On September 5, I interviewed Professor Ted Postol in his office at MIT. Postol is the leading critic of current U.S. plans to build a system of ballistic missile defense. It is difficult now to read about ways that we might defend ourselves from attack without thinking of the slaughter of innocents on September 11, what might have been done to prevent it, and what we might do in the future to ensure that such slaughter is not repeated. In our judgment, the skeptical considerations presented here are at least as relevant now as they were on September 5. So we have chosen to run the interview without modifying it in light of the events of September 11. At the end of the interview, Ted Postol has appended his own reflections on the continuing importance of public debate and critical thinking about plans to build a system of missile defenses.

—Joshua Cohen

Desirability

Joshua Cohen: The United States does not now have any system of missile defense in place; the Bush Administration has announced plans to build one. You have been a leading critic of the feasibility of missile defense. Before getting to your concerns about feasibility, I’d like to ask about whether you think a missile defense system is desirable.

Ted Postol: I am a little hard pressed to answer your question. Desirability and feasibility are very intimately connected. Of course you could imagine the ridiculous fiction of a world in which your defenses essentially remove all security threats, but whether such a defense is desirable or not is an abstract issue that bears no relation to reality. You cannot build such a defense, so it is not really a very interesting thing to analyze.

JC: Why do the proponents think it is desirable?

TP: The ideological assumption of the extreme right of the Republican party is that there are weapons of mass destruction out there that can be delivered over intercontinental range by ballistic missiles, in the hands of people that just aren’t rational like we are (polemically I guess), and that we need to be able to protect ourselves from these irrational people whom we cannot deter and are either out to get us or out to minimize our ability to do what we think we need to do in the world. That is the mindset that seems to me to be behind this push for defense.

JC: These irrational people are the leaders of the so-called “rogue states,” in the current jargon?

TP: Yes, although depending on whom you are talking to it can extend to the Russians, and to the Chinese, although it appears as though the Bush administration is focusing on these states of concern: North Korea, Iran, Iraq, and (sometimes) Libya. That at least is the impression that you get.

JC: What about this threat of rogue states with political leaders who will not be deterred in their use of weapons of mass destruction by the threat of massive retaliation. Do you see a new threat of this kind?

TP: I don’t see any evidence for a new threat of this kind. You can certainly imagine facing an adversary who cannot be deterred: someone who is completely insane and has enough control in the government to do something suicidal for the people of the state that he controls—someone who doesn’t care about the threat of retaliation. I’m not a historian, but I can only think of one example in the twentieth century: Adolf Hitler, toward the end of his reign as leader of Germany still seemed to have a lot of control over the German military forces. Maybe if he had possessed nuclear weapons he would have used them even if he knew that retaliation was probable. But we have dealt with many other nasty characters: Joseph Stalin, Mao Zedong. They certainly showed a high degree of rationality with regard to the dangers these weapons posed, even though I don’t think that many people would regard them as having been relatively gentle with their adversaries. So I don’t really see a sudden emergent threat of this kind: a threat of an irrational political leader, who controls a state, does not care if the state is destroyed, has weapons of mass destruction, and the ability to put those weapons on ballistic missiles. And if you did come up against a threat of this kind, then you would have face the facts about feasibility.

The Decoy Problem

JC: Shifting attention, then, to the issue of feasibility: perhaps you could start by giving us a mental picture of how the system is supposed to work.

TP: The key element of the system that Bush has been pushing hardest is a “mid-course system”: a defense that is designed to try to destroy warheads at high altitudes. By “high altitudes” I mean hundreds of miles of altitude—an environment that is not a perfect but a near vacuum.

JC: So this is distinct from systems that would try to attack missiles in their boost phase?

TP: Yes, distinct from systems that would operate either in the phase of powered flight, when the rocket is in a process of launching the warhead to adequate speed, or in what is called “reentry,” where the effects of the atmosphere are modifying the appearance and/or motion, of the objects of interest. So when one talks about reentry phase missile defenses, you are implicitly indicating that you realize that the near vacuum of space creates a special problem for you in defense.

