In this talk, I offer a new interpretation and evaluation of Hume's theory
of causal inference.  According to my interpretation, Hume argues for his
theory on the grounds that it provides a unified explanation of a wide
variety of experimentally observed phenomena, such as the incremental
nature of causal learning, the "hesitating" confidence which accompanies
probabilistic inferences, and the tendency of human nature to succumb to
the "illusion" of seeing a necessary connection between correlated events.
I then look at some recent evidence from cognitive science which, I claim,
forces us to take Hume's theory seriously.  I argue that this research
gives new plausibility to his explanation of probabilistic inference, yet
it fails to vindicate his "projectivist" account of necessity.  The
philosophical upshot of this is that Hume's naturalistic investigation
into
causal inference does not support the skeptical conclusions he draws from
it.