Robert McNamara
Robert McNamara; drawing by David Levine

The technical ideas which underlie the SDI concept of a defensive screen over the US were generated not by the President or his chiefs of staff, but by scientists and engineers driven along by the vested interests of the weapons laboratories in which they work. Some of them must have provided the information that encouraged Caspar Weinberger to call last December for immediate preparations to deploy and SDI system. Since it was only in October that Lowell Wood, one of the leading figures in the Livermore SDI program, declared publicly that it would be years before anyone could judge whether such a system was feasible, those concerned must have made remarkable progress in the intervening two months. If the call for early deployment turns out not to have been a political gimmick, Mr. Weinberger will have dealt American and Western defense a blow from which it may take long to recover.

Reputable science and bluff do not go together either in the civil or in the defense field. The more prominent scientists and engineers who are working on SDI do not appear to be impressed by what The New York Times called “the incredible pace at which breakthroughs are being made”—as reported to Congress by General James Abrahamson, director of the SDI program.1 Many have expressed considerable disquiet about the sudden shifts in the priorities of the SDI program, and about the campaign for early deployment. Dr. Gerold Yonas, until recently the research director of the SDI organization, is reported as having said that he wonders whether the US is in fact “capable of running a big long-term program.”2

What is clear is that as belief in a perfect space shield has faded, all manner of justifications have been advanced for carrying on with the SDI program.3 Even a leaky defense, it is argued, might reduce the level of destruction the Russians could inflict on the US. An imperfect defense could even enhance deterrence, since the USSR would hesitate before striking at America for fear that a proportion of its warheads might not get through. Another argument is that if the US SDI program continues, the Russians, already in economic difficulties, would feel constrained to follow suit, so adding to their burden, whereas the US, despite the fact that it is now the greatest debtor nation in the world, could afford to carry on regardless. Others assert that the Russians are already spending more than the United States on strategic defense—even, some say, on their own SDI. The US cannot, therefore, slacken its defense effort. The most usual charge is that the USSR has been violating the 1972 ABM Treaty, so that the US would be justified were it to follow suit.

While argument revolves around all these propositions, we seem to have become bewitched by the SDI criterion of “cost-effective at the margin,” as spelled out in 1985 by Paul Nitze at the Philadelphia World Affairs Council. As he put it when explaining the SDI concept:

The technologies must produce defense systems that are survivable…. New defensive systems must also be cost-effective at the margin, that is, it [sic] must be cheap enough to add additional defensive capability so that the other side has no incentive to add additional offensive capability to overcome the defense. If this criterion is not met, the defensive systems could encourage a proliferation of countermeasures and additional offensive weapons to overcome deployed defenses, instead of a redirection of effort from offense to defense. As I said, these criteria are demanding. If the new technologies cannot meet these standards, we are not about to deploy them.4

In the official statement about the goal of SDI that Secretary of State Shultz made last November, the reference to cost-effectiveness was repeated somewhat ambiguously so that, on a broad interpretation, it could also be taken to imply that the advantages that might be derived from an SDI-defensive system would outweigh its political and strategic disadvantages, not just that it would cost the Russians more, in rubles, to destroy or evade the system than it would cost the US, in dollars, to deploy one. Cost-effectiveness would also need to be measured in relation to the consequences of treaties being violated in the process of pursuing a dream. Anyhow, teams of auditors are not going to be running around examining the books of the two sides in order to signal the point at which they judge the criterion to have been violated.

In Blundering Into Disaster, Mr. McNamara guides us back to a real world from the unreal one of the visionaries of SDI. He certainly has the credentials to do so. He was defense secretary during the period of the first major buildup of the nuclear arsenal of the US. He succeeded in checking the three armed services in some of their wilder demands for ever more exotic nuclear weapons. He was one of the first major political figures to realize that the ABM systems that were being developed in the late Sixties and early Seventies were not only technically useless but strategically destabilizing.5 It was he and President Johnson who persuaded Premier Kosygin to take the same view, and so opened the way for the 1972 ABM Treaty.


