Striking First: Preemption and Prevention in International Conflict

JOANNE MYERS: Good afternoon. I'm Joanne Myers, Director of Public Affairs Programs, and on behalf of the Carnegie Council I'd like to thank you all for joining us.

I believe we are all in for a very stimulating presentation, as our podium is graced by two world-class scholars.

Michael Doyle is a political scientist who is currently teaching at Columbia University. Recently, he served as UN Assistant Secretary-General in the Executive Office of Kofi Annan. As a scholar and as a practitioner, he has throughout his career tackled some of the most challenging concerns of the day. This afternoon he will be discussing the difficulties in developing standards to govern the ethics of self-defense of one's country in advance of deadly harm.

Harold Koh is the 15th Dean of the Yale Law School. As one of the world's leading experts on international law, Dean Koh has been widely recognized for his integrity and vision. From 1998-2001 Dean Koh served as Assistant Secretary of State for Democracy, Human Rights, and Labor, and has received international praise for his work as a human rights advocate.

After you've read their CVs, which you all should have picked up at the front desk, I know you will understand why they are both held in such high esteem. Their commitments to issues of international concern, experience, and expertise speak for themselves.

Since 9/11, profound changes have taken place in American foreign policy, reversing many longstanding commitments that had earned us respect around the world. For at least half a century, the United States had, for all intents and purposes, formally agreed to international constraints on the right of any nation, including itself, to start wars. But when George Bush launched his war on Iraq and declared the United States of America has a sovereign authority to use force in assuring its own national security and would seek to preempt emerging threats, he brazenly defied international norms.

In Striking First, Professor Doyle shows how the Bush Doctrine has consistently disregarded a vital distinction in international law between acts of preemption in the face of imminent threats of an armed attack and those of prevention undertaken in order to forestall the growing offensive capability of an enemy.

He addresses not only the underlying moral question of the conditions under which preventive war is justified, but also talks about the complex practical question of how, if at all, international law should be refashioned in the current era of terrorist networks and heightened insecurity to accommodate acts of preventive war or anticipatory self-defense. He contends that international law must rely more completely on the procedures of the United Nations Charter and develop better standards for dealing with serious threats. After explaining how the UN can play an important role in enforcing international law and strengthening international guidelines for responding to threats, he describes the rare circumstances when unilateral action is indeed necessary.

Striking First is based on the 2006 Tanner Lectures on Human Values delivered at Princeton University. Inviting Dean Koh to comment on his presentation provides a lawyerly texture to Michael's argument. Dean Koh will offer an invaluable contribution that will clarify and enrich our understanding about international legal standards and norms involved in the authorization of preventive war.

In conflict situations, strength and might are important—probably most important—but there are good practical reasons to sometimes defer to procedure, law, and the judgment of others. As the post-Bush world starts to take form, there will be a continued need for creative thinking in finding intellectual resources to rethink the challenges of this new century. I believe two of the most gifted thinkers on this subject are with us today.

Please join me in welcoming Michael Doyle and Harold Koh

MICHAEL DOYLE: Thank you, Joanne, for those very kind remarks.

As Joanne mentioned, I will be drawing upon this book, Striking First, which I delivered as lectures at Princeton a couple years ago.

The key question in it is: When should states be permitted to go to war or impose sanctions or blockades before they have been attacked, in their own self-defense?

In March of 2003, we will all recall, the United States and a coalition of the willing invaded Iraq, partly, it was said, to enforce Security Council law and partly to defeat enemies abroad before they attack us at home. Were they right to do so?

Today, some are speculating about how to prevent North Korea from expanding its nuclear arsenal, or Iran from acquiring nuclear weapons.

So my topic is a controversial one. I attempted to answer it straightforwardly, but with full knowledge that anything I would have to say could only be preliminary. Indeed, I also know that it is so controversial that an image comes into my mind. Talking about preventive self-defense today, in the wake of the Iraq fiasco, is something like interviewing the passengers on the lifeboats of the Titanic about their views on ocean travel. This is not exactly the right time to have that full and deep understanding.

But I think we need to do it. We need to deliberate as scholars and policymakers. It is absolutely essential.

I have already benefited from some quite wonderful deliberation. When I delivered these lectures, I had the great privilege of having four wonderful commentators—one seated exactly to my right—and from their comments I was able to, I can assure you, make a number of improvements in the text that you have, which sells for the very reasonable price of $24.95 and may be available this afternoon.

JOANNE MYERS: It is available.

MICHAEL DOYLE: So you will get a better version in print than you would have heard if you went down to Princeton to hear it a couple years ago.

My question focuses on what the law is and what we should do: the ethics, the politics, and law of anticipatory—that is, preemptive and preventive—self-defense.

I start with a problem. In my view, existing international law is inadequate; the Bush Doctrine is inadequate; and the procedures that are embedded in the Security Council have flaws attached to them, they are not perfectly reliable. So we need something better.

We can't rewrite international law. There is no overlapping consensus that allows that to be a reasonable enterprise. But what we can try to do is develop standards that will be useful in making better decisions than the ones that have been constrained by existing international law or inspired recently perhaps by the Bush Doctrine. So that is my aim, to try to develop these standards that, were they to be adopted, would produce better outcomes all around compared to either of those two.

So what are my grumbles with the status quo?

