Translation

The empirical study of translation has traditionally been text-based, typically focusing on the comparison of single originals and their translations. In this way, linguists have looked for examples of translation shifts, or “departures from formal correspondence” in the process of going from the source language (SL) to the target language (TL; Catford 1965 in Venuti 2000: 141ff.), or else have viewed translations as contextualized instances of cross-linguistic correspondences, and analyzed them so as to identify ways in which “the differences in the expressive power of two languages may be overcome” (Doherty 2002: 1). Classicists and literary study scholars have painstakingly combed translations, particularly of canonical works, looking for “rocks on which [past translators] have split, and the right objects on which a translator ... should fix his attention” (Arnold 1861: 2). On this basis, proposals have been made for an analytic of translation based on “deforming tendencies” (tendances déformantes), or “universals of deformation inherent in translation as such” (Berman 1985 in Venuti 2000: 296). This is admittedly an oversimplified account, which nonetheless may help to establish a backdrop for the emergence of the corpus-based approach in the early 1990s. Three factors acted as catalysts. First, the quantitative focus afforded by corpus methods could provide an alternative to “pick and choose” observations − i.e. those “subjective, largely intuitive and impressionist methods” that some researchers had grown dissatisfied with (Holmes 1978 in Weissbort and Eysteinsson 2006: 421). Second, the 1980s and early 1990s saw an upsurge of interest in the specificities of (the language used in) translation, variously termed translationese (Gellerstam 1996) or third code (Frawley 1984). Translation scholars grew interested in the set of “fingerprints” that one language leaves on another when texts are translated between the two (Gellerstam 1996), or in the “third code which arises out of the bilateral consideration of the matrix and target codes” (Frawley 1984: 168). The research agenda thus opened up to consideration of coherent sets of texts (or corpora), as opposed to single originals and their translations. Third, the theoretical shift away from the primacy of the source text (ST) and the notion of equivalence, and toward the target text (TT) and the notion of appropriateness, brought with it the need for studies focusing on translated texts within their contexts of reception, with a view to exploring “the nature of translational norms as compared to those governing non-translational kinds of text production” (Toury 1995: 61). Corpus linguistics provided an appropriate toolkit for the monolingual comparison of translated texts and comparable texts originating from the receiving culture. The first part of this chapter surveys the development of corpus-based translation studies (CBTS), from the programmatic proposals in the early 1990s to recent developments and trends. Section 2.1 briefly outlines the theoretical and methodological bases of CBTS, in particular as concerns the types of corpora and the research questions specific to translation research, as opposed to other uses of corpora in translation (e.g. for translation teaching or practice) and to bordering disciplines such as contrastive linguistics. Section 2.2 digs deeper into five studies that have been especially influential and/or that embody especially sound research practices, while Section 2.3 concludes by summarizing the state of the art in CBTS and the major questions still requiring to be addressed. In the second part (Section 3), a case study is presented which aims to test whether translated texts are more or less collocational than comparable non-translated texts, using a combination of comparable, parallel, and reference corpora.