AN INTRODUCTION TO THE SYMPOSIUM ON RATIONAL CHOICE AND PHILOSOPHY

This symposium contains a selection of the papers that were presented at a conference we organized on Rational Choice and Philosophy that was held at Vanderbilt University on 16 and 17 May 2014. The aim of the conference was to provide an inter-disciplinary forum for philosophical work that uses ideas and tools from rational choice theory, understood broadly to include decision theory, game theory and social choice theory.

In the first of their two papers, Franz Dietrich and Christian List present a novel theory of individual choice designed to overcome what they regard as an important limitation of classical choice theory. Classical theory says nothing about how an agent perceives the options in the choice set, and so cannot capture the intuitive idea that an agent chooses an object because the object possesses certain properties. Dietrich and List introduce the idea that objects have ‘motivationally salient properties’ for agents, which can depend on the context; this allows them to inject an element of psychological realism into classical choice theory. They show how the idea of ‘reason-based choice’ can be elaborated formally and used to solve a number of problems that plague the classical theory.

Richard Bradley’s paper addresses a famous problem in the theory of individual choice under uncertainty, namely how to make sense of agents with ‘Ellsberg preferences’, that is, who are averse to uncertainty about the chances of some good. Ellsberg himself argued that such preferences are incompatible with Bayesian decision theory à la Savage, an argument that many have accepted. Bradley argues, by contrast, that the Ellsberg preferences arise because we value chances non-linearly and, hence, not in accordance with von Neumann–Morgenstern (vNM) utility theory. He concludes that it is possible to reconcile Ellsberg preferences with Savage’s theory so long as we abandon the assumption that the valuation of chances proceeds in accordance with vNM utility theory.

Dietrich and List’s second paper addresses a key foundational issue in decision theory and microeconomics more generally, namely whether we should endorse the behaviourist view that all apparent reference to mental states (e.g. preference or utility) should be understood in a strictly ‘as if’ fashion. This old debate has recently been re-invigorated by Gul and Pesendorfer’s (2008) defence of ‘mindless economics’. Dietrich and List develop a persuasive case against behaviourism by drawing on ideas about explanation, realism, and representation from the general philosophy of science. They also reject the radical neuroeconomic view that seeks to explain economic behaviour directly in terms of brain states.

Matthew Adler’s paper discusses what he calls ‘meta-ethical aggregation’, which is a problem that arises for many ‘ideal advisor’ theories of ethics. On such theories, moral facts are facts about the idealized preferences of a community of advisors. Adler notes an obvious problem for such theories: the advisors may not all share the same preferences. This raises the issue of how to aggregate their preferences into a single community preference. Adler tackles this issue using Arrovian social choice theory, paying particular attention to the question of what the appropriate domain assumption is in meta-ethical aggregation. Adler’s approach nicely illustrates how the formal machinery of social choice theory can be utilized to clarify a traditional philosophical question.

Michael Morreau’s paper is also a novel application of social choice theory, in this case to the problem of ‘collective grading’. This problem arises when a number of evaluators each grade a set of items (e.g. exam scripts, grant proposals or restaurants), and the individual grades are aggregated into a collective grade. Morreau emphasizes that the boundaries (‘thresholds’) between grades are subject to uncertainty because different evaluators may not agree on the exact meaning of the grades. His paper investigates the impact of this uncertainty on the aggregation problem. Using a modification of the Arrow–Sen framework for social choice, Morreau identifies a condition that collective grading should ideally satisfy called ‘soundness’ and shows that whether it can be achieved depends on whether the different evaluators have the same grading thresholds or not.