Revisiting the deconstruction of narratology: master tropes of narrative embedding and symmetry

Style, Spring, 2001 by James J. Paxson

When William Nelles begins his 1992 article "Stories Within Stories" with a string of synonyms customary for his title's infamous effect, ones including "Chinese box" and "Russian doll," he imaginatively champions the effect of narrative framing or embedding. "[S]o widespread among the narrative literature of all cultures and periods as to approach universality," the effect may be, he writes, the greatest "undeveloped resource in literary theory" ("Stories" 79). Although the culmination of Nelles's research on the subject, his 1997 book Frameworks: Narrative Levels and Embedded Narrative, has not retreated from this position of promise or elation, it more nervously proceeds by acknowledging embedding's identity in formal terms; indeed, "no two students of the genre appear to define it in quite the same way" (Frameworks 1). The championed or promising feature I further celebrate, however, is not just embedding in and of itself. I start with excessive embedding, abused embedding, vertiginous embedding for my p articular approach to this variously defined effect in need, as Nelles puts it, of a "comprehensive history." Tzvetan Todorov bestows the multiple-embedding prize conjointly to the Arabian Nights and to The Saragossa Manuscript (71). According to Todorov, however, both of these award-winners suffer diminishment, for he concludes that in each narrative "all the degrees, except for the first, are closely linked and incomprehensible if isolated from one another." In contrast to Todorov, I want to look into those shifts among layered degrees that are quite comprehensible and are predicated precisely on the abusic of embedding. A medievalist by trade, I in fact have to agree with Nelles (Frameworks 2) that Chaucer's Canterbury Tales wears the crown, at least for American literary scholars, as monarch among the most flamboyant and complex multi-degreed embedded narratives.

In this essay, I will ultimately turn to another excessively embedded text, Johannes Kepler's Somnium not because I want to shift the grounds of record-making in narrative embedding while remaining devoted to medieval documents or documents in medievalism. Rather, by looking at a text that stands on the cusp of medieval fantasy and modern science, I may better explain narratology's attention to this effect (favored particularly since the labors of Todorov and Gerard Genette) through an expanded methodological and rhetorical picture. Narrative embedding holds a special allure for theorists and critics, and this allure seems bound up with the legacy of narratology's structuralism--or with structuralism's inherent scientism--and with its variably tacit or implicit reliance on the repertoire of historically sanctioned rhetorical tropes and figures. Embedding evokes mystery, the mystery of hermeneusis (as Nelles would come to conclude; see Frameworks 140-41), while it conjures images of geometrical containment an d even Pythagorean mysticism. While it indeed invokes such aesthetic or psychic arcana, however, it also calls upon the durable elements of eloquence and persuasion, the fundamental bases or units of rhetoric.

First, I locate the neo-structural project of narratology's fascination with embedding more fully in deconstructive critical understanding. This move will be in response to some of Nelles's conclusions about method in Frameworks. While the still-productive structuralism of modern narratology should not be thought of as having been supplanted by deconstruction's corrosive, self-consuming logic or its own anti-scientism, "deconstruction" itself militated against successivist models in philosophical or analytical method. Thus, I generally agree with Nelles's sentiment that deconstruction may "come to be viewed with hindsight as a transitional phase of structuralism rather than as its successor" (Frameworks 162). In taking up the deconstructive legacy of structural narratology and embedding, however, I respond equally to Jeffrey Williams's conclusions in Theory and the Novel, a study that seeks to put paid to narrative embedding, particularly concerning the effect's ultimate deconstructive heritage (i. e., in th e wake of, say, J. Hillis Miller's poststructuralist findings on narrative embedding, multiple focalization, and the anti-linear consciousness seemingly represented in such complex frame novels as Wuthering Heights). The late 1990s seem to have been propitious to the theorizing of narrative embedding. Although Nelles's and Williams's books, published a year apart, provide the fullest discussions to date of narrative embedding, indeed of excessive narrative embedding, their neostructuralist or empirical and formalist models of the effect shorten the needed push into deconstructive terrain and inhibit our sense of the transhistorical linguistic and cognitive elements of embedding.

Second, because these elements are the tropes and figures of rhetoric, I shall also follow Nelles's linking of embedding's structural dynamics to other tropes--to, in the special case of my project, antimetabole and chiasmus. In "Stories Within Stories," Nelles had productively underwritten the structural functionality of embedding via Genette's version of the trope "metalepsis" ("Stories" 92-95; and see Narrative Discourse, hereafter ND, 234-37)-the narratorial confusing of narrative levels or the actual movement of a character from one narrative level into another. But in Frameworks, Nelles upholds not only the flamboyant powers of metalepsis (152-57); he insists that it stands as the signal trope of embedding. He more deeply rethinks the structural processes of embedding by turning to other chief concepts in narratology. Implied author, general narrator, focalization, and narrative temporality, topics comprising the sequence of his actual chapter titles, all serve as prolegomena to a final chapter on embe dding itself. (And this final chapter he rebuilds from his unsurpassed essay, "Stories Within Stories.") These chief concepts he distinguishes by describing their modes of narratoriality in terms of metaphoricity (90, 116), and, addressing one of narratology's longest-lived and most vexing issues, he subordinates actual character discourse in general "textual production" to a function of the foundational "general narrator, who speaks in the character's person using a trope similar to the classical rhetorical device of dialogismus" (121). That narrative focalization should be recalibrated according to the mastering functions of the tropes metaphor and dialogism seems an elegant solution to an old problem. Moreover, it should be recognized that the cognitive primacy of these classical rhetorical tropes enables this solution. This surely incurs a debt in narratological analysis subsequent to Nelles.