Mr. McNamara’s short and powerful book begins with a factual account of the growth of the nuclear forces of the two sides. He reminds us that the United States had no intention of using its superior nuclear firepower when it confronted the USSR in 1961 over Berlin or in 1962 over Cuba. Nor was there any fear of its doing so when the two sides came face to face toward the end of the 1967 Arab–Israeli war. On the other hand, writes Mr. McNamara,

on each of the occasions lack of information, misinformation, and misjudgments led to confrontation. And in each of them, as the crisis evolved, tensions heightened, emotions rose, and the danger of irrational decisions increased…. Today we face a future in which for decades we must contemplate continuing confrontation between East and West. Any one of these confrontations can escalate, through miscalculation, into military conflict. And that conflict will be between blocs that possess fifty thousand nuclear warheads—warheads that are deployed on the battlefields and integrated into the war plans…. In the tense atmosphere of a crisis, each side will feel pressure to delegate authority to fire nuclear weapons to battlefield commanders. As the likelihood of attack increases, these commanders will face a desperate dilemma: use them or lose them.

One should note that despite the enormous growth that has taken place in the nuclear forces of the two sides, the situation was in fact potentially more dangerous in the Fifties than it was after Mr. McNamara became defense secretary. In those early days Europe was seemingly the only battleground where nuclear weapons might be used, on the assumption that greater firepower could compensate for weakness in conventional forces. NATO doctrine then had it that nuclear weapons would be used to stem any assault by Warsaw Pact troops that could not be contained immediately.

There was no novel military thinking behind this so-called strategy. Military doctrine has always had it that when your enemy is not deterred by your own array of force, you have to repel him by using whatever tactics and firepower will, it is hoped, help avert defeat. In the early days of the nuclear era this was the basis of the Dulles doctrine of massive nuclear retaliation, a fiction that survived even after the Russians deployed enough nuclear ballistic missiles to reply in kind, and even against American cities. Faced by what they regarded as a continuing Russian menace, Europeans learned to take comfort in the thought that the United States, the major partner in the Western alliance, would be ready to sacrifice Cincinnati, New York, Chicago, and Washington in defense of London, Paris, and Rome.

By the start of the 1960s, however, serious doubts about NATO strategy and tactics had already started to be voiced. It was becoming clear that were nuclear weapons ever used in Europe as battlefield weapons, not only would there be hundreds of thousands, even millions of civilian as well as military casualties in the zone of conflict, but military operations would grind to a halt because they could not be controlled in an environment of radioactive devastation. There was also the strong likelihood that the conflagration would then develop into an exchange of intercontinental missiles. Obviously nuclear weapons deterred, but, equally obviously, they could not “defend.” Their actual use as battlefield weapons could lead to disaster for both sides.

Mr. McNamara tells us that in 1962 he set out to persuade the NATO Council “to substitute a strategy of ‘flexible response’ for the existing doctrine of ‘massive retaliation.’ ” Nuclear weapons had only “two roles in the NATO context: deterring the Soviet initiation of nuclear war; and as a weapon of last resort, if conventional defense failed, to persuade the aggressor to terminate the conflict on acceptable terms.”

It was five years before this policy was accepted by the European members of NATO, and on paper “flexible response” remains NATO’s policy to this day. But, as Mr. McNamara also points out, a condition of the new policy was that NATO’s nonnuclear forces would be strengthened so that nuclear weapons would be used only in “last resort.” The “nuclear threshold,” as it is styled, would thus be raised. By all accounts this condition has not been met. Given a European war, it is widely held that NATO would be defeated if it did not turn within a matter of days, if not hours, to the use of battlefield nuclear weapons and of so-called tactical or intermediate-range weapons. Were this ever to happen, the chances are that events would swiftly escalate to an all-out nuclear exchange. In such circumstances, flexible response and massive retaliation would in effect become one and the same thing—except that today there are many more nuclear warheads available with which to destroy the world.


Mr. McNamara also points out that Russian views about the military relevance of nuclear weapons evolved in much the same way as did our own. First was the notion that the weapons not only had a “military utility,” but that nuclear battles could be fought and won. Hence the Russian buildup, not only of shorter-range and battlefield nuclear weapons, but also of intermediate-range ballistic missiles such as the SS-20. Gradually, however, more cautious voices started to be heard. Even at the time of the Cuban missile crisis, Khrushchev understood that a war in which nuclear weapons were used could mean total disaster for the USSR. More recently the Russians have formally stated that they will never be the first to use nuclear weapons.

Mr. McNamara is in no doubt that a state of mutual nuclear deterrence was in existence from at least the beginning of the Sixties, and that the political leaders of the two sides were no less deterred then than they are today. “Given the tremendous devastation which those Soviet strategic forces that survived a US first strike would now be able to inflict on this country [the US],” he writes, “it is difficult to imagine any US President, under any circumstances, initiating a strategic strike except in retaliation against a Soviet nuclear strike.” And he goes on to say that “strategic nuclear weapons have lost whatever military utility may once have been attributed to them. Their sole purpose, at present, is to deter the other side’s first use of its strategic forces.”

