What happens in words: Towards a model of incremental event representation

点击翻译按钮获取中文摘要

The central aim of formal semantics is to model how the meanings of sentences are builtfrom the meanings of words, but without concern for the linear order that words appear induring actual speech or writing. Conversely, psycholinguistic research concerns itself withexactly this question - namely, how cognitive processes incrementally (word-by-word) buildmeaning. Given the differences in the methods and goals of these two sub-disciplines,psycholinguistic research has often underutilised the insights from semantic models,particularly with respect to the events described by the words and sentences we use.The way comprehenders construct events when they hear or read a sentence is quitecomplex, and involves several decisions about who is involved in the event, where the eventtakes place, the number of events, and how those events relate to one another. Consider asentence like (1).1) Esther gleefully bought Neil Gaiman's new book with her credit card earlier this month,and Khadija did too - but Jamie sure didn't.Are there two purchasing events and two books described here, such that Esther and Khadijaseparately bought a copy of the new book? Are there two purchasing events and a singlebook, such that Esther bought the book and Khadija bought that very same book at a latertime, perhaps even from Esther? Is the most natural, default interpretation one of a singlepurchasing event where they bought a single copy of the book together? Whose credit carddid Khadija use on each occasion, her own or Esther's? And what of Jamie? Is Jamie's nonpurchasingrepresented as an event in the same way in the minds of comprehenders, or donegated verbs result in different kinds of mental representation?The way we are able to make sense of the worlds we create with language appears to beespecially tied up with how we construct a mental abstraction of the situation described by atext or utterance from the potential underlying events of that sentence or wider discourse.Events themselves have influenced research into topics in semantics as diverse as tense andaspect (how time is encoded in language), adverbial modification (how the syntactic positionof words like 'quickly' subtly affects the meaning of a sentence), and argument realisation(how the interpretive properties and meaning of verbs influence the syntactic contexts theyappear in) - see Bohnemeyer et al. (2007); Lascarides and Asher (1993); Pianesi and Varzi(2000); Rappaport Hovav (2008); Smith (2006); Tenny (2000). Much less has been writtenabout the psychology of these events.While Bott and Sternefeld (2017) showed how insights about events in formal semanticscould, in theory, be applied to incremental interpretations, comparatively little work hasbeen conducted on how comprehenders conceptualise and process event information wordby word, in real-time. The goal of the proposed project is, therefore, to conduct three distinctbut interrelated experiments on the real-time processing of understudied event phenomena,utilising EEG, eye-tracking, and reaction time tasks, respectively. Taking advantage of mybackground in both formal semantics and cognitive processing, and drawing upon thesupervision of researchers with experience in both areas, the present study will bringtogether insights from these often disparate approaches in an interdisciplinary effort toexplore and explain the mechanisms underlying comprehenders' conceptualisations ofevents as speech and writing unfolds.The value of this project is twofold. First, by developing a robust account of how events are incrementally represented in the mind, the groundwork will be laid for future research into a fundamental aspect of language. Second, perceptual psychology has observed that peopleparse continuous streams of visual activity into meaningful event-like chunks, an automatic cognitive process which scaffolds future memory and learning (Zacks and Swallow, 2007).