Agents face serious obstacles to making optimal decisions. For instance, their cognitive limits stand in the way. John Pollock’s book, Thinking about Acting , suggests many ways of revising decision principles to accommodate human limits and to direct limited, artificial agents. The book’s main proposal is to replace optimization, or expected-utility maximization, with locally global planning. This essay describes optimization and locally global planning, and then argues that optimization among salient options has the virtues of locally global planning without certain (...) drawbacks. Although it does not endorse locally global planning, it recommends that decision theory incorporate some of the book’s ideas about settling for improvements when optimization among all options is unrealistic. (shrink)

An agent's options in a decision problem are best understood as the decisions that the agent might make. Taking options this way eliminates the gap between an option's adoption and its execution.

Newcomb's problem and similar cases show the need to incorporate causal distinctions into the theory of rational decision; the usual noncausal decision theory, though simpler, does not always give the right answers. I give my own version of causal decision theory, compare it with versions offered by several other authors, and suggest that the versions have more in common than meets the eye.

Richard Taylor has enlivened various fields of analytical philosophy during the past three decades, especially with his ingenious attacks upon commonly held beliefs. I recall being particularly stimulated to reflection by his challenge to one pair of seeming truisms: our certainty that we no longer have any control over what has already happened; and the complementary assumption that some forthcoming events — notably our own deliberate acts — do remain ‘up to us’. Taylor has argued separately for backwards causation, and (...) for a fatalistic view of what is going to occur. Most significantly, he seems able to transform every objection we produce against fatalism into an unwanted rationale for supposing that we can shape bygone events — and vice versa. -/- . (shrink)