Proceedings of the Annual Meeting of the Cognitive Science Society

A common dependent measure used in visual-world eye-tracking experiments is the proportion of looks to a visuallydepicted object in a certain time window after the onset of thecritical stimulus. When interpreting such data, a common as-sumption is that looks to the object reflect the listener’s beliefthat the object is the intended target referent. While this isintuitively plausible (at least for paradigms in which the taskrequires selecting a referent), relatively little is known abouthow exactly the proportion of looks to an object is related toa listener’s current belief about that object. Here, we test asimple, explicit linking hypothesis: the proportion of looks toan object correlates with the probability that the listener as-signs to the object being the target. To test this hypothesis,we supplement the eye-tracking data from Leffel, Xiang, andKennedy (2016) with an offline incremental decision task tomeasure participants’ beliefs about the intended referent at var-ious points in the unfolding sentence, and assess the extent towhich these beliefs predict the eye-tracking data. The resultssuggest that the degree to which an object is believed to be thereferent is only one factor that affects eye movements in ref-erential tasks. Preliminary free production data we have col-lected for the scenes suggests that utterance expectations alsoplay a role. We discuss methodological implications of theseresults for experimental linguistics.