Well, the substantive doctrine of existing international law goes way back. The most basic principle is that force can only be used in self-defense once an armed attack has occurred—that's Article 51 of the UN Charter—or that it is so imminent that there is nothing but a reflex action that is possible.

That standard of imminence goes right back to a case between the United States and the British Empire, 1837-1841. It was settled in an exchange of notes between Secretary of State Daniel Webster and the British Envoy, Lord Ashburton. They decided to push aside that dispute, irrespective almost of some of the facts. They agreed upon a set of standards, even though they couldn't agree upon the facts, and the set of standards that they agreed upon were designed to make sure that force could almost never be legitimate for these purposes. The threat had to be overwhelming in its necessity, leaving no choice of means and no moment for deliberation. These rules then became embedded in customary international law.

Looking around, these rules are so extreme, merely a reflex action, that I can only find one case that seemed to fit within it when I looked through some of the record. That is the case where the Netherlands preempted an attack on December 8, 1941, that was likely to be launched against it by Japan. Even those facts are a little problematic, but that is one case that is close to it. There just aren't others. The rule has not been doing that good a job of allowing for a genuine use of force on an imminent basis.

But that's not the real reason. It's our own times today, I think, that make it more problematic for us. It doesn't deal well with the threat of mass destruction coming from suicide terrorists. It doesn't deal well with the threat of genocidal projects in the hands of states that are determined to exercise that kind of force, or states planning imperial ingression using weapons of mass destruction. These kinds of threats are different than the kinds of threats that were assessed in 1837 by Secretary of State Webster and Lord Ashburton.

Now, many threats can be safely appeased, and enemies can be turned into cooperating states, as China was because of good diplomacy in the 1970s. For the rest, deterrence will work, or denial—many other standard measures are there. But deterrence does not work comprehensively.

Saddam Hussein was rationally deterred from using WMD [Weapons of Mass Destruction], the chemical weapons they produced, on U.S. forces and Israel in 1990, and there is no reason to believe that he would not have stayed deterred from similar attacks on the United States or Israel. His regime and his own survival were at stake.

Nevertheless, continuing to prevent him from acquiring robust weapons of mass destruction capability, particularly one with an intercontinental reach—missiles that could go further—was quite a worthwhile thing to do. The UN sanctions imposed upon him were designed to do that, and we now know they were quite successful in achieving exactly those purposes.

But still there was a danger from Saddam Hussein. A Saddam Hussein with imperial ambitions and weapons of mass destruction would have been a constant threat—not for the United States or Israel, but for his neighbors, neighbors like Kuwait of course. It would have been a constant source of brinkmanship. He would have been nearly impossible to dislodge from Kuwait in 1990 and 1991 if he had acquired weapons of mass destruction that he could deliver at a great distance.

Would we really have gone into Kuwait if it would have meant the loss to nuclear explosions of Israel or Europe or even the United States? We would have traded it off. And that's too bad.

So the sanctions I think deserve applause. They were a case of good prevention.

And then, when you look at the extremist actors, like al-Qaeda and the Tamil Tigers, they are suicide terrorists. Almost by definition, they cannot be deterred. They need to be, to the extent that one can, prevented. That is, I think, a challenge that still exists.

Now, they kill small numbers. Only 3,000 American citizens were killed on 9/11. That's quite something, but it's still a small number as world politics go. But it's not the same thing as traffic accidents. A state can't sit by and allow its citizens to be the prey of those who would murder them. There is a duty of a state to act in these regards, and one needs to take what measures one can, if one can do so legitimately.

So there is a problem, as I see it, with the existing Caroline standards.

There is a second problem with the Bush Doctrine, if you will, flipping to the other side. The Bush Doctrine seems like a response to these new threats—you know, things like Saddam Hussein and suicide terrorists and others—and they are designed to be preventive, acting before Caroline measures would kick in.

At Ford Bragg, for example, when the President announced the doctrine of prevention, and he said after September 11, "I made a commitment to the American people this nation will not wait to be attacked again. Many terrorists who kill innocent men, women, and children on the streets of Baghdad are followers of the same murderous ideology that took the lives of our citizens in New York and Washington and Pennsylvania." "There is only one course of action," he said, "to defeat them abroad before they attack us at home." There are all sorts of problems piled up into that quote.

But the biggest problem is the danger of allowing one state, even a state such as the United States, to adopt what are inherently subjective standards; that is, the threat that it happens to perceive. This is an invitation to chaos, if all states do pretty much the same. If all states say that whenever they feel threatened they can use force, we'll be in a position that every state will be preempting the preventive responses of every other state. We will be in a world of chaos.

So the problem is the Caroline standards are too strict, the Bush standards are much too loose. The way one deals with this kind of uncertainty in international law and international politics is through procedural standards, procedural rules.

And wisely, in 1945, the founders of the UN Charter set up an authority to engage in prevention, and that is the Security Council. The Security Council has perfect legal right—indeed, one could even argue, something close to a duty—there is a "should" element in it—to identify any threats and to take measures that are adequate to deal with those threats. Now, that depends upon a political process. But there is a responsibility in that institution to engage in prevention when it is justified. Those measures have been binding upon all UN members under Article 25.

So anytime that you can get a vote of nine of the 15 members, including the Permanent Five members (United States, Russia, France, United Kingdom, China), then fully legal preventive action can be undertaken. You can therefore engage in preventive action, ranging in everything from obligatory sanctions, through blockades, through invasions in force, if it is justified and if the vote is there.

