Although much of this roundtable focuses on the legal status of preemptive war, international law has rarely, if ever, constrained governments from initiating hostilities. It may impel decisions to go to war, if the motive is to enforce legal norms against aggression already perpetrated against a third party. But in cases of starting a war altogether, legal conclusions figure mainly as rationalizations or public justifications for decisions made on other grounds. I am aware of no case in which international law has blocked a decision to wage war—that is, a case in which a government decided that strategic necessity required war yet refrained because international law was deemed to forbid it.
Domestic law is a different matter, at least in stable constitutional systems like the United States. American presidents usually adhere to domestic legal constraints, if only because they can be removed from office if Congress decides that they have broken the law. No international court or legislature can enforce judgments of what the law requires. What gives law a causal, as distinct from normative, effect is the existence of effective mechanisms for its adjudication and enforcement apart from the relative power of the parties to the dispute. These mechanisms do not exist in the international system, where the parties (states) are themselves the effective adjudicators and enforcers. Crises managed by negotiation are the analogue to settling out of court in domestic law, and war is the analogue to litigation. In real life, statesmen decide what they believe is right to do, and, in countries like the United States, what domestic law allows. Then they find a lawyer to tell the world that international law allows it.
Every decision for war rests on moral judgment, even if it is the morality of raison d'état—the interest of the community represented by the state.1 What the state should do in the external realm for the safety of its citizens is the moral judgment made by the government. The international legal status of rationales for war may matter for discussions of how governments should decide for or against initiating war. It does not have much to do with how governments do decide. Moral considerations do apply in actual decisions, but they are extruded from perceived strategic necessity. Interpretation of the relevant international law is made to conform to moral instincts. If one is interested in influencing the practice of governments in choosing preventive or preemptive attack as a policy, rather than in how governments justify their choices for public relations, or why they are criticized in diplomatic forums, the key is to analyze the strategic wisdom or error of these military options rather than competing assertions about their legality.
LAUNCHING A WAR
In theory, there are four types of strategy for war. There are offensive strategies driven by aggressive aims—unprovoked and predatory attacks—and defensive strategies—where the victim waits to be struck rather than striking first. The other two involve defensive motivations for strategic offensives: preventive and preemptive wars. Whether a given attack was aggression or self-defense depends on which side is making the characterization. Victims always cite enemy attacks as aggressive, perpetrators always cite them as preventive or preemptive. With few exceptions, governments making war believe that they are acting defensively and legitimately.
Preventive war is almost always a bad choice, strategically as well as morally. Preemption is another matter—legitimate in principle and sometimes advisable in practice. The rationale for preventive war is that conflict with the adversary is so deep and unremitting that war is ultimately inevitable, on worse terms than at present, as the enemy grows stronger over time. Thus it is better to face the music sooner, when chances of military success are greater. It is almost never possible to know with enough certainty that war is inevitable, however, to warrant the certain costs and risks of starting it. Conflicts often cool with time, sometimes even turning enemies into allies.
Preemption is unobjectionable in principle, since it is only an act of anticipatory self-defense in a war effectively initiated by the enemy. If the term is used accurately, rather than in the sloppy or disingenuous manner in which the Bush administration has used it to justify preventive war against Iraq, preemption assumes detection of enemy mobilization of forces to attack, which represents the start of the war2. Beating the enemy to the draw by striking before he launches his attack is reactive, even if it involves firing the first shot. Striking first may be the only way to avoid the consequences of being struck first in the very near future.
In practice, however, it is rarely possible to be sure that enemy preparations for war are definite, or are aggressively motivated, rather than precautionary reactions to rising tension and fear. This uncertainty reflects the standard international relations theory concepts of "security dilemma," "crisis instability," and "reciprocal fear of surprise attack," which made avoiding conditions that could foster mistaken preemption the main focus of concern among Cold War strategists. Nevertheless, one cannot responsibly foreswear preemption. It is rarely possible to be sure that observed enemy preparations make war immediately inevitable, but there is more basis for such a judgment than for the gamble on the more distant future represented by preventive war. Preemptive war is more legitimate than preventive war not because of a moral difference between the two in principle, but because of a practical difference in the weight of evidence that the adversary is bound to attack at some point.
