Key points are not available for this paper at this time.
Minority languages in Europe have a relevant position to strengthen cultural and linguistic communities and preserve societal cohesion. Nowadays, machine translation (MT) technologies have achieved relevant advances using neural networks, which are able to increase language use over communities. Although this is a promising scenario, the latest technological advances in MT technologies are not in-depth tailored and tested for some minority languages in the legal EU domain. Also, due to many challenges related to this domain, such as complex syntax, highly convoluted sentence structures and terminology, the use of MT to translate legal terminology has often been associated with increased risk of errors. To get a better insight to this question, we assess the terminological consistency when using six machine translation engines to translate the acquis communautaire, the accumulated body of EU law and obligations, from English into Spanish and from English into Catalan. This paper presents the results of the assessment, showing that recent advances in MT yield strong performance for English-Spanish. In contrast, terminological consistency is lower when translating from English into Catalan, reflecting the limited resources and support available for this language.
Alvarez-Vidal et al. (Wed,) studied this question.