Flavia Cignini
Roma Tre University
Open-mindedness (OM) is generally regarded as an epistemic virtue, yet its definition remains unstable: it is unclear whether it consists primarily in a willingness to consider alternatives, in the management and assessment of sources, or rather in a constraint on the revision of already acquired beliefs.
This work proposes a reformulation of its normative core within a framework of translation between traditional epistemology and Bayesian epistemology. Its central claim is that OM concerns neither primarily the number of alternatives considered nor simply engagement with different perspectives, but rather the epistemic status retained by discarded hypotheses: to be open-minded is to avoid irreversibly closing the space of revision. This conceptual core is formalized by means of a minimal constraint inspired by the idea of absolute continuity between credence measures. The methodological contribution of the paper consists in translating this normative constraint into a Bayesian toy model designed to make different structures of residual openness explicit and comparable.
More specifically, the model treats open-mindedness as a question of how residual probability mass is distributed over hypotheses that have been provisionally discarded but not eliminated. It then compares three possible distributions: a uniform model, an Occamian model, and a proximity-based model. By varying both the prior confidence in the favoured hypothesis and the strength of incoming disconfirming evidence, the model allows us to visualize, through posterior dynamics, how different allocations of residual mass generate different patterns of corrigibility. The aim is not to provide a definitive formal measure of open-mindedness, but to show how distinct forms of OM can be distinguished by the kinds of revision they make possible: local revision, radical revision, excessive rigidity, or excessive instability.
Chair: tba
Time:
Location:
SOPhiA 2026 