Abstract: The British Academic Written English (BAWE) corpus is a collection of almost three thousand successful university student assignments from four universities in England.
This paper outlines its composition, and explains how 13 genrefamilies were identified in the corpus, informed by ‘top-down’evidence from the local and national contexts, and a holistic reading of the texts, to identify similarities in purpose and staging across disciplines and levels of study.
It then explains how a bespoke multidimensional analysis was conducted that identifies clusters of lexico-grammatical features intexts, using a ‘bottom-up’ multidimensional analysis. Issues inreconciling these two very different approaches to descriptions of academic writing are then discussed.
The BAWE corpus can be freely accessed by teachers and learners through SketchEngine/open, and can be requested for research purposes from the Oxford Text Archive. It is cited in over 100 publications, in addition to student theses and conference presentations (see www.conventry.ac.uk/BAWE for details).