Grammar as validity evidence for assessing L2 interactional competence

Abstract This study examines how L2 learners of English at different pragmatic performance levels implement requests during role-play assessment interaction. In doing so, the role of grammar-for-interaction as validity evidence of assessing interactional competence is explicated. Using qualitative and quantitative data, the following research questions are investigated: (a) How do learners use grammar differently depending on the role-play request situations they are engaged in?; (b) To what extent are learners’ role-play performance levels associated with the ability to use context-sensitive grammar when co-constructing request sequences? The data come from a database of 102 L2 English learners’ role-play interactions with 45 hours of recorded interaction in total. The role-play assessment tasks include various real-life communicative situations, such as requests designed for specific interlocutors in a university context. Using conversation analysis, the role-play interactions at different performance levels were analyzed. Sequential analyses showed that higher-level learners utilized diverse grammatical formats that are sensitive to their sequential positions and contingencies associated with requests. As quantitative evidence, the relationship between selected grammatical constructions and the learners’ role-play performance levels determined by trained raters was examined. With this, I discuss the role of grammar-for-interaction in defining and assessing interactional competence.