It’s true that one-boxing is the strategy that maximizes expected utility, and that it is a fairly uncontroversial maxim in normative decision theory that one should pick the strategy that maximizes expected utility. However, it is also a fairly uncontroversial maxim in normative decision theory that if a dominant strategy exists, one should adopt it. In this case, two-boxing is dominant (if you suppose there is no backwards causation). Usually, these two maxims do not conflict, but they do in Newcomb’s problem. I guess the question you should ask yourself is why you think the one we should adhere to is expected utility maximization.
Not saying it’s the wrong answer (I don’t think it is), but simply saying “We do this sort of math all the time. Why not here?” is insufficient justification because we also do this other sort of math all the time, so why not do that here?
Great, I’ll work on that. That’s exactly what I should ask my self. And if I find that the rule of do that with highest expected utility fails on the smoking lesion problem, I’ll ask why I want to go with the dominant strategy (as I predict I will).
The only reason that I have to trust expected utility particularly is that I have a geometric metaphor, which forces me to believe the rule, if I believe certain basic things about utility.
It’s true that one-boxing is the strategy that maximizes expected utility, and that it is a fairly uncontroversial maxim in normative decision theory that one should pick the strategy that maximizes expected utility. However, it is also a fairly uncontroversial maxim in normative decision theory that if a dominant strategy exists, one should adopt it. In this case, two-boxing is dominant (if you suppose there is no backwards causation). Usually, these two maxims do not conflict, but they do in Newcomb’s problem. I guess the question you should ask yourself is why you think the one we should adhere to is expected utility maximization.
Not saying it’s the wrong answer (I don’t think it is), but simply saying “We do this sort of math all the time. Why not here?” is insufficient justification because we also do this other sort of math all the time, so why not do that here?
Great, I’ll work on that. That’s exactly what I should ask my self. And if I find that the rule of do that with highest expected utility fails on the smoking lesion problem, I’ll ask why I want to go with the dominant strategy (as I predict I will).
The only reason that I have to trust expected utility particularly is that I have a geometric metaphor, which forces me to believe the rule, if I believe certain basic things about utility.