On December 13, 2001, President Bush announced that in six months the United States would withdraw from the 1972 ABM treaty, a treaty that limits the testing and prohibits the deployment of any national missile defense system by Russia or the US. The stated reason for this decision was that the United States needs to develop a system that would protect us from attack by intercontinental ballistic missiles launched by terrorists or by a so-called rogue state. The US has not yet withdrawn from the treaty; this is the formal six months’ advance notice that is required by the treaty, and the President could still decide not to withdraw, but it is hard to imagine that anything could happen before June 2002 that would change his mind.

The arguments by scientists and members of Congress that the US could continue an active program of developing and testing missile defense systems without abrogating the ABM treaty now seem moot. But the issue of whether to actually develop and deploy a national missile defense system is not moot, and will not be settled even after the treaty is abrogated. Requests for missile defense funding will come up again in Congress in mid-2002, and in subsequent years. We can anticipate a continuing national debate about whether the US should seek to develop and deploy a national system of defense against intercontinental ballistic missiles.

Few of the arguments in this debate will be new. Indeed, it is hard to remember a time when the US has not been arguing about a national missile defense program.1 Almost half a century ago, in the Eisenhower administration, the Army proposed to convert the old Nike antiaircraft system to an antimissile system called Nike Zeus, which would send radar-guided nuclear- armed rockets to intercept Soviet warheads as they plunged through the atmosphere toward US cities. It had obvious failings: the nuclear blasts from successful interceptions could put our radars out of action, and the stock of interceptor missiles could be exhausted if the enemy missiles carried several light decoys along with each warhead.

In the Kennedy administration the Nike Zeus plan was upgraded to a two-tier project called Nike X. Long-range nuclear-armed missiles called Spartans would attempt to intercept Soviet missiles while they were still coasting above the earth’s atmosphere; short-range Sprint missiles would then deal in the atmosphere with those warheads that had survived the Spartan attack. As a member of the JASON group of defense consultants, I worked in the 1960s on the problem of discriminating decoys from warheads, and learned how difficult it is. Like others before me, I gradually also became influenced by a powerful argument against deploying any missile defense system: that in the conditions of the times it would simply induce the Soviets to increase their offensive intercontinental missile forces, leaving us worse off than before.

Despite such arguments, the Johnson administration came under powerful political pressure to go ahead with some sort of missile defense. In 1967 Defense Secretary Robert McNamara gave a remarkable speech in which he explained all the reasons against deploying a national missile defense, and then concluded that the Johnson administration would go ahead anyway with a limited antimissile system, now to be called Sentinel, which would protect our cities only from attack either by accident or by what was then considered to be a rogue state, China.

To everyone’s surprise, the most effective opposition to the Sentinel system did not come from experts who criticized its effectiveness or worried about arms control, but rather from citizens who simply did not want nuclear-armed defensive missiles in their neighborhoods. In response to this opposition, the Nixon administration moved the proposed Sprint missile sites away from cities and renamed the system Safeguard. Its declared purpose was now to defend our offensive missile silos instead of our cities against a missile attack. This was intended to defuse worries about strategic stability—protecting our missile silos would not make it necessary for the Soviets to increase their forces in order to maintain their ability to retaliate for a US first strike. And by protecting our own offensive missiles Safeguard would reduce any incentive that we might have to launch missiles in a crisis. As explained by Defense Secretary Melvin Laird, “The original Sentinel plan could be misinterpreted as…and in fact could have been…a first step for the protection of our cities.”2 But in fact there was little technical difference between the Sentinel and Safeguard systems, except that Safeguard would have less effect on suburban real estate values.3

The Safeguard system was scotched by doubts about its effectiveness (especially concerning the vulnerability of its radars) and fears about its cost. In 1972 the Nixon administration and the Soviet Union signed the antiballistic missile (ABM) arms control treaty. It limited defenses against ballistic missiles to one hundred interceptors at each of two sites, later reduced by mutual agreement to one hundred interceptors at one site. The site could be located to protect either the national capital or a field of offensive missiles. This would allow the Soviets to maintain their rather primitive Galosh missile defense system around Moscow, while the US could proceed with the declared aim of the Safeguard system and defend the intercontinental ballistic missile field in North Dakota.


