Chapter 5
"Don't Ask the Soviets. Tell Them"
SDI and Arms Control

Whatever military and strategic impact SDI will ultimately have, it has already had important, if controversial, consequences for arms control. The most zealous advocates of the project have been largely indifferent to these consequences, believing that what matters most in the East-West confrontation is military power, not paper agreements incapable of deterring the Soviets from seeking to exploit any military advantage they can acquire. Some of the most prominent supporters of SDI, however, share the generally accepted view that although arms-control measures alone cannot remove the sources of conflict, verifiable agreements to limit arsenals and regulate development of new weapons can promote stability, ease economic burdens, and possibly lead to more far-reaching forms of cooperation.^[1]    Until the declaration of the SDI, there was also general agreement that the constraints on the deployment of strategic defenses against ballistic missiles and on the development of new types of defenses (agreed to by the superpowers in the 1972 ABM Treaty) were the necessary condition of progress in the limitation and reduction of offensive arms. Since the announcement, the prospect of such progress has been shrouded in the ambiguity of the Reagan administration's vision of a comprehensive defense. Further offensive reductions are likely only if that ambiguity is resolved. An admission that no space-based deployment is feasible for the foreseeable future, coupled with continued adherence to the limitations on testing prescribed by the treaty, is an essential condition for further achievements in arms control. To the extent that either or both superpowers proceed to develop

and deploy comprehensive defenses (not just to conduct research on the possibilities, as a hedge against a breakout), the very premise of strategic arms reduction is called into question. This is the key problem. SALT I and SALT II proceeded on the assumption that the two sides would be equally deterred from attacking each other because of their equal vulnerability to retaliation. The premise of comprehensive defenses shifts the basis of deterrence (insofar as the defenses are reliable) from fear of retaliation to confidence in the ability to survive a nuclear attack of great magnitude. As the advocates like to say, "assured survival" would replace "assured destruction." But while both sides are busy designing and constructing such defenses, they lose all incentive to reduce offensive forces unless and until it becomes pragmatically (and not just theoretically) demonstrable that such defenses can reliably absorb an attack of any size and that additional units of defense can be deployed less expensively than additional units of offense. Until both these demanding criteria, reliability and cost-effectiveness, can be met, the obvious response to the deployment of defenses is to build offenses capable of overcoming and saturating them. Other things being equal, the larger the offensive threat, the larger (and more expensive) defenses must be and the greater the likelihood that a significant fraction of the attack will penetrate the defenses. Either side would therefore be foolish to relinquish any of its offensive punch while the other is building defenses. At most, a sensible counterstrategy would be to trade some offensive power for greater assurance of penetration. Once a commitment is made by either side to develop and deploy comprehensive defenses, therefore, the ABM Treaty would become a dead letter. It could no longer be expected to serve as a foundation for agreed reductions of offensive capacity.

This obvious consequence has been somewhat obscured by developments in the immediate aftermath of the declaration of SDI. At first, Soviet leaders warned that if the United States were to violate the terms of the treaty by testing or deploying "space strike" weapons, they would take whatever measures they considered necessary to maintain strategic parity. The Reagan administration reacted by accusing the Soviets of having themselves violated arms-control agreements but promised U.S. allies that for the time being the terms of the treaty would be respected in SDI experimentation. Arms control negotiations were suspended following a Soviet walkout prompted by the deployment of intermediate-range weapons in Western Europe, so the issue was not joined. In 1985, however, the two sides agreed to resume negotiations over intermediate-range

and strategic weapons and—at Soviet insistence—also over space weapons; the Reagan administration announced a new "broad" interpretation of the treaty's testing provisions. Success in the INF negotiations in 1987, coupled with agreement on the outlines of a strategic reduction agreement, still left SDI as a major unresolved issue. It may not prove to be an insuperable barrier in the Strategic Arms Reduction Talks (START) if both sides can "agree to disagree" about it, or, better still, if they can reach an understanding on what types of testing are permitted under the ABM Treaty. SDI has nevertheless seriously complicated the arms control process. Although some of its supporters claim that it helped to bring the Soviets back to the bargaining table, the president himself insisted all along that SDI was not a bargaining chip. The Soviets, in any event, had other strong motivations for reopening negotiations. In retrospect, the injection of SDI into the already difficult arms control process shows that ill-considered commitments to new military technologies complicate rather than reinforce the pursuit of national security. Unless both sides keep their work on strategic defenses consistent with the ABM Treaty, or some modified version of it, arms-reduction talks could well grind to a halt.

The ABM Treaty And Salt: The Link That Failed

The centrality of the ABM Treaty to strategic arms control was well recognized when the treaty was negotiated in 1972. The premise that seemed to have gained acceptance was that both sides had an interest in restricting deployment of defenses against ballistic missiles. Such deployments, it was reasoned, would lead only to an expansion of offensive arsenals and to uncertainty on each side about the reliability of its retaliatory deterrent. The shapers of U.S. strategy had earlier become convinced that in view of these consequences, in addition to the technical difficulties of achieving a meaningful defense against ballistic missile attack, a mutual decision to forgo such defenses would be advisable.

As we discuss in chapter 2, the effort to forge a link between strategic defense and arms control first emerged in the 1960s when the United States learned of Soviet plans to deploy ground-based defenses. The Soviet defenses resembled those the United States had been working on sporadically, such as the Nike X. The seeming futility of such research efforts, coupled with fear that both sides were being drawn into a spiraling strategic arms race, led the Johnson administration (at the urging of

Secretary of Defense McNamara) to propose a treaty banning the deployment of ABM systems. (This idea was probably introduced in the Pentagon by Jack Ruina, then director of DARPA, in 1962 or 1963. Ruina and the physicist Murray Gell-Mann discussed it at the Pugwash Conference in 1964, the first time the Soviets were made aware of U.S. thinking on this subject.) McNamara's view was that offensive and defensive systems were interactive: The deployment of ABMs by one side would lead the other to develop offensive forces, rendering the defensive effort futile. Because the proliferation of offensive forces was cheaper than the development of defenses, it seemed that the advantage conferred on the offense by the advent of nuclear weapons—coupled with high-speed ICBMs and SLBMs—could not be offset by defenses. The defensive systems of the 1960s were thought to be highly vulnerable to saturation attacks, and the technical experts were far from unanimous in considering them reliable even in simpler situations. As a result, McNamara urged that the issue be raised with the Soviets. It was therefore brought up at the meeting with Premier Kosygin in Glassboro, New Jersey, in June 1967. As we note in chapter 2, Kosygin was at first outraged, arguing that offenses were "bad," and defenses were "good" because they did not harm anyone or serve any aggressive purpose. McNamara tried to persuade him that deployment of defensive systems only complicated the problem of achieving a strategic balance that both sides could afford by forcing each side to proliferate offenses in order to overcome the defenses. In effect, he argued that stability would be enhanced if both sides were to forgo defenses and accept strategic parity so as to promote mutual deterrence based on the possession by both sides of an invulnerable capability to retaliate.^[2]   

McNamara's argument eventually took hold, leading to the SALT I negotiations and the signing of the ABM Treaty in 1972. Although the treaty had only an interim agreement (SALT I) to limit offensive weapons, it was hoped that the signing of the treaty would lead to further agreements to reduce offensive weapons, once the fear of defenses had been eliminated. The results of agreement to limit defenses, however, were mixed. As a result of SALT I, a rough parity was achieved in strategic delivery vehicles—2,200 for the United States, 2,400 for the Soviets—but warheads began to multiply with the mounting of MIRVed warheads on the delivery vehicles. (Ironically, MIRVed warheads were promoted in the 1960s as the answer to ABM, but when ABMs were renounced, MIRVs remained.) Even though both sides proposed major reductions, offensive reductions went unrealized for several reasons—some

having to do with the political context of arms control, others with its inherent limitations.

