace, first of all, tie him into a psychoanalytic notion of a dream’s “latent content” (Riffaterre’s “matrix” and “significance” of the poem), which is encoded in a “manifest content” (Riffaterre’s “stylistic features” and “meaning” of the poem). According to Freud, the two

are presented to us like two versions of the same subject-matter in two different languages, or, more properly, the dream-content seems like a transcript of the dream-thoughts into another mode of expression, whose characters and syntactic laws it is our business to discover by comparing the original and the translation. [Cp. Riffaterre’s “comparison between a text and its generator,” above.] The dream-thoughts are immediately comprehensible, as soon as we have learnt them.[12]

Just as a Freud’s manifest content encodes immediately comprehensible and putatively referential latent dream-thoughts, so Riffaterre’s figural text encodes a matrix in literal language, which has been repressed, its hypograms destroyed, yet its traces marking out decoding directions on the surface of the text. At one moment denying the validity of structuralist concepts of latency, Riffaterre imagines the process of text production as a sublimation generated by a structural repression of a latent literal signification.

Riffaterre is certainly at some level aware of the “intertext” his theory finds in psychoanalysis.[13] It can hardly be an accident that his two modes of text generation are the “expansion” and “conversion” of hypograms, equivalent to displacement and condensation in the dream-work and in symptom-formation,[14] and then, ultimately, to structuralism’s versions of metonymy and metaphor, respectively.[15] In telling the story of the sins of structuralism, he finds himself within the structure of the narrative he discounts.

That structure participates in both a structural unconscious and in the retrogressive logic of the trace. As it turns out, the hypograms—intermediate pointers which refer back to a repressed matrix sentence in “literal language” (p. 19)—are often literary clichés, or even quotations from other texts. Ultimately, the poem is then simply a detour through other texts: other literary texts, clichés, ideological codes, lateral texts, hidden texts, objects always already coded. (Such texts are insistent and operate like a memory trace.[16]) And even if the reader is either not as good a reader as Riffaterre, or is unaware of the source of the codes (the intertext), the text never­theless exerts its pressure by evoking the recognition of a semantic gap :

Implicit intertextuality is highly vulnerable to the erosion of time and cultural change, or to the reader’s unfamiliarity with the corpus of the elite that bred a particular poetic generation. But even when the intertext has been obliterated, the text’s hold on the reader is not affected. The fact that he is unable to decipher the hypogram of reference immediately does affect the content of his reactions, but not his perception of the grid of ungrammatical or nonsense phrases. They function as buoys marking the positions of a sunken meaning. If retrieval is blocked … [t]he reader looks elsewhere for a meaning, as well as for the reason why the text is playing tricks with language—that is, he tries to find within the inner system of references of words to words the justification he cannot find in the semantic system of language at large, in meanings based on referents. (p. 136)

These claims ultimately imply: (1) The intertext, if it arises, arises in a gap, in an empty space produced by a frustration of the hermeneutic enterprise; such a lack arises from within an interpretive compulsion, and is structured like an unconscious. (2) The intertext does not even have to exist, does not ever need to be present. Which further implies that, because the intertext is ultimately produced as the response to a lack, the detour through and beyond referential meaning will never produce a “lexeme” which is not stricken with the trace of its posteriority, its lack. The detour is interminable. The intertext, the matrix, the literal lexeme postulated as prior to and distorted by figuration, is thus produced at the intersection of a drive and a trace: its priority is a posterior figuration. Tracing is both the telos and ontos of (the reading of) the text, its endpoint and essence, its culmination, motivation and constitution.

To briefly summarize: According to Riffaterre, the text is a unitary phenomenological object which encodes its own reading; the text-reader relationship is primary, and is the only basis for valid analysis. But in demonstrating the application of this definition of the literary object, Riffaterre invariably produces readings which have escaped nearly every other reader. The reader has to be a reader as good as Riffaterre to perform his version of reading, and thus to properly constitute the object at all. According to Riffaterre, this reading and this textual constitution, which are semiotic functions of stylistic phenomena, escape a reduction of the text to “latent structures” or “hidden meanings.” Yet the matrix and hypogram turn out to be precisely that—hidden meanings and latent structures. To resolve this tension, Riffaterre claims that the function of the poem is to program a search for the matrix—that this search is in fact the significance of reading and therefore of the poem. Yet such a search is undifferentiated from interpretation, and is ultimately grounded in formal structures prior to any reading of the text. And, contrarily, such a search is ultimately the ongoing production of the very structures which are said to motivate the search itself.

IV : “The Self-Sufficient Text.”

Having sketched out the basic problematics of Riffaterre’s theory of the poetic object, based on texts which predate and postdate his reading of “The Sick Rose,” I shall now turn to that reading to observe how those problematics structure its rhetoric. As we have seen, Riffaterre presents “The Self-sufficient Text” as a minor manifesto, as a test-case for the validity of his claims regarding the general typology of the poem. Although he eschews some of his more technical terminology (e.g. “hypogram”) in this essay—originally delivered orally—Riffaterre retains the term “matrix” and of course the greater part of his theoretical framework.