So it looks like we have a solution. But there are two problems I see with the Security Council as a solution.

One of them is in the past the Security Council has not always behaved responsibly. There have been times when the use of force was justified and the Security Council was not prepared to vote in that way. There have been vetoes that prevented it, even for cases of the use of force which were not purely defensive but in the broader public interest of the world, as in humanitarian interventions. If you think the Kosovo intervention was a justified one, it was not one that could get through the Security Council. And then, in votes on Bosnia and Rwanda, there were also problems of way too late to act, and some people say that is the case in Darfur today. So that is one problem; it is not a completely reliable process.

The second problem is that the Security Council itself lacks the substantive standards that could be used to determine when it should vote for an act of prevention; and when, if it doesn't vote, we might be able to say that it is not acting according to its responsibilities. So process is not enough.

So let me turn to what I think can replace it.

I think we need better standards. We need to go back to the Just War tradition and begin to think through questions like how large is the danger, lethality; how likely is it to occur? And we need to do better than some scholars, some hard realists, who say, "All capacities are dangerous, increases in power, we've got to stop it"; or some liberals who say, "All authoritarian states are danger, they're a legitimate target." We need to do better, get more nuance. We need to look into behavior.

Michael Walzer once wrote a good chapter where he talked about the need for something beyond preemption. He said we need to look at manifest intent, we need to look at active preparations, we need to look at where the danger is increasing. Others have looked at a pattern of rapacious behavior and a failure to give guarantees. Those are the kinds of behavioral standards we need to add to the realist and liberal philosophies.

And we need to take decisions that are legitimate in Just War terms. They need to be proportional—one should exercise no more force than is necessary to deal with the threat. There need to be last clear chances—one shouldn't act if one doesn't need to act now. And there should be deliberation wherever one can find it.

My last standard is legality. We should pay special attention to law. We need to pay attention to whether the threat has actually broken the law. Is it engaged in massive human rights violations? Has it been aggressive? Is it, in acquiring its threatening capabilities, doing so in ways that are illegal, in violation of things like the NPT [Nuclear Non-Proliferation Treaty] and other standards?

And is our response as legal as possible? All of these measures to some degree break the law, but there is a real difference, in terms of respecting sovereignty and self-determination, between an invasion and an occupation, on the one hand, and a blockade on the other.

It is quite important that in the Cuban missile crisis, which I regard as an act of justifiable prevention, that there was a blockade and not an invasion of Cuba. A blockade one can justify. An invasion, in my view, one could not. That's the difference in law and proportionality.

So I think one needs to follow the law. One needs to, number one, go to the Security Council and deliberate there carefully.

If there is an illegitimate veto, if the Security Council should have endorsed an act of prevention and does not, one needs to go through the domestic processes of law, including the Congress. That is absolutely essential.

If in the end a state does engage in unilateral prevention, it needs to have its record assessed by a national commission and an international commission so that, hopefully, it will be deterred from acting irresponsibly by having an incentive, knowing that its record will be weighed, in order to act more responsibly.

In the book I talk about sanctions against South Africa that were fully endorsed by the Security Council in 1977, by the threat posed by apartheid; the Cuban missile crisis, a justifiable case of a blockade; the Israeli attack on the Iraqi reactor in 1981; President Clinton's various attacks on Osama bin Laden in 1998, on where he thought Osama bin Laden was; and then I wind up with a discussion of the invasion of Iraq in March 2003, which I regard as unjustifiable because it was not necessary, the sanctions were working.

I want to stop here by just giving you two little hypotheticals.

One is a case of what I regard as justifiable preemption. The president gets word that an aircraft has been hijacked by terrorists, they have a radioactive bomb onboard, and it is headed toward an American city. This is a clear case of Carolinian preemption. The president should act. It is covered by the inherent right to self-defense, with the Carolinian standards attached of overwhelming necessity, no time for deliberation. And the president can act, should act, even if, tragically necessary, shooting down that aircraft before it drops its radioactive bomb on an American city. The president does not need to consult the Security Council, nor even the U.S. Congress, in these terms. This is within the purview of the duties of the office. That is an act of legitimate preemption.

What is an act of legitimate prevention, looking further down over the horizon? Here is a scenario. The president obtains reliable intelligence that a terrorist group has purchased a nuclear weapon or is developing a nuclear weapon or other forms of weapons of mass destruction that it can deliver. This group has announced publicly its determination to harm the United States, and it has acted in the past to attack some U.S. targets, like U.S. embassies for example. The country that harbors this group is unwilling to do anything to arrest or control the group.

The president then should bring this threat to the attention of the U.S. Congress—by definition, this is prevention (there is time to act)—and he should request their authorization that he is going to take this to the Security Council. He should bring it then to the Security Council and deliberate there and make the case that this is lethal, it's likely; this is going to be a matter of proportionality; he is going to be using legal methods, looking for Security Council approval.

He or she should make an argument that this is well evidenced, present the evidence there: It is not controllable by lesser means, such as economic sanctions; it reaches the level of lethality and likelihood that is relevant; collateral damage can be controlled; it is proportional; and that the terrorist state is acting illegally in harboring these terrorist cells.

Now, in a good world the Security Council will then line up and vote to endorse a preventive action by the United States in this particular scenario. In an ideal world, the state harboring these terrorists will then turn around, knowing it has the world against it, and arrest them, and maybe prosecute them, for these violations.

