Wed Oct 12, 2005
234 Moses, 6–8 PM
|Working Group in the History and Philosophy of Logic, Mathematics, and Science
Branden Fitelson (Berkeley)
Old Evidence, Logical Omniscience, and Bayesianism
In this talk, I will aim to do five things (in order): (1) explain what the problem of old evidence is, (2) explain Garber’s resolution of the problem of old evidence, which involves a novel and subtle Bayesian approach to logical learning (i.e., the learning of logical relations), (3) compare and contrast Garber’s approach to logical learning with a previous approach sketched by I.J. Good, (4) offer an alternative approach to logical learning (with an application to the problem of old evidence), and (5) briefly discuss and critique Jeffrey’s approach to logical learning (and old evidence). Along the way, various important foundational observations will be made about Bayesian models of rational epistemic agents.