To guard against surprises, the treaty also contained a clause that banned developing, testing, or deploying “ABM systems or components which are sea-based, air-based, space-based, or mobile land-based,”4 a clause that later came under special attack by proponents of missile defense. Despite the proclaimed need for defense of our offensive missiles, neither the Nixon administration nor any following administration maintained the ABM defense of the North Dakota missile field that was allowed under the treaty.

There matters remained until the Reagan administration. It is said that President Reagan was converted to missile defense on a visit to the continental defense headquarters at Cheyenne Mountain, when he was surprised to learn that the US had no ability to shoot down enemy missiles attacking our country. Be that as it may, in 1983 he announced plans for a Strategic Defense Initiative, intended to make nuclear weapons “impotent and obsolete.”5 No longer would the system be limited to ground-based interceptor missiles; there were plans for more adventurous technologies, including satellites carrying X-ray lasers that could burn through the skin of an offensive missile booster in the first few minutes after it was launched. The imagined system soon came to be called Star Wars.

Eventually it became clear even to the enthusiasts of the Reagan administration that the X-ray lasers and other features of the Strategic Defense Initiative were beyond current technological capacities. The administration of George Bush Sr. replaced the Strategic Defense Initiative with a system of Global Protection Against Limited Strikes, including about one thousand “brilliant pebbles,” small space-based interceptor missiles, along with more conventional land- or sea-based missiles. This strategy also led nowhere, and was allowed to lapse in the Clinton administration.

Research and development continued at a more leisurely pace. In 1996 the Department of Defense announced a plan to continue further development of a scaled-down missile defense system for three years, after which a decision would be made whether or not to deploy the system within the following three years. The National Missile Defense System under study was now limited to a single kind of interceptor missile. Instead of a nuclear weapon it would carry an “exo-atmospheric kill vehicle” weighing about 120 pounds, which would destroy the enemy warhead above the earth’s atmosphere by a direct hit rather than a nuclear blast. If it worked, it would truly be a bullet hitting a bullet.

Then, on August 31, 1998, North Korea surprised the world by launching a three-stage rocket that carried its third stage over one thousand miles before it broke up into pieces and fell into the Pacific Ocean. The missile did not fly far enough to reach any part of the US, and it could not have carried a nuclear warhead, but its launch put tremendous political pressure on the Clinton administration to do something soon about missile defense.

In July 1999 President Clinton signed a National Missile Defense Act that had been passed by Congress a few months earlier. Like the Johnson administration’s Sentinel initiative, this was more of a defense against Republicans than against external threats. The act committed the US to deploy a national missile defense “as soon as technologically possible.” Later that summer the administration settled on the defense system’s initial (“C-1”) configuration, which remains as a central element of the missile defense system under study by the Bush administration. Twenty (later increased to one hundred) interceptor missiles carrying exo-atmospheric kill vehicles would be based at Fort Greely, Alaska, to be guided to their targets initially by five early-warning radars in Alaska, California, Massachusetts, Greenland, and England. Then, later in their flight, they would be guided by a high-frequency battle management radar on Shemya Island in the Aleutians and finally, in the last six hundred miles of flight, by infrared telescopes carried by the kill vehicles.

This geographical deployment was clearly aimed at defense from North Korean missiles. To better protect the east coast of the US from missiles launched from the Middle East, it would be necessary later to add interceptor missiles at a second site, perhaps in North Dakota or Maine, and also to add additional battle-management radars. The decision to deploy the C-1 system was to have been delayed until 2000, after some of the components of the system had been tested.

