Reflecting on social influence in networks

This paper builds on the logical model of opinion dynamics under social influence in networks proposed by Liu, Seligman, and Girard (2014) as well as on the generalization proposed by Christoff and Hansen (2013). While both accounts of social influence show interesting dynamics, they both assume that agents do not reflect on how they are affected by such influence. This paper shows that, if agents are allowed to reflect upon the very rules of social influence, they may come to know (or "learn") other agents' private opinions, even though they can only observe their public behavior. To represent formally agents who are able to reason about social influence, a logic of social networks, knowledge, influence, and "learning" is introduced.