Algunos problemas de epistemología social: probabilidades grupales dinámicas

ELEONORA CRESTO

Logic and Philosophy of Logic: Recent trend in Latin America and Spain

College Publications

Lugar: Londres; Año: 2019; p. 3 - 23

The article aims to discuss and model a satisfactory concept of group probability; this task can be seen as a particular aspect of the more general question of how to aggregate intentional individual attitudes into a single collective attitude. For the most part my proposal will leave the exact aggregation mechanism for probabilities open. Rather, I aim at solving a previous aspect of the aggregation problem, to wit, I seek to clarify which individual probability functions are to be taken into account at the time of determining the probabilities of the group as a whole. I argue that the probabilities that ought to be aggregated are the individual functions conditionalized on group knowledge. But, as it happens, there are at least two well-defined ways in which we can attribute propositional knowledge to a group: common knowledge, and distributed knowledge. This leads to two different ways in which we can conditionalize. I contend that we should actually focus on the distance between the two; this leads to a dynamic concept of group probability. To address this problem I begin by presenting some standard tools from epistemic logic that enable us to characterize both common and distributed knowledge. I then explain how the machinery of public announcement logics can bridge the gap between the two. Next I present my perspective on dynamic group probabilities in a Kripke framework enriched with probabilities; we will see how to use public announcement logics also in the probabilistic setting. The proposal is further refined by considering a geometric pooling method on the conditional measures, which enables us to establish some connections with technical results from the recent literature.