NATO doctrine and plans have been designed to “couple” the US to its NATO allies through the stationing in Western Europe of powerful American forces armed with nuclear weapons of various kinds. In consequence the chances are that the use of nuclear weapons in a NATO conflict would lead to the destruction not only of Western Europe and the USSR, but of the United States as well. Mr. McNamara therefore asks whether NATO’s supreme commander would ever be authorized by an American president—from whom in the final analysis he takes his orders—to use American nuclear weapons deployed on European soil. Here he agrees with Henry Kissinger who, in 1979, warned America’s European allies not to keep asking the US “to multiply strategic assurances that we cannot possibly mean or if we do mean, we should not want to execute because if we execute, we risk the destruction of civilization.”

What matters now, McNamara writes, is ensuring the continuation of the state of stability in which both sides are deterred from direct hostile action against each other, but at a much lower level of nuclear armaments. (His information is that the United States is producing two thousand new warheads a year.) War must be prevented from breaking out during a period of exacerbated tension and crisis, or inadvertently because of a breakdown, human or mechanical, in the chain of command. Mr. McNamara shares the view that neither side has anything to fear from the existence of nuclear weapons, provided that it accepts that there is no way by which it could avoid being destroyed were they ever used.

He also reminds us that in 1983, the year SDI was launched, President Reagan’s Commission on Strategic Forces, headed by Lieutenant General Brent Scowcroft, judged that it was beyond the capacity of the USSR to launch a first strike that could possibly destroy all of America’s land-based ballistic missiles. Even if that were possible, a rational Russian leadership would surely remember that the greater part of America’s retaliatory force is at sea, and for all practical purposes invulnerable. The submarine-building programs of the superpowers bear witness to the judgment that the seas are too wide, and the depths at which missile-carrying submarines patrol too great, for the odds to change decisively in favor of the anti-submarine warfare technologists. Nonetheless, while “first strike” is not a rational “option” for either side, what matters in a crisis, as Mr. McNamara says, is “what the other side believes—not what is objectively true.”

To Mr. McNamara a leaky SDI defense is just the latest phase in the action-reaction cycle of a technical arms race that is motivated by mutual suspicion and misjudgment (and, I myself would add, ignorance). That is how MIRVs came on the scene to add to the perils we now face. During the years the SALT talks were in progress, and while the diplomats were negotiating, the weaponeers of both sides, with the US ahead, were working intensively to make it possible for a single ballistic missile to carry several independent maneuverable warheads. The result was that the numerical limits to the warheads the two sides were prepared to allow each other in the treaties became far higher than they would otherwise have been. The stable door had been locked—but the horse had bolted. Mr. McNamara quotes Henry Kissinger as saying that in retrospect he wished he had “thought through the implications of a MIRVed world more thoughtfully in 1969 and in 1970” than he had.

While the apparent purpose of SDI is to substitute a regime of defense in place of a peace maintained by the threat of reprisal, it, too, thus threatens to add to our peril. The Russians will react to SDI by increasing their offensive forces and by developing antisatellite weapons.

In an article published before Reykjavik, Mr. McNamara in collaboration with George Kennan, McGeorge Bundy, and Gerard Smith, set out the political arguments that led to the conclusion that the President has to choose between the goal of deploying a defensive screen over America, and that of arms control—he cannot have both. 6 The USSR had already made this plain. But with both sides concerned to deter a first strike, Mr. McNamara and his coauthors warn that any attempt by either side to devise a defense that threatens to make its opponent’s warheads “impotent” would be countered by every means, technical or operational, that seemed practicable. If the US had indisputable evidence that the USSR was on the way to deploying nationwide ABM defenses in defiance of the 1972 ABM Treaty, that is what it would be bound to do. This was the point of the open message that Caspar Weinberger sent the President on the eve of the latter’s Geneva meeting with Mr. Gorbachev at the end of 1985. But this message, it is worth noting, did not so much as imply that the Russians were engaged on any SDI program of their own. It is not surprising that the USSR regards the US program as a process whose goal is the attainment of a first-strike capacity.

