Making a way through the jungle: Exploring learner data in translator and interpreter training

Translator and interpreter (T&I) training is usually practice driven. Practical outputs of learners constitute important resources for the assessment of learner needs, difficulties and competence, and at the same time should feed back into pedagogical or eve curriculum adjustments. However, systematic investigation of learner data has largely been hindered by the difficulty in and lack of consensus on collecting and exploring relevant data. In this regard, the development of corpus analytical tools and means to explore big data has influenced much the data collection and analysis.

This presentation will introduce two T&I learner corpora projects led by the presenters. The first is the Chinese/English Translation & Interpreting Learner Corpus (CETILC), a mega-size learner corpus developed in particular for the study of lexical cohesion (Pan, 2017), and the latter relates to the Hong Kong subset of a university translation learner corpus that constitutes part of an international multilingual student translation corpus initiative (MUST; Granger & Lefer, 2020). The presentation will illustrate the pros and cons of both single-type and multiple error annotations of learner outputs, and the strengths and weaknesses of collecting selected or in-depth learner variables for a T&I learner corpus. It will also discuss the possibilities of machine facilitated human annotation, as well as human edited machine annotation of learner data. The study will show the potentials of large-size learner corpora in T&I training, and shed light on future study of learner data in T&I training.