Reasons and Rational Choice Workshop


Venue: T206 (Second Floor, Lakatos Building, Portugal Street)


 Thursday, 3 February

 15:00 - 15:30 Registration and Welcome

 15:30 - 17:00 Nicholas Southwood (Oxford/ANU), "Reasons, Rational Requirements, and What to Do"


 17:30 - 19:00 Michael Smith (Princeton), "Beyond Belief and Desire, Or: Responsibility and Knowledge of Basic Moral Truths"


Friday, 4 February

 9:30 - 11:00 Niko Kolodny (Berkeley), "Instrumental Reasons"


 11:30 - 13:00 Campbell Brown (Edinburgh), "The Composition of Reasons"


 14:15 - 15:30 Christian List (LSE), "Where do preferences come from?" (with F. Dietrich)

 Short break

 15:45 - 17:00 Franz Dietrich (LSE/Maastricht), "Reasons and Rational Choice" (with C. List)


 17:30 - 19:00 John Broome (Oxford), "Requirements and Basing permissions"


Saturday, 5 February

9:30 - 11:00  Ryan Davis (Harvard), "Reasons without Realism"


11:30 - 13:00 Robert Sugden (University of East Anglia), "Common reasoning in games: a Lewisian analysis of common knowledge of rationality" (with R. Cubitt)



Requirements and Basing permissions

John Broome (Oxford)

 Rationality requires things of you, such as not to have contradictory beliefs and to intend what you believe is a means implied by an end you intend. As a rational creature, you satisfy many of these requirements automatically. But you can also improve your rationality by your own action: you can bring yourself to satisfy some requirements of rationality that you do not satisfy automatically. You can do so by reasoning.

One sort of reasoning is practical reasoning. It is reasoning that concludes in an intention. That is to say, it is a way of making a decision.

Correct reasoning can bring you to satisfy a requirement of rationality that you would not otherwise satisfy. However, what distinguishes correct from incorrect reasoning is not whether or not it brings you to satisfy a requirement of rationality. Correct reasoning is reasoning that follows a correct rule, and a correct rule is derived from a basing permission of rationality. A basing permission is a permission to have one attitude (a belief or an intention, for example) on the basis of other attitudes you have. For instance, rationality permits you to believe q on the basis of believing p and believing that if p then q.

So reasoning is made correct by basing permissions of rationality, yet it brings you to satisfy requirements of rationality. What, then, is the relation between basing permissions and requirements?

The Composition of Reasons

Campbell Brown (Edinburgh)

Abstract:How do reasons combine? How is it that several reasons taken together can have a combined force which exceeds the force of any one alone? This paper proposes an answer in mereological terms: reasons combine by composing a further, complex reason of which they are parts; the combined force of the reasons is the force of the reason they compose. This answer prompts another question. How might the force of a whole reason be related to the forces of its parts? I distinguish two approaches: atomism, which says all reasons supervene on the atomic reasons (those which have no smaller reasons as parts); and holism, which says all reasons supervene on the whole reasons (those which are not parts of any larger reasons). I consider one form of atomism, which I call additivism, and two forms of holism, which I call isolationism and marginalism. I argue (a) that holism is superior to atomism, and (b) that there is nothing to decide between the two forms of holism.

Reasons without Realism

Ryan Davis (Harvard)

Abstract. A persistent worry about objective values is that they would somehow be metaphysically "queer" entities—dissimilar from any other types of object we are aware of. Constructivists respond to this worry by explaining value in terms of evaluative attitudes. The judgments, beliefs, and intentions of actual agents are metaphysically ordinary, and if these attitudes can give rise to values, then values would also be ordinary. However, there is a serious obstacle for the constructivist vindication of value. Many philosophers deny that fundamental practical reasons can ever arise from an agent's attitudes. In this paper, I investigate whether there are any cases of attitude-dependent reasons for action. I argue that for certain classes of value (centrally including symbolic values, some personal projects, and some valuable relationships), the value depends on an appreciative response by an agent. If this appreciative response counts as an instance of valuing, then constructivists are right about the direction of fit between evaluative attitudes and evaluative facts. And if this account can succeed, constructivism about value may help us understand the appeal of toleration for values which may repel us.

Instrumental Reasons

Niko Kolodny (UC Berkeley)

Often our reason for doing something is an "instrumental reason": that doing that is a means to doing something else that we have reason to do. What principles govern this "instrumental transmission" of reasons from ends to means? Negatively, I argue against principles often invoked in the literature, which focus on necessary or sufficient means. Positively, I propose a principle, "General Transmission," which answers to three intuitive desiderata: that reason transmits to means that are "probabilizing," "effective," and "nonsuperfluous" with respect to the relevant end.

"Where do preferences come from?"  |

Christian List (LSE)

Rational choice theory analyzes how an agent can rationally act, given his or her preferences, but says little about where those preferences come from. Instead, preferences are usually assumed to be fixed and exogenously given. We introduce a framework for conceptualizing preference formation and preference change. In our model, an agent's preferences are based on certain 'motivationally salient' properties of the alternatives over which the preferences are held. Preferences may change as new properties of the alternatives become salient or previously salient ones cease to be so. We suggest that our approach captures endogenous preferences in various contexts, and helps to illuminate the distinction between formal and substantive concepts of rationality, as well as the role of perception in rational choice.

