The Good, the Bad, and the Ugly -- Living with Priors in Bayesian confirmation

Abstract

Bayesian confirmation faces a classic problem: where do initial priors come from? In cases with abundant data and repeated updating, different priors tend to converge to the same posterior. However, in frontier research this convergence often fails, and confirmation remains sensitive to priors. We examine how physics practice in the case of gravitational wave research deals with such cases and, normatively, when prior sensitivity should be regarded as epistemically problematic. We offer a practice-based account of the prior problem, so far absent from the philosophical literature on Bayesianism. One upshot is a clearer diagnosis of so-called 'analogue confirmation'.

Contributed paper to PSA2026: The 30th Biennial Meeting of the Philosophy of Science Association (San Diego, CA; 19-22 Nov 2026)