8 November 2016, Computational Social Choice Seminar, Zoi Terzopoulou

Abstract
Judgment aggregation is concerned with the following central questions: Is it possible to construct rules that are able to aggregate sets of rational individual judgments into a 'representative' rational collective judgment? If so, what do these rules look like, and what does it actually mean to be 'representative'? Most literature in the area has been considering Neutrality as one of the uncontroversial properties that a desirable rule should satisfy, by treating all issues under judgment equally. However, this claim has been unavoidably connected with numerous impossibility results that provide a negative answer to the first central question above. This talk is going to contest the Neutrality axiom, by first motivating the need for a relativistic treatment of issues under judgment, and subsequently by providing a formal way out of Neutrality's restrictions. After a family of Neutrality weakenings is introduced, we will show how some primary impossibilities are circumvented, as for example the one connected with the famous 'doctrinal paradox'. Finally, we are going to discuss the behaviour of these new axioms with respect to standard problems of judgment aggregation, such as the Safety of the Agenda.

This talk will be broadly accessible and not assume any prior exposure to judgment aggregation.