A corpus-based approach to human and machine translation: a multidimensional analysis of judgments of the Court of Justice of the European Union

This article describes an exploratory study focusing on the differences between human translation (HT) and machine translation (MT) of legal texts using multidimensional analysis (MDA; Biber, 1988) as a methodological approach. Drawing on a corpus of HT and MT texts of judgments delivered by the Court of Justice of the European Union, the article explores linguistic variations between HT and MT. Using MDA, two main differentiating dimensions were identified: (1) Elaborative argumentation versus informational production; and (2) Discursive dynamism versus institutional formality. Preliminary results indicate that HT is more argumentative and coherent than MT and is characterized by formulaic metadiscursive devices. MT, by contrast, is characterized by less complex grammatical and syntactic structures and a more informational discourse, revealing features commonly associated with the language of MT. These findings highlight the importance of evaluating legal translation quality in terms of genre-sensitive discourse functions that are essential for legal interpretation.