Choice Group

Abstract: TBA

Two of our research students, Sophie Kikkert and Dmitry Ananyev, present their work to the Choice Group.

In this paper, we ask: how should an agent who has incoherent credences update when they learn new evidence? The standard Bayesian answer for coherent agents is that they should conditionalize; however, this updating rule is not defined for incoherent starting credences. We show how one of the main arguments for conditionalization, the Dutch strategy argument, can be extended to devise a target property for updating plans that can apply to them regardless of whether the agent starts out with coherent or incoherent credences. The main idea behind this extension is that the agent should avoid updating plans that increase the possible sure loss from Dutch strategies. This happens to be equivalent to avoiding updating plans that increase incoherence according to a distance-based incoherence measure.

The martingale convergence theorem implies that in certain situations a Bayesian agent is sure to converge to the truth in the limit. Gordon Belot has argued that this constitutes a liability for Bayesian epistemology since it ignores the many ways in which one might fail to identify truth in the limit. In this talk I will study convergence to the truth within a nonstandard probability framework that allows fine-grained distinctions between infinite hypotheses. Within the nonstandard framework, convergence to the truth is expected only for hypotheses that can be finitely approximated. Importantly, this leads not to a revision but a refinement of the standard martingale convergence theorem.

Are there ever good epistemic reasons to misrepresent base rates? I investigate this question in the context of recent legislation restricting the presentation of gender stereotypes, and the representation of minority groups in children’s books. I argue that our hesitancy around certain base rates makes sense in the context of a more general epistemic dilemma we face: between knowledge of acuality and knowledge of possibility. Given this dilemma, there are sound epistemic reasons to behave in ways the may involve wariness or misrepresentation of base rates. This approach has implications for the way in which ethical and epistemic norms interact with one another.

Two of our PhD students, Nick Makins and Nicolas Cote, present their research to the Choice Group.

I propose a unified solution to two natural language modal puzzles: Ross's puzzle (the apparent failure of Ought p to entail Ought(p or q)) and free choice permission (the apparent fact that May(p or q) entails both May p and May q. I begin with a pair of cases from the decision theory literature illustrating the phenomenon of act-dependence, where what an agent ought to do depends on what she does. The notion of permissibility distilled from these cases forms the basis for my analysis of the modals May and Ought. This framework is then combined with a generalization of the classical semantics for disjunction---equivalent to Boolean disjunction on the diagonal, but with a different two-dimensional character---that explains the puzzling facts in terms of semantic consequence.

We study the measurement of well-being when individuals have heterogeneous preferences, including different conceptions of a life worth living. When individuals differ in the conception of a life worth living, the equivalent income can regard an individual whose life is not worth living as being better off than an individual whose life is worth living. In order to avoid this paradoxical result, we reexamine the ethical foundations of well-being measures in such a way as to take into account heterogeneity in the conception of a life worth living. We derive, from simple axioms, an alternative measure of well-being, which is an equivalent income net of the income threshold making lifetime neutral. That new well-being index always ranks an individual whose life is not worth living as worse-off than an individual with a life worth living.

Joint work with G. Ponthière.

The 50th anniversary of John Rawls's classic 1971 A Theory of Justice constitutes an ideal occasion to look back from different theoretical perspectives at the history and trajectory of his work, and also its possible future development in the hands of others. The aftermath of the 2020 summer of U.S. and global protests about racial injustice stimulated by George Floyd's May 2020 killing by the Minneapolis police makes race a particularly pressing perspective from which to conduct such re-examination. Over the years, a very small minority of philosophers, myself included, have raised critical questions about Rawls and race, and what would be required to turn his apparatus to deal properly with racial questions. However, as a result of a recent theoretical epiphany, I now believe that critics (myself included) have all along missed a more foundational and potentially devastating objection: that Rawls's theory was never intended by him to apply to racist societies in the first place. If I am correct, it means that half a century of American (and broader Western) Rawlsianism has basically been misguided, resting on a false assumption about the presumed scope of his theory.

This event will take place online via Zoom.  Everyone is welcome to join using a computer with access to the internet and Zoom. To take part just follow these instructions: Download Zoom Join the event using this link: https://uci.zoom.us/j/97555321293 Please note that these events are routinely recorded, with the edited footage being made publicly available on our website and YouTube channel. We will…

The magnitude and nature of the COVID-19 pandemic prevents public health policies from relying on coercive enforcement. Practicing social distancing, wearing masks and staying at home are voluntary and conditional on the behavior of others. In this talk, Cristina Bicchieri will present and discuss in detail the results of a large-scale survey experiment she conducted, with a team of researchers at the Center for Social Norms and Behavioral Dynamics. Cristina’s survey experiment spanned 9 countries and shows how expectations, both empirical (what others do) and normative (what others approve of), play a significant role in compliance, beyond the effect of any other individual or group characteristic. Results from Cristina’s groundbreaking research in the field are driven by an asymmetric interaction with individuals’ trust in government and science. This talk sheds light on how — at the macro level — country-level trust in science, and not in government, acts as a strong predictor of compliance during the pandemic.

