A Novel Solution to the Problem of Old Evidence

One of the most troubling and persistent challenges for Bayesian Confirmation Theory is the Problem of Old Evidence (POE, Glymour 1980). The problem arises for anyone who wants to model confirmation and theory appraisal in science by means of Bayesian Conditionalization. This paper addresses the problem as follows: First, I clarify the nature and the varieties of the POE, following Eells (1985, 1990). Second, I analyze solution proposals where (i) confirmation is evaluated relative to a counterfactual credence function; or (ii) confirmation occurs through learning the proposition that theory T accounts for evidence E (Garber 1983; Jeffrey 1983; Niiniluoto 1983). Third, I present a novel solution that combines previous attempts while making weaker and more plausible assumptions. Finally, I summarize my findings and put them into the context of the general debate about POE and Bayesian reasoning.