Faculty Scholarly Dissemination Grants


Eye movements reflect the cognitive reality and costs of event structure during reading


Psychology Department


College of Liberal Arts and Sciences


Social and Behavioral Sciences


Research has shown that when readers perceive an event shift to occur while reading narrative texts, they update their representations of what is happening now (Kurby & Zacks, 2008). This shift in representations from one event model to the next implies the existence of boundaries between events in narratives. For example, readers will slow down their reading rate during situational changes, and they lose availability of previous information after a shift in the situation (Zwaan, 1996). Other research has found that when presented with a narrative text presented one clause at a time, reading rate slows down at clauses that had previously been labeled event boundaries, over and above effects accounted for by other situational changes (e.g., a change of character, or goals) (Zacks et al., 2009). The purpose of the present experiments was to use eye movements to explore the moment-to-moment processing consequences of event cognition during more naturalistic reading of larger chunks of text (i.e., more than one clause at a time). In Experiment 1 (N = 34), participants read a narrative account of a boys day while their eye-movements were recorded. In Experiment 2 (N = 84), we replicated the procedures of Experiment 1 but also assessed working memory capacity. Based on previous work on event cognition (Kurby & Zacks, 2008), we hypothesized that reading times (e.g., first run reading time, first fixation duration, total reading time) should be slower for event boundaries than non-boundaries. Readers should also be more likely to make regressive eye-movements into event boundaries. In addition, if readers update their event models at event boundaries and lose accessibility of previous event information at such points, readers will be less likely to make regressions out of event boundaries than non-boundaries. Finally, if event updating incurs a working memory load, individuals with lower working memory capacities should show stronger effects of event boundaries on reading behavior. We conducted mixed effect models with a continuous event segmentation likelihood variable as the main predictor. We also included terminal punctuation, non-terminal punctuation, and number of characters as covariates. In both experiments we found that reading time was slower at event boundaries clauses, as measured by total fixation time, first run fixation time, and even first fixation time. We also found that readers were more likely to make regressive eye movements into event boundary clauses, but were less likely to regress out of an event boundary clause. In Experiment 2, we also found that working memory capacity interacted with event boundary status for total reading time, such that low working memory readers slowed down at event boundaries, but high working memory readers did not (see figure). It appears that eye movements strongly reflect the cognitive reality of event structure during reading, and also that tracking this event structure indeed carries a working memory load. References: Kurby, C. A., & Zacks, J. M. (2008). Segmentation in the perception and memory of events. Trends in Cognitive Sciences, 12, 72 79. Zacks, J. M., Speer, N. K., & Reynolds, J. R. (2009). Segmentation in reading and film comprehension. Journal of Experimental Psychology: General, 138, 307 327. Zwaan, R. A. (1996). Processing narrative time shifts. Journal of Experimental Psychology: Learning, Memory, and Cognition, 22, 1196 1207.

Conference Name

Benjamin Swets

Conference Location

Columbus, OH, United States

This document is currently not available here.