The principle of noncombatant immunity is often regarded as a central moral truth. Tony Coady speaks of the “depth and centrality of the prohibition on intentionally killing the innocent” and says that it “functions in our moral thinking as a sort of touchstone of moral and intellectual health.” I want to preserve its moral centrality, but I have defended it as a derivative principle whose acceptance is connected to the role that it plays in diminishing the suffering caused by war. This particular defense is open to the charge that it reduces noncombatant immunity to a merely useful convention that at best provides weak protection for civilians.

Three aspects of my view invite this charge: my partial reliance on the status conception of innocence; my utilitarian defense of noncombatant immunity; and the resemblance between my defense of noncombatant immunity and a well-known, explicitly conventionalist defense of noncombatant immunity by George Mavrodes. In order to show that I have not reduced noncombatant immunity to a mere convention, I need to show 1) that the distinction between combatants and noncombatants is not purely conventional, 2) that rule utilitarianism is not a conventionalist theory, and 3) that my view differs from Mavrodes's conventionalist view of noncombatant immunity.

IS THE COMBATANT/NONCOMBATANT DISTINCTION MORALLY ARBITRARY?

Defenses of noncombatant immunity often begin by appealing to the idea that it is wrong to kill innocent people. They then infer that killing civilians is wrong.