But in a much less good world, one or more of the P5 members vetoes the authorization, even though, based upon the merits of the case and the past record of other acts of prevention, they should have voted for it. Under those circumstances, the president should go back to the Congress and ask for a declaration of war. And, after a last round of diplomacy with the state that is harboring these terrorists, trying to persuade them to hand them over, he should then engage in a strike that is surgical, designed to prevent any noncombatant casualties.

Having done so, the president then should appoint a national commission to investigate his or her actions, and the Security Council would appoint another commission to investigate the actions of the United States. So a full and accurate record is then put forward.

So that is what I am proposing. These are standards for prevention that look at lethality, likelihood, legitimacy, legality, proportionality. They are not a recipe. They are not something that one can just plug in and have an answer pop out the other end that is absolutely scientifically reality. Life and death and national security do not lend themselves to recipes. At best, they are an invitation to improve the deliberation and develop self-consciously a body of powerful examples that can serve as better guidelines in the very dangerous world we live in.

I'll stop there.

JOANNE MYERS: Thank you very much. We'll now turn it over to Dean Koh for his response.

HAROLD KOH: I'm here because Michael Doyle is both a friend, a colleague, and an inspired scholar who has written a very important and incisive book. I thank him for the book and thank him for inviting me to participate both in the original lectures and also here.

I also want to thank the Carnegie Council as well as Princeton University Press for inviting me. And finally, I want to thank Sarah Palin for reminding everybody what the Bush Doctrine is.

This weekend I was reading Michael's book and thinking about why, as much as I admire it, I don't really buy it all the way. Then I read an op-ed piece by Garrison Keillor from "Prairie Home Companion."

What he had done was he had gone up to Wasilla, Alaska, to try to discover Sarah Palin's views. He described that suddenly, out of the woods, bolts a large moose, running vaguely in his direction. It occurs to him maybe he should apply the Bush Doctrine and shoot the moose.

I laughed, thinking about that. And then it occurred to me: Why am I not comforted by the thought that before he could shoot the moose he would first have to go to the Security Council? If it turned out that the Security Council was not ready to act, that you would balance four L's—lethality, legality, legitimacy, and likelihood—and then if all else failed, you would just go ahead and shoot the moose.

I realized that the reason I wasn't comforted by that analysis, which is at the base of the exception in Michael's pretty much admirable rule, is that the person who would be making the decision is Sarah Palin.

The question is: Can you trust the head of state to make these kinds of ex ante legal judgments? Or are you better off doing what we do in law, which is to say there's a per se rule against; if you end up breaking the rule, you have to defend yourself, you have to put forward some kind of necessity defense.

But when we have a set of activities in the law that we think are just wrong or illegal—like illegal search and seizure, sexual harassment, price fixing, prior restraints—we ban them, clearly. And then, if somebody wants to break that rule because they think they are in a justified situation, they can, but they know they are operating within a zone of illegality and have to defend themselves on a case-by-case basis. Certainly, that is what Harry Truman did when he dropped the atomic bomb on Nagasaki. I think that that ought to be the way we proceed here.

Now, given this, I am three cheers for Michael Doyle, two cheers for his book. The two cheers are as follows.

First, I think he wants to create legal standards to guide use of force. I completely agree. And he wants them to be multilateral. I completely agree.

Second, he wants them to develop a jurisprudence of Article 39 of the UN Charter, which is one where we can compare, like lawyers do, cases with other cases to know which ones are legal and which ones are illegal. And again, with that I completely agree.

But where I get off the train, and where I think he creates a dangerous exception, is the idea that if you can't get the UN to act, that if an individual nation-state (take the United States) decides and goes through the calculation that he goes through—lethality, legality, likelihood, legitimacy—they can just go ahead and do it unilaterally, and that they would be justified in doing so. He would create, in effect, a customary international law exception authorizing unilateral uses of force based on the application of these four principles.

Now, Michael is a political scientist who worked for the UN. I am a lawyer. I worked for the U.S. government twice. I see five problems with his approach, which I'll call, just for shorthand purposes, the trolley problem, the hall of mirrors problem, the manipulability problem, the false analogies problem, and the knowledge-and-prediction problem. Let me just go through those.

First, the trolley problem. We have red lights and green lights. When you see a red light, you are supposed to stop. If you saw if you stopped that a bus full of schoolchildren or a trolley would be hit, then you would have a choice: do you obey the law, the clear, bright-line rule, the red light; or do you drive through, breaking the law, undeniably breaking the law, save the children, and then, if you are prosecuted, defend on the grounds of necessity? Now, obviously, the fact that the red light is a strong per se rule gives you a strong incentive to obey it. But at the appropriate moment, if you felt this overwhelming need, you would act.

My view is that is where we are on this case. We should not preauthorize unilateral strikes under any circumstance, and especially on a claim of self-defense. Remember, as the Bush Administration claimed in the torture memo, they were torturing people in self-defense. We should ban these things. And then, if somebody feels that they have such an extraordinary case that they should be allowed to do it, they should assert a defense on the grounds of necessity, extremist or whatever, but not claim a legal preauthorization.

I think that it is a category error to say that something that should be addressed by legal defense should be addressed by prior legal authorization. If we do that, I think the exception will swallow the rule.

