Feature reassembly as constraint satisfaction

Research on bilingual grammars from a formal perspective has often come under the guise of mainstream generative grammar. Since the inception of Chomsky's (1995 et seq.) Minimalist Program (MP), many scholars have adopted the notion of formal features representing abstract grammatical information that can be associated with lexical items. To model changes in bilingual grammars due to the acquisition of particular forms, the attrition of information by means of incomplete acquisition or the lack of usage throughout the course of the lifespan, or due to intense contact with another grammar, the mechanism known as feature reassembly (e.g. Lardiere 1998)-whereby abstract grammatical information in the form of formal features can be detached and reassigned to other lexical items-has enjoyed a great deal of success in the literature. In this article we argue that in spite of this success, the analysis of aspects of bilingual grammars can be improved upon by replacing the notion of feature reassembly with the satisfaction of constraints. Here we provide conceptual and empirical evidence arguing for the adoption of constraint satisfaction in place of the feature reassembly mechanism. Finally, in addition to constraint satisfaction we also make the case for adopting a parallel model of cognition and language for the bilingual mind, which is strongly supported by recent psycholinguistic and neurolinguistic studies.