Abstract
Imagine you see a video of someone pulling back their leg to kick a soccer ball, and then a soccer ball soaring towards a goal. You would likely infer that these scenes are two parts of the same event, and this inference would likely cause you to remember having seen the moment the person kicked the soccer ball, even if that information was never actually presented (Strickland & Keil, 2011). We tested whether the spontaneous formation of coherent event representations relies on semantic implication, or a more automatic object-tracking system. In Experiment 1 (N=90), participants were more likely to falsely reported seeing the moment of contact/release in unfamiliar computer-generated animations of novel launching or launching-by-expulsion events (M=68.3%, SD=27.0) than in animations in which the second half of the event was an unrelated “non-sequitur” outcome (M=31.7%, SD=31.4), p<.001. Experiment 2 (N=120) independently manipulated whether the object’s trajectory was visible following the cut, and whether there was an event that seemed to be caused by the object in the second half of the video. Results showed that visible object trajectory alone dictated the filling-in effect (visible: M=69.4%, SD=28.2; invisible: M=27.9%; SD=30.9), F(1,110)=55.29, p<.001, partial eta-sq=.334. Experiment 3 (N=200) used naturalistic videos, and independently manipulated whether the features of the object were changed at the cut (e.g., a dart becoming a balled-up piece of paper) and whether the cut placed the object “too far” along its trajectory, disrupting spatiotemporal continuity. We found that participants recognized that the feature change had occurred (77%), but nonetheless filled-in the moment of contact/release provided there was continuity (M=75.0%, SD=28.6), while continuity disruptions significantly diminished the filling-in effect (M=67.5%, SD=30.8), F(1,196)=6.21, p=.014, partial eta-sq=.031. Together, these findings indicate that the spontaneous formation of event representations is driven by object tracking systems that are primarily sensitive to spatiotemporal continuity.