The language of science has been extensively studied by linguists and rhetoricians – as a distinctive register, as a set of genres that students and academics need to master, and as a discourse of powerful social institutions. Most of these studies have been synchronic, focusing on the structures or styles of more or less contemporary texts, particularly research articles. But if we rely on such studies, we may tend to reify some features of text (such as the Introduction–Methods–Results–Discussion form, or the tendency to passive constructions and nominalizations) as inevitable features of scientific communication. We may also treat scientific institutions – such as the lines between disciplines, or between professionals and amateurs – as given by the subject matter, rather than seeing them as changing and as constituted in part by their communicative practices.