Suppose I have friends who are my epistemic peers who have the same abilities that I do, and work with the same knowledge-base, but who disagree with me about some claim that I accept; let's call that claim C. Now, since they are working with the same evidence that I have, and have the same abilities, and they come to a different conclusion, one might say that because of this I should suspend belief in C, or at least not accept it as strongly, because the fact that my epistemic peers deny C makes it plausible that I have gone wrong somehow; given that we are epistemic peers, equal weight should be given to our judgments. Is this right?
No, in the case of suspending belief; because if this reasoning is right, given that the people in question are my epistemic peers, I can apply it to them. For if A is an epistemic peer of B, B is an epistemic peer of A. Since I am their epistemic peer, by the above reasoning, the fact that I, their epistemic peer, accept C makes it plausible that they have gone wrong somehow in rejecting it. Since nothing can undermine a claim if its undermining that claim undermines itself, the reasoning must be wrong.
Thus the mere fact that someone, who is my equal in ability and is looking at the same evidence I am, disagrees with me gives me no reason to suspend belief in my views. The question is, as it always was: Which of us has in fact made the better evaluation in this particular case?
Now, one variant of this might be to say that when I and an epistemic peer disagree in this way, we should both revise our credence to a midpoint between our two original credences. Now I don't think that credences can be revised like this, since I don't think there are degrees of belief, just degrees of a few things sometimes associated with belief for various reasons (e.g., willingness to act). But let's set this aside. Since we are supposing that this other person is a peer and not a better, the mere fact that she is my peer gives me no reason to think her inference from the evidence is likely to be better than mine. But if it is not likely to be better than mine, there is no reason for me to defer to it unless I have reason to think that in fact it was better than mine in this particular case.
A common way to argue that we should revise our credences is to take mathematical examples. Suppose I and my epistemic peer are doing a math calculation in our heads, and I come up with the conclusion, of which I am very certain, that the result is 42, while my peer comes up with the conclusion, of which she is also very certain, that the result is 41. In a situation like this, someone might say, I should revise downward my confidence that it's 42, and upward my confidence that it's 41 (and she should do the reverse). I think this is obviously false. In a case like this, what has happened is that a checking mechanism (my epistemic peer) has failed to confirm an inference. The rational thing to do is to go back over the calculation; and nothing about this requires that I change my confidence unless I find clear evidence that I went wrong in the first place. Thus, my disagreement with my epistemic peer simply raises the question of whether I accidentally made a mistake of which I was not aware; if I'm pretty certain that I did not (e.g., if I go back and check), this question is answered. So, again, the question is what it always was: Which one of us has made the better evaluation in this particular case?
Let's take a different case. Suppose I and my peer are watching the races at Ruidoso Downs. We're both good perceivers, we're both watching closely, and in a very close finish you think Horse A won while I think Horse B won. Shouldn't we think, given that we are epistemic peers, that we are equally likely to be correct?
No. The plausibility of the argument lies in this: that if someone had no other evidence for whether Horse A or Horse B won than the fact that I said B and my peer said A, then he should conclude that we are equally likely to be correct. This is precisely because we are peers. However, I am not deciding the question of whether Horse A or Horse B won on this evidence; I am deciding it on the fact that my best judgment as to what I saw is that I saw Horse B winning. The only question raised by the disagreement of my peer is whether I might have overlooked some possible arena of mistake. If, on serious reflection, I'm fairly certain I didn't, I have no reason to revise my confidence in the claim that Horse B won.
Thus epistemic peers are useful for checking my answer; but they are no more (and no less) relevant to my level of confidence in my answer than if I were to check my answer soberly and carefully myself. Suppose I'm doing a math problem, and get the answer 41; and then I do it again and get the answer 42. The reasonable thing is not to say, "Oh, they're equally likely, so I should revise my confidence in the first so that it's halfway beteen its initial state and my confidence in the check," but to do it yet again. If the double-check comes up 41, I ignore the 42 as a glitch. If the double-check comes up 42, I ignore the 41 as a glitch. Perhaps, if the situation really warrants being very careful, I might triple-check before I do this. In each case the revision arises only from the further inquiry into whether I made a mistake. The mere fact of disagreement is not relevant.