Greimas, Bremond, and the Miller's Tale
Mosher, Harold F., Jr., Style
In the late twentieth century, the application of modern and contemporary critical approaches to literature has become widespread especially in regard to modern literature. David Lodge, for example, has used Roman Jakobson's distinction between metaphor and metonymy and the Formalists' concept of defamiliarization to explain the relations between modernism, antimodernism, and postmodernism (3-16, 74). The James Joyce Quarterly has devoted part of a number to the results and critique of a graduate-class exercise on the Seymour Chatman-Roland Barthes analysis of James Joyce's "Eveline" (Chatman, "New Ways"; Sosnoski), and Jennie Skerl has also resorted to the same story to test Vladimir Propp's model for Russian fairy tales. Wallace Martin has provided a convenient bibliography of such studies on a number of nineteenth- and twentiethcentury writers of narrative (189-90n5). Since 1990 Style's annual bibliographical issue (number 4) has published William Baker and Kenneth Womack's annotated listings of literary theory and criticism as well as specialized studies by David Gorman on Russian Formalism (26.4), Bakhtin and his circle (27.4; 28.4; 30.4), Jonathan Culler (29.4), and Gerard Genette (30.4). The theories of the French semiotician A. J. Greimas have most frequently served critics of nineteenth- and twentieth-century literature: Henri Mitterand on the nineteenth-century French novel (203-12), Susan Suleiman on the bildungsroman and the thesis novel (65; 273n39; 279n25), Lodge on Hemingway (32), John N. Duvall on Faulkner, and perhaps most persistently Fredric Jameson on Balzac, Gissing, and Conrad. Jameson's interpretation of the latter's Lord Jim and Nostromo by means of Greimas's semiotic square is exemplary (277). Despite this tendency to concentrate on writers of the last two centuries, some semioticians have devoted their attentions to ancient and medieval writers, including Ovid, Petronius, Boccaccio, and Chaucer (Martin and Conrad; Allen),' and a few have applied Greimas to medieval narrative, notably the Arthurian corpus (Maddox; Collins).
Not all of this practical criticism is successful, however, and the failure may not, at least in some cases, be ascribed to the critics. Not many theoreticians have hastened to adopt the Greimasian method, which, in the words of Robert Scholes, may be "more interesting than satisfying" (Structuralism 107) because it is incomplete and confusing (103, 106). In his second book on structuralism, Semiotics and Interpretation, Scholes avoids Greimas although he attempts to apply Tzvetan Todorov, Genette, and Roland Barthes's codes again to Joyce's "Eveline" (89-104). And in his later, pedagogical book, Textual Power, the structuralist model is veiled, appearing, for instance, in the method of examining oppositions in the manner of Levi-Strauss (32). This suggests that Greimas's semiotic square, based on contraries and contradictions, is simply redundant since most narratives proceed on these bases (Segre 51). Generally, though, the opposition to Greimas and to Claude Bremond, another French semiotician, has complained of their mistaken claim to create a system valid for all narrative when the model has actually been derived inductively from a limited corpus of fairy tales (Kloepfer 117n 1) or cultural cliches (Ricoeur 264), or opponents have criticized these systems' reductiveness, inapplicable abstraction (Culler 232, 235; Segre 27, 34; Riffaterre 31), or their failure to take into account the temporal aspect of plot (Brooks xiii).
In the analysis that follows, I shall try to reply to these criticisms (though occasionally acknowledging their justification) by demonstrating how Greimas and especially Bremond might successfully be applied to a narrative. What we shall discover, I believe, is that Bremond's system does account for the forward movement of action, though perhaps not in the way Peter Brooks envisages it-psychologically-and although both Greimas's and Bremond's taxonomies abstract character and action, these very abstractions may reveal similarities and differences not immediately apparent at the so-called surface level of plot. …