by Ian Apperly and Stephen A. Butterfill
We argue that mindreading involves not only multiple systems but also multiple models (or theories) of the mind. The argument turns on three questions. First, how could mindreading be both flexible and efficient? We suggest that, in a sense it can't. Instead mindreading involves two or more systems; some trade efficiency for flexibility and others make the converse trade-off. But how could mindreading--which paradigmatically involves constructing reason-giving, causal explanations for actions by appeal to mental states with arbitrarily nestable contents and uncodifiably complex functional roles--ever be efficient? One possibility is that mindreading sometimes involves unsophistcated but useful models of the mind. If this is right, we face a third question. How can hypotheses about which model of the mind a mindreader is using be tested? Unsophisticated models have signature limits, and these limits may make it possible to identify the operation of a given model across contexts and across types of subject.