A1: Script-based Surprisal in On-Line Comprehension
Project A1 builds on recent behavioral and neurophysiological evidence suggesting that comprehenders rapidly exploit knowledge of typical events to predict upcoming words. Specifically, A1 examines the extent to which more structured knowledge of event sequences determines expectations about what is likely to take place next in the discourse and thus influence surprisal, as indexed by both reading time and ERP measures. The planned experiments firstly seek to shed light on the mechanisms that integrate script knowledge with more explicit linguistic constraints, as revealed by differences in the time-course and relative priority of their influence. Secondly, the project investigates the hypothesis that such script knowledge is highly structured, by examining whether linguistic cues – such as temporal and spatial shifts – modulate comprehender’s beliefs and expectations about which sub-events are likely to be encountered next.