NMT influence on the use of dictionaries by second-year MA legal translator trainees

This dissertation aims to analyse the degree of influence neural machine translation (NMT) has in the translation process of a legal text for second-year MA legal translator trainees, considering only dictionaries as a type of source employed by participants of an empirical study, the “LeMaTTT Project” (Quinci, forthcoming). Here the study will be presented, after a general overview of translation competence and post-editing competence, focusing on specific competencies required and the impact Machine Translation and CAT tools (Computer-Assisted Translation Tools) have in legal translation. Then dictionaries as a translation source will be explored, relying on recent empirical studies with the aim of outlining their role and usage for legal translation and its training. The analysis presented in this dissertation will consider two different task groups of the “LeMaTTT Project”, i.e., those doing post-editing and those translating from scratch as the control group. It will address the role dictionaries play in terms of translators’ usage, time, pattern of searches and management of given solutions. The ultimate aim of the analysis is to detect whether said aspects might affect the accuracy of the translation process when comparing the post-editing of a machine-translated text with the from-scratch translation process. The methodology employed was based on data extracted from video recordings provided by participants translating a given source text, a “Power of Attorney” from English into Italian, and aimed at calculating the frequency of usage and triangulating specific given aspects through charts and ‘box and whiskers’ plots.