The second issue is what I call the false analogy problem between humanitarian intervention and preemptive intervention. I was Assistant Secretary of State for Human Rights during Kosovo. In a multilateral situation with regard to humanitarian intervention, there are external checks. In other words, for the United States to bomb Kosovo in a humanitarian case, they had to persuade the other NATO nations to go along; and, if they didn't, they couldn't do the joint bombing. So even though they didn't operate through the UN, they did operate multilaterally.

But in a unilateral preemption situation of the kind that Michael would ultimately authorize, they don't have to consult anybody except their own internal legal standards. That's where we have what I think I would call the problem of proportionality.

When you are in a situation in which you are engaging in humanitarian intervention to stop some gross atrocity, there is something which it is proportionate to: You have to do enough to stop the killing. But on preemption, if you think someone is going to kill you, there is no notion of proportionality. You kill them first. If you think they are going to kill you, you strike out at them and you shock and awe them, you use overwhelming force.

Now, you tell me, doesn't that look a lot like the justification for Pearl Harbor? Remember, as Bobby Kennedy said during the Cuban missile crisis, "Now I know how Tojo feels. We feel like, on the one hand, we are responding to an imminent threat preemptively. On the other hand, what we are proposing to do is to strike them first with overwhelming force."

Now, I think the problem, as Michael recognizes, leads to the third problem, which I call the hall of mirrors problem, or cascades of preemption. It's not just that you think they are going to hit you first so you hit them. It's that they think you think they're going to hit them, so then they react. Pretty soon you are not responding to a threat; you are responding to an inkling of a premonition that they will think that you are threatening them, and then you might as well hit them first. That is destabilizing, and you have a lot of people striking first, justifying it later, and only after the fact do you know who was right and was not.

That leads to what I think is the third problem, which is what I'd call the problem of manipulability. These four factors that Michael says we should look at—lethality, legality, legitimacy, and likelihood—can be manipulated, particularly in unilateral situations.

I know this from experience. In 1984, when I was an attorney in the Justice Department, I got a call saying, "Have you heard of a place called Grenada?" I said, "I thought it was Grenada." He said, "You say Grenada, I say Grenada. You say invada, I say why nada."

Now, the rationale was at the time was because there was an unfriendly government that had just been put in place, but pretty soon the rationale that was put forward was that there were some medical students who needed rescuing. They were rescued, but nevertheless the troops stayed there. It was a highly manipulable doctrine, especially when being exercised unilaterally by people who weren't listening ex ante to their lawyers.

Now, why is a per se rule better? Because when you know you are operating in a zone of illegality, you get out of there. Have you ever driven in the breakdown lane? You're not supposed to. Have you ever done it? Everybody here has done it. But when you're doing it you want to get out of there as quickly as possible. Do you exceed the speed limit? Sometimes. But if you know it's illegal, if it's clearly illegal, if there is no defense really, then as soon as you can you return to a state of legality, as opposed to illegality. That's why we have per se rules.

If you think I'm making this up, just look at the Iraq invasion. The Iraq invasion was premised on a whole series of shifting rationales. Everybody said at first they thought it was preemption because of WMD; but then, pretty soon, it was disarmament; pretty soon the rationale was humanitarian intervention; then it was democracy promotion; then it was regime change; then it was long-term peace and security. These rationales can all shift over time.

Another reason why there is manipulability, it seems to me, is because you end up having the last problem, the fifth problem, the problem of prediction and proof. If you read Fiasco by Tom Ricks, you know that what they say is the best case, the optimistic scenario, and worst case, the pessimistic scenario. As a result of which, everyone said at the time, "Well, we'll go in and we'll do Iraq very quickly. We'll be out of there for $50 billion." Well, you know, whether they were accurately predicting it or not, they miscalculated. The real reason is because they were interested to do this and interested to do it with a coalition of the willing.

Now, I think it's a real problem, even if you are trying to do this legitimately. I gave in the book the example of Kim Jung-Il. In December 2000, I went to North Korea with Madeleine Albright and I sat in a meeting for three days with Kim Jung-Il. I speak Korean and I understood what he was saying. Is he crazy? He's not crazy. He's very, very smart. Did he make threats? He made a lot of threats. Would he carry out on those threats? I'm not sure.

Now, could I, as a clever lawyer, take what I heard and say, "It's lethal—he's got weapons of mass destruction; it's likely—he made some threats; it's legitimate—we've got to protect ourselves; it's legal—we could construct a legal rationale." It fits within the four-part exception.

We could hit Kim Jung-Il the day we went out. Would that have been a good idea? No, as we've now seen, because many more years have gone by, there have been diplomatic negotiations, et cetera, et cetera.

But you will always have somebody who is making the argument, "We can't risk it." If you've read Ron Suskind's book The One Percent Doctrine, he says what the current administration did was two things. It said that when there's a 1 percent certainty, you treat it like a total certainty. Then there's the famous quote by Donald Rumsfeld, "The absence of evidence is not evidence of absence."

Now, put those two things together. If there's a scintilla of information, you should treat it as highly likely you are about to be attacked, in which case you are justified, if you do the calculation, of going in with overwhelming preemptive force. That is very, very dangerous and very, very destabilizing.

So why do I then say at the end of the day two cheers for Michael's approach? I think that we do need legal standards. I think that operating through the Security Council is great. I think that operating and developing a jurisprudence of Article 39 of the UN Charter makes a lot of sense.

