Two people approach the same evidence on a contested question with comparable intelligence, comparable training, and access to comparable sources, and reach opposing conclusions. Neither has obviously made a mistake; neither is obviously irrational. The case has come, in recent epistemology, to be discussed under the heading of 'peer disagreement,' and the question it raises is awkward enough to have spawned a literature of its own: ought one, on noticing such a disagreement, to lower one's confidence in one's own view by way of accommodation?
The 'conciliationist' answer is that one should, and substantially. If a peer one has independent reason to regard as a peer disagrees, the disagreement is itself evidence that one's own reasoning may have gone astray; treating the peer's testimony as carrying the weight one would otherwise have attached to it, one rationally lowers confidence. The 'steadfast' answer is that one ought not, or at least not nearly to the extent the conciliationist demands. To revise one's view down whenever an equally credentialed person disagrees is to make confidence depend on a sociological feature of the field rather than on one's own grasp of the reasons; a community of confident peers, on the conciliationist principle, would compel mutual revision to a position none of them, individually, would have any reason to hold.
The steadfast intuition has more force than is sometimes conceded, but it is not without difficulty either. Someone who routinely persists in his view in the face of qualified disagreement, while presenting no further evidence and revising no part of his argument, looks less like a careful thinker than like an obstinate one. The proper response, more careful authors have suggested, is neither strict conciliationism nor strict steadfastness but a calibration to the kind of question at stake: where the issue turns on technical detail that can in principle be checked, peer disagreement should occasion serious revision; where the issue turns on differences of judgement that competent reasoners can hold without contradiction, the demand for revision is weaker.
The disagreement-about-disagreement, then, may have less to do with epistemology in the strict sense than with what one takes the typical contested question to look like. Those who picture the typical case as a technical one tend toward conciliationism; those who picture it as a matter of judgement tend toward steadfastness. The careful position resists picking the picture first and the principle after, and asks instead which picture the case in front of one actually fits. That sort of attention is harder than either pure position offers; it is also what each position, on its more reflective formulation, has already begun to acknowledge.
It can be inferred from the passage that a steadfast theorist would be most likely to regard the following move as illegitimate: