Recent advances in economic theory, largely motivated by experimental findings, have led to the adoption of models of human behavior where a decision-maker not only takes into consideration her own payoff but also others’ payoffs and any potential consequences of these payoffs. Investigations of deontological motivations, where a decision-maker makes her choice not only based on the consequences of a decision but also the decision per se have been rare. We propose an experimental method that can detect an individual’s deontological motivations by varying the probability of the decision-maker’s decision having consequences. It uses two states of the world, one where the decision has consequences and one where it has none. We show that a purely consequentialist decision-maker whose preferences satisfy first-order stochastic dominance will choose the decision that leads to the best consequences regardless of the probability of the consequential state. A purely deontological decision-maker is also invariant to the probability. However, a mixed consequentialist-deontological decision-maker’s choice changes with the probability. The direction of change gives insight into the location of the optimand for one’s duty. We provide a formal interpretation of major moral philosophies and a revealed preference method to detect deontological motivations and discuss the relevance of the theory and method for economics and law.