The Discourse Function of Collocation in Research Article Introductions

Abstract : The increasing use of computer-held text corpora containing many millions of words has allowed linguists to establish lexico-grammatical patterns in language that were previously unavailable to observers. Such patterns range from lexical collocations and idioms to the phraseology of grammatical items. Recently, collocations of high frequency words in medical research abstracts and articles have been found to be useful indicators of the prototypical phraseology of the genre. In this article we characterize the phraseology of Introductions from a corpus of 150 cancer research articles. We explain the fixedness and idiosyncratic nature of scientific phraseology in terms of discourse processes such as reformulation. We argue for the design of a representative and specialized corpus of the research article and a contextual approach to corpus work that is appropriate to the teaching of languages for specific purposes (LSP) and the ethnographic aims of genre analysis in general.