“Does EMI Improve Students' English?” and Other Unanswerable Questions

Abstract Questions about EMI often begin from an apparently straightforward premise—does EMI improve students' English? This article argues that the question itself is the source of the main analytic problem. Using EMI as an exemplar of a wider problem in applied linguistics, we show why binary claims about language gain travel poorly across contexts. Existing evidence is constrained by the limits of quasi‐experimental comparison, the ideological load carried by self‐report data, the instability of ‘English proficiency’ as a construct, and the hypercontextualized conditions through which EMI is enacted. The issue is therefore a mismatch between the question being asked and the phenomenon under investigation. EMI outcomes emerge through configurations of pedagogy, language support, institutional policy, disciplinary practice, learner histories, and unequal access to English. We argue for a shift from asking whether EMI works to examining what forms of language development become possible, for whom, through which mechanisms, and under which conditions. The article concludes by outlining implications for EMI research design and for policy claims that treat incidental English development as an automatic benefit of EMI.