Thanks to everyone for the extremely helpful comments on what I’ve posted so far. Now that I’ve laid out the groundwork, I’m going to try in this post to defend some of my claims in more detail.
My proposal is that in situations where a thinker faces a conflict between what the evidence seemed previously to support and what it now seems to support, she is not psychologically, conceptually, or rationally required to go with her present perspective. In other words, the deliberative question “should I continue to believe that P?” is not as a matter of necessity transparent to the evidence as one now sees it. To exercise doxastic self-control, on my view, is to consciously continue believing that P when you have received no significant new evidence since forming that belief, even if the same evidence now seems to you not to sufficiently support P – even if P no longer seems to you to be true. I think this is possible if the thinker conceives of both the past and present perspective on the evidence as equally hers, and elects to treat the past perspective rather than the present as authoritative in determining what she believes.
Let me take a second to address a couple of responses that might be tempting people. In some cases where a thinker’s judgment has been compromised, there will be higher-order evidence available to her that this has occurred. She will know that she took a judgment-corrupting drug, or is severely sleep- or oxygen-deprived, and will therefore have good evidence that she should not trust her current judgment. It is plausible that cases like this have a purely evidential solution, and so are no threat to belief transparency. But I’m imagining cases that are not like this; in the cases I’m interested in, nothing highly out of the ordinary has happened, and the thinker doesn’t take herself to have conclusive evidence as to whether her past or her present judgment is mistaken. She's just navigating the world with its normal emotional, social, environmental, and professional pressures.
A second tempting response is to suppose that one’s past judgment can simply be incorporated into the present perspective by treating it as testimonial evidence, as if from any other person. But I think this won’t do. If the thinker treats this conflicting testimony as decisive, even though she does not take her past self to be a greater expert or in possession of better evidence, this would be an odd and perhaps impossible instance of epistemic akrasia – it would be a matter of believing that P against her own best judgment, just because someone else with the same evidence believes it. On the other hand, if she merely treats the testimony as some evidence to be weighed into fresh deliberation, there is no guarantee that it will suffice to counteract the distortions of the epistemic temptation she is experiencing.
So, I think purely evidential solutions won’t do. This is why I characterize doxastic self-control as a partly volitional matter. Think about cases that are underdetermined by normative practical reasons, either because the values at stake seem to be incomparable, or because they are equally good (i.e. Buridan cases). I don’t want to argue over the word ‘volition’, but I want to suggest that the latitude we have to choose what to do in cases that seem to be underdetermined by their practical reasons also exists in the epistemic cases I’ve described. When the evidence doesn’t tell us whether to go with the past or the present perspective, we can exercise something worth calling a choice. It’s not a choice we can make for pragmatic reasons, but we don’t make it for any particular epistemic reason either.
Does this mean that in exercising doxastic self-control, the thinker ends up believing that P for no reason? I don’t think that follows. Rather, in identifying with her past perspective rather than her current one, she ends up believing that P on the basis of the reasons that she took to be sufficient when she formed the belief.
Still, in the moment, P doesn’t seem true or sufficiently supported by evidence. How can it be rational to continue believing it rather than simply withholding judgment? I want to defend only the modest claim that it is rationally permissible to maintain belief rather than withholding judgment; I think the latter is generally also permissible, and in some cases it might even be permissible to switch to believing not-P. I’m still trying to figure out exactly how to argue for this, but the basic thought is that a rational requirement to withhold judgment whenever one is experiencing doubt about a belief – even serious doubt – would be too costly from the perspective of acquiring important true beliefs. Whenever the doubt is a result of epistemic temptation, it would require the thinker to lose a true belief that she has (with potentially serious consequences, if an important action is riding on that belief). It might be that for a thinker who values the avoidance of error extremely highly, even at the cost of having true beliefs, withholding judgment is always the right option. But I think most of us are not like this, and we don’t have to be in order to count as believers.
Finally, it’s worth noting that the advocate of transparency as a rational ideal can potentially agree that doxastic self-control is possible – it just exhibits a problematic form of alienation. If I have to treat my past judgment as having any bearing on what I believe now, and cannot avow my belief by reflecting only on my reasons, then I am estranged from my belief; it is as if I have programmed myself. My question to you guys is: how much should this bother us? Perhaps there is a sense in which it would be better if we were thinkers who had no need of capacities like doxastic self-control, but we are such creatures. And I think there is a sense in which we are most autonomous both in our doxastic lives and in our practical lives precisely when we have the latitude to choose which of our own perspectives to identify with over time. If I’m right, it can be the key to weathering episodes that might otherwise lead one to quit a long-term project, e.g. because one no longer believes one’s efforts will be rewarded. So if it is alienating – is that so bad?