While subjects of transitive action verbs in English and Dutch are typically realized as human agents (see Comrie 1989), both languages also feature instances of nonhuman agents in subject position. However, Vandepitte and Hartsuiker (2011) have shown that there are fewer options in Dutch and that translation issues present themselves in cases where both languages do not overlap. This paper wants to document overlap and differences in terms of non-prototypical subject realization by focussing on the strategies that are used in Dutch translations of six actions verbs (give, demonstrate, show, suggest, offer and tell) in combination with non-human subjects. Results reveal that a fair share of non-human subjects are also translated as such in the target language. Other strategies include occasional humanization of the non-human source text subjects, reduction of valency patterns with reduced agentivity vis-à-vis the English source-text sentences and shifts in the mapping of semantic roles onto syntactic functions.
This paper provides an overview of state-of-the-art research in translation studies as represented in this special issue, with a special focus on corpus-based approaches that (re-)connect translation studies with other fields of corpus-based research in linguistics or which explore new types of translation data in the broadest possible sense of the term. It does so by singling out papers that illustrate different methods of data harvesting, on the one hand, particularly in areas that are currently underrepresented in the field, i.e. interpreting and subtitling, and by presenting studies that approach translation data from a perspective other than that of “translation universals”.