None of the arguments deployed by Mr. McNamara is new. It is the fact that he makes them, and the way in which he makes them, that are so important. Their force has become all the greater in the light of Mr. Gorbachev’s adamant opposition to agreeing to any deep cuts in strategic nuclear forces so long as there is a possibility that the R&D associated with SDI would entail “tests” outside the laboratory—by which it is assumed he means any testing that breaches the ABM Treaty. But in revealing the logic of Mr. Gorbachev’s position, Mr. McNamara insists, as I have already said, that the state of mutual deterrence could be maintained at a much lower level of nuclear armament—his figure is one thousand warheads per side. Leading members of the administration, from the President right down into the SDI organization, agree that mutual deterrence would not be weakened were both sides to make considerable reductions in the size of their respective arsenals. But Mr. McNamara also quotes Dr. Robert S. Cooper, who was an assistant secretary of defense and the director of the Pentagon’s Advanced Research Projects Agency during Mr. Reagan’s first term, as saying that whatever might emerge from the US “technology arsenal” in the attempt to devise a space-based ABM system, America could never afford to do away with its ICBM forces.

Dr. Cooper has no doubts about this. In a contribution to a US–German workshop on strategic defense that took place in Germany on December 15, 1986, which he entitled “Technical Realism and Strategic Defense,” he wrote that it would take up to ten years of vigorous R&D before it could be known whether “the goal of a reliable ballistic missile defense system is feasible,” and that “only then will an actual system design and development be possible” (in effect thus echoing what Lowell Wood of Livermore had stated in October). Since the possibility of moving from an offense-dominated to a defense-dominated strategy is so speculative and would take so long, he holds that the US will have “to continue to rely on its retaliatory forces as the principle means of deterrence in the foreseeable future.”

So there we have it. According to the President and Mr. Shultz, even if one could do away with all nuclear weapons, the US would still need the insurance of whatever emerges from SDI “to hedge against cheating and other contingencies.” According to the technical people, even were a defensive screen to materialize—and no one believes that a nonleaky defense is remotely possible—the US could not afford to do away with its nuclear weapons. In short, Star Wars I, the term that Mr. McNamara uses when he refers to the President’s vision of a defensive screen that would protect all American citizens, is likely to remain a dream forever—mainly because Russian scientists and engineers will see that it does so. The most likely outcome of the efforts of the American weaponeers is what he calls Star Wars II, a “leaky defense,” particularly of land-based ICBMs, together with a continuing buildup of all offensive strategic forces.

There are those who believe that it is worth spending as much as Mr. Weinberger is now seeking—$6 billion this year, $8 billion the next, up to a total of $70 billion for a ten-year R&D program—even if at the end of the day all that is achieved is Star Wars II. What the cost of deployment would be, if it ever came to that, no one knows. Mr. McNamara offers no guess of his own, but he cites estimates of one trillion dollars independently made by James Schlesinger and Harold Brown, two of his successors in the office of secretary of defense. What he does do, however, is warn, as does Charles Zraket in Weapons in Space,7 that experience shows that whatever the assumed price tag, the United States would “for the rest of time” have to add $100 to $200 billion a year to upgrade the system to meet new developments that would be bound to occur in the Russian attack system.

Among those who regard the prospect of Star Wars II as a not unfavorable outcome of the SDI program are Henry Kissinger, Zbigniew Brzezinski, President Carter’s national security adviser, and Max Kampelman, at present the US chief arms control negotiator in Geneva. “Retaliatory offense in reverse,” as they put it, “enhanced deterrence.” Mr. McNamara also refers to a 1984 article by Henry Kissinger, in which he said that “even granting—as I do—that a perfect defense of our population is almost certainly unattainable, the existence of some defense means that the attacker must plan on saturating it…. Anything that magnifies doubt inspires hesitation and adds to deterrence.”

“The case grows stronger,” Kissinger adds, “if one considers the defense of Intercontinental Ballistic Missiles (ICBM) launchers…. The incentive for a first strike would be sharply, perhaps decisively, reduced if an aggressor knew that half of the opponent’s ICBMs would survive any foreseeable attack.” He should have added that the surviving missiles “would of course complement the US submarine missile force.”

