Reasons, Coherence, and Group Rationality

If groups can have beliefs and other attitudes of their own, what determines which such attitudes the group rationally ought to have? A widespread presupposition is that group-level beliefs should be a function of the beliefs of the group's members, and similarly for other attitudes. But a host of impossibility theorems show that no such aggregation function can satisfy intuitively attractive constraints while ensuring coherent group-level attitudes. I argue that this presupposition is false. Group-level attitudes should be a function of group-level reasons (evidence, in the epistemic case), not individual-level attitudes. This allows for a theory of group rationality that (i) bypasses a host of pessimistic results in the literature on judgment aggregation and (ii) treats rational individual-level attitudes and rational group-level attitudes in parallel.