Designing a corpus for translation and language teaching: The CEXI experience

Building resources for the education of future language mediators is a widely felt need in Europe today. Translation and interpreting schools have experienced a boom in the past two decades (Caminade & Pym, 1995), with the consequent need to adapt language teaching practices and develop translation teaching approaches tuned to these specializa- tions. A number of researchers working in this setting (see, e.g., the collection edited by Aston, 2001) have suggested that an inductive, data-driven (Johns, 1991) approach to learning with the aid of corpora may be effective, motivating, and well tuned to such settings. They have also, however, warned against the risk of adapting pedagogy to technology rather than technology to pedagogy--largely sharing Widdowson's (1984) view that "it is the responsibility of applied linguists to consider the criteria for 'an educationally relevant approach to language' and to avoid the uncritical assumption that applied linguistics must necessarily be the application of linguistics" (p. 19), even corpus linguistics. Recently, multilingual corpora have started to become more widely available, complementing monolingual ones. More costly to build and arguably less relevant to the production of dictionaries and grammars, multilingual corpora tend to be smaller than monolingual ones and more difficult to access, both technically (e.g., in the case of parallel, aligned texts, in which each sentence in the original has to be made accessible in parallel with its translation) and legally (copyright problems are doubled). However, institutions in various parts of Europe (e.g., Finland, Germany, Norway, Portugal, and Sweden) have started to set up multilingual corpora for research purposes (to study translation practices and to do contrastive linguistics analysis) that can also be used for didactic ones. The approaches to building multilingual corpora are as varied as corpus typologies and uses relevant to these fields. This report describes the approach of the Corpus of English X Italian (CEXI) project (see Aston, Bernardini, & Zanettin, n.d.), whose aim is to provide a resource for the study and teaching of the English and Italian languages and of translation between them. In particular, it focuses on issues of corpus typology and design, relating these to intended uses and to current views of language and translation pedagogy.