Talk (Work in Progress): Hannah Pillin (MCMP)

Title:

Fragmentation and Imprecise Probabilities

Abstract:

In recent epistemology, fragmented accounts of doxastic attitudes such as belief or credence have been receiving attention. The basic idea is that agents’ doxatic states are allowed to be fragmented instead of global. Assuming a broadly Bayesian epistemology, this means that, instead of modeling an agent’s doxastic state via one single credence function that adheres to the central norms of Bayesianism, the agent may be represented as having a set of multiple credence functions, each representing a different fragment of the agent.

Interestingly, there is another stream of literature that has nothing to do with the idea of fragmentation but uses exactly the same formal framework - imprecise probabilism. Imprecise probabilism has the goal of providing a tool for modeling whenever agents cannot assign a precise credence to a proposition.

In this talk, I will present both accounts, and work out interesting similarities and differences between those two theories that are identical in formalism but aim to model significantly different phenomena (or maybe not so significantly different, we’ll see).