
Cog Lunch: Vicente Vivanco Cepeda
Description
Zoom Link: https://mit.zoom.us/j/98426780032
Speaker: Vicente Vivanco Cepeda
Affiliation: Tenenbaum (CoCoSci Lab)
Title: Ensemble Physics: Perceiving the Mass of Groups of Objects is More Than the Sum of Its Parts
Abstract:
From playing with marbles, pouring cereal into a bowl, or watching leaves cascade through the air in autumn, humans frequently encounter collections of objects that behave as cohesive groups. Despite the improbability of perceiving and simulating each object individually, people intuitively and accurately predict the behavior of these ensembles. This ability suggests the existence of a perceptual mechanism that extracts ensemble-level properties, enabling judgments about a group’s physical characteristics without relying on detailed representations of its individual components. Research on ensemble perception has demonstrated that people can efficiently encode summary statistics like average size, motion, or emotion from groups of objects. However, it remains unclear whether this ability extends to intuitive physics, where physical properties are not directly observable but must be inferred from interactions in a scene. In this talk, I will present evidence that people do form ensemble representations of mass. Using tasks in which participants judge the mass of a single marble or a set of marbles falling onto an elastic cloth, we find that people judge the average mass of a group more accurately than the mass of an individual. Furthermore, we show that this ability is not merely the result of aggregating individual object properties but instead reflects a distinct perception of group-level properties. These findings suggest that people’s ability to reason about physical systems relies not only on tracking individual objects, but also on forming representations of groups as cohesive entities.