JC: What is that special problem?

TP: In the near vacuum of space, a rock and a feather will travel along together. If the rock is slowly tumbling, and the feather is slowly tumbling, they will both slowly tumble with motion unmodified by aerodynamic drag. In the atmosphere if you have a rock and a feather traveling at high speed, the aerodynamic drag on the feather will quickly strip the feather from the rock. In the near vacuum of space, they travel along together. By analogy with the rock and the feather, think of a heavy warhead and a very light balloon that is inflated in the shape of a warhead; they would also travel along together in space. There would be no aerodynamic drag to strip the light balloon away from the heavy warhead. If the light balloon were tumbling end-over-end, and the warhead were tumbling end-over-end, they would be totally indistinguishable—even if you were floating along in a spacesuit—as long as you could not touch them, or push one relative to the other. If you were simply using your eyes, the motion of these two objects would be unrelated to their mass. As a result, the simplest, easiest to implement, most trivial kind of technical countermeasures can result in vast numbers of credible decoys that you have no hope of distinguishing from the warheads.

So when General Kadish, the director of the Ballistic Missile Defense Organization, says, “we don’t think that the decoy threat that Ted Postol argues for is a serious problem,” what is he saying? He is saying that there is an adversary out there—let’s not worry whether it is North Korea or Bolivia—with the vast scientific and industrial base to build intercontinental ballistic missiles. Not a trivial task. The adversary also has the vast scientific and industrial base that allows it to build compact and light nuclear warheads that can fly on these ballistic missiles. Also not a trivial task. But the adversary can’t figure out how to build a balloon.

JC: So the essence of the feasibility problem is: decoys. And the decoy problem is that you can’t distinguish between the reentry vehicle with a warhead on it and, say, a balloon.

TP: More precisely, the problem is that you can distinguish different objects, but you don’t know why things look different. Let me give you an example: imagine I have a balloon in space, and I construct a balloon around the warhead. (Someone from the Pentagon actually called us up and asked us how you do this. This is ridiculous: you take a scissors, cut open the balloon, re-assemble it around the warhead, and put on some one-way valves.) Then you can deploy dozens of balloons. You don’t try to make them look alike; in fact you try to make them look different. Each one is painted with a different surface so it will have a different infrared brightness. When the kill vehicle homes in on this complex of objects, it sees dozens of points of light. It can tell that one point of light is brighter than another, that some are scintillating and some are not, but it doesn’t know why. When the sensor is sixty seconds away from trying to hit the warhead, its range from the warhead is going to be roughly three or four hundred miles. In sixty seconds it will close to fractions of a yard from the object it wants to destroy. And in this period of sixty seconds it has to look at all of these objects and decide which one to destroy. The kill vehicle can tell that the objects look different from one another, but it has no knowledge of why, so it can’t choose the right object.

In fact, when you look at the inadvertent things that an adversary can do, you find out that the system could fail even against things that are not intended to be countermeasures, but are just unexpected. For example, the warhead looks to the kill vehicle like a smeared-out point light in the distance. The kill vehicle looks at this smeared point of light and it tries to find the center. We know it can do that. But suppose the warhead is attached to an upper stage of a rocket that happens to have stayed attached for whatever reason, and let’s say the upper stage of the rocket and the warhead are oriented roughly perpendicular to the line of sight of the homing kill vehicle. It still sees a point of light, but when it homes on the center of that point of light, it passes through the upper stage, and does not get the warhead. So here you have the example of not even a conscious effort of an adversary causing you to miss, simply because you don’t have complete knowledge of the target you are trying to hit.

JC: The New York Times recently published an article about the problem of “tumbling warheads”—warheads that are not stabilized in flight by having them spin. Is the ability to hit the target with high reliability influenced if the target is either tumbling or spin-stabilized? That is, irrespective of decoys, are tumbling warheads harder to hit reliably than non-tumbling?

