David Braddon-Mitchell

The paper makes three points about the modularity of folk psychology and the significance of metarepresentation: The hope that metarepresentation may provide a principled divide between intentional and merely representational systems focuses on a divide of mechanism. I suggest that we also look for a divide of task: the difference could be a principled difference in the task performed by the systems, not in how the task is performed. There is no incompatibility between the hypothesis that folk psychology is modular and the observation that we have conscious access to some of the principles of folk psychology on even a slightly moderated understanding of modularity. Much of the debate about the theoretical nature of FP hangs on what counts as a theory. Different conceptions of theory form a continuum, and it is unclear what hangs on the position that FP occupies within it.

How can the reference of theoretical terms be stable over changes of theory? I defend an approach to this that does not depend on substantive metasemantic theories of reference. It relies on the idea that in contexts of use, terms may play a role in a theory that in turn points to a further (possibly unknown) theory. Empirical claims are claims about the nature of the further theories, and the falsification of these further theories is understood not as showing that a term in the original theory fails to refer, but rather that a scientific hypothesis encapsulated by the further theory is mistaken. Introduction 1.1 Two theories involving ‘cat’ 1.2 Some objections 1.3 Doing it with matrices 1.4 The theory so far Approximate truth 2.1 Modes of reference 2.2 Papineau's solution Meaning, analyticity and verbal dispositions 3.1 Theory 1 and the folk theory 3.2 Is the armchair too comfortable? 3.3 What role for the armchair? 3.4 How can we be mistaken about our dispositions? 3.5 The theory of reference and the theory of meaning.