Language and translation on the Web

Corpus linguistics is an approach to the study of language that developed in the late 20th century. Opposing analytical methods that rely exclusively on introspection and decontextualized, artificial examples, corpus linguistics assumes that language should be studied by looking at genuine texts, or text samples, collected according to explicitly defined criteria, stored electronically and searched using dedicated software (corpus query tools). This chapter reviews the origins and basic assumptions of this approach, drawing attention in particular to the interest in phraseological and lexical perspectives on language that it brought about. Surveying the main analytical methods, insights, and practical applications of corpora in the fields of translation and interpreting studies, the chapter focuses on widely used corpus designs (monolingual comparable, parallel, and bidirectional). It identifies three issues of special relevance to corpus-based translation studies, namely typical features of translated language, translation shifts, and translator style, and the specificities of corpus-based interpreting studies, especially the triangulation of methods from corpus linguistics and other disciplines such as sociolinguistics and ethnography of communication. The chapter identifies challenges, debates and future directions in both research areas, and concludes with an outline of uses of corpus research in the translation and interpreting professions, including corpus use for machine and computer-aided translation, translator learner corpora, and corpora for the training of translators and interpreters.