Testimony

My recent trip to Chicago was a success, and I had a great time at the Disagreement Conference¹. While there were many excellent presentations–with Roger White, Sherri Roush, and Branden Fitelson’s presentations (in order of appearance) being my favorites–, I have one persistent worry that I noticed unaddressed throughout the conference. In other words, when considering the epistemic significance of disagreement, the presenters tended to bracket the considerations by excluding those who were not considered epistemic peers in some form or fashion. This kind of bracketing took many different forms, from considering the general reliability of the testimonial reports of the agent to whether the agent was epistemically virtuous. The worry I have is this: when considering whether one is an epistemic peer, it would seem that the primary method of analysis is that of attributing a dispositional property, and when attributing such a property, one tends to run into the problems of the conditional fallacy. In what follows I will take an unrefined notion of epistemic peerhood, requiring only that it be considered some form of dispositional property of an agent, and look at several possible sharpenings of a familiar thought experiment in the disagreement literature which will serve to develop some problems for such a dispositional analysis.

A dispositional analysis of epistemic peerhood can be stated thus:

P. Necessarily, a subject S is considered an epistemic peer iff were S to be in some context of disagreement, some dispositional property M would emerge.

The dispositional property involved here will differ with the particulars of the theory in question. To simplify the exposition, I will consider the manifestation property of peerhood to be reliability of testimonial reporting relative to the kind of dispute in question.

Now, for the example, I will use the classical disagreement over a restaurant bill:

Suppose agent A and agent B are splitting the bill at a restaurant. Both agents see the price of the bill, and, on the basis of some quick mental math, come to the conclusion that a certain amount is owed. Agent A asserts that the amount owed is $42, while agent B asserts that the amount owed is $45.

In this case both A and B have access to a shared set of evidence, the cost of the bill, and, given that the numbers do not match, they are in a context of disagreement. To demonstrait the finkishness of epistemic peerhood, it will not be necessary to stipulate which member of the disagreement is correct, but we will consider A to have both sound mathematical reasoning abilities and to be a generally reliable purveyor of testimonial knowledge. In what follows, I will provide three separate cases in which B’s disposition to be A’s epistemic peer is finked. First, B’s disposition will be masked, and then two separate cases of dispositional mimicking will be considered.

For the first case, suppose B is generally reliable in mathematical calculations, has no problems with mental math, and in fact preforms calculations of the sort required for splitting the bill every day without error. Further, B is able to record her calculations reliably in her mathematics notebook. It just so happens, however, that due to randomly occurring crippling social anxiety, the majority of the time B is interacting with other people, her mathematical acuity diminishes severely. When this social anxiety occurs, B can only provide approximate yet unreliable guesses when considering mathematical problems such as that provided above. Further, it just so happens that in all actual cases of disagreement, B has been beset by said social anxiety.

For the second case, suppose that B is actually an incompetent mathematician, such that even the most simple arithmetic problems are incomprehensible. B has, however, survived thus far by a series of incredibly lucky random guesses. B has never mentioned to her so-called peers that she is incompetent, and has a demonstrably better than average track record in producing solutions to arithmetic problems in cases of disagreement.

In the final case, suppose that B is a pathological liar, and while her mathematical abilities are functioning properly, she is not a reliable purveyor of testimony. As it so happens, however, in all previous contexts of disagreement she has based her lies on false beliefs due to accidental error, and the majority of the testimony itself has been true.

Thus, in the first case, while B has the disposition transmit reliable testimony regarding her mathematical calculations, it just so happens that in contexts of disagreement she fails to do so. In the second case, while B does not have the disposition to transmit reliable mathematical testimony, she has so far been succesfull and reliable in such transmission. And finally, in the third case, while B has the disposition to transmit unreliable testimony, she has heretofore been unsuccesfull in her endeavors.