Opinions and Preferences as Socially Distributed Attitudes

The dissertation focuses on how to best represent the consensus and attitude dynamic of a group given the attitudes of its individuals. This is done in the Bayesian epistemology framework using pooling with imprecise probabilities, and in utility theory extending Harsanyi's aggregation theorem to characterize other directed attitudes like spite and altruism. The final part of the dissertation considers attitudes within social networks and provides explanations and simulation models for online segregation and tribalism as well as the spread of rumors through contagion. The dissertation hopes to contribute to foundational issues like that of epistemic consensus, but also to new emerging phenomena in social epistemology.