Corpora, Translation Studies, and Contrastive Linguistics

Today, both corpus-based translation studies (CBTS) and corpus-based contrastive studies (CBCS) actively use text corpora as sources of data. During the past decade, corpora have grown considerably in size and number, and morphosyntactic parsing became available for many languages. Most corpus resources can be accessed online, and compiling ad hoc corpora is no longer an issue. Processing and querying data have become much easier, and statistical tools give fantastic opportunities for using current quantitative methods. Both CBTS and CBCS deal with cross-linguistic correspondences; only their approaches differ. CBCS concentrate on differences between languages, while CBTS pay special attention to the translation process. CBCS use parallel and comparable corpora to study correspondences between certain lexemes, grammar forms, grammatical constructions, and meanings in different languages, trying to take into account interference caused by translation. CBTS use parallel and translation corpora to detect the overuse or underuse of grammatical forms and constructions in translations, and these findings can be used in the practical translator’s work. In this chapter, various important projects from the CBTS and CBCS fields are referred to (e.g. INTERSECT, ENPC, CroCo, etc.) and the basic methods of research are explained. The challenges of using web corpora and universal dependencies parsing when studying language structure are also addressed.