TP: Yes, but the problem of the tumbling warhead is not that it is harder to hit. The problem is this: if the warhead is tumbling in space and you have almost any object that is somewhat credible that you can make tumble at roughly the same rate (and when I say “roughly the same rate,” I mean it could be tumbling two times as fast, or a third as fast) then you have a credible decoy. That is why the tumbling warhead is so problematic. It is totally unpredictable. The rate of tumbling could be anything. One of the arguments some people make is that the warhead is heavier, the decoys are lighter; so the light decoys when they get pushed off will tumble at a different rate. Well that is just a ridiculous argument because if I can’t control the deployment of the decoys (and incidentally I believe that you can control the deployment of the decoys) I can just put a very tiny rocket motor on the warhead when I deploy it, and kick it into a tumbling state that is in the range of variability of my decoys. So I can always make a tumbling object tumble at an arbitrary rate just like a bunch of decoys, and there is no reason why it will tumble faster or slower other than due to an accident of the way I deployed it. There is no physical phenomenon that says: this must tumble at this rate, hence this must be the warhead; the decoy must be at this different rate, hence it must be the decoy. So you don’t have any of the physical phenomena that will allow you to say, “Okay, this scintillating object I see in front of me, three hundred miles away, must be a warhead.” Again, you have a horrendous, fundamentally unresolvable problem, of picking the warhead from the decoy.

JC: You said earlier that the proponents of these system have to think that a country with the scientific, technical know-how to build a ballistic missile, put a warhead on it, and get the warhead up into space would nevertheless be unable to build a decoy. But as I understand it, the tumbling warheads are easier to build than non-tumbling warheads; it is harder to achieve spin-stabilization.

TP: It’s harder. I believe it is not very difficult, but it is true that it is harder.

JC: So the very places that are presented as the focus of concern—the rogue states—are the ones that are the most likely to have tumbling warheads, and therefore the ones that will probably have the most effective decoys.

TP: Yes. What the adversary will do—the most primitive adversary—is push the warhead off the upper stage without spinning it up to stabilize it. And the only reason to spin-stabilize it is if you are trying to achieve relatively high accuracy. If you just want it to land on people and make them upset, you will not worry about precision. I think it is very hard to believe that you have an adversary who has all of these irrational characteristics that are postulated by the people who advocate these defenses, and that this adversary is going to behave in a rational way with respect to everything else. If the adversary knows that you’re not going to be able to distinguish between tumbling warhead and tumbling decoy, and all they care about is trying to create political pressure on you and your allies, then why would they help you by spin-stabilizing the warhead when they don’t have to do it anyway?

JC: Does each decoy need to be launched separately, or can warheads and decoys all be launched on a single missile?

TP: A single missile. The upper stage of the rocket that launches the warhead would have decoys on it, and there are lots of ways to build these decoys. You can stack them like ice-cream cones. You can put a little rocket motor, a charge of explosive, on the inside of a hollow cone, and when the charge goes off it just pushes off the decoy, and then you could have a sequence of charges going off and these objects coming off.

JC: Is there any limit to the number of vehicles—of decoys—that you could put on a single missile? You could put on ten balloons and a balloon-covered warhead, and that would not be difficult?

TP: That would not be difficult. If you start talking about five hundred balloons, then it becomes trickier, but tens of balloons—which already is a very problematic situation for the kill vehicle—that is trivial.

JC: The advantage, then, is very much on the side of the offense, because it would be prohibitively costly to try to kill every warhead and decoy since you can multiply the number of decoys.

TP: Yes.

JC: You have been talking about the kill vehicle seeing points of light and needing to find its target on the basis of its observations of smeared points of light. Why is it that sensors that guide the kill vehicle to the warhead have to be on the kill vehicle? Why not use ground radar?

