Abstract The analysis of language use in real-world contexts poses particular methodological challenges. We codify responses to these challenges as a series of methodological imperatives. To demonstrate the relevance of these imperatives to clinical investigation, we present analyses of single episodes of interaction where one participant has a speech and/or language impairment: atypical prosody, echolalia and dysarthria. We demonstrate there is considerable heuristic and analytic value in taking this approach to analysing the organisation of interaction involving individuals with a speech and/or language impairment.