One explanation for the lack of progress on offensive reductions is that the political climate was uncongenial. In effect, arms control came to be held hostage to other aspects of the superpower relationship. The failure of the United States to ratify the SALT II agreement resulted from its displeasure with the Soviet Union's invasion of Afghanistan and suppression of discontent in Poland; in addition, President Carter's leadership had been badly weakened by the hostage crisis in Iran. From the Soviet point of view, the failure of the United States to ratify SALT II was evidence of a stubborn refusal to accept the military parity resulting from Soviet investments and other changes in the global "correlation of forces." The collapse of the détente of the 1970s was made final by the advent to power of Ronald Reagan, who was opposed not only to SALT II, which he described as "fatally flawed," but also, it seemed, to the entire process of arms control as it had been conducted by his predecessors. Those who believed in arms control, he implied, were naïve in supposing that a hostile and expansionist "evil empire" could be trusted to enter into and observe equitable agreements.

Equally important in the failure to go beyond the ABM Treaty and the SALT II agreement was the Soviet Union's insistence on adding to and diversifying its strategic forces and the United States' modernization of those it already had. In SALT II both sides agreed to limits that did not affect expansion and modernization in areas where both were seeking to gain certain military advantages. Thus, the Soviets were not inhibited from deploying intermediate-range missiles in Europe (hardly demonstrating the spirit of détente), maintaining heavy ICBMs that threatened the land-based leg of the U.S. triad, developing new strategic bombers capable of attacking the United States, continuing to expand and improve their strategic air defenses, and designing and testing more accurate and mobile missiles to complicate U.S. targeting policies. For its part, the United States proceeded with the development and deployment of the MX—a more accurate missile than the Minuteman and regarded as a first-strike weapon by many observers. The development of the Trident submarine and the Trident II D-5 missile gave the United States a new and potent capability against Soviet silos. In addition, two new bombers—the B-1 and the ATB (Stealth, or B-2, bomber)—along with a new generation of cruise missiles capable of air- and sea-launch, strengthened the third leg of the U.S. triad. These offensive developments, like those of the Soviets, hardly indicated a real willingness on

the part of military planners and their political supporters to wind down the arms race or to accept any kind of freeze on potentially destabilizing developments.

Arms-control progress has also been affected by the refusal of both the U.S.S.R. and the United States, despite the ABM Treaty, to abandon efforts to develop defenses against nuclear attack. As we note in chapter 2, the Soviet Union has mounted the most sustained and elaborate effort in strategic defense. They have also exploited the two most glaring loopholes in the ABM Treaty, which does not forbid the development either of tactical ABM (ATBM) weapons capable of intercepting short-and intermediate-range ballistic missiles or of anti-satellite (ASAT) weapons. When the Soviets tested (anti-aircraft) SA-X-12 against an RV with the characteristics of an intermediate-range weapon, they stepped into a gray area. U.S. analysts were bound to think the move might lead to a full-scale breakout from the treaty. If the Soviet ATBM could intercept intermediate-range missiles, it could also intercept submarine-launched missiles, considered by both sides as strategic weapons. If the SA-X-12 could be further upgraded to intercept ICBMs, the U.S.S.R. would quietly acquire a territorial defense while supposedly complying with the treaty. The Soviets also developed and tested a co-orbital ASAT, which poses a threat to U.S. satellites in low orbits and could lead to a new generation of ASATs capable of threatening communications satellites in high orbit. Although ASATs could not provide an ABM defense, they threaten the U.S. offensive capability by jeopardizing U.S. command, control, and communications. The Soviets, whose emphasis on land-based missiles enables them to put more reliance on ground communications, are not as threatened by the development of ASATs. The development of ASATs also strongly indicated the Soviets' readiness to exploit the loopholes in the ABM Treaty. The Soviets were also known to be working on advanced lasers that could conceivably serve as ASAT and ABM weapons.

The most egregious instance of Soviet noncompliance with the treaty was the construction of a phased-array radar near Krasnoyarsk, in central Siberia, hundreds of miles from the nearest coastal region. In order to prevent the creation of a radar network that could serve as a basis for a nationwide territorial defense, Article VI(b) of the ABM Treaty limited the emplacement of large phased-array radars (LPARs) to positions "along the periphery" of both countries and required that they be "oriented outward." The Soviets, however, pointed to Agreed Statement F, which allows for the inland location of phased-array radars "for the

purposes of tracking objects in outer space," which, they claimed, was to be the function of the Krasnoyarsk facility. U.S. specialists responded that in its outward dimensions the radar is identical to the Pechora-type radar used for ABM purposes and that its location violates an explicit treaty provision. It is much more elaborate than would be needed for space tracking and is wrongly oriented for that purpose in any event. They therefore concluded that the Soviets were deliberately building an expensive facility in full knowledge that it was incompatible with the treaty regime.

During the Nixon, Ford, and Carter administrations, questionable Soviet activities were queried by the U.S. representative to the Standing Consultative Commission established under the treaty to allow for clarification of its provisions and to deal with problems of implementation of ABM limits and the limitation of offensive weapons. Robert W. Buchheim, the U.S. commissioner until 1981, has asserted that despite difficulties, "overall Soviet fulfillment of major obligations under agreements, whether ratified or in effect on a de facto basis, has been generally good."^[3]    The Reagan administration, however, made only limited use of this confidential forum and instead brought allegations of Soviet violations of this and other treaties to public notice. This policy mirrored a larger one of treating the post–SALT II negotiations as opportunities to tell the U.S.S.R. that the United States did not intend to be as accommodating as the Carter administration had been. Indeed, the United States' only interest in the negotiations, according to this new policy, was to secure Soviet concessions, specifically the withdrawal of SS-20s from Europe and the elimination of the threat posed by the heavy land-based ICBMs, especially the SS-18s.

The Soviets saw more to gain by withdrawing from the negotiations. They hoped to score propaganda gains and to drive a wedge between Western Europe and the United States by denouncing Reagan as a cold warrior interested only in restoring U.S. military superiority and in dictating to the Soviet Union. The Soviets argued, often effectively, that this U.S. policy represented not only a departure from the U.S. policy of pursuing peaceful coexistence but also a return to a full-fledged arms race, leading inexorably to the abandonment of the SALT II Treaty—which the United States was observing on the basis of an informal understanding with the Soviets—and of the whole effort to achieve progress in arms control.

