After attempting - and largely failing - to delimit a distinct field of interpersonal pragmatics, this paper explores what is distinctive about interpersonal pragmatic practice; that is, what makes it different from the scholarly tradition of pragmatics. Three facets of practice are discussed: its aims, its approach to data (what aspects are brought into relatively clear focus) and its analysis of data. The common thread running through what is found is a changed, more modest, place for language, the understanding of which is no longer the assumed goal of scholarship, the size of examples of which for analysis have become larger and the use of which is no longer the single focus of analysis. It is argued that this last development should point the way to a particular procedure for analysing interaction. Accordingly, the paper proceeds to an example analysis of one piece of data.