Spoken Discourse

Abstract

When second or foreign language (L2) learners take a speaking proficiency test, their scores are really the only outcome in which they are interested, especially in high stakes assessment contexts—where a sufficient score is necessary for university admission, citizenship, or employment. So it has been with language testers: the emphasis on test scores (or ratings, as they are likely to be in speaking assessment) leads to the primary concern about score as well as rater reliability—the degree to which a rater is consistent over time, across language samples, and with other raters. More recently, however, attention has shifted to issues of test validity, which considers the processes by which test scores lead to inferences about their meaning. This is increasingly the case for L2 speaking tests, where investigation of the process of speaking assessment has become almost as important as questions about outcome scores. Two primary questions underlie the emphasis on test process: first, how can the language that is produced in a speaking test be characterized? Second, how can careful analyses of that language lead to improvements in test design and rating processes? To answer these questions, the analysis of spoken discourse has become a key strand of research in the language assessment community. The chapter begins with a definition of discourse and discourse analysis; subsequent sections present facts about the nature of spoken discourse in both monologic and interactive contexts and empirical findings on the co‐construction of discourse in the speaking assessment process involving test takers, interviewers, and peers. The chapter concludes with a look at the challenges facing spoken discourse analysts and future directions in L2 speaking assessment, including the impact of digital technologies and large‐scale language corpora, the systematic development of rating scales, and the importance of formative, classroom‐based assessment in understanding the nature of L2 speaking proficiency.