Introduction

This chapter examines the role of functional change within the Jespersen Cycle. It investigateswhether the diachronic stages identified in Chapters 2–5 perform distinct pragmatic or discourse functions during the periods when they compete, and whether changes in their functions explain changes in the distributions of ne, ne…not and not during the course of the Jespersen Cycle. In particular, I examine how functional change contributes to the grammaticalisation of not. Using logistic regression, we can model pathways of functional change within the Jespersen Cycle predicted by different functional analyses. The distributions of negative markers in diachronic corpus data provide an empirical basis against which to test these models, just as they did when we modelled the formal changes within the cycle in Chapter 2.

Within competition models of change, the null hypothesis is that functional change should arise simply as a consequence of morphosyntactic competition, without stipulating any further mechanisms of functional change. Functional constraints should observe the Constant Rate Effect just like the grammatical constraints in Chapter 2 did. Thus the relationship between functional change and the formal model of the Jespersen Cycle as morphosyntactic competition I proposed in Chapter 5 is crucial. I argue once we dissociate functional and structural competitors, in the way a morphosyntactic feature-based analysis allows us to do, functional change arises through competition between forms that are structurally equivalent but functionally distinct.

The idea that functional change plays a role in the grammaticalisation of new negative markers dates back to Gardiner (1904) who says in respect of the French negatives pas and point:

More recently, this idea has been taken up by Detges and Waltereit (2002), Kiparsky and Condoravdi (2006), van der Auwera (2009), Hansen (2009) and Hansen and Visconti (2009).