Introduction

Recent years have seen an enormous growth of interest in task-based language learning and teaching (see, for example, Bygate et al., 2001a; Skehan, 1998b; Willis, 1996). This interest has been motivated to a considerable extent by the fact that ‘task’ is seen as a construct of equal importance to second language acquisition (SLA) researchers and to language teachers (Pica, 1997). ‘Task’ is both a means of clinically eliciting samples of learner language for purposes of research (Corder, 1981) and a device for organizing the content and methodology of language teaching (Prabhu, 1987). However, as Bygate et al. (2001b) point out, ‘task’ is viewed differently depending on whether the perspective is that of research or pedagogy. Researchers, for example, may view a task in terms of a set of variables that impact on performance and language acquisition whereas teachers see it as a unit of work in an overall scheme of work. Of course, there are also points of contact between the two views – for example, information about significant task variables acquired through research can assist teachers in deciding what tasks to use and when.

The purpose of this chapter is to examine theoretical views of language use, learning and teaching that underlie the work on tasks that has taken place to date. Two broad and disparate views will be identified and discussed. The first of these draws on what Lantolf (1996) has referred to as computational models of second language (L2) acquisition, which treat acquisition as the product of processing input and output. In accordance with this model, researchers have sought to identify ‘psycholinguistically motivated dimensions’ of tasks (Long & Crookes, 1987) – that is, to establish the task features that have a significant impact on the way learners process language in performance and, therefore, potentially, on how they acquire an L2. The identification of such dimensions can be used to select and grade tasks for teaching and learning.