Abstract
It is now widely recognised and has been repeatedly shown (e.g. Hyland, 2004, 2008; Gray, 2015) that professional academic writing differs considerably between individual disciplines.
Detailed description of the language use within a given specialisation is an invaluable source for university courses of English for specific purposes. However, mathematical texts in general, and mathematical research articles in particular, are underrepresented in research into phraseology of scientific texts, making the teaching of English for mathematicians a challenging task.
The present paper focuses on the distributional phraseology of mathematical texts. Using corpus-driven methodology, it demonstrates identification and constructional description of key elements of mathematical research articles, which can be seen as the basic building blocks of this genre.
The method of function keywords is used for the primary identification. On the example of the keyword 'we' the paper shows the possibilities of a detailed description of the individual constructions, their mutual interactions, and the use of relational networks in uncovering a bigger picture of the discourse of mathematics.