Comparing across languages in corpus and discourse analysis: some issues and approaches

Corpus-assisted discourse studies is, by its nature, interdisciplinary. However, this need to reach across borders becomes even more salient when we study discourses across languages, and this represents a natural intersection with translation studies. The aim of this paper is to reflect on the issue of comparison in cross-linguistic corpus-assisted discourse studies, positing a series of key questions including: How do we compare across or within corpora containing different languages? How do we identify meaningful language units for comparison in this context? How do we know that we are comparing like with like? Using a series of case studies, we start by addressing how we can approach comparison at the lexical level. We then move on to consider methods which allow us to abstract above the lexical level using three case studies which illustrate the use of semantic fields, discourse frames and rhetorical features. By presenting some issues and partial solutions regarding comparison across and within multilingual corpora, we hope to initiate a productive discussion in which we will also be able to collectively enrich and inform this set of resources.