Modelling eye-movements made in the course of reading syntactically ambiguous sentences

A central goal for work on human language processing is to spell out the characteristics of the transitory mental operations that enable people to make sense of sentences and passages of prose. In this field, much of the core empirical evidence has come from detailed measurements of the eye-movements that individual readers make as they work through carefully crafted samples of text often based on the use of temporarily ambiguous sentences. Using this approach, fine-scale observations of the temporal and spatial patterns of such movements are used to draw inferences about the cognitive and linguistic operations that contribute to the process of comprehension. The present investigation aims to remain securely within this broad research tradition, and in particular to use standard eye-tracking methods for data-gathering purposes. However, in contrast with most previous work, the project will play down the use of informal verbal models, instead assigning a central role to computational models that are precise enough to derive detailed numerical predictions of eye-tracking data. Irrespective of the accuracy of any data-fits achieved, the work is expected to provide an indication of novel features that will eventually have to be built into any successful overarching model of comprehension and eye-movement control