Precisely because antagonists divided by a political conflict of interest are rarely certain whether each other's mobilization and preparations for war are aggressive or defensive, preemption is extremely rare. In a study of the concept, Dan Reiter counts only three preemptive wars in the past century: World War I, the Chinese intervention in Korea, and the Six Day War of 19673. Preventive wars, however, are common, if one looks at the rationales of those who start wars, since most countries that launch an attack without an immediate provocation believe their actions are preventive.
MODELS FOR STRIKING FIRST
Are there good examples of preemptive or preventive war—that is, ones that were proper to fight? Taking the most promising of the two categories—preemption—only one actual case seems clearly right: the Israeli attack on Egypt and Syria in June 1967. (Approving preemption at that time in no way implies approval of Israeli policy in the years since. Colonization of the West Bank has been strategically disastrous as well as illegitimate.) The closure of the Strait of Tiran and political rhetoric at the time made the circumstantial evidence that the Arabs were preparing to attack Israel as good as such evidence ever gets. Israel could not rely on a defensive strategy for two reasons. First, within the 1967 borders it was vulnerable to being dismembered by an Arab offensive. Second, Israel had to fully mobilize reserve forces in order to match Arab military manpower and could not maintain that level of mobilization for long without risking economic collapse, as reservists were kept out of the civilian labor force. The surprise attack of June 5 enabled numerically inferior Israeli forces to thwart the Arab military threat.
Reiter's other examples are not easy to approve. Although the Chinese feared that U.S. forces would have moved against them if the advance to the Yalu River on the Korean border had been completed, it is almost certain that would not have happened. And World War I makes the negative case most starkly. It is the best example of the danger in preemption that has been identified by deterrence theory and literature on crisis stability—a catastrophe brought on by mistaken hopes for a short, successful war. The European great powers relied on rapid mobilization and offensive operations in their war plans, so once any of them began mobilizing, all had to follow suit to avoid being victimized by their adversary's first strike or leaving their allies in the lurch.
There are ample cases of preemptive actions that would have been justifiable but were not undertaken, and countries fell victim to surprise attack. For example, if American forces had been able to detect and hit the Japanese carriers while they were on the way to Pearl Harbor, it would have been reasonable to do so. But measured against other cases that could have turned out like World War I or worse—the Cuban Missile Crisis, for example, when the U.S. Strategic Air Command was on hair-trigger alert—these missed opportunities do not warrant a more general endorsement of preemption as a strategy.
Are there good past examples of preventive war? Perhaps, but I cannot find one. In contrast to the preemptive Israeli attack in 1967, the preventive Sinai campaign in 1956 accomplished little, since the Israelis, along with the British and French, evacuated the peninsula soon after capturing it and did nothing to suppress the long-term threat to Israel from Arab countries. If anything, the short-lived military success further embittered the Egyptians and spurred the actions a decade later that produced the Six Day War. Many consider the 1981 Israeli air strike on the Osirak nuclear reactor a good preventive action, on grounds that it set back Iraq's attempts to develop nuclear weapons. It is not clear, however, that by 1981 Iraq had come close to developing plutonium reprocessing or uranium enrichment facilities, which, in contrast to a simple reactor, are the critical facilities for a weapons program. It is clear, though, that within a decade Iraq did develop an enrichment capability. Thus, the Israeli attack on the reactor did not destroy the crucial ingredients in a weapons effort, nor did it interfere with subsequent Iraqi efforts in the 1980s. It is hard to determine in fact whether the strike against Osirak retarded Iraq's nuclear weapons program or spurred it4.
Are there preventive wars that were not fought but should have been? The one most often cited is the French decision not to attack Germany when Hitler remilitarized the Rhineland in 1936. If the Nazis had been stopped then, when they were weak, Europe might have been spared the apocalypse of World War II and the penetration of Soviet power into the heart of the continent during the forty years of the Cold War. Even if the Nazi regime had been overthrown, however, the problem of disproportionate German power and unresolved German grievances would not have been settled, and the potential for conflict and eventual wider war would have remained. Nevertheless, this is the best example imaginable to justify preventive war.
