This paper describes an empirical investigation as to the utility of a principles and parameters model of grammar in describing the interlanguage grammars of second language learners. I argue that the framework of parameterized universal grammar is useful in accounting for second language learner knowledge and behaviour. The vast majority of their errors can be accounted for by the transfer of their first language (L1) parameter setting into the second language (L2).

In investigations of the question of whether or not Universal Grammar (UG) is available in non-primary language acquisition, a number of researchers have tried to isolate situations where the way principles of UG operate in the first language (L1) could not help the learner acquire the relevant properties of the second language (L2). If learners show evidence of acquiring properties of the L2 that could not be acquired from the input alone and could not be reconstructed via the L1, this suggests that UG is available in non-primary acquisition; in contrast, if learners fail under such circumstances, this supports the claim that UG is no longer directly accessible. In particular, there has been a tradition of looking at island constraints in this light, using L1s and L2s which differ radically in terms of the surface effects of principles like Subjacency and the Empty Category Principle (ECP) (e.g., Bley-Vroman et al 1988; Johnson and Newport 1991; Martohardjono 1991a, 1991b; Martohardjono and Gair 1992; Schachter 1989, 1990; White 1989, 1992).