Hypal4MUST: A community-based web interface for translation teaching
(2019) EUROCALL 2019 — Location: UCLouvain (28.August.2019)
Files
No attached file found for this publication.
Details
- Authors
Granger, Sylviane 
- Abstract
- One of the tasks that translation teachers are called upon to carry out on a regular basis is the preparation, processing and annotation of translation exercises. This task involves the following workflow: identification of suitable source texts (STs); submission of the texts to students; collection and annotation of the students' translations (also called ‘target texts’ or TTs); return of the annotated translations to the students. These activities are, notoriously, highly time-consuming and provide very little return on investment, as the teachers' and students' work is scattered and essentially lost. Surprisingly, while computer-aided translation tools such as SDL Trados Studio are ever-present in translator training, especially at master’s level, technology is largely absent from the workflow mentioned above. A recent project, called Multilingual Student Translation (MUST) , aims to remedy this deficit by providing a web-based environment which allows translation teachers to share source texts and turn student translations into a rich searchable database. The Hypal4MUST interface, based on the Hypal tool (Obrusnik 2014), is a user-friendly tool for the collection, alignment and linguistic annotation of student translations. The student-teacher interactions take place via two seamlessly interconnected parts of the software: the teacher interface and the student interface. One particularly noteworthy feature of the teacher interface is the source text database, a collaborative repository of STs where teachers can submit their own STs and/or choose existing ones from the shared database. The ST database is a perfect example of "community sourcing", i.e. the cumulative development of resources resulting from the active participation of virtual community members (Branzov 2016). Another key feature is the semi-automatic alignment and subsequent concordancing of the translations, which allows teachers, for example, to visualize how different students have translated the same ST segment. As regards translation correction, the standardized Translation Annotation System integrated into Hypal4MUST guarantees a high degree of comparability across translation tasks and across teachers, and gives teachers the possibility of drawing up profiles of common translation and language errors produced by individual learners or learner populations. The MUST community currently includes 34 partner teams from 17 countries. The database contains 110 STs, representing both general and specialized language, and 1,100+ student translations, representing 20 language pairs, together with rich metadata. TTs, for instance, contain source-text-, translation-task- and student-related metadata. As the interface is used by teachers and students a rich learner translation corpus is effortlessly created and keeps growing. It currently contains 425,000+ TT tokens. Such a corpus is a most valuable resource for translator training, which teachers can use, for example, to design exercises that are tailored to their students' attested difficulties (see e.g. Espunya 2014).
- Affiliations
Citations
Granger, S., Lefer, M.-A., & Adam Obrusnik. (2019). Hypal4MUST: A community-based web interface for translation teaching. EUROCALL 2019, UCLouvain. https://hdl.handle.net/2078.5/216539