In this study, the mini-corpus consisted of ten abstracts of research articles from the field of discourse analysis. The authors of the articles, or at least one of the authors for multi-authored texts, were native speakers of English, as indicated by biographical data in their CVs found on the internet. Abstracts were chosen for our present purposes because they are short, presumably cohesive units. Care was taken to choose texts from the same genre and the same topic: abstracts of research articles written in the field of applied linguistics, mainly in discourse analysis (see Appendix D).