Agents in Discord. On Preference Aggregation under Uncertainty

Ulrich Metschl


Harsanyi’s Aggregation Theorem states that, under some modest assumptions, it is possible to aggregate individual preferences into an collective preference ordering by a utilitarian rule. For many, this restored some of the confidence in collective decision-making that was challenged by Arrow’s result on the impossibility of preference aggregation under weak, though minimally desirable constraints. However, recent results indicate that Harsanyi’s theorem does not extend to situations of collective choice under uncertainty, where agents simultaneously disagree on their preferences and their probability assessments. Besides their technical interest, these results seem to be pertinent for any account of public decision-making that accepts that in pluralistic societies public deliberation on vital issues is subject to conflict and uncertainty. Our contribution will present the difficulties of Bayesian aggregation and the available options to bypass these difficulties. Special consideration is given to proposals which separate the requirements for collective rationality from those for individual decision-making.


20th century philosophy; philosophy; Wittgenstein Ludwig; Bayesianism; collective choice; preference aggregation

Full Text:



  • There are currently no refbacks.