Majority Judgment for Collective Beliefs and Science

When attributing a claim to a group, should one consider that the larger number of agents who believe P rather than Q is epistemically relevant, and thus "go with the numbers"? The classical view says one should. The thesis defended here is one should not, as soon as agents are able to assign qualitative grades to propositions. At the group level, the aggregation should be made with the Majority Judgment rule. This paper extends it from preferences to belief and science, with significant consequences for the philosophy of expertise and testimony.