The first test of the exo-atmospheric kill vehicle (EKV) was made on October 2, 1999. A dummy warhead that had been sent into space by a Minuteman intercontinental ballistic missile launched from Vandenberg Air Force Base in California was hit over the Pacific by an EKV from an interceptor missile fired from Kwajalein Atoll in the Marshall Islands. But there was less to this success than met the eye. The EKV was at first off course, so that its telescopes did not pick up the Minuteman warhead. When the EKV widened its field of view, at first it saw a large bright balloon decoy, and corrected its course, after which it saw the warhead and managed to hit it. If the warhead had not been accompanied by a decoy it might have escaped detection, and if the decoy had looked more like a warhead the EKV would have hit the decoy instead of the warhead. Even so, it seemed that under the right conditions a bullet could hit a bullet.


Then in January 2000 the EKV failed a second test. The krypton gas needed to cool the EKV’s infrared telescopes had been blocked by ice in the plumbing, so that the EKV never saw the warhead, and missed it by over two hundred feet. A third test in July also failed, when the EKV failed to separate from its booster. Further, even if all these tests had been successful, three tests were not nearly adequate to test the system.6 In August 2000 President Clinton finally decided that the Department of Defense should not start preparing the Alaska site for the battle management radar, and he announced that he would leave the decision whether to deploy the missile defense system to the next administration.

President Bush has taken the movement toward national missile defense in a new direction. Where the Clinton plan called for spending $5.75 billion in 2002 for all forms of ballistic missile defense, the Bush plan calls for spending $8.3 billion on the same tasks. The Bush administration assumed that an antimissile system much like the Clinton administration’s National Missile Defense C-1 system would be tested not only by engagements between rockets fired from Vandenberg and Kwajelein, but also by interceptors fired from Alaska sites, which could later be converted to operational missile defense sites. Also, the Bush administration proposed to supplement this land-based midcourse interception system with an ill-defined mixture of other systems, including possible airborne or spaceborne lasers that would attack enemy missiles during the initial boost phase of their flight.

The Alaska interceptor test site might violate the 1972 ABM treaty; the development and testing of airborne or space-based missile defense systems surely would, long before any actual deployment. But where President Clinton had ruled out a unilateral abrogation of the treaty, President Bush has from the first been eager to free the US from its restrictions. In August 2001 he said that the US will withdraw from the treaty at a “time convenient to America.”7 Since then the disaster of September 11 has brought Presidents Bush and Putin into closer collaboration, but the Russians have refused to agree to major changes in the ABM treaty, and now the President has given notice of his intention to withdraw from it.


So here we are again, arguing the pros and cons of missile defense. The debate raises three main issues:

Would a missile defense system actually protect the US against even the sort of attack that might be launched by rogue states like North Korea or Iraq?

It seems to me likely that the problems that bedeviled the early tests of the exo-atmospheric kill vehicle can all be solved. The fourth and fifth tests in July and December 2001 were successful, though the kill vehicle booster failed in a test later in December. The big problem, as it has been since the days of Nike X, is that any number of interceptor missiles could be used up in attacking decoys that had been sent by the attacker along with its warheads.

This is a particularly acute problem for such missile defense systems as that planned as the first phase of the Clinton-Bush National Missile Defense, which rely on intercepting warheads in midcourse, above the earth’s atmosphere. Balloons that are deployed in space at the same time as warheads will follow the same trajectory as the warheads until they reenter the earth’s atmosphere. They can easily be shaped to look much like warheads to ordinary telescopes, and heated to look like warheads to infrared sensors. It is also possible and probably even easier to make the warhead look like a decoy by putting it in a decoy balloon, or make it invisible by hiding it in a cooled shroud. There has been no realistic test of the ability of an exo-atmospheric kill vehicle to hit a warhead that is accompanied by such penetration aids.

Much of the technical argument over several decades about the effectiveness of antimissile systems has focused on the question of whether the US can solve this problem.8 It may be that at great cost we could develop a midcourse interception system that could defeat any particular group of penetration aids and warheads, but of course we are not likely to know what aids are chosen by the attacker, and I don’t see how we could ever have confidence in our ability to deal with an unknown threat. The attacker always has the last move.