Riffaterre commences his reading with scornful caricatures of allegorical, archetypal, genetic, contextual, anthropomorphic, historical and psychological interpretations of “The Sick Rose”—singling out T.R. Henn, Kathleen Raine and Wolf Mankowitz in his demonstration of the “ambiguity,” “verbal acrobatics” and “silly comments” on which such “humanistic” readings are constructed. Riffaterre finds Henn’s brand of valorizing ambiguity wishy-washy, willful and critically irresponsible.[17] He claims to show not only that the poem has a single meaning, and that this meaning inheres in a purely formal level of language, but further that the first line of the poem—its (unrepressed) matrix (p. 44)—encodes the entire semantic content of “The Sick Rose.” In demonstrating his claim that “a knowledge of the language itself” adequately accounts for the proper frame of reference—in a reading that “must be aimed inward” (p. 40)—Riffaterre must perform certain (epistemological) acrobatics of his own:

I am not suggesting that external models for the text do not in fact exist, and that there is no intertextuality. What I am saying is that external models function, not as literary topoi, but as language stereotypes .… They are … codes, … verbal structures that have no meaning per se, but serve as a lexicon and even as prefabricated syntactic structures for whatever meaning may be demanded by the context. (p. 40)

Riffaterre, in proposing an interplay, in a (semantically determinant) linguistic context, of (semantically void) stereotypes and syntactic sequences, wishes to banish any “extrinsic” meaning that might be imported along with the codes to the text at hand. Meaning, he proposes, is determined by “context” prior to the deployment of semantically blank codes within the text (there is a context for the language before there is the language). In carrying on his construction of meaning, Riffaterre problematizes his own appeal to a simple “knowledge of language itself,” as that knowledge becomes rather the situation within ideological codes, within language as a specific social institution, replete with inherent meanings. For example:

The word flower, is higher on a scale of functional synonyms than any name of fruit. And this, of course, is even truer of rose, since rose is to flowers what lion is to animals [i.e., the monarch]. Flower or rose is therefore, within this function, a hyperbole for “fragility” or “frailty.” (p. 41)

This granted, how is the function of “rose” self-sufficient in “The Sick Rose,” as Riffaterre asserts, if its function as an hyperbole (as an “envelope of appearance containing or masking a ‘reality’”) demands that its “rule of opposition must be verifiable by other hyperboles of beauty and purity of floral codes” (p. 42, I underline)? How is a “topos” of evidence distinguished from “semantic” topoi?

Although the tricky (in)definition of Riffaterre’s concepts of “language,” “knowledge,” “meaning” and “codes” ratifies one’s nervousness about the underlying axioms of his formalism, in isolation from his more outrageous claims, his reading of “The Sick Rose” is nonetheless fairly ingenious. In short: Since the “word, worm is meaningful only in the context of flower,” and vice versa—that because “[p]olarity then is what makes the two words a semantic unit [and] conversely, [because] their common significance cannot be derived from one without the other”—then the “rose must therefore contain a worm,” that is, “beauty [is] the latent destruction of itself, purity the germ of impurity” (p. 41). “Thus interiorization … is the overdeter­mined component of the poem’s structure, the motor of its effectiveness” (p. 42). The implication of this proto-deconstructionist reading is that a proper hermeneutic is an effect of the poem’s central (overdetermined) trope; this claim is far from novel, but Riffaterre’s version is a little slicker than most.

The problem with Riffaterre’s reading is, again, not a lack of ingenuity, but an incapacity to resist extending absolute claims that he cannot abso­lutely honor. It is certainly fair for Riffaterre to reject “thematic readings” and “genetic interpretations”—for example, the claim that worm and rose are Cupid and Psyche in disguise—on the grounds that they import alien narratives to serve as the “underlying story” that clarifies ambiguous details in the primary text (p. 40). But Riffaterre does not stop at the point of saying that unsupported claims for direct allusion are methodologically suspect; he recklessly claims that his reading restores the poem to its proper place in isolation from any prior use of the language and images it comprises. In short, “All we need is the text—references to outside uses of its components can only blur its meaning” (pp. 43–4).

This claim is directly at odds with his insistence elsewhere that reading necessitates a detour through “outside uses” of a text’s language. And even if Riffaterre would have us construe the poem only by citation of a dictionary, he would still put us in a problematic situation—because dictionaries, of course, are collations of paradigmatic uses of language. What does it mean to say that “all we need is the text” or “a knowledge of the language itself”? Riffaterre’s argument relies heavily on the notion of “transformation” or “inversion” of “semantic features” inherent in “codes” or “language stereotypes,” yet how is he to account for what is a “more fundamental” semantic feature (p. 44) without appealing to a notion of semantic “use value”? If the fundamental semantic features (“semes”) of a code are a sort of lowest common denominator of all its uses, how are those features in any way “inherent” in the code? If not, by what authority are some semantic features more fundamental than others? His perfectly acceptable point that words function only within contexts (signs are arbitrary) cedes to his quite arguable point that “semes” have at some point become natural properties of words that are merely made available by context.