Should we leave the exception for unilateral action based on a unilateral actor's evaluation of these four standards? I think no. We should close that loophole, because it will be the exception that swallows the rule.

What we are much better off doing is to create a per se rule. If a leader believes they must act, they should act and take the consequences.

And think about the moose before you adopt the idea that unilateral exceptions are a good idea.

Thank you.

JOANNE MYERS: Thank you very much. I think we'll have Michael take five minutes to respond, if you'd like, and then open it up to discussion.

MICHAEL DOYLE: I'll be very brief, because Harold and I have had these exchanges a couple times.

What worries me more about the moose is Sarah Palin in the moose scenario. But let us leave that aside.

For my case, I think the choice comes down to the following. In the scenario that I set forth, do you want the president in this case to be going to Congress and making the case, and going to the Security Council and debating and putting forth a series of standards, then going back to the Congress? Or do you want him sitting in the Oval Office to decide, "That's a danger out there, and we'll go try to explain ourselves afterwards"? I think that's what it comes down to.

If you want to have the deliberation up front, this is a schema to do so. The law is in place. We have the prohibition against the use of force other than in self-defense. We have the Article 39 visions of the UN Charter. That's the law.

We're talking about how to break the law. If we come to the situation where the law should be broken, my view is it needs to be done after a lot of deliberation, not just to go do it and try to justify yourself later.

This argument is designed to put in place a series of procedures and a series of standards, standards that would be used by the Security Council and standards that would be used when you're going to break the law and take the consequences of that, that would nonetheless allow you to continue to deliberate, to be able to have your decision challenged, so you are not just taking it arbitrarily.

In short, I think what I'm describing here would reduce, rather than increase, the very legitimate dangers that Harold put forth in his critique of what I regard as the Bush Doctrine at this point.
So that's my side.

QUESTION: The discussion here revolves primarily around what the United States is going to do, the president, be it either one of the two candidates right now. But let's take it away from there and go into other areas of initiative.

1996, the USSR into Hungary—no UN action, no action from the West; 1968, USSR into Czechoslovakia—no action; 2008, Russia into Georgia—no UN action, no NATO action; a lot of talk, no action.

That's a certain kind of intervention by another country that really nothing that we are going to do is going to compel them to get out. Now, we may talk with them.

So in the case where you have this, whether it's China or Russia or other countries, where the intent has been to occupy a territory until such time as their own policy is satisfied, then what does the United States do, what does the UN do, what does NATO do? How do your standards really apply to those areas where they don't talk ad infinitum, like we do here, in all of the kinds of ways that you're talking about?

MICHAEL DOYLE: This particular study is designed to deal with the hard ethical question of using force legitimately in self-defense. It doesn't deal with every use of force. The record of invasion, the record of imperialism, interventions by the Soviets, by the United States, during the Cold War, many of them don't fit within this whatsoever; they were simply motivated by other concerns, sometimes motivated by profit, sometimes motivated by politics. This is looking for a set of standards that will show when you can legitimately use force in self-defense. It's not talking about other cases of force, where many of them are illegitimate.

And if you were to apply some of these standards and say, "Were those legitimate uses of force?" I think it would be very hard to get them to qualify. And they certainly didn't go through the kinds of procedures that I am setting forth here, partly because the good reason being that they didn't think that they could get that kind of justification.

And this is not just a Soviet problem. The United States has done many of those similar kinds of interventions throughout the period of the Cold War, including Grenada and others, such as the ones that Harold has mentioned.

QUESTIONER: We gaze at ourselves, but the Russians don't necessarily do that.

MICHAEL DOYLE: We do, and that's a sign that we still maintain some of our democratic principles. Not all countries do it, but many do. We are not the first. The most articulate debate on humanitarian intervention in Kosovo came out of the United Kingdom rather than here. But there are problems still.

HAROLD KOH: I think you make an important point, which I think illustrates what I'm trying to say, which is it's not just the United States that is subject to this unilateral rule. In other words, if they do this calculation, they can unilaterally attack. It's everybody. Guido Calabresi, my great former dean, now the judge, says the cheapest cost of order is how you solve problems of cost of access. Well, what you're doing here is putting the use of force unilaterally into the hands of the least rational calculator.

So whoever thinks that they are going to be attacked does this calculation and attacks. Now, if that is so, then everybody else who is responding on the issue of likelihood, which is one of Michael's key criteria, has to weigh the likelihood that somebody irrational will make the calculation wrong. That will affect how legitimate it is for them to do it.

Suddenly, you have a world in which everyone feels that they have some limited capacity to strike first. Then a lot of people are going to be striking first.

Now, why is it better to just have a per se "no" rule? It's not because there won't be deliberation. I give this example: My son is a good boy. He learned how to drive a car. He loved to drive the car. My wife and I had to go away one weekend. The question was: Should we pre-authorize him to drive the car? My wife wanted to say, "If it's a medical emergency and you have to go get some medicine, if you want to go with a movie with a friend that's a clean movie, we pre-authorize you to take the car."

I said, "It's better for us to just say, 'Don't take the car.' But if I come back and I find out that you did take the car, I want you to be able to give me a good reason why it was so compelling and urgent that you had to do it," in which case he had to come back and justify it under one of these grounds anyway, but as an ex post facto defense for having broken a rule that we had clearly specified.

Now, it turned out, as we found out over time, that a per se rule is something that limits the use of the car to real emergencies.