Beyond Belief and Desire, or: Responsibility and Knowledge of Basic Moral Truths

Michael Smith (Princeton)

Abstract. The standard belief-desire account of the explanation of action, at least in the form in which that account was put forward by Donald Davidson, is inadequate to the task of explaining even the very simplest actions.  If some version of the standard account is correct, then we must suppose that is some variation of the version put forward by Carl G. Hempel prior to Davidson.  According to Hempel, three psychological states are in play when we explain action, not as Davidson supposes, just two: desire and belief are part of the explanation of every action, but so too is the capacity to be  instrumentally rational, a capacity that is but one among many capacities rational agents possess.  Once we enrich our understanding of action explanation to acknowledge the causal role played by an agent's exercise of his rational capacities, much richer accounts of action explanation come into view, accounts that highlight the many different ways in which agents' actions can be explained by their rational capacities.  Of special interest are cases in which agents' actions are explained by their failure to exercise their rational capacities, where these are capacities that they possess, and cases in which their actions are explained by their failure to exercise their rational capacities, because these are capacities that they do not  possess.  Richer accounts of action explanation such as these suggest a distinctive story about the conditions under which people are responsible for wrongdoing, a story with surprising implications for our understanding of what it is for an agent's moral beliefs to be justified.

Reasons, Rational Requirements, and What to Do

Nicholas Southwood (Oxford/ANU)

Abstract:  Rational requirements generate what is by now a familiar puzzle. On the one hand, they appear to be normative in some deep and important sense. On the other hand, it is far from obvious what their normativity amounts to. My aim in this paper is to offer a characterisation of rational requirements that can do something to resolve this puzzle. The key will be to offer a functionalist characterisation of rational requirements. A number of philosophers have recently offered functionalist characterisations of reasons . I shall suggest that we may also usefully approach the task of understanding rational requirements by thinking of them in functional terms. According to the account I propose, whereas reasons are considerations that help determine what one ought to do , rational requirements are, roughly, principles that help determine the correct answer to the quite different question of what to do . Once we understand the different functional roles played by reasons and rational requirements, we can see clearly two things. First, there is a way in which reasons are normative that rational requirements clearly aren't inasmuch as the former help to explain, in a way the latter don't, what we ought to do. Second, there is nonetheless an important way in which rational requirements are normative that other kinds of principles, such as principles of etiquette and grammar, aren't.

Common Reasoning in Games: A Lewisian Analysis of Common Knowledge of Rationality

Robert Sugden (East Anglia)

Based on joint work with Robin Cubitt

Abstract. It is a fundamental assumption of standard game theory that each player of a game acts according to Bayesian subjective expected utility theory and that this is common knowledge.  Intuitively, this conception of common knowledge of rationality (CKR) appears to be a meaningful idealization, in the same sense that perfect competition is a meaningful idealization in price theory, or frictionless surfaces are in theoretical mechanics.  However, attempts to formalize the assumption have persistently generated paradoxical implications that seem to call into question the conceptual coherence of CKR.  In this paper, we offer a diagnosis of these paradoxes and, by means of a new class of models of 'common reasoning', show how the intuitive idea of CKR can be formulated without creating paradoxes.

The canonical approach to modelling CKR is due to Aumann (1987).  Aumann offers a Bayesian modelling framework which can be used to represent any given noncooperative game, which he interprets as a representation of CKR.  Although Aumann's model is logically consistent, apparently natural extensions of it turn out to imply contradictions.  We conjecture that most game theorists are aware that paradoxes arise when particular restrictions are imposed on the Bayesian model, but are still inclined to respond by proposing their favoured sets of 'intuitively reasonable' conditions.  However, this approach provides no formal method of resolving disagreements between alternative proposals or intuitions.  This paper takes the more radical approach of developing an explicit model of players' reasoning.

The underlying reason why the Bayesian approach runs into problems, we will argue, is that it seeks to represent a situation in which the players of a game have common knowledge about which strategy each of them will choose in every state of the world, and in which each individual's choices are expected utility maximising relative to her beliefs.  This implies a binary partition of the set of strategies, which is common knowledge: one element of this partition contains those strategies that are played in some state(s) of the world, while the other contains those that are played in none.  However, the Bayesian approach does not attempt to describe the modes of reasoning by which the players might discover this partition for themselves.

Our approach is inspired by Lewis (1969).  Lewis is widely credited with the first formal definition of common knowledge, but it is less well known that this definition is only one component of a detailed analysis of a mechanism by which common knowledge can be generated through interlocking processes of individual reasoning.  Building on an earlier analysis of some of the distinctive features of Lewis's game theory (Cubitt and Sugden, 2003), we extend Lewis's approach to allow individuals to reason about the standards of practical rationality that other individuals endorse, and in this way to reach conclusions about whether specific strategies are or are not rationally playable.  We capture the idea of CKR by assuming that players have common access to the specific modes of reasoning that constitute the common rationality being modelled.  We show that this representation of CKR is internally consistent for any of a very wide class of standards of practical rationality.   

Please register for this event by emailing e.p.senchaudhuri @ lse.ac.uk