A corpus-based investigation of the phraseology in various genres of written English with applications to the teaching of English for academic purposes

This thesis describes an investigation into the various kinds of fixed expressions or prefabricated language which occur in certain genres of written academic English. A
basic premise is that language users, when they write, remember phrases as much as, or more than, they compose them. Although the grammar of a language licenses the
use of a variety of forms to express any proposition only a small subset of these grammatically possible locutions are considered natural and native-like. It is demonstrated that prefabs, my preferred term for prefabricated language, serve many functions in written academic discourse. Their use helps the writer to generate idiomatic text which meets the expectations of the reader and shows clearly the
writer's discourse community. Prefabs also enable the configuration of writer's stance more explicitly. At the same time, thanks to prefabs the reader is better able to
navigate within the text.

The main focus of the investigation is the use by non-native speakers of English of prefabs in writing. Corpus linguistics is presented as the most appropriate
methodology for this investigation. Two main kinds of corpora are constructed: an experimental corpus of argumentative essays produced by Portuguese `apprentice'
English for Academic Purposes (EAP) writers and three control corpora of texts of comparable length, by different categories of native speakers of English. The kinds
and frequencies of prefabs in the corpora are measured and tabulated. Finally, recommendations are made on how to use the findings of the research to improve EAP teaching and learning programmes.