TP: Think of a radar that is tracking the complex of objects from a great distance. The radar continually measures the location of the object at different times. But the measurements are uncertain, and if I project the trajectory forward, it gets more and more uncertain as the range increases. Imagine I have a bunch of objects that are close together (by close together I mean hundreds of feet, or even miles apart). I can only track them to a precision that will allow me to estimate their location to several, or tens of miles, a thousand seconds later, when the kill vehicle encounters them. Even if radar could identify the warhead relative to the decoy (which incidentally can be denied) the objects will have been mixed up again by the time the kill vehicle is closing in. So the kill vehicle must be able to pick out the warhead on its own—as the Ballistic Missile Defense Organization and its contractors acknowledge.

JC: So you have got a system of sensors on the kill vehicle, and those sensors can detect brightness, and fluctuations in brightness. And for any brightness and fluctuation in the signal from the warhead, you can build a decoy that looks different, and that might also be the warhead.

TP: I could make a tumbling warhead look like a non-tumbling target by just enclosing it in a featureless balloon; I could make a balloon look like a tumbling warhead by painting half of it with one surface and the other half with another surface, so as it tumbles in space it gets brighter and dimmer in a way that is similar to a tumbling warhead. I can essentially manipulate any of the parameters I want, using the most trivial physical principles. It is a very, very simple thing to do relative to distinguishing these objects from great distance.

JC: So it is technically easy and economically cheap to overwhelm the kill vehicle with decoys that are as likely to be the warhead as the warhead itself. And it is too expensive to go after all the objects: both warheads and decoys.

TP: Even the advocates don’t claim that you can.

JC: Scientists and engineers often disagree, and they fight about an issue, but they recognize that reasonable people disagree. Is this such an issue? Is there anyone who disagrees with you about the difficulty of solving this problem who you think knows what he is taking about?

TP: No. It is that clean. I want to be clear that there are many issues where I would not answer this way, where a matter of technical judgment could lead you to have different opinions about the situation; there are enough ambiguities in many technical situations where well-informed, honest people could have different judgments about what the possibilities are. The issue we are discussing is not in that category. This is so straightforward that in my view, anybody who is at all knowledgeable and understood this problem would find that this is a hopeless task.

JC: In principle unsolvable?

TP: Right: getting a faster computer, getting a better sensor, doesn’t help you. I can measure the scintillating properties of these objects better with a better sensor; I can more precisely measure the brightness of one object relative to another. But such improvements don’t tell me anything that I need to know that allows me to select one object relative to the other. Getting better information that is irrelevant doesn’t help you.

Testing

JC: Could you explain the history of missile defense testing over the past few years, starting with the basics? How many tests have been done?

TP: There have been six tests, beginning in June 1997. They are called Integrated Flight Tests (IFTs). There was an Integrated Flight Test 1A because in the Integrated Flight Test 1, the complex of targets was launched but the kill-vehicle failed to be launched, so 1 became 1A . So we have six Integrated Flight Tests, 1A through 6. The first two flight tests were not intended to try to hit a target; the purpose of those flight tests was to understand what different objects would look like as the kill vehicle approached them in space, in order to know how you would go about separating warheads from decoys. We have a lot of information about Integrated Flight Test 1A. That test consisted of a mock warhead (a mock warhead being a cone-shaped object about two yards in length), eight decoys, and the upper stage of a rocket that deploys them, which is potentially an infrared target and could be mistaken for a warhead. In that experiment, a bunch of the objects looked very much like the warhead, and they were unable to reliably distinguish the warhead from the decoys. They could do it after the fact, because the experiment was highly instrumented so that they knew where each object was in space, and since they knew where the kill vehicle was they could reconstruct the location of different objects, which is a perfectly legitimate thing to do in the process of running an experiment. But when they tried to pick the warhead using infrared-observed signals, they were unable to do that reliably. To deal with this problem they removed the data from the experiment where they were having problems selecting the warhead.

JC: Explain what you mean.

