Building bridges within and beyond corpus linguistics. Methodological and theoretical issues

Gilquin, Gaëtanelle [UCL]

With its focus on authentic data, corpus linguistics was enthusiastically welcomed by those linguists who had lost faith in intuition as the sole means to investigate language and who found with this new trend a way of empirically grounding their analyses. Too often, however, corpus linguistics has been seen as an end in itself, with no consideration for its potential combination with other types of data and other frameworks. Using a corpus-based study of English periphrastic causative constructions as an illustration, it will be demonstrated that better descriptions of linguistic phenomena can be arrived at by ‘building bridges’, both within and beyond corpus linguistics. With respect to methodology, it will be shown that automatic methods of extraction can – and sometimes should – be combined with manual ones, but also that corpus data can fruitfully be supplemented with other types of data (in this case, elicitation data). From a theoretical point of view, the proposed model integrates a top-down and a bottom-up approach and, in response to criticism against the a-theoretical status of corpus linguistics (see Grabe & Kaplan 1996: 46), suggests that bridges should also be built with other theoretical frameworks, such as discourse analysis or cognitive linguistics. While such attempts do exist among corpus linguists (e.g. Flowerdew 1998 or Mukherjee 2004), they are still too rare and should be encouraged, for reconciling these seemingly opposed approaches (manual vs. automatic, top-down vs. bottom-up, etc.) leads to a coherent and productive model that capitalises on the strengths of each approach and makes up for their weaknesses.