Introduction

In Chapters 3 to 5 we have surveyed twentieth-century changes in the structural inventory of standard English. We have noted statistical shifts in speakers' and writers' preferences in those cases in which the system provides options, and in a number of cases – of course, more so in the lexicon than in grammar or pronunciation – we have observed the emergence of new options altogether. We have approached change through contextualized corpus data, but the aim of the description was the reconstruction of changes in the decontextualized underlying system. For example, changing trends in the use of the progressive form were described without systematic reference to contexts of use for fairly abstract constructs such as “American English,” “spoken English,” or “spoken British English,” and not with regard to specific groups of speakers operating in specific communicative contexts, for example, young people trying to formulate polite requests.

The abstract, decontextualized perspective on change is fully justified theoretically and also very useful presentationally, because it has allowed us to present the phenomena in an orderly fashion, moving from the lexicon through the grammar to pronunciation. It is, however, incomplete for at least two reasons. First, the orderly sequence of the presentation has obscured an important fact; namely, that in actual discourse the levels of structural organization constantly interact. Second, it is in discourse, in actual language use, that the experiments leading to structural innovation first take shape.