Always Conditionalize?
Let P be a proposition of which you neither believe that it's true nor that it's false, say Goldbach's Conjecture. Since you know that you don't believe P (otherwise you couldn't have chosen it), your conditional subjective probability for [P and I don't believe P] given P should be close to 1. However, if you were to learn that P, your subjective probability for [P and I don't believe P] shouldn't be close to 1, but close to 0. So is this a case were you shouldn't conditionalize?
I think the fault in that argument probably is that it doesn't work any more if "believe" is time-indexed, as it should. Let t1 be the present, t2 the time at which you learn that P. There are two easy cases. If "believe" is "believe at t1", there is nothing wrong with conditionalizing at t2 and assigning [P and I didn't believe P at t1] a probability close to 1. If "believe" is "believe at t2", and it is known at t1 that t2 is the time at which you will find out whether P, your conditional subjective probability at t1 for [P and I won't believe P at t2] given P shouldn't be close to 1 but already close to 0.
The tricky case is where it isn't known at t1 that t2 is the time at which you will find out whether P. You just happen to find it out at t2. Then although your conditional probability at t1 for [P and I won't believe P at t2] given P shouldn't be close to 1, it shouldn't be close to 0 either. But at t2, it seems that your unconditional probability for [P and I don't believe P at at t2] should really be close to 0. So were's the fault here? Or, if there is none, why can't this be plugged into a diachronic Dutch Book argument?