Phraseology in multilingual EU legislation: a corpus-based study of translated multi-word terms

Phraseology is a key element to take into account when considering the quality of a translated text. This paper investigates the translation of multi-word terms in a bilingual (English-Italian) parallel corpus of European Union (EU) labour law legislation. The aim of the research is to identify the recurrent patterns of translational behaviour for these terms in view of the complexities of legal translation in the multilingual context of the EU. The study was carried out with a frequency-based approach, combining quantitative analysis with the aid of corpus linguistics tools and qualitative analysis of the parallel sections where the collocations occur. The data were compared with a reference corpus of Italian national labour law legislation and the results were considered in the light of the contextual factors that may have influenced the translators’ choices. The findings reveal inconsistencies in the translational behaviour for the multi-word terms analysed, strong interference from the source texts on the collocational patterns used and marked divergence from the conventions of legal language in the target culture. In the context of universals of translation, the study also provides evidence of the untypical collocations hypothesis as a linguistic feature that characterises legal translation.