Telltale Signs:  What Attention to Representation Reveals about Scientific Explanation

Dorit Ganson
Oberlin College

Full Paper

One of the most devastating objections to explantionism (the view that explanatoriness is a potentially epistemic virtue) is van Fraassen's Bayesian Peter argument, in which he tries to show that allowing explanatoriness to enhance your degree of credence leads to probabilistic incoherence, and is hence irrational.  Though some vulnerabilities of the argument in its stated form have been captured in recent criticisms, in general the extent to which explanationism is susceptible to charges of probabilistic incoherence have been underestimated.  Ultimately, this deep worry about the incompatibility of explanationism and Bayes' theorem is somewhat answerable by making information about explanatory superiority an added constraint on our evolving probability function, in addition to the raw evidence.