The Logic of Reflective Altruism

ELEONORA CRESTO

San Francisco (EEUU)

Congreso; APA, Pacific Meeting; 2020

American Philosophical Association; Association of Symbolic Logic

Higher order likes and desires sometimes lead agents to have ungrounded or paradoxical preferences. In this talk I develop a dynamic logic of preferences that can help us gain insight into this phenomenon, in the context of games. In particular, I examine cases in which payoffs are interdependent and cannot be fixed, and hence the overall assessment of particular courses of action becomes ungrounded. Paradigmatic examples of this situation occur when players are 'reflective altruists' or ?reflective haters?, in a sense to be explained. I begin by describing the nature of interactions between reflective altruists and haters. In previous work I offered a framework to model such interactions successfully; I recall some of its main results here. I?ve argued that ungrounded payoffs cannot be captured by standard games with incomplete information. Rather, we need to rely on the concept of an underspecified game, in which the matrix of the game is radically under-determined. Players can then provide the necessary specifications through a second order coordination game for subjective probabilities. Players locked into ungrounded, but not paradoxical, preferences, may (but need not) succeed at the time of fixing a unique matrix for the first order game; players locked into paradoxical preferences, by contrast, can never fix a matrix.Next, I propose a dynamic preference logic that can mimic the search for a suitable matrix. Updates are triggered by conjectures on other players? utilities, which can in turn be based on behavioral cues. We can prove that pairs of agents with paradoxical preferences eventually come to believe that they are not able to interact in a game. As a result I hope to provide a better understanding of game-theoretic ungroundedness, and, more generally, of the structure of higher order preferences and desires.