Bitexte, bi-concordance et collocation.

Abstract

Corpus analysis has become an essential part of unilingual lexicography and its usefulness has been shown time and again. Unfortunately, few researchers have actually written about the usefulness of bilingual corpora in bilingual lexicography, perhaps because bitext and bilingual concordancers are still quite new and quite rare. This thesis will try to bridge the gap. In writing this thesis, we had two main objectives: to show the importance of bilingual corpora in bilingual lexicography and to design a program that extracts collocations and their possible translations. We will thus highlight the role and usefulness of corpora--and of bilingual corpora in particular--in bilingual lexicography first by clarifying the terminology and then by showing how unilingual and bilingual corpora are used differently. The second part of the thesis deals with the automatic extraction of collocations from a bitext. We will show how biconcordancers can be useful in extracting collocations, present two programs specially designed to extract collocations and propose a method to automatically extract collocations from bitext. Finally, we will compare our results to those extracted by software that already exists.