A formal approach to the peer disagreement debate in epistemology

Abstract

The newborn field in philosophy known as the epistemology of disagreement emerged about fifteen years ago as a quest for guidance: guidance as to how to handle in the most rational way the many controversies that we encounter in our lives. In this work, I assess the credentials of the most promising candidate for the job: Conciliationism. More precisely: I claim that it does not stand on as solid foundations as one would expect and I put forward a principled way for amending it. The first pillar of Conciliationism is the so-called Defeat assumption according to which a peer disagreement on the proposition Q is evidence that one of the peers has reached the wrong conclusion on it. The second pillar of Conciliationism is Symmetry according to which neither of the peers has any reason for thinking that it was the other agent who made the mistake rather than herself. In this work, I show how on closer inspection both assumptions for Conciliationism are problematic. About the latter: peerhood is crucial for the fulfilment of Symmetry. Yet, the notion of peerhood stands on shaky foundations, undercuts the Symmetry condition and thereby threatens the rationale for Conciliationism. In the first part of this work, I come to rescue its rationale by embarking on an exhaustive investigation of the notion in a precise fashion. About the former: Defeat is puzzling when translated to a credence framework on which there is no such a thing as a wrong or mistaken credence. Of what is the disagreement evidence then, in a graded model? The rationale for Conciliationism is threatened once again. In the second part of the work, I put forward a credence version of Defeat on which the disagreement is taken to be evidence of a significantly inaccurate credence. Not only my proposal restores the conciliatory argument, but it even strengthens it. Does the rationale for Conciliationism so amended stand on more solid grounds? Unlike what one might think, the reply is negative, because a further challenge emerges. I consider it in the third part of the work. The challenge in question comes from the field of information economics in the form of a case of peer disagreement that I call the ‘AB-case’. In the AB-case we know with mathematical precision what two rational agents would do. And they would not conciliate. All the conditions for the straightforward application of the conciliatory solution hold true in it. The agents at stake can be considered as competent epistemic peers for whom the symmetry condition holds. Because their disagreement is evidence of a significantly inaccurate credence, the defeat condition is satisfied. The timeline of their disagreement matches the story from philosophy. Yet, the blind endorsement of the conciliatory solution in the case leads to a contradiction.