Once NATO overrode domestic Western European objections to the deployment of longer intermediate-range weapons, the Soviets decided

to resume negotiations. The announcement of SDI in 1983 may not have been intended to promote arms control, but it nevertheless gave the Soviets an excuse to return to the negotiating table. They almost certainly wanted to return for other reasons, including the end of the interregnum created by Brezhnev's extended illness and eventual death. SDI enabled the Soviets to claim that a new and ominous challenge had been added, and that they had compelled the United States to address the issue of the militarization of space. A previous Soviet proposal to extend the ban on weapons in space had been rebuffed by the Reagan administration. The revived START talks could thus cover three subjects—intermediate-range weapons in Europe, the proposed reduction of offensive strategic weapons, and the issue of weapons in space (both ABMs and ASATs). SDI was at least the catalyst for the renewal of the negotiations, even though it was later to prove a serious obstacle.

From Deadlock To Dashed Hopes: The Impasse At Reykjavík

The unexpected resumption of negotiations put domestic critics of SDI in an awkward position. Those in favor of SALT II and of arms control in general had denounced SDI as a tactic designed to undermine adherence to the ABM Treaty, at the very least. They saw in it the hand of those who had opposed the treaty in the first place. These critics argued plausibly that the president and his supporters favored SDI because it seemed to bypass the need to negotiate with the Soviets and relied instead on unilateral U.S. initiatives. It was being advanced, they also argued, to disguise the fact that the United States had deliberately refused to negotiate seriously in the Geneva talks, preferring to use the failure of the talks as a way of putting the onus on the Soviets while proceeding with the administration's modernization efforts. Instead of seeking to reassure the Soviets about the United States' peaceful intentions, the Reagan administration's approach (illustrated by the SDI) was to threaten them with (1) an expensive arms race (in the expectation that if it did not bankrupt the U.S.S.R. it would at least expose the weakness of its economy) and (2) a new military capacity that would confer military superiority by seizing the high ground of space before the Soviets could. The proponents of the ABM Treaty and of arms control in general therefore saw the entire effort as one designed to subvert negotiations in favor of space-age saber-rattling.

But when the Soviets agreed to resume negotiations in 1985, despite

their denunciations of SDI, many of these proponents of the ABM Treaty and of arms control took a somewhat different line. Although still convinced of the folly of SDI, they admitted, if only grudgingly, that it had been useful in bringing the Soviets back to the bargaining table and began to view it as a valuable bargaining chip. Although the president himself denied that he would allow SDI to be traded away, those in favor of the treaty and of further arms control argued that SDI could well present a historic opportunity. If, as seemed to be the case, the president had changed his tune and was now eager to achieve radical reductions of strategic weapons (having been concerned in his first term to win reelection, but in the second to win a place in history), then he might recognize that restraint on SDI was a small price to pay for achieving a breakthrough on arms control.

Gradually, the pieces of such a historic breakthrough seemed to fall into place. When Reagan and Gorbachev met at the Soviet leader's request in what was billed as a "pre-summit" at Reykjavík in 1986, this breakthrough was actually within grasp. There the Soviets agreed in principle to the zero–zero INF proposal, to the equal astonishment of those who had first proposed it and critics who had denounced the president's offer as nonnegotiable and therefore insincere. The Soviets also seemed to agree to a proposal for a 50 percent reduction in offensive warheads that was not very far from the administration's formula, especially because it would have reduced the Soviet advantage in the heavy missiles, which the administration had particularly emphasized. But the price of such concessions, as the Soviets made more or less clear in the final session at Reykjavík, was that the United States agree to restrict research on SDI to the laboratory and to forgo testing of any kind in space. They may not have intended to be quite so categorical and may have been insisting only on a reaffirmation of the traditional interpretation of the ABM Treaty. Whatever the intention, their adamantly stated view was enough to lead the president to abort the negotiations.

Postmortems of Reykjavík suggested that the Soviets were deliberately trying to entrap a gullible president ill-equipped to appreciate the nuances of arms control and to cause further difficulties in relations between the United States and its allies. Although he refused in the end to accept the Soviet proposals, Reagan aroused the ire of many U.S. and allied strategists by apparently agreeing to a Soviet proposal to eliminate all ballistic missiles (or, possibly, all nuclear weapons) over a ten-year period. Among most mainstream Western strategists, it is taken for granted that U.S. possession of nuclear weapons is a vital counterweight

to Soviet conventional strength, especially in Europe. Shock over Reagan's readiness to abandon the nuclear deterrent, as well as the prospect of an INF accord, made Western European political leaders uneasy. Without nuclear weapons based on European soil and with the prospect of a reduced U.S. retaliatory force, they feared that the military balance in Europe would swing toward the Soviets unless agreements on strategic weapons were accompanied by reductions in Soviet short-range nuclear forces and in conventional forces—subjects neglected in the Reykjavík talks.

The Political Effort To Shore Up SDI

When Reagan terminated the Reykjavík meeting without agreement, his staunch conservative supporters expressed relief, though they often couched their feelings in the form of praise for the president for having the courage to say "no" to a deal the Soviets had put forward only as a trap. But the conservatives had been unhappy to start with—that the president appeared to be edging closer to an agreement. Just before the first Reagan-Gorbachev summit earlier in 1985, Secretary of Defense Weinberger took the extraordinary step of writing a "top-unclassified" letter to the president in which he warned the president not to agree to limit or abandon SDI or to make other concessions. Even more extraordinary was the leak of the letter, evidently by forces in the White House who shared Weinberger's opposition to arms control and were convinced that Secretary of State Shultz was leading the president in the direction of détente. After Reykjavík, these forces were on the defensive, but they could at least argue that the president had come to his senses when he saw what the Soviet design really was, however much his instinctive desire for peace had tempted him to suppose that there might be something serious in the deal the Soviets were proposing.

After Reykjavík these same conservatives lost little time in deciding that the only way to prevent the president and his secretary of state from making concessions on SDI was to demand early deployment of a strategic-defensive system. This demand was couched in terms of the unexpectedly early success of SDI research—a success that informed critics saw as dubious—with the evident intention of nullifying any attempt in Congress to kill or emasculate SDI. The proponents sought to maintain momentum by showing that the initiative was not simply a research program to be pursued for decades. Still another motive in proposing "early deployment" was to force the hands of those who had been

able to avoid open opposition to SDI by supporting research. A call for deployment would compel them to take a stand either for or against SDI (and the president). Finally, if successful, a move for early deployment would force abrogation of the ABM Treaty by mandating testing in space and deployment of space-based systems. The same forces conveniently argued that the Soviets had already violated the treaty, so it had already ceased to be an operative restraint on the Soviets' own undeclared SDI.

