Retour vers liste

Détail de la contribution

Auteur: Phillip WALLAGE

Pragmatically marked ‘emphatic’ negation and its contribution to the Middle English Jespersen Cycle

Abstract/Résumé: Many accounts of the Jespersen Cycle suggest that secondary negative markers start out as emphatic. Van der Auwera (2009) proposes to distinguish emphatic and neutral negatives in the Jespersen Cycle. He suggests that the Jespersen Cycle involves a pragmatic change - an emphatic negative marker becomes the default marker of sentential negation, as shown in (1) for English. (1) Stage one ne (neutral) Stage two ne...not (emphatic) Stage three ne...not (neutral) Stage four not (neutral) In order to test this model of the Jespersen Cycle against historical data, we first need to define what it means for a negative marker to be emphatic in pragmatic terms. I adopt an approach based on a distinction between propositions that are given or accessible to the addressee and those which are new or inaccessible to the addressee, following Schwenter (2006), Hansen (2009), Hansen & Visconti (2009) and Larivee (2010). Quantitative data from diachronic corpora of early English –the PPCME2 (Kroch & Taylor 2000) and the YCOE (Taylor et al. 2003) – provide empirical evidence that this pragmatic distinction underpins the early stages of the Jespersen Cycle. This is a tendency which can be quantified. Therefore, we can estimate the constraining effect of these pragmatic functions on a writer’s choice of ne or ne...not probabilistically, using logistic regression analysis. Crucially, these estimates of probability are independent of change in the overall frequencies of ne and ne...not over time. Therefore, the probabilistic effect of pragmatic constraints at successive periods can be compared directly, irrespective of the overall increase in the frequency of ne...not during the 13th and 14th centuries. Under this analysis, pragmatic unmarking of ne...not will manifest itself as weakening of the pragmatic constraints on its use over time. However, the regression analysis indicates that these constraints on ne...not remain highly statistically significant and their effect highly consistent while ever ne and ne...not are in competition. Two conclusions follow. First, the increase in overall frequency of ne...not, and the corresponding decline in the frequency of ne are independent of the pragmatic constraints I describe here. Second, a comparison of the constraints on stage two ne...not and stage three not indicates that not does not become pragmatically unmarked until it becomes independent of ne at stage three. Hence the loss of ne in the 13th and 14th centuries plays a central role in the grammaticalisation of not as a pragmatically unmarked negative marker, such that in some pragmatic environments, stage three not may replace stage one ne directly without an intervening bipartite ne...not stage.