The aim of this course is to introduce the basics of decision theory, and to critically discuss the philosophical foundations and implications of the theory. We start by discussing different uses of decision theory and the formulation of decision problems. After introducing the expected utility maxim, we discuss where the utilities are meant to come from and how to measure them according to the von Neumann-Morgenstern (vNM) expected utility theory. We go through the justification of the vNM axioms and consider criticism of the theory based on seemingly rational attitudes to risk that conflict with these axioms. A limitation of the vNM theory is that it assumes that alternatives give rise to an objective probability distribution over outcomes. In most realistic decision problems, however, the decision-maker has to use her own estimates of how likely an option is to lead to each of its possible outcomes (i.e., she needs to consult her subjective probabilities). Savage’s decision theory is meant to accounts for this. We discuss Savage’s axioms and the idea that a rational person’s preferences determine a unique subjective probability function (as Savage’s theory entails). We consider whether Savage’s representation theorem provides a convincing argument for Probabilism, the view that beliefs should be probabilistic. We conclude by discussing the so-called Dutch Book Argument for probabilism.
Teacher: H. Orri Stefánsson
Literature:
- Martin Peterson: An introduction to decision theory (Cambridge University Press 2017)