Another way of defeating a US midcourse interception system was mentioned in the report of a “blue ribbon” panel on missile defense convened by Congress in 1998, which was headed by Donald Rumsfeld, and has been emphasized several times since then by Richard Garwin, one of the panel members.9 Instead of using a rocket that would launch at most a few nuclear warheads, an attacker could use the same sort of rocket to launch hundreds of “bomblets,” containing biological warfare agents, such as anthrax spores. Once deployed, the bomblets would be immune to any sort of missile defense now contemplated. This sort of missile could kill even more people than one carrying a nuclear weapon.

None of these objections applies to a missile defense that can damage an attacker’s missile while it is still in “boost phase,” i.e., during the brief period when it is being accelerated upward, before it has time to deploy warheads, decoys, or bomblets.10 But the boost phase lasts only a few minutes. A missile attempting to intercept another missile during the boost phase would have to be launched within about six hundred miles of the intercontinental ballistic missile launch site; so this sort of missile defense system would have to be targeted only at one or at most a few particular potential attackers. For example, a sea-based system that would target missiles in North Korea would not protect against missiles launched from China or Russia. Likewise, an airborne laser would not be effective against missiles that at the end of boost phase are still beyond the horizon, which for a missile at an altitude of 120 miles is about one thousand miles away. (The actual range of the laser would be substantially less than this.) An air- or sea-based boost phase intercept system could be vulnerable to preemptive attack (as also would the radars of the Clinton-Bush National Missile Defense system), but unless this attack were very carefully timed, it would trigger a counterattack that would destroy the enemy’s offensive missiles while they were on the ground. If it could be made to work, a system of space-based lasers (or “brilliant pebbles”) might be able to provide protection from threats coming from a much larger area, but this technology does not yet exist, and in any case no specific space-based laser system has been proposed by any administration.11

Is it plausible that the US would be attacked by intercontinental ballistic missiles launched by terrorists or a rogue state, or by accident?

The attack of September 11 made it clear (though it was pretty clear before) that there are people in the world who want to damage us. This seems to have shifted public opinion in favor of missile defense, and it stopped moves in the Senate to deny funding for missile defense tests that would violate the ABM treaty. Hearings on this issue before the Senate Foreign Relations Committee at which I had been asked by the staff to testify were canceled soon after September 11. But the attack also demonstrated that there are ways to hurt the US that do not involve the launch of intercontinental ballistic missiles. Even nuclear weapons could be delivered in many ways, for instance by using trucks or freighters or (as suggested by the Rumsfeld panel) ship-launched short-range missiles. But the intercontinental ballistic missile is not just one among the many vehicles that might be used by terrorists or a rogue state to attack us with nuclear weapons—it is the least likely vehicle. Though some terrorists are willing to commit suicide in their attacks, the heads of the nations that harbor them never have been. The leaders of the Taliban did not publicly acknowledge that the September 11 attacks were organized in Afghanistan, and Qaddafi has never admitted that the explosion of a Pan American airliner over Lockerbie was planned in Libya.

But unlike such terrorist attacks, an attack by intercontinental ballistic missiles carries an indelible return address. Every launch of such missiles is inevitably detected and its source identified by the fleet of American Defense Support Program satellites. Even granting that a state like North Korea or Iraq might eventually be able to deploy nuclear-armed intercontinental missiles, why would any head of government, however much he may hate us, attack us with intercontinental ballistic missiles, or allow terrorists on his soil to launch such an attack, when he and they could use many other means to deliver nuclear weapons anonymously?

On the other hand, there are circumstances in which the very visibility of intercontinental ballistic missiles might be an advantage. For instance, the US might not be deterred from its recent actions in Afghanistan or from trying to overthrow Saddam Hussein by a mere threat of nuclear terrorism; but would it risk trying to overthrow a state that had nuclear-armed intercontinental ballistic missiles?

This is a real problem for America, but it is not clear that anything but a perfect antimissile defense would make much difference. If the US had an antimissile system that had never been used in action, would this give us sufficient confidence to attack a regime that possessed intercontinental ballistic weapons, especially when we did not know what sorts of decoys their missiles carried?

