Boundedness in event cognition: Viewers spontaneously represent the temporal texture of events

A long philosophical and linguistic literature on events going back to Aristotle distinguishes between events that are internally structured in terms of distinct temporal stages leading to culmination (bounded events; e.g., a girl folded up a handkerchief) and events that are internally unstructured and lack an inherent endpoint (unbounded events; e.g., a girl waved a handkerchief). Here we show that event cognition spontaneously computes this foundational dimension of the temporal texture of events. People watched videos of either bounded or unbounded events that included a visual interruption lasting either 0.13 s (Experiment 1) or 0.03 s (Experiments 2 and 3). The interruption was placed at either the midpoint or close to the endpoint of the event stimulus. People had to indicate whether they saw an interruption after watching each video (Experiments 1 and 2) or respond as soon as they detected an interruption while watching each video (Experiment 3). When people responded after the video, they were more likely to ignore interruptions placed close to event endpoints compared to event midpoints (Experiment 1); similarly, when they responded during the video, they reacted more slowly to endpoint compared to midpoint interruptions (Experiment 3). Crucially, across the three experiments, there was an interaction between event type and interruption timing: the endpoint-midpoint difference depended on whether participants were watching an event that was bounded or unbounded. These results suggest that, as people perceive dynamic events, they spontaneously track boundedness, or the temporal texture of events. This finding has implications for current models of event cognition and the language-cognition interface.