In the debate on determinism and decision-making (deliberation) the major views are usually grouped under the simple titles deliberation-compatibilism and deliberation-incompatibilism. Deliberation-compatibilism being the view that engaging in rational decision-making is compatible with believing in determinism. But this division is, I think, far too course. There are many possible variations on these views, as there are many different kinds of compatibilism and incompatiblism in the free will debate. We can start to get a grip on this using the following flow chart:
Here question 0 is labelled as such since its answer is, I think, trivially yes (the Belief in Ability Thesis), Clarke notwithstanding. Anyone who thinks that anything like decision theory is the correct account of rational decision-making will also have to agree with this claim, since all decision theories that we have work on the assumption that decisions are made by comparing the relative merits of different, presumably epistemically possible, actions.
Leeway-deliberation-compatibilists are so named in analogy to leeway-compatibilists in the free will literature who agree that some principle of alternative possibilities applies to the possession of moral responsibility. These leeway-deliberation-compatibilists agree that we must have the epistemic leeway to perform more than one action in order to engage in rational decision-making, but they are compatibilists because they do not require that we believe that we could actually perform more than one action in the actual world. So long as we are unsure which action we will perform then we can engage in decision-making about them. They will then have to employ counterfactuals to assess utilities in some cases where a state, for example “I am determined not to take the job at the university” is incompatible with some action I am considering, for example “Take the job at the university”.
Taking the right hand fork here indicates that one thinks that for an act-state pair to be admissible in decision-making it must be the case that the act could occur in the given state at the time of decision-making. This will be true of any (otherwise possible) indeterministic state and (otherwise possible) action, but not of all act-state pairs where the state is determinisitic.
Question 2 challenges those who take the right fork above on how one can engage in rational decision-making when there is some world one considers possible that is deterministic and rules out at least one of your otherwise possible actions. Given that taking the right hand fork above has committed us to the inadmissibility of such act-state pairs it appears we are faced with a challenge: how do we make a decision in this case? The two options, it seems to me, are either to deny that rational decision-making is possible in such a case, or to rule out the problematic states from consideration. Taking the extreme-deliberation-incompatibilist position, then, is tantamount to saying that if determinism is even epistemically possible it is impossible to engage in rational decision-making. It seems no surprise that no one that I can find in the literature has taken this position.
But if we take the other option and claim that rational decision-making presupposes that determinism is false then we are still faced with another question: how does this bear on our epistemic attitudes (given that we are trying to give a categorisation of views with regards to the compatibility of belief in determinism and rational decision-making)? Answers to this question probably form a spectrum, rather than breaking simply into the three categories I have given, but this will do for an initial broad brush pass.
The perspectivist-deliberation-compatbilist position basically says that while we must presuppose determinism is false in our decision-making (we must act under the idea of freedom, in a common slogan), this does not mean anything for our beliefs because we can separate out a theoretical or objective perspective, which our beliefs are sensitive to, and a practical or agential perspective, which assumes/presupposes freedom/the failure of determinism. I am not sure Bok would be entirely happy with my characterising her view in exactly this way, since she might disagree that rational decision-making presupposes indeterminism as opposed to simply freedom.
Finally we can distinguish two answers from deliberation-incompatibilists here: the weak reading requires only that we not believe in determinism in order to allow us to make the presupposition that rational decision-making requires, while the strong reading would require us to actively reject determinism. The latter would likely be endorsed, I assume, by, for example, epistemic consequentialists who endorse the idea that epistemology should be done with decision theory (assuming, of course, that they did not get off the bus somewhere earlier).