I’ll defer to the paper here, which states
Denition 1 (Coherence). We say that P is coherent if there is a probability measure over models of L such that P(phi)=mu({M:M models phi})
That said, I’m not aware of a reason why we should require P be backed by a finitely additive probability measure.
I’ll defer to the paper here, which states
That said, I’m not aware of a reason why we should require P be backed by a finitely additive probability measure.