Proceedings of the Annual Meeting of the Cognitive Science Society

Functioning in the world requires information about objects properties. People perceive object mass using perceptualcues when the material is observable. Here, we examine how people predict an objects motion when its material isunobservable, but predictable from cues learned via category learning. When given an ambiguous object, people tend topredict properties based on the propertys propensity in the most likely category. But, recent work has found that givenan ambiguous cue, people will integrate over categories (as rational agents should) in a variety of contexts. In our study,we investigate how uncertainty in categorization affects continuous judgments in the domain of intuitive physics. Weincorporate real materials (like wood and iron) into a category learning framework and test peoples judgments about thedistance a payload travels in two scenarios before and after category learning. Our results are equivocal, but suggest thatpeople do integrate in these scenarios.