When Weinberger endorsed the proposal for early deployment, he once again exposed the disagreement within the administration over the virtues of arms control. At Reykjavík, the president had offered to refrain from deployment (though not necessarily from testing in space) for a period of ten years. The Weinberger proposal ran counter to this offer. Further underscoring the disagreement, the president's arms-control negotiator in Geneva, Max Kampelman, reiterated that the U.S. proposal at Reykjavík still stood. In his January 1987 State of the Union address, the president straddled both positions. He reaffirmed support for the SDI but reiterated his interest in achieving an arms-control agreement with the Soviets. As the two sides edged closer to a treaty, a reaffirmation of the ABM Treaty came to seem essential.

Despite the confusions, the debate over SDI at least had the virtue of illustrating the pivotal role of the ABM Treaty in arms reductions. A decision to deploy strategic defenses not permitted under the 1972 ABM Treaty would require either the treaty's abrogation or amendment; even the effort to develop such systems would at some point entail one or the other. Paradoxically, some supporters of SDI acknowledged that U.S.-Soviet cooperation is essential because no defense could possibly prove feasible unless accompanied by constraints on offensive deployments. Nevertheless, the relationship between SDI and arms control remained politically vexing. On the one hand, supporters of arms control argued passionately that to tamper with the ABM Treaty is to remove the cornerstone of the whole process; on the other, supporters of SDI contended that arms control is a means to an end and that if technology can obtain that same end—a peaceful world—then arms-control measures would have to accommodate the technology. Other supporters saw SDI as a substitute for arms control, contending either that arms control had failed or that it is essentially a Soviet device to achieve equivalent status as a great power with the United States, and nothing more. They therefore argued that the United States would do better to rely on its technical, economic, and military capacities to thwart aggressive

moves by the U.S.S.R.—moves that are inevitable given the conflict between the two systems and the expansionist character of the Soviet Union. In this debate, the precise terms of the ABM Treaty assumed a new importance.

The Terms Of The ABM Treaty

As amended by a 1974 Protocol, the first three articles of the treaty forbid both parties from deploying ABM systems "for a defense of the territory of its country" or for more than one region with a radius of 150 km. Both sides are allowed only one ABM site of one hundred interceptors and one hundred launchers positioned either around the national capital or at one ICBM deployment area. Strict limits are put on the number of ABM radars at the permitted site (the numbers differ, depending on whether the ABMs are protecting missiles or area). Article V limits efforts to develop, test, and deploy weapons and components except for those that are fixed and ground-based. Article VII specifically permits "modernization and replacement" of permitted ground-based systems and components.

Several of the final articles of the treaty spell out its binding character. According to Article XV, the treaty is "of unlimited duration." Article XIII provides that disputes over compliance and other relevant matters are to be referred to a Standing Consultative Commission (SCC), composed of representatives and advisors appointed by both parties. Article XIV allows for amendments in the event of new developments. Article XV allows withdrawal on six months' notice if either party determines that its "supreme interests" are no longer served.

The treaty was accompanied by an interim agreement establishing launcher limits on ICBMs and SLBMs. The expectation was that a new treaty would be worked out to limit offensive weapons comprehensively, i.e., to include bombers and/or cruise missiles that could reach the territory of the two superpowers. The guidelines for such a treaty were established in the Vladivostok accords between President Gerald Ford and the Soviet leader, Leonid Brezhnev, in 1974. The SALT II agreement was signed in 1979 by representatives of the Carter administration, but it was withdrawn from consideration in the Senate after the Soviet invasion of Afghanistan. It had also aroused considerable opposition in and of itself because of doubts about its verifiability and because it allowed the Soviets to retain certain numerical advantages and throw-weight advantages in land-based missiles considered to be the most accurate available systems.

In the course of negotiations, certain interpretations were spelled out in the form either of "agreed statements" or "common understandings" appended to the treaty along with "unilateral statements," added by the U.S. side only. Additional protocols and understandings were later reached in the SCC, but some ambiguities remained. "Some key definitions and concepts used in the treaty," several of its defenders admit, "were not clarified by agreed interpretations during or subsequent to SALT I,"^[4]    but even a leading critic of its ambiguities has acknowledged that "it is concise and elegant as treaties go."^[5]   

In essence, the treaty bans the deployment of ABM systems that would provide or prepare a base for a defense for the entire territory of one of the parties. It also limits testing and development of ABM systems and components. Even a regional defense is banned (except for a radius of 150 km, presumably for the national capital area, which is considered subregional). An ABM system is defined (in Article II) as any designed to counter strategic ballistic missiles or their elements in flight. The article describes such systems as "currently consisting of" ABM missiles, launchers, and radars. Testing is permitted (by Article IV) at "mutually agreed upon test ranges," at which no more than fifteen ABM launchers are permitted. Article V explicitly prohibits not only deployment of systems designed to be based in the air, in space, at sea, or to be mobile land-based but also the testing and development of such systems. Article VI bars the application of ABM capabilities to non-ABM systems or the testing of non-ABM systems in an "ABM mode." (These systems, not defined in the treaty, include air defense, anti-tactical ballistic missiles, strategic offensive missiles, and anti-satellite missiles.) The same article stipulates that ballistic missile early-warning radars are to be located only along the periphery of the national territory and oriented outward, so that they could not serve as the basis of a territorial defense system. Article IX prohibits the transfer to other states and deployment outside national territory of ABM systems or their components.

The agreed statements amount to negotiated clarifications of treaty language. One of the statements makes clear that the parties intend to ban launchers with multiple interceptors from being mounted on the allowable launchers. Another stipulates that transfers of technology to other states include technical descriptions and blueprints. In 1978 the SCC also developed protocols specifying criteria for "testing in an ABM mode." The criteria are classified. The one that has proved to be most controversial, Agreed Statement D, stipulates that in the event ABM systems "based on other physical principles" are "created in the future," limitations on such systems would be "subject to discussion in accordance

with Article XIII and agreement in accordance with Article XIV," which covers amendments.

The Testing Issue

When the treaty was submitted to the Senate in the summer of 1972 for "advice and consent," testimony concerning its various provisions made clear, at least to those especially concerned with the issue, that development and testing of "exotic" technologies such as lasers would be permitted only in the case of ground-based systems. The journalist John Newhouse later reported in a book on the negotiations that the U.S. delegation had been under instructions not to agree to any clauses that might have precluded the development of ground-based lasers in an army project.^[6]    In a legal analysis prepared on May 24, just two days before the signing of the treaty, John B. Rhinelander stated unambiguously that Article II and Agreed Statement D prohibited all efforts to develop "future systems" that might be substituted for those then in use: "An ABM system is described in paragraph 1 of Article II in terms of 'current' ABM components. This does not, however, limit the generality of the term ABM systems as used in the Treaty to systems composed of 'current' ABM components, but would also include 'future systems' based on physical principles other than those used for 'current' ABM components and capable of substituting for 'current' ABM components."^[7]   

