Semiotics of poetry By Michael Riffaterre (review)

456LANGUAGE, VOLUME 56, NUMBER 2 (1980) for the complex process of early second-language acquisition. Schumann's book fails because its conclusions are drawn without sufficient methodological and conceptual rigor. REFERENCES Bickerton, Derek, and Carol Odo. 1976. Change and variation in Hawaiian English: General phonology and Pidgin syntax. Honolulu: University of Hawaii. Mühlhaüsler, Peter. 1979. Structural expansion and the process of creolization. Paper presented at the Conference on Theoretical Orientations in Creole Studies, St. Thomas, Virgin Islands. Schumann, John. 1974a. The implications of interlanguage, pidginization and creolization for the study of adult second language acquisition. TESOL Quarterly 8.145-52. -----. 1974b. The implications of pidginization and creolization for the study of adult second language acquisition. New frontiers in second language learning, ed. by John Schumann & N. Stenson, 137-52. Rowley, MA: Newbury House. ------. 1975a. Second language acquisition: The pidginization hypothesis. Harvard dissertation. -----. 1975b. Affective factors and the problem of age in second language acquisition. Language Learning 25.209-35. -----. 1976a. Social distance as a factor in second language acquisition. Language Learning 26.135-43. -----. 1976b. Second language acquisition: The pidginization hypothesis. Language Learning 26.391-408. Valdman, Albert (ed.) 1977. Pidgin and Creole linguistics. Bloomington: Indiana University Press. [Received 20 June 1979.] Semiotics of poetry. By Michael Riffaterre. Bloomington : Indiana University Press, 1978. Pp. x, 213. $15.00. Reviewed by Marianne Shapiro, Los Angeles This book argues that poetic discourse means nothing but itself, and that all else about it follows. At a time that is witnessing the climacteric of modernism, Riffaterre's persistence in viewing poetic creation as a predominantly deviant form of language, in which everything happens at the level of signifiers, moves him to a kind of theoretical idiolect that appears paradoxically and crucially dependent on a presupposition of theoretical ignorance on the reader's part. Poems themselves are 'evidently a game' and 'an exercise in form'—a statement subject to no graded assessment (115). But, understandably, R's analyses are almost entirely confined to works with an especially substantial quotient of introversive semiosis. Texts retain an authoritative status within R's concentric circles of intertextuality; however, they dominate other texts not as productions of significance, but as classificatory or taxonomic 'devices', balancing collision with collusion in a regressive literary play. R's theory ofsemantic over-determination is implemented by a transformational methodology that takes off from the principal presupposition that semiosis cancels mimesis: 'The poetic sign is a word or phrase pertinent to the poem's significance' (23)—a swath too wide for the phoneme, and too narrow for syntax. Signs are 'poeticized' when they refer to pre-existent word groups. R constantly emphasizes modalities of derivation which are governed over- REVIEWS457 whelmingly by laws of analogy. Poems result from the transformation of a literal 'utterance' into a complex periphrasis. To cite an example of R's complicated terminology, the ' previous utterance' or the connotative field of its origin is a 'hypogram'. The decoding of the invariant structure takes place pursuant to the cancelation of the 'representing, mimetic variants' (15). The hypogram, already a sign system comprising 'at least a predication', may be 'potential, therefore observable in language', or 'actual, therefore observable in a previous text' (23). Verbal 'structures of incompatibility' that emerge from the transfer multiply exponentially into further 'matrices for figurative expansion on the grammatical geometry of language' (56). However, transfer is most frequently accomplished at the sentence level, according to three derivational modes that correspond to three types of analogy: tautological, polarizing, and hypogrammatic (the latter comprising the threefold classification of semes and presuppositions, clichés or quotations, and descriptive systems). The 'prior utterance' is often an earlier text. Transfer and 'transform' emerge as gleefully interchangeable. Of the multiple objections that could be raised against the TG model, two chiefly concern me here. First, it depends upon underlying assumptions about deep structure which R is rarely (if ever) able to corroborate by evidence from linguistic change. The TG grammarian is likely to limit himself to formulating rules that do not take account of informational changes, while failing to characterize explicitly the semiotic status of all the primitive elements and rules that the account of language comprises . As a result, the TG approach...