TP: They have a template that reflects expectations about the brightness of the warhead relative to the balloon, and the rate of fluctuation of the brightness relative to the balloon, and so on. But the template did not match the critical objects they were trying to select among. So they took a small interval of time where they had data that they felt was stable enough that they could use it, and they altered the template based on the observed data—as though you were throwing a dart at the wall, and you put a bull’s-eye over where the dart hit. They altered the template and then claimed that they were able to make a match. But they cut the data both before and after the time interval, because it looks like in these other time intervals the match changed from what they were trying to get.

JC: So the template is a set of predictions, or expectations about what the different objects will look like in space: say, the warhead will be the reddest object, while the others will look more green.

TP: But, suddenly, they see yellow.

JC: And when things didn’t look as expected, they took one set of observations, re-designed the template, and claimed to have successfully discriminated the warhead. But even that after-the-fact success also required cutting off earlier and later data because the signals sent by the warhead and the decoys changed.

TP: That’s right. The signals they were seeing changed as the kill vehicle closed in on the target complex. And you can understand why this can occur: imagine that I am expecting to be looking at a rod in space, and the rod is tumbling end-over-end so sometimes it is nose-on looking at me, and other times it is side-on. That rod would fluctuate with a certain pattern—a certain sinusoidal brightness that I would see. Because even though I can’t see that it is side on, I can see that it would be brighter when it is side-on relative to when it is nose-on. So, in this geometry I have a clearly fluctuating signal that is oscillating with a certain period and amplitude. Now, if instead this rod is turned so that it is like the propeller of an airplane that is flying at me, then when it tumbles in space the orientation of tumbling is different, and the brightness will not fluctuate at all. So the orientation of the object is critical in determining how it appears to the kill vehicle. What happened, then, is the signals were a very complex mix of radiant energy from the object (by radiant energy I mean that they are lit up like light bulbs in the infrared), and radiant energy from the earth. The object is like a hot, tumbling mirror near the sun. As it is tumbling you not only see some of the luminosity from the hot object, but you also see some of the reflected sunlight off it. And these complicated changes—subtle changes in the signal—are what they were hoping to exploit to tell one object from another. But the changes in the signal were not subtle; they were significant. And so they fundamentally did not have a stable signal they could exploit during the entire homing process. So what they first did was remove data that undermined their claim that they could select the warhead from among the objects they saw. They removed the data from the end of the experiment—because the evidence we have is that another object was selected as the warhead—and they just cut the time period off. For roughly eleven or twelve seconds, out of sixty seconds, they were artificially able to make it look like they were able to select the warhead.

JC: If I understand this right, if they had taken another period of time and redesigned the template to fit the observations during an earlier or later time period, they would have had a different template.

TP: Right. So their methodology for selecting objects was fundamentally flawed. And what did they do? They removed all the decoys that gave them the problem from the follow-up experiments. As of Integrated Flight Test 3, they removed all the decoys except one. And the decoy they used in tests 3 to 6 was a very large balloon, ten times brighter in the infrared than the warhead. So they went from trying to pick a warhead out of a complex of credible decoys, to having two objects in front of the kill vehicle, one roughly ten times brighter than the other. Now remember: we can see the differences, we just can’t tell why they are different. But they knew. They just robotically programmed the kill vehicle to say: you are going to see two objects in front of you, and one is going to be brighter than the other; home on the less bright. This is not a credible discrimination capability because I could make the warhead look as bright as the balloon by putting it inside the balloon.

JC: If somebody was trying to defeat the test, and they knew the kill vehicle had been programmed that way—you will see two things, one brighter than the other, and go after the dimmer one—they could have designed the warhead to be the brighter and the balloon to be the dimmer.

TP: Right.

JC: A series of tests, then, initially failed to solve the discrimination problem, and then subsequent tests gave up even on addressing the discrimination problem.