In Senate hearings, several Nixon administration officials, including Secretary of Defense Melvin Laird and DOD Director of Research and Engineering John Foster, testified that although development of ground-based lasers was permissible, no other development and testing of such unconventional systems would be allowed. In a written reply to a question from Sen. Barry M. Goldwater (R., Ariz.), regarding the effect of the treaty on the development of boost-phase missile defenses (presumably relying on satellite-based lasers), Secretary Laird expressed the same understanding. He noted that there is "a prohibition on the development, testing and deployment of ABM systems which are space-based" but "no restrictions on the development of lasers for fixed, land-based ABM systems" and that the prohibition of the development and testing of ABM systems other than those that would be ground-based does not apply "to basic and advanced research and exploratory development of technology which could be associated with such systems, or their components."^[8]    Sen. Sam Nunn (D., Ga.) was to point out in 1987 that this

"formal, written executive branch response … clearly sets forth the traditional interpretation of the treaty with respect to exotics, permitting development, and testing only in a fixed, land-based mode. The reply makes it clear that mobile/space-based exotics are subject to the comprehensive ban on development, testing and deployment, with the understanding—as stated in Secretary Laird's reply—that the treaty only permits 'basic and advanced research and exploratory development.'"^[9]    Foster, under questioning from Sen. Henry M. Jackson (D., Wash.), was even more explicit in this exchange:

To clarify the testing issue, in response to a direct question from Senator Jackson, Foster subsequently inserted into the record of the hearings a fuller explanation:

Article V prohibits the development and testing of ABM systems or components that are sea-based, air-based, space-based, or mobile land-based. Constraints imposed by the phrase "development and testing" would be applicable only to that portion of the advanced development stage following laboratory testing, i.e., that stage which is verifiable by national means. Therefore, a prohibition on development—the Russian word is "creation"—would begin only at the stage where laboratory testing ended on ABM components, on either a prototype or bread-board model.^[11]   

As Senator Nunn also later pointed out, testimony from military officers presented the same understanding of the restrictions on testing. One of the officers specifically indicated that the Joint Chiefs of Staff "were aware of the limits on development and testing of laser ABM's, had agreed to them, and recognized that this was a fundamental part of the final agreements."^[12]   

In response to a question from Senator Jackson during testimony before the Senate Armed Services Committee on July 18, 1972, Amb. Gerard C. Smith, the chief U.S. negotiator, submitted a written clarification

(based on a review of the SALT delegation's reporting cables) of the treaty's provisions restricting development and testing. According to Smith, the treaty prohibits

that stage of development which follows laboratory development and testing. The prohibitions on development contained in the ABM Treaty would start at that part of the development process where field testing is initiated on either a prototype or breadboard model. It was understood by both sides that the prohibition on "development" applies to activities involved after a component moves from the laboratory development and testing stage to the field testing stage, wherever performed. The fact that early stages of the development process, such as laboratory testing, would pose problems for verification by national technical means is an important consideration in reaching this definition. Exchanges with the Soviet delegation made clear that this definition is also the Soviet interpretation of the term "development."^[13]   

As Smith noted, the understanding of the ban on mobile systems, including those based in space, had been clarified in a meeting of the Ad Hoc Working Group headed by U.S. representative Sidney Graybeal and his Soviet counterpart, Viktor Karpov, in Helsinki on September 15, 1971. A U.S. "memorandum of conversation" ("memcon"), made public in 1987, recorded that "Karpov agreed with Graybeal's interpretation that the Soviet text meant 'any type of present or future components' of ABM systems."^[14]   

When the treaty was ratified in the Senate by a vote of 88–2, two of the senators explained their differing votes on ratification by referring to testimony on the testing issue. Sen. James Buckley (R., N.Y.) said he had decided to vote against ratification because the treaty ruled out the development and testing of space-based laser weapons:

The agreement goes so far as to prohibit the development, test or deployment of sea-, air- or space-based ballistic missile defense systems. This clause, in Article V of the ABM treaty, would have the effect, for example, of prohibiting the development and testing of a laser type system based in space which could at least in principle provide an extremely reliable and effective system of defenses against ballistic missiles. The technological possibility has been formally excluded by this agreement.^[15]   

Sen. Strom Thurmond (R., S.C.) announced that he was voting to confirm the treaty only reluctantly because he objected to the ban on the development of space-based lasers, which he thought "effectively prevents us from ever having the means to protect our population from a Soviet first strike."^[16]   

In the years following ratification of the treaty, this understanding became official policy (reflected, for example, in the ACDA's annual arms-control impact statement prepared in conjunction with the presentation of the defense budget to Congress). As late as 1985 the ACDA report declared that "the ABM Treaty prohibition on development, testing and deployment of space-based ABM systems, or components for such systems applies to directed energy technology … used for this purpose."^[17]   

Reinterpreting The Text: The Rise Of The "Legally Correct" Interpretation

Despite such clarifications and pronouncements, the issue of what the treaty permits became problematic in 1985. Reagan's national security advisor, Robert McFarlane, revealed during a "Meet the Press" appearance on October 6 that a review of the treaty language had indicated that development and testing of systems based on "other physical principles" was actually allowable: "Research involving new physical concepts … as well as testing, as well as development indeed, are approved and authorized by the treaty. Only deployment is foreclosed."^[18]    McFarlane's announcement was the outgrowth of a legal review of the treaty initiated in the Department of Defense by Richard Perle, assistant secretary for international security policy, and Fred C. Iklé, under secretary for policy.

The question of the permissibility of developing new technologies not incorporated in ABM systems in 1972 had first been raised in 1975 by Donald G. Brennan of the Hudson Institute. Abraham S. Becker and William R. Harris, both affiliated with the Rand Corporation, raised it two years later—independently. Brennan's query stimulated a correspondence with several of the negotiators, at the end of which he pronounced himself satisfied that the text did in fact ban the development of "exotic" systems, except for those that would be ground-based and could be tested at the agreed test ranges. In a letter in 1977 to Rhinelander, Brennan made a full and gracious concession, noting that "any further insistence that the Treaty does not necessarily ban the development of (among others) space-based exotic ABM systems would have to be reckoned willful, indeed obstinate, stupidity."^[19]    Becker raised the issue publicly in a letter to the editor of the journal International Security , commenting on an article by Raymond L. Garthoff, who, as executive secretary of the U.S. delegation, had played a key role in negotiating the language dealing with testing. In terms that anticipated the line of

reasoning later used in the Reagan administration, Becker contended that Agreed Statement D could plausibly be read to permit testing of new systems based on "other physical principles":

One might ask why Agreed Interpretation [D] was necessary at all. Does not the introductory phrase of Article III—"Each Party undertakes not to deploy ABM systems or their components except that"—rule out any deployments other than those permitted by the two following paragraphs? Why then is that special provision necessary for the contingency of exotic systems? The answer seems to be that the Treaty's core limitations in Article III relate to a specific form of ABM technology. Thus there was a need to adapt the limitations of Article III to possible future systems using alternative technologies.

However, this raises the more general problem that Article II, Paragraph 1 defines ABM systems for the purpose of this Treaty as consisting of ABM interceptor missiles, ABM interceptor missile launchers, and ABM radars. Presumably, Article V … also refers to such systems. There would, therefore, appear to be no prohibitions against developing, testing, or deploying any system … that does not employ the canonical ABM triad. The only bar to such an interpretation consists of one word in Article II, Paragraph 1—"currently."

