The Universalizability Trap: Rational Agency, the Authority of Ought-Claims, and the Limits of Utilitarian Justifications for Killing

This article defends a formal argument concerning whether a group of rational agents ( A , B , and C ) can be rationally justified in killing a fellow rational agent ( D ) on utilitarian grounds. I do not contest the utilitarian calculus, and I do not claim that utilitarians derive morality from instrumental rationality. I argue instead for a general condition on when a normative claim binds an agent. A claim of the form “ S ought to φ” gives S a reason that S cannot rationally reject only if S already shares the end from which the claim is generated or if S’s acting otherwise is demonstrably incoherent. Utilitarian prescriptions to eliminate a nonconsenting agent satisfy neither condition with respect to that agent. There is no contradiction in D ’s continuing to pursue survival, and D is under no rational compulsion to adopt the end of aggregate welfare maximization. The utilitarian conclusion, however sound within its own framework, therefore carries no authority over D . Since the lethal coercion of a rational agent requires a reason that binds that agent, and the utilitarian prescription supplies none, A , B , and C are not rationally justified in killing D . I call this result the universalizability trap, and I distinguish it carefully from Kantian universalization and from Alan Gewirth’s argument to the principle of generic consistency.