Introduction

Chapters 2 and 3 focused on changes to ne at successive stages of the Jespersen Cycle. This chapter focuses on changes affecting not. It characterises the grammar and functions of not at successive stages of the Jespersen Cycle and integrates them into a model of the Jespersen Cycle as grammatical competition.

The negative marker not develops through grammaticalisation of the negative argument na wiht ‘no creature’ (Jespersen, 1917).Many accounts postulate an intermediate adverbial not stage within the grammaticalisation process so: negative argument nawiht > negative adverb not > negative marker not, as in (62). What functional and syntactic steps comprise this reanalysis, and how are these manifest in the distribution of not in corpus data?

The analysis of bipartite negative forms, such as Middle English ne…not, proposed in Chapters 2 and 3 allows for two potential configurations – (62b) in which ne is a negativemarker and not a concordant negative item, and (62c) in which not is a negative marker and ne is a concordant negative item.

From an empirical standpoint, the issues here are first, what evidence there is to distinguish stage two from stage three of the Jespersen Cycle; and second, how the grammatical properties or functions of not change as it becomes a negative marker at stage three. Here, I argue there are two forms of adverbial not at stage two whose distributions distinguish them from the sentential negator not at stage three: one a focus marking adverb, the other an adverbial minimiser. Both the focus marker not and the sentential negator not represent reanalyses of the OE minimiser not.

These changes in the distribution of not correlate with competition between ne1 and ne2 in exactly the way (62) predicts. Chapters 2 and 3 argued that the changes in the distribution of ne are consistent with the loss of its ability to mark sentential scope negation at stage three. The reanalysis of not and competition between ne1 and ne2 are interlinked: the reanalysis of not brings not into competition with ne1, resulting in changes to the distribution of both ne and not which are entirely predictable if ne1 and not are functionally equivalent and ne2 is functionally distinct from both.