In a response in the same issue, Garthoff, who had drafted Agreed Statement D, denied that text was ambiguous with respect to systems based on "other physical principles":

Mr. Becker incorrectly interprets Article V as not applying to futuristic types of systems including components capable of substituting for ABM interceptor missiles, ABM launchers, or ABM radars. The reason for his erroneous interpretation is that he curiously assumes that "the only bar to such an interpretation consists of one word." The same could be said of each of the Ten Commandments. One word can indeed make a critical difference, and the word "currently" was deliberately inserted into a previously adopted text of Article II at the time agreement was reached on the future systems ban in order to have the very effect of closing a loophole to the ban on futures in both Articles III and V (and several others). The wording of the key introductory sentence of Article III was also agreed on at the time and for that purpose.

While admittedly the result has a Rube Goldberg air to it, the interlocking effects of the final wording of Articles II and III and Agreed Interpretation [D] was [sic ] intentionally devised and clearly understood—by both Delegations—to ban future "ABM systems based on other physical principles and including components capable of substituting for ABM interceptor missiles, ABM launchers, or ABM radars" unless specific limitations short of a ban were agreed on under the amendment procedures. The negotiating history fully supports the interpretation given by the Delegation, Mr. Rhinelander, and myself.^[20]   

In 1980 Harris advanced a view much like Becker's in discussions with Michael E. Pillsbury and David S. Sullivan, who were on the Reagan administration transition team responsible for ACDA and had hired him as a consultant to help prepare a study of purported Soviet armscontrol treaty violations.^[21]   

The issue surfaced again when Perle asked a DOD lawyer with no previous background in arms-control matters, Philip Kunsberg, to reexamine the text of the treaty in the light of a paper credited to an anonymous administration official (identified later as Bretton Sciaroni) and published April 14, 1985, by the Heritage Foundation. Kunsberg spent less than a week reviewing the text before producing a nineteen-page report in which he concluded that the prevailing interpretation was incorrect. He is reported to have drawn on the view expressed by Harris in a paper submitted at the request of T. K. Jones of the Office of the Secretary of Defense, who knew that Harris had been promoting an interpretation that would have allowed for space testing of the beam weapons being considered under SDI.^[22]    Perle then asked Kunsberg to review the classified record of the negotiations, and this review became the basis of a second report in which Kunsberg found that although U.S. negotiators had proposed a total ban on ABM systems of any kind, the Soviets had rejected the proposal. Under the terms of the treaty as negotiated, he argued, both sides were free to test technologies based on other physical principles than those incorporated into the ABM components specifically mentioned in the treaty (launchers, missiles, and radars). Kunsberg even went so far as to suggest that Agreed Statement D could be read to allow for the unilateral deployment of such technologies, in the event that the parties did not agree on what to do about them after the discussions called for in the statement.^[23]   

Kunsberg's two reports set off a kind of chain reaction. Perle, who said he "almost fell off the chair" when he saw the second report,^[24]    passed on the findings to the Special Arms Control Policy Group set up by the National Security Council to review SDI-related matters. When Secretary Shultz, a member of the group, heard the news, he asked the newly appointed legal advisor to the State Department, Abraham D. Sofaer, who had been a judge on the federal district court in New York, to review the findings. Sofaer assigned the task to several staff members, and their study essentially corroborated the new interpretation. Then Sofaer looked at the issue himself and discussed it with Nitze, the one member of the ABM negotiating team in the Reagan administration,

and reported to Shultz on October 3 that in his opinion the testing of technologies based on other physical principles was permissible under the treaty but that deployment would violate the language and aim of the agreement.^[25]    Sofaer's view was quickly adopted as that of the DOD, and the question of how the issue should be handled was discussed on October 4 by the interagency Special Arms Control Policy Group, chaired by McFarlane, and then with the president on October 11.^[26]    The issue assumed special urgency once McFarlane revealed the new interpretation during an appearance on NBC's "Meet the Press." Asked his opinion of McFarlane's remarks, Gerard C. Smith replied: "All that is permitted for space-based systems is research. That means in a laboratory."^[27]    McFarlane's response was unyielding: "It's a simple matter of reading the text. The fact of the matter is that when you are dealing with new principles—and by that we mean principles other than ground-to-air missiles—research other than pure research in the laboratory is allowed."^[28]   

McFarlane's comments aroused a storm of controversy. The chairman of the House Foreign Affairs Committee, Dante B. Fascell (D., Fla.), reacted angrily and promised to summon administration officials to explain the apparent decision to reinterpret the treaty. Rep. Norman D. Dicks (D., Wash.) said he and Sen. Albert Gore, Jr. (D., Tenn.), would amend the SDI authorization bill to assure that none of the funds could be spent on activities that would erode the treaty's ban on development of ABMs. Hearings were held later in October in both the House and Senate. Rhinelander testified that McFarlane's interpretation was "in effect a repudiation and abrogation of the Treaty."^[29]    Smith testified that the new view makes the treaty "a dead letter." The reinterpretation, he charged, was designed to evade the treaty's intention and express prohibitions: "It is one thing to have research programs as we now do for insurance against a Soviet technological breakthrough in defensive systems; it is quite another for an American president to announce a major quest to develop the technological basis for systems presently banned by international law."^[30]    Spurgeon Keeney, former deputy director of ACDA, said the administration's position meant that the United States was "no longer going to respect the treaty."^[31]    Urgent inquiries and warnings also reached the White House from allied capitals.

An attempt to deflect the criticism was made after the meeting with the president on October 11, at which Shultz was reported to have expressed strong reservations about the wisdom of adopting the new interpretation as official policy. He was scheduled to meet with representatives

of the other NATO countries on October 15, after addressing the North Atlantic Assembly in San Francisco. The day before that meeting, Shultz issued a statement indicating that a compromise had been reached within the administration. The new, or "broad," interpretation, he said, was "fully justified," but the question involved a "moot point" because "our S.D.I. research program has been structured and, as the President has reaffirmed on Friday, will continue to be conducted in accordance with a restrictive interpretation of the Treaty's obligations."^[32]   

Despite Shultz's effort to put the controversy to rest, it continued to simmer. In response to congressional demands, Sofaer made available under strict rules of access excerpts from the negotiating record and a new classified report he had prepared. On December 1, however, Sen. Carl Levin (D., Mich.) criticized the new interpretation as unjustified by the text or the negotiating record, and Senator Nunn, the leading Democratic spokesman on defense issues, publicly warned Reagan that by reinterpreting a treaty ratified by the Senate without consultation he was courting a "constitutional confrontation of profound dimensions."^[33]    On May 13, 1987, Sofaer released a declassified and revised report, quoting excerpts from the negotiating record. Six days later Senator Nunn delivered a detailed rebuttal of the report on the floor of the Senate.

In his declassified report, Sofaer marshaled the arguments and evidence in favor of the broad interpretation, which by now was being referred to within the administration as the "legally correct" interpretation. The report is quite revealing in offering a glimpse at the complexities and confusions of the negotiations, even though the case Sofaer makes has been persuasively rebutted.