But it need not come to this. There is another way that the US can avoid being subject to nuclear blackmail by states like Iraq or North Korea. It is occasionally mentioned in discussions of missile defense, though briefly and perhaps with some embarrassment. It is preemption. (Or, as it is sometimes called, pre–boost phase interception.) If a country like Iraq or North Korea were suspected of having nuclear weapons, and we saw that it had tested a ballistic missile of intercontinental range, would we really watch them begin to erect these missiles without taking steps to destroy them on the ground? These steps need not involve our use of nuclear weapons; cruise missiles are now sufficiently accurate to do the job with conventional explosives. I very much doubt if intercontinental ballistic missiles could be put in place by any state without the US knowing it, and indeed they would be of no use for nuclear blackmail unless they were known to us.

This leaves a mistaken launch by Russia or China as the only plausible way that intercontinental ballistic missiles might threaten the US. Here “mistake” might mean anything from a purely mechanical malfunction in a single rocket, to an unauthorized launch by a few madmen of all of the missiles in a submarine or a land-based missile field, all the way up to the launch of a whole arsenal of missiles ordered by a national leader who is under the mistaken impression that his country is under attack.

Launch by mistake is a serious danger, and although it was not mentioned by President Bush at the time he announced his intention to withdraw from the ABM treaty, it had frequently been cited as one reason for building a national missile defense system. Indeed, a large-scale mistaken attack by Russia is the only plausible threat that could not only damage our country but destroy it beyond our ability to recover. Such an attack would be far more devastating than anything terrorists could manage. Russia has some 3,900 strategic nuclear warheads, of which over one thousand are on land- or submarine-based intercontinental ballistic missiles that are ready to be launched at a moment’s notice.12 These missiles are increasingly vulnerable to an American first strike, as the Russian early-warning capabilities get progressively weaker. At least twice the Russians have mistakenly thought that they were under missile attack: once in March 1983 because a Soviet satellite mistook bright reflections for the launch of five missiles, and again in January 1995 because a Norwegian research rocket that had been detected by a Russian radar was interpreted as an incoming American missile. In both cases the Russian launch process came within minutes of the point where they would make a decision whether or not to launch a nuclear retaliation.

The danger of a launch during a crisis is made worse by the fact that most Russian and American warheads are MIRVs—multiple independently targeted reentry vehicles. For example, each Russian SS-18 missile carries ten warheads, each of which can be directed to a separate target. Without MIRVs, and with equal numbers of missiles on each side, there would be a disadvantage in striking first. Even if every missile had a 90 percent chance of destroying its target, the side that struck first even with all its missile forces targeted at its adversary’s missiles would leave its own arsenal empty while its adversary would still have 10 percent of its forces left. But with, say, ten warheads on each missile, the side that struck first with just 10 percent of its forces could destroy 90 percent of the adversary’s forces and still have 90 percent of its own forces left.13 Of course, this reasoning is insane. No one today thinks that either Russia or the US would plan such an attack. But in some future crisis, with different Russian leaders and with misleading data coming in from early-warning radars—who knows?

The sort of missile defense planned by the Bush administration would not protect us against a massive attack by mistake. Indeed, it has been specifically advertised not to be able to defend us from a large-scale Russian attack. (It might not even protect us against a mistaken launch of a few missiles, since Russian missiles are presumably accompanied by sophisticated decoys or other penetration aids; they surely would be if we were to deploy a missile defense system.) With or without the Bush missile defense plan, we have to face the danger of annihilation by Russian nuclear-armed missiles. With the degradation of Russian early-warning capacities and the general loosening of Russian society, this danger may be even greater than it was during the cold war. Which brings me to the third and most important issue.

Would a national missile defense system of the sort proposed by the Bush administration help or hurt our national security?

At first sight, this question seems to answer itself. Isn’t any missile defense, however ineffective, better than no missile defense at all?

