Foundations of Decision Theory

June 28, 2024
Munich Center for Mathematical Philosophy, LMU Munich

Geschwister-Scholl-Platz 1
München 80539

This event is available both online and in-person


Centre National de la Recherche Scientifique
Massachusetts Institute of Technology
University of Exeter
University of Paris 1 Panthéon-Sorbonne
Universität Konstanz
Carnegie Mellon University

Topic areas

Talks at this conference

Add a talk


Registration required: [email protected]


Wolfgang Spohn, “Foundations of Decision Theory and the Problem of Free Will”
Of course, our actions are caused, ideally and thus actually, more or less, precisely as described by the right kind of decision theory. This entails two comments on the problem of free will. First, the right kind of decision theory is causal decision theory, the mark of which is that the agent perceives her actions under consideration as uncaused, i.e., as free. Here, it is important to understand that the agent’s first-person perspective is not inferior, but on a par with the observer’s third-person perspective, in which those actions are perceived as caused. Second, the right kind of decision theory is reflexive decision theory, which gives an account of how second-order desires rationally influence first-order desires. This connects up with our ways of holding agents accountable, which we usually associate with their’s having free will.

Franz Dietrich, “Welfare vs. choice-theoretic utility”
Ever since the Harsanyi-Sen debate, it is controversial whether someone's welfare should be measured by her von-Neumann-Morgenstern (VNM) utility, for instance when analysing welfare intensity, social welfare, interpersonal welfare comparisons, or welfare inequality. As we show, natural working assumptions lead to a different welfare measure. The new measure addresses familiar concerns about using VNM utility, while requiring only ordinal evidence, such as observed choices or self-reported comparisons. The measure is determined by the VNM utility function and a single additional parameter, the person's revealed 'pure risk attitude'. Using this measure instead of VNM utility leads to considerably different social welfare assessments (and policy recommendations), in riskless or risky contexts.

Marcus Pivato, “Autonomy and metapreferences”
The standard model of rational choice in economics treats the preferences of the agent as exogenous. This raises interesting philosophical problems: if an agent cannot choose her own preferences, then she is not really "autonomous" ---she is condemned to slavishly maximize the preferences which have been "imposed" on her from the outside. Likewise, we cannot hold her morally responsible for her choices (good or bad), if these choices are simply the result of maximizing an (unchosen) preference order. But suppose instead that an agent could choose her preferences. On what basis would she make such a choice? Presumably, on the basis of "second order" preferences. But how does she choose these second-order preferences? This leads to an obvious infinite regress. Furthermore, what does rational choice mean when the agent must simultaneously optimize with respect to first-order, second-order, and higher-order preferences ? What happens when her higher-order preferences come into conflict with her lower-order preferences? In this talk, I will introduce two mathematical models of such "metapreferences", and discuss possible solutions to these problems.

Kevin Zollmann, “Reflections on the Independence Thesis"
Rational individual epistemic behavior might be distinct from rational group epistemic behavior. Groups of rational individuals may not make for the epistemically best groups. This idea was called "The Independence Thesis" by Mayo-Wilson, Danks, and me. We originally focused on one particular model of group problem solving and with one particular definition of what counts as "rational." In this talk, I will explore other ways of defining notions like individual and group rationality, and explore versions of the independence thesis for them. I will briefly introduce several models that illustrate independence-thesis-like behavior. Finally, I consider ways that one might rebut the independence thesis that parallel attempts to make cooperation rational in the prisoner's dilemma. Ultimately, I conclude that epistemic rationality is like instrumental rationality: there can sometimes be a conflict between what's good for the group and what's good for the individual.

Kevin Dorst, “Ambiguity Drives Hindsight Bias”
Some of our uncertain judgments are clear—How likely is a fair coin to land heads? Others are ambiguous—How likely am I to own a dozen spoons? Ambiguous judgments are subject to higher-order uncertainty: we are unsure what our own opinions are. I argue that the contrast between clarity and ambiguity helps explain hindsight bias: the tendency for learning something to lead us to increase our estimate for our prior credence in it. (“I knew it all along.”) I’ll show, theoretically, that for Bayesians (1) hindsight bias is rational when and only when your prior credence is ambiguous, and (2) such hindsight bias can lead to rational forms of predictable “overconfidence” (i.e. miscalibration). I’ll then present empirical evidence that people exhibit hindsight bias in a way that matches ambiguous-Bayesian predictions.

Silvia Milano, “Rational Updating at the Crossroads”
This talk, based on joint work with A. Perea, explores the absentminded driver problem and the rationality of forming commitments over time. We employ centred possibilities, modelling the states (i.e. the events about which the driver is uncertain) as the possible final destinations indexed by a time period. The optimal probability we find for continuing at an exit is different from almost all papers in the literature. In this scenario, act-state independence is violated, but states are mutually exclusive and the driver arrives at his optimal choice probability via Bayesian updating. We show that our solution is the only one guaranteeing immunity from sure loss via a Dutch strategy, and that -- despite initial appearances -- it is time consistent. This raises important implications for the rationality of commitment in such scenarios.

Supporting material

Add supporting material (slides, programs, etc.)




June 27, 2024, 9:00am CET

Who is attending?

No one has said they will attend yet.

Will you attend this event?

Let us know so we can notify you of any change of plan.

RSVPing on PhilEvents is not sufficient to register for this event.