Secondary Predication in L2 Grammar

A persistent question in the SLA (second language acquisition) research concerns the status of the involvement of the native language (i.e. L1 transfer). Inquiry into the question has been made on the basis of a variety of linguistic frameworks, such as contrastive analysis (Lado, 1957), L2 (second language) generative parametric approach (Flynn, 1987; White, 1989; Schwartz, 1996), and the more recent generative inquiry into the initial state of L2 acquisition (e.g., Vainikka and Young-Scholten, 1994, 1996, 1998; Eubank, 1994; Schwartz and Sprouse, 1994, 1996; Schwartz, 1998). A clear settlement does not seem to be forthcoming soon, although some advances have been made. There appears to be very little consensus among L2 researchers on the nature of L1 transfer in L2 acquisition. The question itself is so puzzling that it qualifies as what Pinker (1989: 1) eloquently describes as a research scenario, and one in which we must have missed some profound principles because “nature does not go out of its way to befuddle us”. In this paper we make an attempt that departs fundamentally from the current view on L1 transfer. Central to this proposal is that L1 transfer constitutes the L1 performance systems, external to L1 grammar (i.e. I-language). We employ Distributed Morphology (DM) (Halle and Marantz, 1993; Marantz, 1997; Harley and Noyer, 1998, 1999), a version of the minimalist approach (Chomsky, 1995, 1999, 2001), to account for the results from two experiments designed to examine the function of L1 properties in L2 grammar. Two specific questions were addressed: What is the nature of L1 transfer? Assuming that the final state constitutes an I-language and an E-language system, do the L1 properties observed in L2 spontaneous and experimental data come from the E-system or the I-system? A hypothesis called the Relativized L1 Transfer Condition (RTL1C) was tested in the experiments. This hypothesis conjectures that an L1 property is subject to transfer if and only if there is no such item that consists of more canonical morphosyntactic features. The RTL1C predicts that non-canonical L1 features are often left untouched. Furthermore, the RTL1C, if verified empirically, indicates that L1 transfer is external to L1 grammar in the sense as defined in Chomsky (1991: 9). Consequently, the ultimate L2 attainment (i.e. L2 Ilanguage) is incomplete, or a proper subset of that mastered by the native speakers, in that it falls short of the exhaustive paradigms of all the possible L2 SDs (structural descriptions, Chomsky, 1995), 3