Universal aspects of L2 learner varieties are currently being discussed with respect to the issue of the 'L2-initial state'. Minimal Tree (Vainikka & Young-Scholten, 1996), Full Transfer/ Full Access (Schwartz & Sprouse, 1996) and Valueless Features approach (Eubank, 1996) are three approaches that provide theoretical linguistic explanations of L2 acquisition based on UG. They differ with respect to the way in which LI-instantiated knowledge is assumed to influence L2 acquisition. The aim of the Basic Variety approach (Klein & Perdue, 1997) is to explain cross-linguistically similar phenomena of L2 learning on the basis of universal of semantic, pragmatic and configurational constraints. It claims to account for both the simplicity and stability of learner varieties. One can expect these approaches to contribute to our knowledge of L2 developmental processes in the following areas of research.
First, there is the question of the relation between the data and linguistic theory. Are native-speaker and nonnative-speaker knowledge of the same knowledge type? Schwartz (1994) argues that both LI and L2 acquisition are UG based. The question, however, is how to determine UG involvement in L2 acquisition on empirical grounds.
Second, universality of developmental processes of L2 acquisition is a matter of discussion. UG-based approaches differ with respect to the degree to which the presence of the instantiated LI grammar influences the acquisition of L2 syntax. In Basic Variety approach, universal principles are determined crosslinguistically. They are of a pragmatic, semantic, and configurational nature.
Third, Basic Variety provides an explicit account of what constitutes a simple language system. It represents a 'potential fossilization point' because it is not only a simple but also a stable system. Conflicts between constraints, i.e., possible sources of instability, are avoided. In the Full Transfer/Full Access approach, fossilization is due to problems of learnability. Restructuring an LI-induced grammar is problematic because negative evidence seems to be irrelevant, and in some cases, L2 input is highly obscure, i.e., very complex and/or very rare.
Fourth, UG-based approaches differ with respect to the role of the LI. In the Minimal Tree and the Valueless Features approach, it is restricted to the initial stages, i.e., the structure of lexical projections. For Full Transfer/ Full Access, it may be relevant in every stage of L2 acquisition. Basic Variety leaves room for crosslinguistic influence only in the early stages of acquisition. When more options are available, L2 learners seem to take the alternative closest to their LI.
Fifth, the question of the driving force has never been explicitly posed within UG-based approaches. This is probably because triggering, parameter setting, and feature strengthening exposure to the target language is regarded as trivial. Development in the Basic Variety approach is caused by intrinsic factors, such as discourse contexts where constraints come into conflict. In such situations, the learner has two possibilities: either to 'override' one of the constraints or develop specific means to accommodate the 'competition'. In addition, another motivation is suggested: the need to solve structural ambiguities. Categorization in terms of syntactic functions such as subject, object, predicate, noun, verb, adjective, etc., solves ambiguity through hierarchical structuring.
Transfer Effects at the Syntax-Semantics Interface: The Case of Double-que Questions in Heritage Spanish