One trouble with this reasoning is that we do not face a fixed threat, a threat independent of what we do about missile defense. True, we are not now in the position we were in during the 1960s and 1970s, when we could reasonably expect that any US missile defense system would be countered by an increase in Soviet offensive missile forces. The current Russian economy would not support an increase in Russia’s missile forces and, indeed, the Russians have been eager to reduce their forces, reportedly down to some two thousand or so strategic nuclear warheads. But this is still a force that could destroy the US, and much else in the world besides. The large size of their arsenal also increases the danger that Russian nuclear weapons or even long-range missiles might be stolen or sold to terrorists or rogue states. I am told that Russia now maintains tight control over its strategic nuclear weapons, but this wasn’t true in the early 1990s and it may not be true in future. There is nothing more important to American security than to get nuclear forces on both sides down at least to hundreds or even dozens rather than thousands of warheads, and especially to get rid of MIRVs, but this is not going to happen if the US is committed to a national missile defense.

The Russian nuclear force is the sole remnant of its status as a superpower. Whatever good feelings may exist now between us and Russia, any US system that might defend our country against even a few Russian intercontinental ballistic missiles therefore sets a limit below which the Russians will not go in reducing their strategic nuclear forces. Even if Russia is forced by economic pressures to continue reducing its missile forces, it can cheaply maintain its deterrent, although in ways that are dangerous. It could, for example, remove whatever inhibitions it may now have from launching its missiles on a moment’s notice. Nor is Russia likely to eliminate its MIRVs if the US goes ahead with missile defense. The START II treaty was to have eliminated all land-based MIRVs on both sides, but the Russians have already indicated that they will not go through with implementing this treaty if the US abrogates the ABM treaty.

As for China, it has right now about twenty nuclear-armed intercontinental ballistic missiles, enough for a significant deterrent against any attack from Russia or the US. A recent National Intelligence Estimate that was leaked to The New York Times and The Washington Post predicted that if the US develops a national missile defense, then the Chinese will increase their forces from about twenty to about two hundred missiles.14 And if China makes this sort of increase in its missile force, then what will Japan and India do? And then what will Pakistan do?

It may seem contradictory to argue that the proposed national missile defense system would probably be ineffective against even a small attack by a rogue state, while also arguing that it would prevent needed reductions in Russian missile forces and promote increases in Chinese missile forces. But each country “prudently” tends to overestimate the effectiveness of any other country’s defenses, especially as they may develop in future. The Soviet deployment of a primitive antimissile defense of Moscow was a major factor in America’s decision to multiply its warheads by deploying MIRVs, so that Moscow was in more danger after it was defended than before. I remember how in the early 1970s US defense planners became terrified that Soviet antiaircraft missiles might be given a role in defense against intercontinental ballistic missiles, something that never happened. Imagine then how Russians and the Chinese defense planners will take account of the unilateral American withdrawal from the 1972 ABM treaty.

If it were possible tomorrow to switch on a missile defense system that would make the US invulnerable to any missile attack, then I and most other opponents of missile defense would be all for it. But that is not the choice we face. What is at issue is a missile defense system that will take almost a decade to deploy in its initial phase, and then many more years to upgrade to the point where at best it would have some effectiveness against some plausible threats. During all of this time, however, American security will be damaged by measures taken by Russia or China to preserve or enlarge their strategic capability in response to our missile defense.

There is one sort of missile defense that would not raise these problems. As already mentioned, a defense that targets intercontinental ballistic missiles during the boost phase with either missiles or airborne laser beams could only defend against missile launches within a limited geographical area; and it would also be immune to decoys and other penetration aids. We could defend against the launch of North Korean missiles by using short-range missiles based on ships in the Sea of Japan, though to defend against a launch from Iraq or Iran would require cooperation from Turkey or some republic of the former Soviet Union, respectively. Such a defense would have no effectiveness against missiles launched from sites in Russia or China, which during the boost phase would be beyond the range of any missiles or airborne lasers we might deploy. For this reason, although this sort of defense would violate the 1972 ABM treaty, President Putin has already indicated that he would consider revising the treaty to allow it. But a boost phase intercept system would have all the destabilizing effects of other missile defense systems if it were based on satellites, or if it were combined with exo-atmospheric midcourse interceptors like those of the Clinton-Bush National Missile Defense proposal.