TP: Yes, they removed the decoys from the subsequent tests; yet they continued to misleadingly claim that the kill vehicle is discriminating. Every test they described looking at the large balloon, discriminating the large balloon from the warhead, basically choosing the dimmer of the objects and calling it discrimination—and then in the subsequent tests, after Integrated Flight Test 6, we know what they were planning to do.

Now they have changed their schedule because they got caught. A patriotic American handed me a briefing when I was in the Pentagon last year, which showed the decoy suites that they were planning up until Flight Test 21, when the system would be declared operational. This document showed that they had removed all the credible decoys from all the subsequent tests. They were planning a test series that would hide the fact that they could not tell the difference between tumbling decoys and tumbling warheads. This was carefully orchestrated. These tests were not meaningful in terms of building a defense that has any chance of working.

JC: The decoy issue is, as I understand it, not new. People have been worried about this problem in ballistic missile defense from day one.

TP: Yes, any ballistic missile defense system that is going to work at high altitudes has the problem of decoys. If you work at lower altitudes, during reentry, you have a series of physical phenomena that you may be able to exploit to tell the warheads from the decoys, depending on the sophistication of your adversary. The United States can build decoys that the adversary will not know are decoys until they hit the ground. A rogue state, I think, would have hard time doing that. But building a decoy that is credible to thirty or forty miles altitude is a low-technology effort. And if you can get your decoys to work below fifty or sixty miles, you will defeat every system that the United States is currently claiming it will use in the reentry phase. If the adversaries we need to defend our cities from get to ninety miles altitude with a decoy—not a balloon, but with another decoy—we’re lost.

JC: General Kadish says that we are doing walk-before-you-run, one-step-at-a-time tests—tests that prepare us to move to the next level of sophistication. What is he saying?

TP: He is claiming that he is going somewhere, when in fact he is going nowhere. Not only do the flight tests fail to address the problems of decoys; they have no technology on the drawing-boards, or under development, that has any promise at all of dealing with this problem. When The New York Times published an article recently about tumbling decoys, none of these people were able to provide a credible statement as to why they had any hope at all of dealing with a tumbling decoy.

So the bottom line is that they have no technology at all, no program, no hope of being able to tell the warheads from the decoys, and yet they claim that they can go ahead and build a system that will protect our families from nuclear attack by these rogue states.

Consequences

JC: And the idea is that if we don’t have a ballistic missile defense system in place, we are subject to blackmail from Iran, Iraq, North Korea; that in a situation of conventional conflict they could credibly say: “we are going to launch weapons of mass-destruction.” If we don’t have any way of protecting ourselves from mass-destruction, we will be subject to such blackmail. But if we have ballistic missile defense, we put ourselves in a very different situation: if they say, “we are going to launch these missiles,” we can say, “go ahead and launch them, we can take care of them with our defense.”

TP: Yes, but you know this argument is even stranger than the one that you are presenting. For example, let’s take the argument that we have this virulent, unstoppable Iraq—which is a pretty good example, given that Iraq is pretty virulent and so far has been relatively unstoppable—and that they have this miraculous capability of building an intercontinental ballistic missile. Which I think is ridiculous. States like Iraq are a problem, but I do not think they are ICBM-threats to the United States. The scenario as these people construct it is that in some future time we have a maniacal leader—Saddam Hussein is a good choice—and his people will construct this ICBM out in the desert, or wherever, and we stand there paralyzed watching them do this. We attack their surface-to-air-missile sites daily when they turn on the radar, but when they put the ICBM out there, we don’t attack it. We are afraid to do it. And then the net result is that they launch this ICBM and we are without a defense. That is the scenario. The scenario is ridiculous because it is unimaginable that we wouldn’t attack this ICBM first.

But let’s take that scenario at face value. Why not have the scenario where you have these maniacal adversaries who know that they are up against the United States and that the United States is going to oppose them, and they sneak a nuclear weapon into the United States, and then tell us that they have snuck this nuclear weapon into the United States at some point: it is somewhere in some American city. They don’t even have to tell us what city. Or they can sneak two nuclear warheads in. Blow one of them up, and then say the next one is really going to go off. This nightmare scenario is no less plausible than the nightmare scenario invented by missile defense advocates. I would rather have no missile defense and have some lunatic assembling an ICBM in front of me, because I know what I’ll do about it if I am in a position of authority: I am going to take that thing out. But the other scenario is something that I don’t know what to do about.

