Instance theory and Universal Grammar in second language research

Abstract

This article considers the possibility of applying instance theory to the study of language, second language in particular. Instance theory de-emphasizes the role of abstract principles in knowledge and its acquisition and use, focusing instead on the storage and retrieval of specific experiences, or instances. I argue that the application is feasible only if one also adopts a restrictive theory of Universal Grammar. I then present a sketch of a combined UG–instance theory approach, in which invariant aspects of UG are maintained and variability is allowed in exactly the same areas as in standard theories, but the variation occurs in pools of stored instances, not in abstract parameter values. This approach can be productively applied to various problems in language learning research, including noisy input to learners, undoing of errors during the learning process, transfer and fossilization, and the nondiscrete character of learning.