In a more recent article Kissinger suggests that a coherent US negotiating position for the arms control talks in which it is engaged should include “a commitment to proceed with SDI testing and deployment,” and that the Russians should not be allowed to propose any “qualitative restrictions.”8 Kissinger, the practitioner of diplomacy, seems to have returned to his roots as an armchair strategist. He turns the debate into a confrontational battle of words. No mention is made of what the weaponeers and the arms industry on both sides would be doing during the years the negotiators would be facing each other across the table, and of the consequent intensification of the arms race if the Russians added thousands of nuclear warheads to the thousands already targeted on the US. Kissinger also implies that both Americans and Russians still treat “first strikes”—nuclear bolts from the blue—as options relevant to the achievement of political objectives. He seems to have forgotten his regret that he had not given the thought he should have given, during the period of the SALT negotiations, to the deleterious implications of going ahead with the technology of MIRVing. For reasons that I fully share, and that I have already outlined, Mr. McNamara dismisses his arguments.

What I also find extraordinary about the proposition that a leaky space defense system would be of value is the assumption that it is a tolerable and realistic option, and that it would become even more tolerable were the Russians to agree to “deep cuts” in their ICBM armory—despite their insistence that SDI is a critical obstacle to so doing. What is being suggested is that an ineffective spacebased ABM system might become an effective one if “deep cuts” were made.

The underlying issues were posed by a poll, recently carried out by a Cornell University team, of members of the National Academy of Sciences who work in the physical sciences. It reported that of the large number who responded, 80 percent were opposed to SDI. Alvin Weinberg, a member of the academy and formerly the director of the Oak Ridge nuclear plant, objected to the poll because he felt that the wrong question had been put.9 While he agrees that a defensive screen that could give “effective” protection to the American population is not within sight, he says that he would have answered “yes” had he been asked whether an effective defensive shield could be built, given that both sides reduced the numbers of their respective ICBMs to “a few hundreds.” “I would guess,” he writes, “that many of my fellow academicians who denied the feasibility of a defensive system against the current threat would concede that defense is feasible against a lower offensive threat.”

Had I, in my professional capacity as a scientist, been asked the question put to the members of the academy, I should have replied first, that it was not up to me whether it was an acceptable risk to deploy a “leaky” defensive system that would allow, say, a few score of the warheads carried on a few hundreds of missiles to get through, and kill a few millions of my fellow citizens. That is a question for our political leaders to ponder.

Second, since the question implies that the US had carried on with SDI to the point, however many years ahead, when it had deployed a “leaky” space defense, and by all accounts had at the same time retained a retaliatory force, I would have asked whether it was assumed that during those years the USSR had not gone ahead with its antisatellite program, with research to discover the weak points of SDI, and with the elaboration of ways of delivering nuclear warheads against US targets—including cities—that did not entail an ICBM offensive. (As Representative Charles Bennett of Florida recently asked, “What can be expected from potential enemies in the deployment of weapons not endangered by SDI, such as cruise missiles and low-trajectory submarine-launched missiles?”10 )

As a citizen, however, I am mostly concerned with the political perception of what constitutes an acceptable risk. A single ground-burst megaton warhead could lay enduring waste to Manhattan or Washington. How many such strikes could the US absorb before its total breakdown? In 1979 the US Arms Control and Disarmament Agency published an analysis of “US Population Vulnerability.” It indicated that given an all-out ICBM offensive of the kind assumed by Kissinger and discussed in the Marshall Institute Report, a missile defense that was 90 percent effective would still “leak” enough Russian warheads to cause between 80 and 100 million “prompt deaths.” The figure would be between 45 and 70 million for a defense that was 95 percent effective.

But this scenario leaves out NATO Europe, which surely should be included. It would not take many warheads to convert Berlin, Bonn, London, Brussels, and any other city one cares to name, into radioactive wastelands to match those that would simultaneously be created in the USSR. No rules enforcing restraint apply to the kind of all-out nuclear exchange that is the scenario of SDI.

Europe cannot afford a breakdown in the state of nuclear deterrence that now helps keep the peace. The nuclear destabilization that SDI threatens also threatens the stability of the Western alliance. In his recent article, Henry Kissinger wrote that the US “has a right to expect” its allies to agree with the negotiating position that he has now formulated, and that, by calling for a commitment to develop and deploy a space-based ABM system, entails the abrogation of the ABM Treaty, a treaty that he himself helped to negotiate when he was President Nixon’s national security adviser.

I fear that he is likely to be disappointed. The European allies of the US have not only declared themselves as favoring a strict reading of the treaty but are aware of the powerful opposition that is mounting in Congress to any other interpretation. They must also realize that the tentative understandings about intermediate-range ballistic weapons could be jeopardized were the advice now being tendered by Henry Kissinger to become administration policy. These are matters with which I deal in a third article.

This is the second of three articles on SDI.

This Issue

April 23, 1987