JC: Surely you have to have very high confidence that the system is going to work in order to be willing to hit an adversary who has weapons of mass destruction on ballistic missiles, and not worry that the adversary is going to come back after you with those weapons. Suppose you thought that there was a twenty percent chance that the rogue state would be successful: you tell them to do something and they say, “we’re not going to do it, we’re going to attack you with weapons of mass destruction,” and you say, “go ahead and attack, we’ve got ballistic missile defense and so your threat doesn’t carry any weight with us.” If you thought that there was a twenty percent chance that the missiles were going to get through the defense, that would be a crazy thing to do.

TP: The argument that people use is that missile defense creates uncertainty for the adversary, which I think is true. The problem is that if you have this adversary whom you have defined as reckless and irrational, whose uncertainty is more important to you? The uncertainty created for that adversary, or the uncertainty created for you? And so the problem here is that again, no matter how you look at the arguments that these people construct, they don’t make any logical sense.

JC: If they are not building something that has any chance of success, and that any rational person who knows anything about this believes has any chance of success, what are they doing?

TP: I have to fall back to the description that I talked to you about earlier, which is basically that these people have a belief system that I can only describe as best as I understand it, but that I don’t believe is logical. For example, you hear Condi Rice say that the ABM Treaty constrains us. I don’t know how it constrains us. There is no defense system worth doing that I can think of that the ABM Treaty basically doesn’t allow us to continue doing. So the ABM Treaty doesn’t constrain us in any way.

For me, the closest thing to an explanation occurred recently in the newspapers, where it appears that there was a serious effort to basically tell the Chinese that it is okay with us for them to go ahead and expand a nuclear force in response to our missile defense, because we’re concerned about rogue states, not them. And the way that I read this bizarre, and almost laughable—if it were not so reckless—position, was that they have somehow convinced themselves that this imaginary threat from these rogue states is so serious, so overwhelming, so overriding, that they have the technology to deal with it, (so there are a lot of caveats here); and that because they have both the technology to deal with it and the threat is so large, they are willing to put up with the increased threat of a larger Chinese nuclear arsenal. They are willing to say this to the Chinese even as the Chinese leadership wants to emphasize economic growth, not military expansion. They are doing modernization that is prudent given their national interest, but they haven’t been emphasizing nuclear expansion—other than to make a modest expansion as they modernize their force. Now what we’re doing is pushing them to move their force beyond this modest expansion.

Vulnerability

JC: So you believe that they are responding to a virtually non-existent threat, with a system that is in principle not feasible, and that threatens enormous destabilization internationally.

TP: Yes, though I actually don’t even think for sure that it is going to create a sense of security in the political leadership. My guess is that the American people will also not believe it is very viable, at least from my recent experience of the last half year. I think that people are pretty sensible when you explain the details to them. I think that people are worried, understandably, about their family situation, their jobs—things that most normal people worry about. But if you engage people who are not especially interested in these problems—as most people understandably are not—they show very good common sense when you point out these problems. So my guess is that if we were to go ahead and build such a system, not only would the population not be especially confident that the system would work, but the political leadership wouldn’t have any confidence in it, so you would get none of the benefits that these people claim you’re going to get, because even the political leadership would not be in a position to rely on this defense if they felt that they needed it.

JC: But you would have in the meantime provoked an arms race in South Asia.

TP: Right. So this behavior is just hard to understand, and it is totally unexplained in terms of the logic or the factual base from which the logic is supposedly derived.

JC: Maybe they just don’t believe that you can be both secure and vulnerable.

