Introduction

Throughout this book so far we have discussed how corpus evidence can be used to draw attention to features and patterns of words that may not always be noticed by relying on our intuition, however extensive this may be. For example, we have seen in Chapter 1 how information from the concordances for words such as bargain or way may display patterns that tell us about the key partnerships a word has with other words, about the most frequent prepositions it takes or about the kinds of idiomatic functions revealed by its usage. We have also seen that, although we conventionally regard words as single items, they habitually occupy the territory of other words or of strings of words. Sometimes these patterns, if they occur regularly, force us to speak of common collocations, idiomatic expressions and chunks (see chapters 2, 3 and 4). In this chapter, we take an important next step and consider the ways in which words combine to form particular grammatical patterns. A corpus can once again assist us in this endeavour.

A corpus can tell us different things about grammar. It can extend our understanding of traditional grammatical notions and categories, in particular by giving us more information about the distribution of these categories (see below the example of 's not and isn't) or, for example, across specific spoken and written registers of the language (Biber et al. 1999, is a very good example of this latter kind of information).