The formulaicity of translations across EU institutional genres

This paper explores the formulaicity of EU translations into Polish across four institutional genres (legislation, judgments, reports, websites) with reference to the corresponding EU English corpora in order to understand how the degree of formulaicity is affected by the variable of genre. Formulaicity is operationalised as lexical bundles – high-frequency multi-word sequences (Biber and Barbieri 2007). The study shows a strong correlation between formulaicity and genres, as well as multiple facets of formulaicity (e.g. tokens vs. types). Our findings generally confirm the increased aggregate formulaicity of translations as regards bundle tokens for all EU genres, except for judgments, and the increased variation of bundles (types) for all the genres at the 40 occurrences per million words (pmw) threshold, although the “micro-formulaicity” threshold yielded inconclusive results. Another finding reveals a consistently low overlap of bundles between translations and non-translations. We argue that translations develop their own formulaic profiles which are levelled out compared to EU English corpora and which minimally overlap with formulaic profiles of domestic genres.