The final question, which is my own experience in this, when Madeleine Albright and the Clinton Administration were trying to decide whether to bomb Kosovo, a couple of people asked me, "Are we going to be prosecuted for this before some international tribunal?" I said, "Well, I'll represent you. I think it's justifiable. You can justify it after the fact. If you have acted, and you have acted in accordance with what seems to be an emerging customary norm, you will not be prosecuted and convicted. Now, you might have to go through some hassle for it, but do the right thing. But don't do it because you think you've gotten a free pass to do it ahead of time. Do it because it's the right thing."

Now, I think that continues to be the case. It's like my trolley problem. If you think it's important to do it, do it, and be prepared to defend yourself on these grounds later on. But don't think that it was pre-authorized, and don't think it's pre-authorized because of your juggling of these categories.

QUESTION: Professor Doyle, you were talking about preemptive war, and you said you thought there was probably only one instance of a legitimate preventive war, and that was the Dutch in the Dutch East Indies on December 8, 1941. Many people would have said that Israel's attack and the beginning of the Six Day War was another example of a legitimate preventive war, when the Arab armies were mobilized, the UN observer force was ordered out, and in fact the Straits of Tehran were blockaded. Apparently you don't think so. I just wonder if you could explain why you didn't think that the Israel attack on June 6 was a case of legitimate preventive war.

MICHAEL DOYLE: I say that I do think that the Israel case in 1967 was a case of legitimate preventive war. I just don't think that it meets the standards of the Caroline doctrine, the doctrine going back to 1837, where the necessity has to be overwhelming, there has to be no time for deliberation, no choice of means. Those weren't the circumstances in 1967. There was a legitimate reason for Israel to attack; that is, it saw the manifest intent that was being directed toward itself of hostile action. It saw preparations out there. The threat seemed to be growing. There were no other options. It couldn't go to the Security Council during that period of time. I think it is a case that meets the kinds of standards that I am talking about, and therefore it was a legitimate act of prevention. But it doesn't meet the kinds of standards in traditional international law.

What I would have preferred to see done was, to the extent that it could engage in this kind of deliberation (1) to convince its own cabinet; (2) to convince the world that its act was justifiable. In the circumstances then, it couldn't go to the Security Council. But we now live in a world where you can go to the Security Council.

And you owe it to the rest of the community of mankind to justify use of force. This is where I differ from Harold. You shouldn't just act and try to excuse it later. You should make an effort to try to make the Security Council function, to ask them to come on to your side, to accept the quality of your arguments that it's a danger that's fundamental. We owe that duty to the rest of mankind, to operate within the law. We could in this case, with Security Council approval.

If we can't get a right decision out of the Security Council, we owe it to the American public—we owe that kind of deliberation to the Congress. When we have time to act, we should do so. And so this is a case in point of a justifiable act of preventive war, in my view.

In our own time, we should do whatever deliberation we can to stay within the framework of the law and not act arbitrarily and plan to think up some good excuses later, even if they are good excuses.

HAROLD KOH: But let me point out what I think are two examples we all lived through which show the way procedures can really affect this. One is Desert Storm, and the other is the Iraq war that occurred in 2003.

Now, in Desert Storm, what happened, which I thought was correct, was they negotiated first and did diplomacy first. Jim Baker went and was negotiating, and then they got a Security Council resolution. That was multilaterally done at the international level. Then they went to Congress and they said, "Vote us authorization to enforce the UN Security Council resolution." So there was already a restriction set by the Security Council resolution, and Congress votes the resolution that enforces that. So it was multilateral at both levels and it was considered to be legitimate, and it set certain limitations, and that's where they were stopped.

Case number 2, the Iraq war of 2003. What happened there was the opposite. Instead of going to the UN first, they went to Congress first. Congress approves using all necessary means. At that point, the Bush Administration had no reason to worry about the Security Council. They could go in whichever way they wanted—with the Security Council, with a coalition of the willing, by themselves, as necessary. There was no incentive for them to negotiate with the French, if the French were going to be difficult. At a certain point, they voted one resolution, 1441, which the United States argued later was sufficient and everybody else thought was insufficient.

Now, my own view on this is that this is where procedure really comes in. If you want the two legal constraints in our system to work, the UN restraint and the constitutional restraint, then do it so that Congress is actually voting to enforce our international obligations.

But the mistake in 2003 was that the way in which the political decision was placed, and the way in which Congress responded, which was essentially "let's vote this resolution and then get out for the mid-term election," ended up giving the president absolutely no incentive to go multilaterally to the Security Council. That's what led to the situation we are in now.

QUESTION: The scenario that you developed, of the dirty bomb being developed in another country and us going to the Security Council and then going back to Congress, et cetera, et cetera—in today's world, do you think that is realistic, when people can just dissolve into the woodwork once they know that you know what they are going to do and when you are going to plan an attack against them? It's like Tora Bora, when we took six months to go into Afghanistan. They were all gone. It was useless. So what do you do in the circumstances like that—when you're trying to be legitimate, you are actually going against your self-interest?

MICHAEL DOYLE: Yes, there are some costs. What I am describing is not cost-less. That is, by revealing that we have information that some terrorist group is purchasing or has acquired nuclear weapons or other forms of weapons of mass destruction that are deliverable, we are going to give them incentives to hide it. There is no question about that. And so there are costs involved in having a process of persuasion, deliberation, and transparency. There is that danger.