The problem in 1936 was underestimation of the long-term threat. The issue then was not blocking immediate aggression—the Rhineland was German territory—it was whether the legal restrictions imposed by the victors of World War I on German military presence in their own territory under the punitive Versailles treaty should continue to be maintained by force. If international law was to be determinative in this case, it would have been an impetus to war rather than a restraint. Moreover, no one then considered another war inevitable, much less predicted the full horror of 1939-45.
In the Cold War, however, many in the West did believe that a third world war was a significant possibility and believed that Stalin in 1950 was comparable to Hitler in 1936. Some argued in favor of preventive war against the Soviet Union, and later against China. Secretary of the Navy Francis Matthews, Senator John McClellan, and General Orvil Anderson, to name a few, promoted an attack against the USSR in the early 1950s, and such a move was considered in studies early in the Eisenhower administration. Destruction of developing Chinese nuclear facilities was also weighed by a number of figures in the national security community in the 1960s5. In hindsight such proposals appear fatally reckless, based on overestimates of the communist powers' propensity to undertake direct military aggression. At the time, however, Stalin and Mao were widely viewed in the same light as Saddam Hussein has been early in the twenty-first century: wildly aggressive fanatics whose lust for conquest could not be deterred. In these cases, proposals for preventive attacks never got very far, but they were seriously considered at levels not far below the top of the government.
Within a few short years of recommendations for preventive war against Stalin, the threat that he posed had changed: he was dead. The case for preventive war against China in the 1960s, in turn, was washed away when Richard Nixon's secret diplomacy produced the rapprochement of the early 1970s. Overnight, Mao went from being a dire threat to a tacit ally against the Soviet Union. The peaceful end of the Cold War in both cases demonstrated the wisdom of waiting the adversary out, relying on containment and deterrence rather than precipitating a showdown that turned out to be unnecessary.
LESSONS AFTER THE COLD WAR
If a government could ever know with absolute certainty what the future would bring, decisions for or against striking first would be easy. Uncertainty forces choices that pose risks either way. In the war against terrorism, these questions are not at issue. Having already been attacked, it is logical for the United States and other victims to strike first against al-Qaeda and similar groups whenever doing so is militarily feasible and effective, and not politically counterproductive. The issue arises in regard to states who have not attacked us—at least not yet. This distinction between Iraq and al-Qaeda, obscured in much discussion of this issue, must be clearly maintained.
What should be the model for how to deal with the threats posed by the dangerous regimes in Baghdad and Pyongyang, or for that matter, Tehran, the third member of Bush's patchwork "axis of evil"? Hitler in 1936, or Stalin and Mao during the Cold War? If any of the "axis of evil" were great and growing powers, there might be more of an argument for preventive war. But even if they assumed Nazi proportions, there is one crucial reassuring fact that differentiates the current situation from the 1930s: none of these regimes has any allies of consequence, and all the great powers are arrayed against them. If Hitler had been faced at Munich in 1938, or even in September 1939, by a united front of all the other great powers, indefinite containment of Germany until the passing of the Nazi regime would have been plausible. Instead the United States remained uninvolved, and the USSR struck a short-lived nonaggression pact with Germany, leaving the French and British to face Germany, soon to be supported by Italy and Japan, alone. In any case, the failure of containment in the Munich crisis is not an argument for striking first. The British and French might have gone to war in 1938, but in defense of Czechoslovakia when it was attacked. It is hard to imagine them striking the first blow, before the Germans initiated war.
But of course there is now another difference that points in the other direction. What of the unique threat posed by the rogue states' weapons of mass destruction, likely to grow more potent with time, and to function as something of an equalizer? This indeed is a worrisome difference from the 1930s-but it is not different from the Cold War. Antagonistic great powers survived more than four decades of confrontation in the shadow of nuclear war on a vastly more destructive scale. It is not comforting to rely on deterrence to contain aggression, but it is better than precipitating precisely the clash that is feared. Despite common fears in the early Cold War that the threat of nuclear retaliation would not suffice to prevent a Soviet or Chinese attack, it did. Is the Cold War record irrelevant? Despite common assertions in recent years that Saddam Hussein and Kim Jong-il are crazy and undeterrable, evidence has yet to demonstrate this. Yes, rogue state leaders have been risk-prone and have frequently miscalculated. But North Korea and Iraq attacked their neighbors in 1950, 1980, and 1990 only when the United States failed to deter them. Indeed, Washington gave them a green light in all three cases. In 1950 Secretary of State Dean Acheson's speech to the National Press Club excluded South Korea from the U.S. defense perimeter in Asia six months before North Korea struck. In 1980 the United States was engaged in a vicious struggle with Iran, and did nothing to discourage Saddam from attacking. In 1990 the Bush administration instructed Ambassador April Glaspie to tell him that the United States had no position on Iraq's dispute with Kuwait.
