SOPhiA 2019

Salzburgiense Concilium Omnibus Philosophis Analyticis

SOPhiA ToolsDE-pageEN-page

Programme - Talk

Inference to the Best Bayesian: On the Prospects of an Abductive Updating Rule
(Epistemology, English)

How should we update our degrees of belief in hypotheses in response to new evidence? Bayesians maintain we should update via Bayes' Rule (BR), whilst proponents of inference to the best explanation (IBE) contend explanatory considerations should play a role. Recently, some proponents of IBE have formulated IBE such that it is incompatible with Bayesianism. The question is then: which updating rule should we accept? This paper argues we ought to update via (BR) over the existing incompatibilist formulation.

Following Schupbach (2017), __2 shows there is a single adequate formulation of the existing incompatibilist proposal (henceforth, '(Schupbach)'). __3 presents Climenhaga's (2017) objection that updating via (Schupbach) leads to incoherent credences when we consider multiple levels of explanation. In response, I show his argument relies on agents considering the hypotheses at one level of explanation independently of the hypotheses at the other levels of explanation, even though the two levels are interdependent. I independently motivate the claim that agents ought to consider a single complex partition comprised of conjunctions of hypotheses from the different levels of explanation and show, if agents consider this complex partition, Climenhaga's objection no longer arises.

___4 presents data collected by simulating two agents -- one of whom updates via (BR), and the other by (Schupbach) which shows (BR), as an updating rule, is both more accurate and faster to converge to the truth than (Schupbach). Finally, despite some of its proponents (cf. Lipton, 2004, p. 92) claiming IBE avoids the confirmation problems which face Bayesianism, such as the problems of irrelevant conjuncts and grue, __5 proves (Schupbach) is equally susceptible. I therefore conclude IBE's proponents must either find an alternative incompatibilist precisification, or pursue a compatibilist approach to IBE's relation to Bayesianism.

Time: 14:00-14:30, 18 September 2019 (Wednesday)
Location: HS E.002

James Skinner 
(University of St Andrews, United Kingdom)

I am currently an MPhil student at the University of St Andrews, and I studied at Worcester College, University of Oxford for my undergraduate degree. I am presently interested in formal epistemology, justice, and animal ethics.

Testability and Meaning deco