I think that, in my view, those costs are worth paying for the opportunity to stay within the law through the Security Council, through the potential that will arise that this may provide incentives for the state that has then been identified as harboring this terrorist activity to itself clamp down upon it, and in order to make sure that fewer mistakes are made. So there are costs of deliberation.

It might be the case that it will be more difficult to take out some of those capabilities. But that is the cost you pay to operate within the framework of the law.

QUESTION: Congress requires the president to do a national security paper every year. I happen to have gone through four or five them a couple of years ago. Both Clinton and Bush say that they will use preventive action, sometimes if required or "we reserve the right," and so forth like that. On a couple of them, they specifically said Security Council—no matter, because the Security Council has Russia and China and you can't always get that Okay. So I have a couple of questions.

How can you call it the "Bush Doctrine"—and I really don't blame Sarah for not knowing what it was, because every president that I can see out of the last few have said they can use it. And didn't Bush explain quite carefully to Congress what he was going to do in Iraq? It seems they did quite a job of explaining what he intended to do. If the Congress didn't see it, at least they must have seen 150,000 troops sitting around in Kuwait. So didn't he deal with the problem of explaining what he was going to do? I'm not saying he did the right thing.

MICHAEL DOYLE: With regard to the Bush Doctrine, indeed President Bush is not the first president who has announced that he will take actions to protect the United States, and he is not the first president to say that he will not be governed by the Security Council. You may remember back in 2004 the "no permission slips" issue that arose repeatedly. Both candidates are on board with—well, McCain is on board with the Bush Doctrine, and Senator Obama has said that under circumstances that he perceived—the details were lost in a speech—as a threat from a terrorist group in Pakistan and if the government of Pakistan was doing nothing, he would not regard himself as barred from taking action if he thought the necessity was there. So all U.S. presidents in recent history have signed up to one version of this.

It is called the "Bush Doctrine" because there is nobody who has enjoyed it as much as this particular President has or has made such hay out of it. I think that is why it has that label. Whether Governor Palin should have known it or not, I'll leave that to the political campaign.

With regard to Congress, I share some of the same concerns that my colleague Dean Koh has mentioned about the vote that was taken there prior to the Iraq war. There were some senators who were of the view that they were voting for taking it to the Security Council, with the view that—and this is what New York Senator Hillary Clinton has said—the president would then bring it back for a second authorization. They were voting, they thought, to authorize him to raise it at the Security Council and as an expression of concern. They didn't think that they had signed up in advance to the Iraq war.

Now, I think that some of them were snookered. I think some of them were manipulated. The resolution was very ambiguous as it came through, and I think that is a misfortune.

That is why I would prefer that before one goes to war preventively—I'm not talking about an act of preemption, where you just don't have time; the president has to act—but in a preventive war I'd like to see a declaration of war. I don't see any reason not to follow the Constitution. The president should come back in and make the case to the House of Representatives that this is something that is so dangerous, there are no other options, nothing better, he has tried everything else, and that he would be irresponsible not to act. He should convince them and ask them to vote for it. I think that's part of his job.

HAROLD KOH: With regard to Iraq, I think the real problem is not preemption, it's the aura of preemption. If you actually look at the legal opinions that were issued to justify the Iraq war in 2003, they rely on revived Security Council resolutions. Both the U.S. government's rationale and the British rationale said there were these resolutions there that essentially had come back to life. So if you looked at what their pure narrow legal rationale was, it was Security Council resolutions.

But what actually they invoked and kind of waved a wand over was, "We need to do it because of the aura of preemption." So that's the way preemption will actually be used. They will always find some other narrow legalistic rationale, but the preemption argument will be the one in the background, and then it won't be the basis on which it is justified.

Point two is when you create a unilateral option as a last resort, it will become the first resort. Everybody knows this. If you say, "I can do it the difficult way, the really difficult way, the second-most-difficult way, the third-most-difficult way, but when it comes right down to it, you can do it unilaterally," they are not going to go to the most difficult way, or they'll go to it only in the most casual way.

The third point, which I think Michael would probably agree with, is institutional reform is really what this is about, two kinds of institutional reform. The Security Council needs to be reformed. Michael, as much as any human being, has worked on this, and this whole effort was stalled in the last couple of years but has to be now revived. The Security Council should simply be more representative of who the powerful nations are. For us to be—you know, where are the Latin countries? Why is it that two European nations are on?—how to make it more representative, et cetera.

The other kind of reform that needs to occur is within Congress. In the old days, the president would be talking to a small group, a core consultative group, of senators who were really expert on the issues and fully informed. So they couldn't be snookered in the same way. They actually had all the information.

One advantage there might be in having a vice president who is also a senator—you have two senators who are running for president; one of them picked as his vice presidential candidate someone who is also a senator—is that there are actually kinds of relations between the president and the key figures that would have to be consulted on these things, which would make consultation more likely.

I think one thing that the new Congress ought to do, whoever is the president, is create a core group of leaders of the key committees who have security clearance and meet regularly with the president and get updates so none of these things come as a big surprise. They can evaluate a mounting threat and give quick authorization, and then other people defer to them.

JOANNE MYERS: Well, we've been talking about rules and procedures. The rule of the Carnegie Council is to end at 6:30. But before we do that, we have to thank Dean Koh and Michael Doyle for a very important discussion. The procedure now is to continue the conversation and have a drink.

Thank you very much.