When choosing between two courses of action having the same two possible outcomes, we should choose the act for which the conditional chance of the preferred outcome, given its performance, is higher. This simple principle is, I argue, both a basic condition of instrumental rationality and the core of our conception of chance. To support this latter claim, I show that it in the presence of very weak rationality conditions on choice, the principle implies:

A version of Lewis’ Principal Principle (and that the converse is false).
Stochastic Dominance: a principle endorsed by all main theories of rational decision making under risk.
That the evidential and causal decision values of an act, conditional on the chances, are the same.
Finally, I argue that its appeal as a principle of instrumental rationality derives from the fact that higher chances of a good outcome are better than lower ones together with the fact that conditional chance of an outcome O, given some act A, measure the efficacy of A in bringing it about that O.

Can a group be an orthodox rational agent? This requires the group’s aggregate preferences to follow expected utility (static rationality) and to evolve by Bayesian updating (dynamic rationality). Group rationality is possible, but the only preference aggregation rules which achieve it (and are minimally Paretian and continuous) are the linear-geometric rules, which combine individual values linearly and individual beliefs geometrically. Linear-geometric preference aggregation contrasts with classic linear preference aggregation, which combines both values and beliefs linearly, and achieves only static rationality. Our characterisation of linear-geometric preference aggregation implies as corollaries a characterisation of linear value aggregation (Harsanyi’s Theorem) and a characterisation of geometric belief aggregation.

When the number of people one can save from harm is fixed, and the difference in harm one can save people from is substantial, standard principles for health resource allocation prioritize by severity. For example, if death is a substantially greater harm than paraplegia, then saving one from death takes priority over saving a different person from paraplegia. Standard principles are also fully aggregative: one death can be outweighed not merely by a large number of cases of paraplegia, but also by a sufficiently large number of very minor burdens (such as a case of toenail fungus). While prioritization by severity in fixed-number cases involving substantially different harms is uncontroversial among leading thinkers, many have challenged full aggregation. Instead, they have proposed a limited form of aggregation, on which a large number of cases of paraplegia can outweigh one death, but no number of cases of toenail fungus can outweigh one death. One reason cited in favour of limited aggregation is its supposed better fit with people’s considered moral judgments. There is, however, a lack of evidence on the public’s views. Here, we report the responses of a representative sample of the UK population to priority-setting dilemmas. We find that around half of respondents do not always adhere to prioritization by severity, primarily because of an apparent dislike for selecting one group facing death over another same-sized group in lesser, but still grave, need. However, among the remaining respondents, limited aggregation is much more popular than full aggregation. Our central results are robust to a test for status quo bias. They indicate that an overwhelming majority of people’s views do not align with standard priority setting principles, and that among those who prioritize by severity, limited aggregation is by far the most common.

Inferring welfare from choice has long been a central issue in choice theory and provided foundations for public policy interventions. Rational choice theory assumes that choices are induced by optimizing a preference where the preference also represents welfare. However, Behavioral Economics robustly rejects the preference optimization assumption while often attempting to retain welfare-based justifications for public policy. For example, nudge theory imposes a definitional requirement that nudges raise welfare. Thus, the foundational question in behavioral welfare economics is: What can be inferred about welfare from choice alone without assuming anything about the determinants of choice behavior? Bernheim & Rangel (2009), after offering two related answers, reject one on the grounds that it will sometimes fail to establish which available alternative has the highest revealed welfare. However, it does not always fail. Thus, the question arises: When it does not fail, is it then OK or does it have other problems? A simple model of boundedly rational choice is developed to address this question, based on some remarks in Sunstein (2015). It is shown that, in this Bounded Rationality model, the choice-based welfare ranking rejected by Bernheim & Rangel may reverse the “true ranking” of the chooser. This preference reversal has some general significance for behavioral welfare economics, and a particular one in nudge theory for the following reason: Definitionally, an intervention fails to be a nudge unless it it may be confirmed that it is welfare enhancing. But that requires welfare revelation from choices, and the preference reversal draws attention to a possible problem in using some of Bernheim-Rangel's analysis for this purpose.

Abstract: TBA

Two of our PhD students, Charles Beasley and Fabian Beigang, present their research to the Choice Group.