Quantitative Physical Modeling of Physics Cognition

Abstract

Current theories of physics cognition require specification of complex mechanisms for explaining knowledge acquisition. I demonstrate that a quantitative model of physics perception can be constructed by assuming that physics problem perception provides an optimal summary of a set of physics problems. In so doing, I offer the first model of physics cognition determined from a single optimization principle.

I use the model to produce categorizations of physics problems according to surface features. These categorizations suggest, contrary to previous claims, that surface feature perception may in fact be a productive resource for novices: it may provide access to "deep" knowledge originally considered by influential studies as accessible only to experts. The model suggests a potential explanation for why novices often focus on surface features: novices may simply be responding to a set of physics problems in which the surface features of those problems provide relevant information for problem solving.

The model predicts that the initial perception of a physics problem is characterized by the identification of the "surface feature context" of which the problem is a particular example. I use the model to predict potential contexts that individuals may perceive when confronted with a physics problem. Experiments that focus on the initial perception of a physics problem have not been considered previously; I use the model to encourage the construction of such experiments. I speculate that the large differences in novice and expert physics problem classification originally highlighted in influential experiments could reduce substantially when experimental observation is restricted to the initial, perceptual stage of physics problem classification.