The acquisition of preposition stranding by second language learners and parametric variation

This paper discusses the application of Government Binding Theory to second language acquisition in the context of a project which is looking into the acquisition of preposition stranding in English and Dutch. The bulk of the discussion focuses on the theoretical problems involved. Firstly, the potential value of Government Binding Theory in principle is considered both in terms of the formulation of linguistic questions per se and also in terms of more specifically acquisitional questions having to do with the speed and order of acquisition. Secondly, some results in the pilot studies conducted so far in Utrecht are examined with respect to the theoretical usefulness of the framework adopted. The potential of the framework to generate sophisticated linguistic research questions is found to be undeniable. The acquisitional aspects need to be elaborated and adapted to cope with the special features of second, as opposed to first, language acquisition. This involves an elaboration of scenarios to be investigated: one in which the learner's initial assumption is that the unmarked setting of a given parameter of Universal Grammar holds for the target system, one in which the settings of parameters shared by the target and native systems are assumed to be identical, the second being a 'cross linguistic' scenario. These possibilities are considered in the light of the nature of evidence derived from the input and in the light of a set of possible learning strategies derived from the scenarios. The scenarios, the types of evidence and the strategies are spelled out in terms of the specific problem of preposition stranding in Universal Grammar, in Dutch and in English.