LEGIT · Cognitive foundations of legitimate authority

When do people perceive authority as morally legitimate? Surprisingly little research in cognitive science has examined how the mind makes moral judgments about authority. Instead, most research in moral cognition has focused on non-hierarchical domains, such as when it is permissible to harm, or which distributions of resources are fair. This is a major gap in the literature, since moral obligations across human societies are structured by hierarchical relationships (e.g. parent-child, leader-follower, governor-governed). Why do human minds perceive that some individuals have a right to command, and others a duty to obey? Why do they see some uses of power as rightful and others as abusive? Why do people think that leaders are entitled to resources and deference? Answering these questions is essential, not only to understand the workings of our moral minds, but also to organize societies in ways people see as just. And it is also timely, given the crisis in legitimacy facing governments and democratic systems across European countries. To fill this critical gap, we will propose, formalize, and test a new theory of the cognitive algorithms underlying moral intuitions of legitimacy. First, we will build the first computational cognitive models of legitimacy intuitions, formalizing their mental operations as implicit negotiations over fair terms of hierarchy. Second, we will test quantitative predictions of these cognitive models against participant moral judgments of experimental scenarios about authority and obedience. Third, we will identify universal and culturally variable dimensions of these moral computations by running our experiments across 20 countries on five continents. This project will train the fellow in computational cognitive methods developed at MIT and Harvard (under Prof. Rebecca Saxe and Fiery Cushman) and transfer this expertise back to Europe, in a host institution specializing in cross-cultural research (Toulouse School of Economics, France).