Using Corpora as a Resource in Language Teaching

The paper begins by considering the need to provide a learner-directed computer package to help mature students to perfect their English. The information they require to do this has two components: expert procedural knowledge and expert domain knowledge. Expert procedural knowledge encompasses the lexical, grammatical and discourse skills involved in writing good English. Expert domain knowledge is the collection of documents from different genres that is accepted by the native speaking community. This collection is dynamic. Three components of special interest to the implementors of Computer Assisted Language Learning (CALL) packages are dictionaries and thesauri, tagged corpora, and discipline-specific documents generated by scholars or professionals in a specialist field. The paper considers, with examples, how corpora of each type can best contribute examples and exercises to suit learners of a wide range of abilities. Examples are drawn from the Oxford Advanced Learner's Dictionary (OALD) and the Susanne Corpus. Dictionaries excel in the provision of general lexical exercises, but are inadequate for grammar or discourse. Tagged corpora are unreliable for some lexical examples but are superb for clause level grammatical exercises and simple discourse exercises that rely on easily recognizable frameworks such as enumeration. Neither of these prime sources is adequate for specialist vocabulary or domain-specific document structures such as laboratory reports which require resort to specialist dictionaries and texts.