Classroom talk is organised for the controlled transmission of knowledge. In this chapter I want to consider whether sociological investigation of the resulting discourse has identified a ‘basic’ structure to the communicative strategies normally employed. Has it detected a ‘relatively stable, socially structured series of events’ even where the participants themselves may be more immediately aware of a ‘fluid, transient and fragile situation’?1 Is such an analysis irredeemably static and normative, or can it meet Bernstein's criterion for an ‘exciting’ sociological account by relating structural features and interactional practices?2 Although these questions have particular meaning for me in the context of a recent study of ‘the language of teaching’, my purpose here is not to summarise the findings of that research but to reflect on some of the issues which it raised.3