A Computational Theory of Dignity

People seem to hold a variety of conflicting intuitions about the concept of dignity. Here, we seek to reverse-engineer the computational basis of these intuitions. We propose a Bayesian model that casts intuitions about dignity as computations about what agents' actions imply about their own and others' social rank. We show through two behavioral experiments that our model captures people's intuitions well, reconciling seemingly conflicting intuitions with a common computational framework.