Abstract
Humans are remarkably adept at intuiting whether an object will fall or not. Adults, and even infants, judge object stability across various physical scenarios. Yet the computations involved in making such judgments remain poorly understood. Much research has focused on model-based simulation approaches (e.g., Battaglia, Hamrick, & Tenenbaum, 2013), but the perceptual primitives required as input to such models have been relatively unexplored. Here we tested whether the geometric centroid of objects plays a role in stability judgments. In three experiments, children (ages: 5 to 7 years) and adults (college students) judged the direction of falling for an unstable object. The Point of Subjective Equality (PSE)—the critical angle at which the object is equally likely to fall to either side—was modeled for each object. In Experiment 1, participants judged stability for objects of different aspect ratios. Participants’ PSEs scaled according to the centroids of the objects. However, because height and base size also differed across objects, these cues, not the centroid, may have allowed for judging stability. In Experiment 2, we equated base size across objects to rule out the exclusive use of this cue. Nevertheless, participants’ PSEs varied across objects, suggesting that the centroid (or height) of the objects was used for judging stability. In Experiment 3, we used objects that differed in height (identical base size and centroid) and we found no difference in participant’s PSEs, consistent with participants judging stability via the centroids. Moreover, even when objects differed only slightly in their centroids, adults’, but not children’s, PSEs for the objects differed, suggesting a developmental difference in the precision of the estimated centroid for object stability judgments. Taken together, our results demonstrate a perceptual mechanism underlying physical judgments that improves in precision across development.