The computational model of L2 acquisition and its implications for second language instruction

Although the study of second language acquisition (SLA) is not much more than 30 years old, it has already generated a number of different theories that purport to account for the processes underlying language acquisition. These theories, however, can generally be divided into two broad categories : those based on psycholinguistic approaches that explore the cognitive processes at work, and those that concentrate on the social factors and processes involved in language acquisition. This paper will concern itself with the former category, and the computational model of SLA in particular. While « in particular » usually denotes the narrowingdown of a topic area, it must be kept in mind that the computational model is a fairly broad account of SLA that accommodates a number of mutually exclusive or at least opposing positions within its framework. We are fully aware that the computer metaphor in SLA can be perceived as reductionist. The present paper tries to put the model in perspective without taking sides. Outlining this framework and drawing out its most controversial and hotly debated aspects, and the discussion of some empirical research that has contributed to the formulations of the model will be the aim of section 2 of this paper. Section 3 will look at the theory from a more critical perspective while section 4 will consider the pedagogical consequences of the model.