TP: Well if you could build a defense that truly allowed you to intercept attacks of this kind with a reasonably high level of confidence, at a cost that was within your reach, my attitude would be, why not. If the defense can truly not be countered, and your adversaries have any brains, they will not drastically alter their behavior to deal with it. The reason they will alter their behavior is that they can defeat it using very modest resources that are well within their reach. If we truly had an overpowering defense capability, there would be nothing these guys could do about it.

JC: So in the world we live in, there is no alternative to achieving security under conditions of vulnerability. That’s life.

TP: Right, it is an existential condition. The idea that mutual assured destruction is a policy choice shows a complete misunderstanding of the basic facts of existence. Mutual assured destruction is a consequence of the existential condition created by the incredible destructive capabilities of nuclear weapons. If it were possible to avoid this kind of vulnerability, anybody in their right mind would choose to avoid it. But this was not a decision that someone made. This is life. This is an existential condition. We basically have to depend on the rationality, sanity, and desire to live on the part of other people who could potentially do damage to us.

JC: And that is not theory; that is life.

Postol: It is a fact of life. When I get on the road, I still have to depend on other drivers to stay on the right side. It doesn’t always happen, but you really don’t have any choice, because the world in which you drive anywhere you feel like driving, using any rules you choose, is far more dangerous than the world where occasionally somebody drives on the wrong side because that person is drunk or has lost control of the car.

Effectiveness

JC: You’ve been very involved in this issue and related issues for some time, and it has been said that you are too strident to be maximally effective. Is that true? Do you think that you would be more effective if you were less direct?

TP: I’ve been surprised by the very positive response I’ve gotten from people. I didn’t expect it. I expected people to react against what I was saying, but I felt that it was important for someone to speak clearly and rationally about this problem. For me that was the most important thing: to be honest, and to be able to defend the statements that I make. And I believe I have done it, and what surprised me is how many people have contacted me and said how refreshing it is to hear somebody speak plainly. So my experience, which was not my expectation, is that people are relieved that somebody that they (I hope accurately) perceive has real knowledge of these issues is speaking plainly about them.

JC: You are also outraged by what the Ballistic Missile Defense Organization’s supporters are doing.

TP: I am certainly outraged. To me there is a profound ethical question involved here. I am trained as a scientist and engineer, and to me one of the most profoundly immoral things you could do as a knowledgeable scientist or engineer is to say to somebody “go ahead, it is safe” when you know it is not. This is so basic an ethical issue that I find it astonishing that a military officer, who commands people whose lives are in his hands when they go into combat, would be involved in this outrageous lie to the American people. I am outraged, and I think it’s outrageous.

JC: You believe you are right to be outraged.

TP: I think anybody who sees this sort of immorality in front of them and who isn’t outraged, should be.

Afterword

It is difficult to review this interview without the greatest sadness and sobriety given recent events.

The shoddy air security that made the attacks on the World Trade Center twin towers and the Pentagon possible are simply tragic examples of what can happen when the nation depends on defenses that can only work against adversaries who are assumed to be massively incompetent. The only good thing that we can hope for from these events is that it will now catalyze the nation to face the real threats to us.

Depending on missile defenses that cannot deal with even simple predictable countermeasures will only repeat the tragic mistakes we have already made with our air security defense. In addition, proceeding with such missile defenses will take critically needed resources away from our national effort to solve the real problem that now confronts us. We need to instead commit the nation’s highest caliber professionals, both military and civil, to the task of systematically identifying, analyzing, and responding to the threats made clear by the September 11 attacks.

The evidence I have so far seen indicates that the Administration will use these tragic events to obtain full funding from the Congress for their ill-defined and ill-considered missile defense programs. However, I believe that the recent events will result in the average American becoming aware of the debate over missile defense and the trade-offs associated with missile defenses versus our society’s real security needs. The net result I expect is that when most people become informed about the issues being debated they will have the common sense to focus on the real problems and to question the utility of missile defenses.

               —Ted Postol, September 18