Abstract

When multilingual corpora are used in translation studies, it is usually assumed that they are either parallel or comparable, or both; and that their size and text composition are analogous. As general reference corpora become more widely available, it is inevitable that these too should be used to compare and contrast SL norms, thus extending the definition of comparability to include text collections whose size and content may vary considerably, and which are nevertheless considered representative of their languages. This paper addresses the contribution of comparable reference corpora to the identification of translation equivalence. Focusing in particular on native-speaker norms, it demonstrates how the effect of creative and idiosyncratic language can be identified and reproduced by the translator.