ABSTRACT
This chapter focuses on the connectionist approach in the context of second language acquisition. After the publication of the PDP volumes in 1986, connectionism was quickly introduced to the field of SLA, first in a review article by Richard Schmidt, and then by a series of connectionist modeling by two UCLA students at the time (Michael Gasser and Magi Sokolik), and a series of conceptual papers by other students (Shirai, Fantuzzi, Yap, Duff) which spurred the initial momentum/excitement for a connectionist approach to SLA. Next, the connectionist approach has been primarily advocated by Nick Ellis, and later DST researchers joined forces, to provide a theoretical alternative to the then dominant universal grammar approach.
In the second half of the chapter, I review core issues in SLA, namely the input–output issue and the role of explicit–implicit knowledge that has been theoretically important in SLA, and provide a connectionist view to these important questions in the field. In essence, I have argued that although L1 acquisition can be primarily achieved (without negative evidence) through connectionist-type input processing that results in implicit knowledge of language, for various reasons, adult L2 acquisition, in most cases, falls short of native-like attainment, and requires involvement of explicit processes. How connectionist architectures can handle the interface of implicit and explicit processes is as yet unknown, but promising attempts have begun to be explored.
