A Direction for Contrastive Analysis: The Comparative Study of Developmental Sequences

Abstract

This paper compares developmental sequences for mature L2 structures with a view to determining influences of the learner's L1 on them. Two kinds of L1 influence are identified: the pace with which a sequence is traversed and the number of developmental structures in a sequence. The paper concludes with some tentative proposals concerning the interaction of prior L1 knowledge and the creative construction process. Second language acquisition research of the last several years has contributed much toward a refinement of the contrastive analysis hypothesis as originally proposed by Lado in 1957. For one, given the dynamic nature of second language acquisition, it has become evident that contrastive descriptions based on mature structures of the L2 are of doubtful validity in attempts to predict points of facilitation or interference. In this approach, the contrastive researcher is comparing an ideal terminal state, equivalent to quasi-native speakers competence in the L2, with the learner's L1. Structures of the L2, however, are acquired by progressing through a sequence of developmental stages. Hence, if the contrastive analysis hypothesis is to provide a theoretical perspective for second language acquisition study, it is paramount that the role of prior L1 knowledge be conceptualized as a variable which may introduce variation into a developmental sequence. A further refinement of the contrastive analysis hypothesis has been made possible by mounting evidence pointing to the indirect and highly constrained nature of Li influence. Marked structural contrast, for example, tends to promote avoidance or underrepresentation rather than a mechanical imposition of Li structures (Schachter 1974, Hakuta 1975). Work by Kellerman (1977), Gass (1979), and Gundel, Sheldon, and Tarone (1981) converges on the insight that linguistic features which are marked by virtue of their limited distribution cross-linguistically will tend not to be eligible candidates for Ll influence. Also, now that some of the dust surrounding the morpheme order studies has settled, a picture that has * An earlier version of this paper bearing a different title was presented at the 15th Annual TESOL Conference, Detroit, March 3-8. The author is indebted to Stephen Krashen for some valuable discussion and correspondence.