Our results so far have been largely negative—that the argument from self-defense to national-defense, in both its analogical and reductive forms, fails. It might be supposed that the only conclusion we can draw from this result will be a form of pacifism. In this chapter, however, I shall develop an argument for the positive conclusion that military action against an aggressive state could potentially be justified, not in terms of self-defense, but as a form of law enforcement. I shall argue, however, that such a justification is unlikely to prove effective without the establishment of something like a minimal universal state, and I shall attempt to show that there is a plausible contractarian argument for the justice of such a state. But before I can argue for this conclusion there is an important obstacle that must be navigated. If any form of military action (be it a defensive war, or a legitimate act of law enforcement) is to be justified, then we must provide a moral explanation of why the soldiers against whom one fights are appropriate objects of violence.
Nowhere does the phenomenon of the two levels of war generate greater difficulties and paradoxes for the Just War Theory than in the task of explaining how it can be permissible, in a defensive war, to kill the soldiers of an aggressive state. In other words, how does the Just War Theorist get from a right of national-defense held against a state to a right to kill held against a particular person? From our investigation of defensive rights in Part I , we have learned that if one is to be justified in harming a person in defense, then it must be the case that the object of the defensive force has forfeited or fails to possess a right against being harmed. I argued that an aggressor can only forfeit or fail to possess rights against being harmed if there is an appropriate normative connection between him as a subject and