Justifying preventive force: reply to Steven Lee.

Buchanan, Allen ; Keohane, Robert O.

Steven Lee offers a probing and fair-minded critique of our effort to show that the preventive use of military force could be justified if the decision to undertake preventive action was reached through an appropriate institutional process. Our responses to his thoughtful criticisms will help to clarify both the moral-philosophical arguments and the institutional proposal presented in our earlier paper.

THE DISTINCTION BETWEEN PREVENTIVE WAR AND LIMITED PREVENTIVE FORCE

Lee correctly points out that there are likely to be significant moral differences between preventive war and more limited uses of preventive force, such as a missile strike to destroy a nuclear facility. (1) We heartily agree, and concede that our paper may not have adequately emphasized this important point. However, we also believe that the institutional decision-making process we recommend can take this distinction into account, by requiring credible evidence, subject to vigorous contestation, if the claim is made that war and territorial occupation rather than a more limited use of force is necessary to avert a threat. In addition, as Lee notes, the decision-making process we propose would include measures to help ensure that the just war requirement of proportionality is satisfied, which would in turn require a careful consideration, ex ante, of the considerable risks that toppling a regime might engender.

THE DISTINCTION BETWEEN PREVENTIVE INTERVENTION AND HUMANITARIAN INTERVENTION

According to Lee, we "run together preventive intervention and humanitarian intervention, two categories that should be kept distinct" because "humanitarian intervention ... like traditional self-defense, involves avoiding harms that have already begun to occur" (p. 106).

On the contrary, humanitarian intervention, as usually understood, is military action undertaken not only to stop harms that are already occurring, but also to prevent new harms--to different individuals--that are about to be committed. For example, if UN forces had intervened to stop the genocide in Rwanda, they would not have limited their action to interrupting the harmful actions that were actually occurring at that time, but also would have acted to disarm and detain genocidaires who were poised to attack still more people. However, it is still useful, perhaps, to distinguish between humanitarian intervention in the usual sense and preventive military intervention to protect people in other countries from massive human rights violations, using the latter term to refer only to cases where the massive violations of human rights to be averted are expected to occur sometime in the future, but not imminently. Nothing we say in the article is inconsistent with this suggestion.

MEETING THE DEONTOLOGICAL OBJECTION

Lee claims that the most compelling objection to our argument is deontological in character: we have not rebutted the claim that "preventive action cannot be justified because the target of the intervention has done nothing to make itself morally liable for the intervention. Actual aggression can make the aggressor liable to attack, but expected aggression cannot" (p. 104). Our argument that prevention can be justified was twofold. First, many preventive actions are directed against actions already planned, in which case the planning constitutes an act justifying an armed response so long as concrete steps have been taken to begin to implement the plan. We invoked the analogy of the law of conspiracy to refer to such cases. Second, preventive action may be justified even in the absence of a conspiracy, when the potential victim has reasonable grounds for believing that unless he acts he will be subject to an unjust attack that will result in his death or serious bodily injury. This is also the case with respect to individual self-defense.

Lee seems to accept our proposition that planning an aggressive attack itself constitutes conduct that could justify intervention. His objection to our analogy with conspiracy law is merely that evidence for such conduct might be difficult to find. This objection, however, would be an argument for institutions that could more effectively produce reliable evidence, not an argument against intervention per se. His more fundamental objection relates to the second case, under which there may be a justified expectation of future attack but there has been no conspiracy. Lee's position is that the potential attacker cannot be morally liable to the use of force against him because he has not yet acted wrongfully.

But this argument fails to note that there are two distinct ways in which the potential attacker (A) can be morally liable. First, A can have done something, in a fairly robust sense, that makes him morally liable: A has either attacked B or is imminently attacking B, or, as we have argued, has wrongfully imposed a dire risk on B by forming a plan to attack B and taking the first steps to execute the plan, even if the harm is not imminent. Second--and more controversially--it may be the case that B reasonably believes that A has wrongfully imposed a dire risk on him even though A has not yet even formed a plan of attack, much less begun to execute it.

Admittedly, when it is reasonable for B to believe that A has wrongfully imposed a dire risk on him, it is normally the case that A will have undertaken some threatening course of conduct (for example, formed a plan of attack and begun to execute it). But this need not always be the case. A may have done things that are not in fact aimed at B's destruction, but that B may nonetheless reasonably interpret as imposing a dire risk upon him, perhaps because A has often behaved aggressively toward him or others in the past. In such cases, because of the risk that his past aggressive behavior will cause B and others to mistakenly suppose that he intends to act aggressively toward B in the future, A may have a duty to provide B with assurances that he does not intend to attack. The nature and extent of this duty would be determined by a consideration of the costs that it is reasonable for A to bear with respect to providing such assurances. In that sense there is an essential symmetry: the nature and scope of the right of self-defense is determined by the burdens of self-restraint that it is reasonable for individuals to bear, both as potential actors in self-defense and as potential targets of self-defensive action.

This account effectively addresses the deontotogical objection. It shows that in virtually any actual case of justified self-defense the target will in fact have done something wrong, such as forming and beginning to execute a plan of attack. It is not the case, however, that to be morally liable to the use of defensive force one must have done something wrong.

JUSTIFYING PREVENTION, NOT JUST PREEMPTION

In arguing that there is no categorical moral prohibition against engaging in the preventive use of force, we invoke a hypothetical example. A state has reliable information, from multiple sources, that a terrorist group has a deadly virus in its remote base camp, that it plans to release the virus in a major city, and that if the virus is not destroyed in its present location, later efforts to thwart the attack will almost certainly fail. Lee writes, "If the use of this weapon is imminent, as the story suggests, then a military strike to destroy the weapon would count as preemption rather than prevention. If its use is in fact not imminent, then it is not clear that force is justified, given the possibility of other options" (p. 106). In the example as we describe it, the harm is not imminent--for example, it may take several weeks for the virus to reach the target city. Furthermore, we described the example to make clear that there are not "other options" than a preventive strike; if the virus is not destroyed in the terrorists' base camp, and allowed to be moved from that location, there will be no way of determining where it has gone and little prospect of intercepting it at a point where the threat has become imminent. So the example stands as a case where preventive action is prima facie justified.

After praising our institutional scheme for its ingenuity, Lee asserts that it would be inapplicable "in just those situations where preventive intervention would be called for." Preventive intervention would be justified "only when the expected aggression to be prevented would involve a sudden and massive violation of human rights. In many or most cases of this sort, however, the exigencies of the situation would not allow for the ex ante part of the process [prescribed in our institutional proposal]." The difficulty, he thinks, is that the "publicity that would attend that part of the process would forewarn the expected aggressor ... allowing a dispersal its forces" (p. 101). But if the potential aggressor were deterred from continued depredations by fear of intervention, the objective of action would be achieved without the costs of war. This would seem not to be a cost but a clear benefit of our proposal for instituting mechanisms for ex ante accountability.

There could be a significant problem, here: in some cases, there might not be time to go through the entire deliberative process we outline. Under such circumstances, a state or group of states might nonetheless be justified in undertaking preventive action without prior approval--but only if there were adequate provisions for an impartial evaluation ex post as to whether invoking the "emergency exception" to prior approval was justified. A fully fleshed-out proposal would need to include such provisions.

(1) Steven Lee, "A Moral Critique of the Cosmopolitan Institutional Proposal," Ethics & International Affairs 19, no. 2 (2005), pp. 99-107. All in-text citation references are to this article.