When the United States has posed deterrent threats against Iraq and North Korea, they have worked. Going to war in 1991, the Bush administration warned Saddam of dire consequences if he used weapons of mass destruction during the war. Despite humiliating defeat, Saddam held those weapons back. Nor has Iraq attacked any of its neighbors since American deterrence has been made clear over the past decade. North Korea, in turn, has not repeated the mistake of 1950 in the fifty years since the Korean War ended and an American trip wire was institutionalized along the 38th parallel.
If fully reliable intelligence is ever obtained that an adversary is preparing to attack, and if striking first can reduce the damage that will otherwise be absorbed as a result of waiting to defend against the blow, preemption is the moral decision for any responsible guardian of national security. But those are two big ifs. Neither of the conditions is often met. Some strategists hope that arms control arrangements that discriminate against forces with offensive capability, and in favor of those best configured for defensive military operations, can promote crisis stability. In theory this would make defense rather than preemption the more efficient strategy for reducing losses in the event of enemy attack. In practice, however, this aim has proved difficult if not impossible to implement.
The conditions for legitimate preventive war are met even more rarely, if ever. Because costs in initiating either preemptive or preventive war are certain, while the probability that the enemy will eventually strike is less than 100 percent, the burden of proof is on the case for striking first. That is essentially a matter of careful analysis of the nature of the threat. The international legal validity of military action flowing from that strategic analysis will be disputed. The legal opinions that could affect the decision would be those of American officials. Since U.S. courts traditionally avoid ruling on national security matters because of the "political question doctrine," the only opinions likely to figure in the process would be those of State Department lawyers, and more specifically, those among them whose interpretation best serves the decision taken on strategic grounds.
1 Of course, corrupt leaders or particularistic groups in poorly institutionalized political systems may effectively substitute their own interests for those of society at large.
2 For a discussion of preventive and preemptive war, see Richard K. Betts, Surprise Attack: Lessons for Defense Planning (Washington, D.C.: Brookings Press, 1982), pp. 141-46; Betts, "Surprise Attack and Preemption," in Graham T. Allison, Albert Carnesale, and Joseph S. Nye, Jr., eds., Hawks, Doves, and Owls: An Agenda for Avoiding Nuclear War (New York: W. W. Norton, 1985), pp. 54-79; and Betts, "Suicide from Fear of Death?" Foreign Affairs 82 (January/February 2003), pp. 34-43.
3 Dan Reiter, "Exploding the Powder Keg Myth: Preemptive Wars Almost Never Happen," International Security 20 (Fall 1995), pp. 5-34.
4 Richard K. Betts, "Nuclear Proliferation After Osirak," Arms Control Today 11 (September 1981), pp. 1, 2, 7.
5 Marc Trachtenberg, History and Strategy (Princeton: Princeton University Press, 1991), pp. 103-18, 132-46; C. L. Sulzberger, An Age of Mediocrity: Memoirs and Diaries, 1963-1972 (New York: Macmillan, 1973), p. 463; Robert S. Litwak, "The New Calculus of Pre-emption," Survival 44 (Winter 2002-03), pp. 61-62; William Burr and Jeffrey T. Richelson, "Whether to 'Strangle the Baby in the Cradle': The United States and the Chinese Nuclear Program, 1960-64," International Security 25 (Winter 2000/01), pp. 54-99; and Robert M. Lawrence and William R. Van Cleave, "Assertive Disarmament," National Review, September 10, 1968, pp. 898-905.
6 Louis Henkin, Constitutionalism, Democracy, and Foreign Affairs (New York: Columbia University Press, 1990), ch. 6