ABSTRACT
What are people doing when they use a mental model? As psychologists, we are attempting to understand how people learn and reason about physical systems. We are trying to describe in detail the event of human reasoning. We believe that such a description is necessary to guide the development of theories of human reasoning and of formalisms intended to capture the salient aspects of the way people reason about physical mechanisms. This chapter is exploratory and descriptive. We derive a definition of mental models from a set of distinctions made by de Kleer and Brown (de Kleer, 1979; de Kleer & Brown, 1981) and then employ it to examine protocols taken as a subject answers questions about a heat exchanger. Our concern is primarily with the descriptive and predictive power of the models we evolve based on this definition. Do they predict the subject's errors, the deductions he makes, and the objects and parameters he refers to? Although we formulate a detailed model, which accounts for a substantial portion of our subject's reasoning behavior, we do not intend this model to be the primary content of our exploration. We use modeling as a discipline to help us pull apart the structure of the subject's reasoning.
