Not if they’re sufficiently different. Even within Bayesian probability (technically) we have an example in the hypothetical lemming race with a strong Gambler’s Fallacy prior. (“Lemming” because you’d never meet a species like that unless someone had played games with them.)
On the other hand, if an epistemological dispute actually stems from factual disagreements, we might approach the problem by looking for the actual reasons people adopted their different beliefs before having an explicit epistemology. Discussing a religious believer’s faith in their parents may not be productive, but at least progress seems mathematically possible.