Developing a national missile defense system would also harm our foreign relations. It would add to the general perception that the US is unwilling to be bound by international agreements, such as comprehensive test ban treaties or environmental agreements. It would weaken Putin’s hand in dealing with Russian ultranationalists. By trying to defend the US from missile attacks while leaving our allies defenseless, missile defense would tend to undermine alliances like NATO. A boost-phase intercept system would not really be an exception; it is true that the interception of a long-range missile in the boost phase does not depend much on the destination of the missile, but by interrupting the boost it would probably only cause the warhead to fall short, perhaps on an ally, such as Canada or Germany.

A missile defense system would hurt our security in another important way, by taking money away from other forms of defense. We are simply unable to do everything we can imagine that might defend us. We need to upgrade our hospitals to deal with biological attack; improve security along our border with Canada and in our ports; upgrade the FBI computer system; and so on. All of our activities along these lines are constrained by a lack of funds. Legislation to increase funding for homeland defense was blocked in the House of Representatives because the amounts of money requested exceeded the administration’s guidelines.

If we are particularly worried (as we should be worried) about terrorist nuclear attacks on the US, then we ought to give a very high priority to working with Russia and other countries to get rid of the large stocks of weapons-grade plutonium and uranium that are produced by their power reactors. Russia now holds about 150 tons of plutonium and one thousand tons of highly enriched uranium. This material could be used not only to make nuclear bombs, which can be delivered to the US in all sorts of ways; even a technically unsophisticated terrorist could instead use it to make so-called “dirty” bombs, in which an ordinary high explosive is surrounded with highly radioactive material that when dispersed in an explosion would make large urban areas uninhabitable.

Unfortunately, this material is not under tight control.15 Since 1991 there has been a bipartisan Nunn-Lugar Cooperative Threat Reduction Program that among other things aims at improving the security of Russian control over fissionable materials and making Russian plutonium and uranium unusable as nuclear explosives, but this too is not being adequately funded. A bipartisan panel headed by Howard Baker and Lloyd Cutler has called for spending at an average level over the next decade of about $3 billion a year for securing, monitoring, and reducing Russian nuclear weapons, materials, and expertise.16 The amount in the 2002 budget for these activities is only about $750 million, even after substantial increases by Congress. Comparison of these figures with the $60 billion quoted cost (certain to be greatly exceeded) of a minimum missile defense system gives a powerful impression that the Bush administration and some in Congress are not entirely serious about national security.

I was at a press conference in Washington in November 2001, when the Federation of American Scientists released a letter signed by fifty-one Nobel laureates that opposed spending on national missile defense programs that would violate the 1972 ABM treaty. One of the reporters present asked me why, if the arguments against national missile defense are so cogent, many people in and out of government are for it? It was a good question, and one to which I am not sure I know the answer. There are the usual pressures for large military programs that come from defense contractors and from politicians trading on patriotism. The arguments for national missile defense may seem simpler and more straightforward than the arguments against it. But I think there is also a peculiar fascination with anything that projects American power into space. How else explain the idiocy of the International Space Station, or the card tables that were set up at airports during the Reagan administration by people advocating a “high frontier” missile defense program? I have to admit that thoughtlessness is not a monopoly of missile defense advocates. Some opponents of missile defense are automatically against any large military program. In assessing missile defense, or anything else for that matter, there is no substitute for actually thinking through the issues.

In my own field of physics, we make a distinction between applied physics, which is motivated by some social need, and pure physics, the search for knowledge for its own sake. Both kinds of physics are valuable, but not everything pure is desirable. In seeking to deploy a national missile defense aimed at an implausible threat, a defense that would have dubious effectiveness against even that threat, and that on balance would harm our security more than it helps it, the Bush administration seems to be pursuing a pure rather than applied missile defense—a missile defense that is undertaken for its own sake, rather than for any application it may have in defending our country.

This Issue

February 14, 2002