In this book the consequences of the frame problem for the relation
between cognitive science and folk psychology are investigated. I focus
on classical AI in relation to J.A. Fodor's intentional realism
and connectionism in relation to P.M. Churchland's eliminative
materialism.
The frame problem arises in the context of trying to model the
human ability for non-demonstrative inference (e.g. the ability to
quickly grasp the relevant consequences of an event). I argue that
a negative appraisal of, especially, the 'scruffy' approach
within classical AI to the frame problem is unjustified because it
depends on a failure to distinguish between descriptive and normative
aspects of non-demonstrative reasoning.
Connectionism's use of distributed representations has been
regarded as promising with respect to the frame problem. However,
I argue that doubts can be raised about connectionism's potential
to adequately represent realistically large amounts of complexly
structured information.
Because both the pessimism regarding classical AI and the optimism
regarding connectionism concerning the frame problem have been found
wanting, I conclude that folk psychology still has a fair chance of
being vindicated by cognitive science.
See for some reviews:
T. Clowry (1999).
Seán Ó Nualláin (2000),
Theory & Psychology, 10(1)