Pietro Cibinel

Many people accept pluralism about welfare or well-being: there is more than one thing that is finally good for us. Philosophers have discussed extensively the question of how welfare should be distributed, whether across people or states of nature. But how should we distribute individual welfare goods across these dimensions? I argue that certain principles connecting the distribution of welfare to the distribution of welfare goods fail, and explore the connection between these principles and the project of devising an appropriate quantitative measure of welfare. As it turns out, the failure of these principles corresponds to a conflict between different approaches to constructing a quantitative measure of welfare. I identify one such approach, based on tradeoffs across welfare goods, as superior to its alternatives, which are based on tradeoffs across people and states of nature respectively. I explain how this conclusion provides a measurement-theoretic argument for prioritarianism over utilitarianism, and for risk-sensitivity in welfare over risk-neutrality.

This paper develops and defends a non-utilitarian interpretation of John Harsanyi’s social aggregation theorem and sum of vNM utilities approach. On this interpretation, vNM utilities transform an independently available cardinal measure of fully comparable individual well-being. The resulting proposal for ranking well-being distributions—the Risk-Priority View— is not welfare-anonymous and can favor a smaller increase in well-being for one individual rather than a larger increase in well-being for another, equally well-off individual. I argue here that such counterintuitive implications can be defended, and that impartiality can still be secured through the imposition of an alternative, interprofile anonymity axiom.

This dissertation presents and defends an ethical outlook based on two guiding ideas: first, there is a plurality of rationally permissible attitudes towards risks that involve our welfare or well-being; and second, an ethical deliberator should defer to rational individuals’ attitudes towards risk when assessing policies with uncertain consequences that affect them. John Harsanyi famously argued that deferring to people’s preferences under risk leads to utilitarianism. Amartya Sen and John Weymark pointed out that Harsanyi’s argument relies on a contentious interpretation of decision theory under risk—according to which, effectively, only one attitude towards risks involving our well-being, that of risk neutrality, is rationally permissible. As the Sen-Weymark critique clarifies, Harsanyi’s stance is untenable if our quantitative notion of well-being is independent of preferences formed in conditions of risk. Chapter 1 argues that this is indeed the case: our quantitative notion of well-being depends on comparisons between outcomes with no uncertainty. It defends a method for deriving quantities of well-being based on conjoint measurement between objective goods. Chapter 2 explores the space of possibilities that our two guiding ideas leave us with; it shows that under weak assumptions they are incompatible with welfarism, the view that all we need to know in order to evaluate outcomes is how people’s well-being is distributed. It then presents a dilemma for non-welfarist views: they either entail that risk attitudes affect the moral significance of well-being in contexts of zero risk, or they struggle to vindicate the thought that improving everyone’s well-being is an important ethical concern. Chapter 3 sides with the first horn of this dilemma. It offers a defense of a position I call the Risk-Priority View: on this theory, the ethical import of an individual ending up with a certain amount of well-being is affected by that individual’s risk attitude. Roughly, more risk-averse individuals get priority in dire circumstances, while more risk-seeking individuals get priority in favorable ones.

This paper studies the relationship between promoting people's welfare and respecting their autonomy of choice under risk. I highlight a conflict between these two aims. Given compelling assumptions, welfarists end up disregarding people's unanimous preference, even when everyone involved is entirely rational and only concerned with maximizing their own welfare. Non‐welfarist theories of social choice are then considered. They are shown to face difficulties, too: either they fail to respect the value of welfare in at least one important sense, or they end up prioritizing different people's welfare differently in non‐risky choices, on the basis of their attitudes to risk, which are intuitively irrelevant in this context.

Orthodox decision theory is fanatical in the way it treats small probabilities of enormous value, if unbounded utility functions are allowed. Some have suggested a fix, Nicolausian discounting, according to which outcomes with small enough probabilities should be ignored when making decisions. However, there are lotteries involving only small-probability outcomes, none of which should intuitively be ignored. So the Nicolausian discounter needs a procedure for distinguishing the problematic cases of small-probability outcomes from the unproblematic ones. In this paper, I present a dilemma for Nicolausian discounting: the view must be coupled either with a procedure that delivers fanatical verdicts of its own, as bad as those of orthodox decision theory, or with one that entails intransitive cyclic preferences.