Tying it all together: Observers increasingly tune in to structure across repeated viewings of novel activity

Kosie, J. . & Baldwin, D.

University of Oregon

Everyday intentional activity possesses inherent structure, in that actions or events succeed one another with varying degrees of predictability. For example, an act of vegetable chopping is more likely to follow knife-grasping than refrigerator-door opening. When predictability is high, small-scale acts cohere into larger event units during processing. Conversely, low-predictability junctures yield a sense of boundaries between distinct events. Previous research with novel, artificial activity sequences documents that observers' processing rapidly reorganizes across viewings; attention increasingly targets boundary regions where one event transitions to the next. Use of such artificially designed sequences offers the benefit of complete control of predictability structure; however, it is not clear how such findings generalize to naturalistic activity sequences that are considerably more complex. Using Hard, Recchia, and Tversky's 'Dwell-time Paradigm,' we investigated attentional reorganization when participants learned to enact a novel, naturalistic activity sequence across repeated self-paced viewings. Preliminary evidence indicates that observers a) amplified attention to a specific region depicting motion critical to goal attainment, b) reduced attention to the rest of the sequence, and c) progressively reorganized attention to target event boundaries. These findings illuminate how observers rapidly tune-in to structure as they gain fluency in processing complex novel action sequences.