From naive physics to naive Physics

The goal of the naive physics program may be succinctly stated by the following two tenets: (1) formalize the knowledge an intelligent agent needs to get around, and (2) accomplish this using mathematical logic as a notation. Although the two parts of the program are on the surface independent, the notion is that unless commonsense reasoning is modeled as logic and deduction then the knowledge used in reasoning cannot be separated from the particulars of the reasoning process. McDermott’s article argues primarily against the appropriateness of modeling reasoning as deduction; the assumption is that if reasoning isn’t deductive then the use of logic is inappropriate and the program as a whole fails. Although I accept on psychological grounds, independent of McDermott’s arguments the inappropriateness of modeling human reasoning as deduction, I will argue that this in no way invalidates the naive physics program. It merely changes it to my mind, for the better.