Abstract

In this talk I will present work being done by Ye Tian and myself that attempts to bring together ideas from theoretical semantics/pragmatics and mainstream sentence processing research. Theory says that we should consider interpretation in terms of an update to a kind of shared information state. Psycholinguistics says that language processing is probabilistic and incremental. Incrementalism means that automatic processes take linguistic input together with information in utterance situation to yield (anticipatory) hypotheses about interpretation. Thus, on-line interpretation processes make hypotheses about how shared information is being updated. We argue that such a view is necessary given recent results concerning particularised implicatures. In this talk, I will show how this perspective on language processing sheds new light on well-known results concerning how negation is processed. We show that classic results in verification tasks due to Clark & Chase (1972) and many others results result from the way context update is affected by negation. Our results undermine classic proposition-comparison models proposed to explain verification results. We also show that ‘rejection’ models of negation favoured among embodiment theorists (e.g. Kaup et al 2007) do not make correct predictions in all cases. We propose instead that we can account for how negative sentences are processed by adopting a situation-theoretic approach to negation (e.g. in Cooper 1998) combined with the dynamic-incremental view.