Abstract

Newcomb's problem, first presented by Robert Nozick in 1969, has aroused much interest among philosophers because it appears to involve a conflict between two intuitively attractive principles of decision: the Principle of Dominance and the decision-rule of Richard Jeffrey's decision theory, the Principle of Maximizing Conditional Expected Utility. I believe that the Principle of Dominance makes the rational prescription in Newcomb's problem and, consequently, that the problem constitutes a counterexample to Jeffrey's theory. ;Setting aside the question whether Jeffrey's decision theory makes the right or wrong prescription in Newcomb's problem, it has many merits as a decision theory. In Chapter One, I describe some of these merits by comparing it with the 'standard' decision theory--namely, Savage's theory. In Chapter Two, I describe some other decision problems, inspired by Newcomb's problem, which have been presented as more plausible counterexamples to Jeffrey's theory. Following custom, I call these Newcomb decision problems. ;Philosophers who believe that Newcomb decision problems are genuine counterexamples to Jeffrey's theory diagnose the deficiency in Jeffrey's theory as consisting in the fact that it does not employ any causal notion in the calculation of expected utility: for this reason it does not accurately reflect the agent's causal beliefs. In the light of this diagnosis, several so-called 'causal decision theories' have been proposed as replacements for Jeffrey's theory. They are called 'causal decision theories' because causal notions enter into their calculations of expected utility. In Chapter Three, I describe two such causal decision theories; and in Chapters Four and Five, I consider the question whether they are equivalent. ;Shortly after the various causal decision theories were developed, several subtle and ingenious arguments were proposed to the effect that Newcomb's problem and Newcomb decision problems are not genuine counterexamples to Jeffrey's theory. These defences of Jeffrey's theory try to show that Jeffrey's decision-rule makes the same prescriptions as the causal decision-rules in the alleged counterexamples. In Chapters Six and Seven, I analyze, in detail, two defences along these lines; and in Chapter Eight, I present a general argument against any defence of Jeffrey's theory which attempts to reconcile Jeffrey's theory with causal decision theories.