Probabilistic Empiricism

Theme: Metametaphysics and the Sciences

Speaker: Quentin Ruyant (Complutense University of Madrid)

Title: Probabilistic Empiricism (co-author: Mauricio Suárez)

Abstract: Ruyant (2021, ch. 3–5) has proposed a modal epistemology, called Modal Empiricism (ME), that aims at accounting for the possibility to acquire modal knowledge from our past experiences. The view is roughly the following: (1) we should think of objective possibilities not in terms of possible worlds, but in terms of alternative ways local, coarse-grained situations could be in relevant respects, “all else being equal”, and (2) then assuming that observed situations are representative of merely possible ones, claims of (relative) necessity can be justified by simple enumerative induction from experience, which is an “empiricist-friendly” form of induction. However, ME does not say much about probabilities. This is problematic, because the fact that many scientific models are probabilistic is one of the reasons to assume that there are objective possibilities in the world in the first place. Our aim is to fill this gap by extending ME to the probabilistic case, and incidentally, to provide a coherent empiricist epistemology for objective probabilities. After reviewing the main features of ME, we first discuss various interpretations of probability and argue that single-case propensity accounts are more compatible with the spirit of ME. Then we explain that an extension of ME to the probabilistic case faces a difficulty: since the empirical manifestation of probabilities are frequencies, which are characteristics of collections and not of individuals, a simple enumerative induction seems to be blocked. Finally, we show a way out of this puzzle, based on the notion of composite model, and present our final account.

Contact: Florian Marion & Kévin Chalas