Riffaterre does finally allow that the semantic content of a code is determined—and, more often than not, overdetermined—by the literary contexts in which its “components” have become clichés; for example:

[The code “dark secret love”] is made up of two clichés: dark secret, and secret love. Now first of all, the sensed familiarity of the components gives the new phrase a ring of truth,—it seems to be especially apt. Second, dark secret confers on secret the heavy connotations of Gothic novels, novels which are built around precisely this cliché. In the same manner, secret love brings sinfulness to the fore. But this again would appeal only to a contingent interest of the reader, and would involve nothing but narrative structures. (p. 44)

There are a number of questions to be asked of this account—such as: What is the force, or even the meaning, of “In the same manner,” or of “nothing but narrative structures”? Furthermore, if the familiarity of a code is somehow rhetorically essential—in providing a “ring of truth” and mitigating the reader’s resistance—how is the specific weight of its connotations, “heavy” or “light,” only an irrelevant contingency? Riffaterre breaks down familiarity into an essential force (pacification) and an irrelevant force that is undefined but surrounded by words such as “only,” “con­tingent,” and “nothing but.” In so doing, he ultimately splits his “literarily competent” reader into a subject who is both sensitive to allusiveness and invulnerable to the erotics of textual slippage.

It may prove fruitful to look at Riffaterre’s initial account of familiarity and its effects, presented at the awkward moment in his argument when he must acknowledge but protect himself from what he calls “topoi.”

The text simply uses a lexicon and a collection of images, a sequence of representations that have been prearranged, ready-made, and tested for use. In terms of verbal artifact, the assem­bled product has the advantage of being familiar. Whereas an untested sequence might be objected to because of esthetic or ethic preconceptions, the tested sequence gives the reader a feeling of déjà vu. He readily translates this feeling into accep­tability and efficacy of communication, despite his preconceived ideas. The verbal components of the topos retain their prestige even if the topos is no more than a storage bin for these usable prefabricated formulas. In occidental literatures, one cannot speak indifferently of the rose or of the cankerworm …. (p. 40)

Riffaterre seems to be claiming here that the effective force of the “ready-made” represen­tation is primarily delusional: the reader, in a fit of nostalgia, succumbs to the truth-effect of the stereotype. The implications of this delusion are threatening. Comforted by the sense of having already seen, the reader is unable to resist the seeming efficacy of meanings produced by a text as an aggressive act against his “esthetic or ethic preconceptions.” But a new level of probematicity is opened in this account: how are “preconceptions” distinguished from the force of a “stereotype”? Isn’t one sort of blindness being displaced by another? On the other hand, and more crucial to Riffaterre’s argument in general, how does this latently aggressive form of provoking nostalgia empty out the content of that nostalgia? Once again, we’re dealing with the problem of the literary competence presupposed for the evocation of this déjà vu. Such awareness of prior use is certainly something more than a “knowledge of the language itself.”

In defending his reading against the content of the already-seen, Riffaterre is curiously dependent on that content. As he continues to unpack the “meanings” identification of the central trope (interiorization) entails, he appeals more and more to inter­textualities: to the prior-use “semantic components” (the “semes”) of words (e.g., “love,” p. 43), which are maintained or “overturned”; and to literary history and convention (“Shakespeare had already implied [in Twelfth Night] that an important semantic feature of the worm is precisely its ‘hiddenness,’” p. 42).[18] Ultimately, Riffaterre’s “meaning,” or his account of it, is mutually interior with a semantic “topos” that exceeds the “language” of “The Sick Rose” “itself.” Riffaterre offers, in mediocre faith, only this take-it-or-leave-it excuse:[19]

One might object that I have repeatedly quoted from other texts and that I have alluded to traditional verbal conventions. These references beyond the text would seem to present a departure from my own principles. However, their sole purpose was to prove that a given phrase was indeed a stereotype. (p. 45)[20]

In reality, extensive damage has been done. In concluding that “The Sick Rose” evidences “radical” “inversions of semantic relations,” relations organized in a literary “network,” Riffaterre gives away the game. His text may be self-sufficient, but his critique isn’t. There is simply no way to describe self-sufficiency without defining interiority as the negation of exteriority and, in effect, ending up with the same sort of referential account which is ostensibly rejected. Such is the effect of Riffaterre’s numerous claims for the “transformations” and “overturnings” of prior, use-defined semantic contents. As many “humanists” before him, Riffaterre projects an account of the text that claims to introject (resituate) its power within the boundaries of the “text itself.” Yet in his demonstration of a purely hermetic semiosis, he disseminates the text, externalizes it, and grounds his demonstration in significations that bleed back through the boundaries. No demonstration of self-referentiality can escape being framed in the act of disseminating the text.