Chapter 2 - Decision Problems

Summary

This chapter sets the stage for future developments by providing a formal characterization of the sorts of decision problems that agents face when choosing among alternative courses of action at a given time. The classic treatment of this subject, and the one still endorsed by most economists, can be found in Leonard Savage's masterpiece The Foundations of Statistics. A somewhat different characterization of decision making situations, which tends to find favor with philosophers, has been offered by Richard Jeffrey in The Logic of Decision. The model to be presented here will be a kind of hybrid of the Savage and Jeffrey approaches. It agrees with Savage in retaining a clear distinction among actions, states, and outcomes but follows Jeffrey in treating all three kinds of entities as species of propositions. The chief advantage of this mixed approach is that it allows us to be much more specific than Savage was about what the elements of decision problems actually are and how they relate to one another. It also gives us a general formal framework within which both evidential and causal decision theories can be conveniently expressed.

SAVAGE'S MODEL OF DECISION PROBLEMS

Savage's model envisions a decision maker, or agent, who chooses among risky or uncertain prospects, called actions, whose consequences or outcomes depend on the state of the world.