ABSTRACT
In this chapter, the discourse term will be limited to two basic senses – the formal sense: discourse as connected text and the functional one: discourse as language in use. But the main focus of the chapter will be less on what corpus analysis can tell the reader about discourse as language in use. The descriptive findings are generated by searching for particular discourse features in a corpus – typically a collection of texts of a specific register, but possibly a single extended text, such as a textbook or a novel – using computational means. However, the corpus-based approach at least helps to counter some of the bias, by providing quantitative evidence of patterns that may be more difficult to ignore. Discourse analysis, on the other hand, requires whole texts, often of the same type, as its database. Nevertheless, analysts still lack tools that will perform many of the kinds of operations that are traditionally done manually.
