A Limitation of Evidence-Backed Communication

A conventional starting point in disclosure games is that messages can be verified by evidence. Yet, as is often the case in political debate, when the same evidence is consistent with multiple hypothetical assertions, disclosure never conclusively verifies any of them. Using an evidence-specific communication model, the paper characterizes this limitation on verifiability in terms of the players' prior information. The main result is that, from a state of ignorance, with little prior information about how pieces of evidence correlate, a receiver should optimally ignore hypothetical supersets and accept evidence at face value. This principle extends to equilibrium in strategic disclosure games, both with and without sender competition. Finally, the paper maps accepted evidence into probability spaces, situating the results within a broader theoretical framework of belief formation.

Link to the paper: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5186252
Avinat Gottesman's homepage: https://en-econ.tau.ac.il/profile/avinatg_03