Revisiting the notion of translation universals through L2 written production: theoretical and methodological issues

Taking as a starting point an overview of the notion of ‘translation universals’, this article describes the resources developed for an ongoing research project which adopts a corpus-based methodology to investigate the extent to which phenomena that are claimed to be typical of translated texts – e.g. explicitation, normalisation, levelling out and simplification – are also found in L2 written production. The article argues in favour of a re-examination of the concept of universals narrowly understood as properties possessed uniquely by translated texts (whereby universal features are seen as superficial traces resulting from the constraints imposed by interlingual textual transfer). It is hypothesised, instead, that for a given language pair (Italian and English in the case of our project) these patterns may be similar to those found in non-mediated non-native written language production. The testing of this research hypothesis poses a number of novel challenges in terms of corpus data collection and experimental design, calling for a cross-fertilisation of approaches elaborated in the related disciplines of translation studies and second language acquisition. The article focuses on the theoretical and methodological issues that have emerged so far with regard to research design, corpus structure and corpus compilation.