While research on English-medium instruction (EMI) has grown considerably in recent years, we still know relatively little about what happens inside EMI classrooms, particularly in terms of interactional practices. There remains a shortage of empirical resources, such as corpora, that can offer systematic insights into the features of EMI classroom discourse. This paper addresses some of the gaps by outlining the methodological and ethical procedures involved in building and annotating a corpus of digitally mediated EMI interactions, drawn from 24 university seminars in the social sciences across two EMI contexts. Using the Multimodal EMI Corpus (MEMIC) as a case study, we demonstrate how such a resource can be used to examine lexical and discursive features of EMI interaction across modes and registers. We also show how these insights can inform the development of EMI-relevant teaching materials and assessment tasks. The overarching aim of this contribution is to encourage other scholars and practitioners to collect and analyse data from their own local EMI contexts, in order to advance corpus-based research into EMI classroom practices and support more effective, evidence-informed pedagogy.
Jaworska et al. (Thu,) studied this question.