Towards a Firthian Notion of Collocation
Sabine Bartsch, Stefan Evert, Friedrich-Alexander-Universität Erlangen-Nürnberg · 2014
Collocations are pervasive in language. According to Altenberg (1991: 128), “roughly 70% of the running words in the corpus form part of recurrent word combinations of some kind.” The investigation of such word combinations in corpora of authentic language dates back to the earliest studies of collocations by J. R. Firth (1957), who is commonly credited with introducing the concept within British Contextualism. However, serious corpus-based exploration of collocations on a larger scale has only become feasible with the arrival of the computer in the linguist’s workspace in the late 20th century. Since then, a substantial number of corpora of different sizes have become available, opening up new possibilities for collocation studies and many other linguistic applications. Progress has been made in particular by harnessing ever larger corpora, a growing range of statistical measures of association (cf. Evert 2004), and state-of-the-art software tools for automatic linguistic annotation and analysis.