Sofaer's Case For The "Broad" Interpretation

Sofaer's analysis begins by referring to generally accepted principles for interpreting treaties. Under the usual standards of international law, he notes, the meaning of treaties depends in the first instance on the treaty language itself. When the language is ambiguous, other considerations may be taken into account, notably the parties' understandings as reflected in subsequent agreements, the practical construction of the terms in subsequent conduct, and the record of the negotiating process. Sofaer argues that the review conducted by his office had confirmed his earlier finding that "the Treaty text is ambiguous, and that the negotiating record establishes that the Soviet Union had refused to agree to prohibit

the development and testing of mobile ABM devices based on OPP [other physical principles]."^[34]   

Sofaer's case is based on the interpretation of several provisions in the treaty text—the definition of "ABM system" in Article II(1) of the treaty; the prohibition of development, testing, and deployment of nonfixed land-based systems and components in Article V(1); and Agreed Understanding D, which deals with future systems based on "other physical principles." He contends that there are three possible interpretations of the testing issue as it is dealt with in these provisions. The "restrictive" interpretation holds that the definition in Article II refers to all ABM systems and components, whether they were in existence when the treaty was negotiated or are substitutes introduced later. In this view, the only new systems that can be developed (in the sense that they are brought to the stage of field testing, and not merely worked on in preliminary phases of advanced research) are those that are designed to be used in fixed, land-based systems. Agreed Statement D is only a clarification that establishes no exception for exotic technologies. The "broad" interpretation holds that the language in the text refers only to systems and components using then available technology, and that Agreed Statement D implicitly permits the development of new technologies that could substitute for existing ABM systems by referring to ABM systems and components "created in the future." Deployment of such weapons is prohibited pending consultation and agreement between the parties, and might require amendment. A third even broader interpretation adds that deployment is permissible, according to Agreed Statement D, if the parties discuss the question but cannot agree on deployment restrictions.

These are the relevant passages Sofaer cites:

Article II

Article V

Agreed Statement D

In order to insure fulfillment of the obligation not to deploy ABM systems and their components except as provided in Article III of the Treaty, the Parties agree that in the event ABM systems based on other physical principles and including components capable of substituting for ABM interceptor missiles, ABM launchers, or ABM radars are created in the future, specific limitations on such systems and their components would be subject to discussion in accordance with Article XIII and agreement in accordance with Article XIV of the Treaty.

Sofaer dismisses the third, or "broadest," interpretation as very doubtful in view of the reference in Agreed Statement D to the obligation not to deploy ABM systems except as provided for in the treaty and in view of the fact that the treaty is of unlimited duration. If Agreed Statement D were read to permit the parties to proceed with the deployment of ABMs based on other physical principles, the essential stated aim of the treaty in barring any deployments other than those expressly allowed would be undermined. Although the statement does not explicitly ban such deployments but only calls for discussion prior to a decision to deploy, the fact that the discussion is to be held in accordance with the provision of articles XIII and XIV is significant, he points out. Because those provisions call for periodic review, clarification of ambiguities, and amendment in the light of new developments, the intention could not have been to give the parties a way of circumventing the prohibition on deployment. But Sofaer also argues that the "restrictive" interpretation, though plausible and acknowledged in some officially adopted U.S. statements, is not the most persuasive interpretation of the treaty language or a reflection of what the negotiating record reveals about the parties' understanding of the provisions.^[35]   

The negotiating record, as revealed selectively in his study, suggests that the U.S. delegation made a considerable effort to win Soviet agreement to language that would specifically and unambiguously ban both the testing and deployment of all possible "devices" that could be used as ABM systems or components, except for those components that might substitute for existing fixed, ground-based ABM components or systems. The record also indicates that the Soviet delegates balked at such all-embracing language, ostensibly on the ground that there was no sense in banning nonexistent weapons and that new developments could best be dealt with in the SCC, in its role of monitoring compliance. Acting on the basis of National Security Decision Memorandum (NSDM) 117 (July 2, 1971), the U.S. delegation was instructed to propose that the agreement should ban "any systems" capable of countering ballistic

missiles unless allowed by the treaty—though in the same month it received another instruction to hold off, pending an internal review.

The Soviet position, by Sofaer's account, was ambiguous. For example, on September 15, 1971, Soviet delegate Viktor Karpov stated that the term "ABM systems and components" in the Soviet draft was intended to cover "future components." Sofaer argues that this does not indicate Soviet recognition that testing of systems based on other physical principles was to be covered by this understanding. More likely, he argues, it referred to new components designed to substitute for existing ones relying on conventionally applied physical principles. If the Soviet understanding had been otherwise, he contends, the U.S. delegation would not have felt a continuing need to press the Soviets for more express language covering future systems and technologies. Between that date and December the Soviets rejected definitions that would have unambiguously included unknown devices based on other physical principles. Finally, however, an agreement was struck during the period between December 1971 and February 1972. The language of Article II was changed (to include the phrase "currently consisting of") and the Agreed Statement was adopted to address future technologies. Contrary to the views of all the leading negotiators except Nitze, however, Sofaer interprets the Soviet assent to the revision and the Agreed Statement as a refusal to treat new technologies in the same restrictive way as existing ones were treated. Presumably, the Soviet delegation (perhaps in response to promptings from the military) wished the treaty language to permit testing of exotic systems. Why else would they agree that such systems might be "created," to use the language of the Agreed Statement—a term clearly meaning development and testing outside a laboratory?^[36]   

According to Sofaer, the record indicates that at first the decision makers in Washington and the negotiating team were divided on the issue. Some negotiators wanted a total ban on ABM systems. These included the lead negotiator, Gerard Smith. Gen. Royal Allison and Amb. J. Graham Parsons doubted that such a ban was practical because of verification difficulties. On July 20, 1971, President Nixon instructed the delegation, in NSDM 120, not to agree to prohibit deployment of future systems other than those currently in question. In a back-channel cable to National Security Advisor Henry Kissinger, Smith expressed his belief that the United States should try to obtain a ban on future devices, but Kissinger was impatient at the lack of progress over the main issues of deployment owing to controversies over "a plethora of esoteric issues." Smith was sensitive to Kissinger's wish for more rapid progress,

but asked that the decision makers provide him with necessary guidance. Did they want to limit all ABM systems or only those that were then known and in use? Smith queried Kissinger on August 7, 1971: "Do we seek an ABM constraint to provide greater stability by … maintenance of retaliatory capability, halting a buildup of defensive systems that could threaten that capability and lessening pressures for buildup of offensive systems—or just a temporary truce in ABMs—until such time as more effective futuristic ABMs are developed and deployed?"^[37]   

In Washington, the Office of the Secretary of Defense wanted to prohibit only the deployment of future systems; indeed, the Joint Chiefs of Staff opposed even this policy, fearing it would be hard to obtain support for R&D if deployment were prohibited. As the JCS chairman, Adm. Elmo R. Zumwalt, Jr., wrote:

The Joint Chiefs of Staff strongly recommend that futuristic ABM systems not be banned. If deployment of futuristic systems were to be banned but research and development permitted, as advocated by some government agencies, such an approach would make it extremely difficult to get funds for such research and development (R&D) and could lead to unilateral U.S. neglect of the field. In all likelihood, the Soviets would proceed with R&D on such systems. It seems imprudent to foreclose options on future systems that cannot now be defined or envisioned or which may be unverifiable, particularly in view of the numerical super[iority] of Soviet offensive missiles involved in the proposed interim strategic offensive agreement. Under any circumstances, R&D programs must be kept viable to avoid technological surprise.^[38]   

Kissinger was annoyed by this effort to clarify the status of "exotic" devices. In a meeting of principals of the NSC Verification Panel, he argued that the negotiation was becoming stalled because the U.S. delegation was raising "nuances" and "academic" issues. He urged that the issue of seeking a ban on exotics be deferred, given that it would not become a problem until at least the 1980s and could be handled by other existing provisions for review and withdrawal. In August, in response to another request for guidance from Smith, Kissinger cabled back that although the United States had made clear to the Soviets its desire for a comprehensive agreement, it was important to take advantage of Soviet willingness to work out even a partial agreement, and that Soviet reluctance to deal with all possible issues should not be a deterrent: "In matters affecting so directly their vital interests, it is understandable that the Soviet leaders have preferred to move to an initial agreement of limited scope."^[39]    Kissinger's statement seemed to imply

that the delegation need not be concerned about such matters as the "exotics" issue, which might be dealt with at a later stage. Nevertheless, the next day, August 12, 1971, an instruction was issued to the U.S. negotiators calling for inclusion of potential new weapons in the terms of the agreement:

The agreement should contain a provision whereby neither side shall deploy ABM systems using devices other than ABM interceptor missiles, ABM launchers, or ABM radars to perform the function of these components. (This provision along with that in the next paragraph, should not prohibit the development and testing of future ABM components in a fixed, land-based mode.) …

… The agreement should contain a provision whereby neither party shall develop, produce, test, or deploy: (a) sea-based, air-based, space-based, or mobile land-based ABM launchers, ABM missiles, or ABM radars; (b) ABM components other than ABM interceptor missiles, ABM launchers, or ABM radars to perform the functions of these components.

But, the instruction continued, the delegation should not "invite a detailed negotiation or discussion of future ABM systems. Our objective is to reach agreement on the broad principle that the agreement should not be interpreted in such a way that either side could circumvent its provisions through future ABM systems or components."^[40]   

Pursuant to this instruction, the U.S. delegation proposed two new sections for inclusion in the draft then under consideration. One provided that "each party undertakes not to deploy ABM systems using devices other than ABM interceptor missiles, ABM launchers, or ABM radars to perform the functions of these components." The other added to the provisions banning testing and deployment a specific reference to "other devices to perform the functions of these components." Smith argued that the U.S. draft proposed a more complete prohibition, which would be in the interest of both parties specifically because it makes clear that the obligations in the relevant provision apply "also to possible future types of devices capable of performing the functions of these components."^[41]   

The Soviets rejected the U.S. wording, arguing that it would add an unnecessary element of vagueness to the treaty and that existing provisions for review and amendment would be sufficient. Karpov was reported to have expressed the belief that "it was wrong to limit means not known to anyone." The representative of the Soviet Ministry of Defense, Lt. Gen. Konstantin A. Trusov, objected that the United States was proposing limitations on "conjectural" systems.^[42]   

Smith replied firmly in support of the U.S. position that if the proposed changes were not adopted, "it would be a cruel illusion to the people of both nations to say that we had concluded an agreement on ABM systems. We should more properly say that there had been an agreement to limit ABM launchers, interceptors and radars." He "had a higher regard for Soviet weapon designers than to believe that they are content with ABM technology which dates back to the early 1950s."^[43]   

In Sofaer's view, then, the Soviet negotiators never abandoned their persistent opposition to a total ban on the development of future "devices." In accepting Agreed Statement D, they merely agreed that if such devices were developed (or "created") in the future, the intent of the parties to the Treaty was that such devices not be deployed before discussions between them, and that such discussions should be based on the express aim of the treaty to prevent the deployment of territorial defenses. This provision—coupled with the limiting language of Article II, which referred to ABM systems "currently consisting of" launchers, radars, and missiles—is therefore best understood not as a restriction on the testing of space-based systems incorporating new physical principles, but, on the contrary, as acknowledging that such testing would not be subject to the restrictions imposed on conventional technologies.

The Rebuttal Of The "New Interpretation"

A number of compelling, and in our view persuasive, responses have been made to Sofaer's reinterpretation.^[44]    Raymond L. Garthoff has responded in meticulous detail. The principal negotiator of the key changes made at the end of the negotiations on the "exotics" issue, Garthoff makes a particularly significant point by noting that, in developing the case for the reinterpretation, all of its advocates ignored the arguments advanced against it within the administration itself. In deliberately ignoring or bypassing opposition from responsible officials, the style of policy-making was, as Garthoff notes, disturbingly similar to that which produced the Reagan administration's Iran-contra fiasco. When Harris was invited by T. K. Jones to submit a study of the treaty testing provisions, two former negotiators, Col. Charles L. Fitzgerald and Sidney Graybeal, were also asked to submit a report. Their study, which reaffirmed the traditional interpretation, was not brought to Sofaer's attention when he was asked to do his original review. Earlier in 1985, moreover, the Office of the General Counsel of the Defense Department

had reexamined the validity of the traditional interpretation and had also reaffirmed it. Even at the time Kunsberg's study was under review, the assistant general counsel of the Defense Department, John H. McNeill, "had written a memorandum rejecting it,"^[45]    a fact apparently not mentioned by Perle and Iklé at the October 11 meeting with the president. The reinterpretation was also rejected by the legal counsel of ACDA, Thomas Graham, Jr., and by one of Sofaer's own researchers, William J. Sims III, who subsequently moved to ACDA and submitted a rebuttal to Sofaer's 1986 classified report to Congress, resigning from his government position at the same time.^[46]   

The case against the reinterpretation rests essentially on several contentions. They are, first, that the text of the treaty supports the traditional interpretation; second, that the negotiators on both sides understood and explained the terms in accordance with the traditional interpretation both during and after the negotiations; third, that the terms were explained to the Senate in accordance with this interpretation at the time the treaty was ratified, and they were so understood by the senators who voted for and against; fourth, that subsequent official statements affirmed the traditional interpretation; and fifth, that the adoption of the so-called broad interpretation would vitiate the intent of the treaty and invite its violation, in express contradiction of the understanding in the treaty that it was to be permanently binding.

Of these, the single most important consideration is, as Sofaer himself points out, the treaty language itself, inasmuch as it expresses the intentions and understandings of the parties. Sofaer makes much of the fact that the phrase "currently consisting of," inserted into Article II, is used to describe the systems and components expressly regulated by the treaty. But as the critics of the reinterpretation point out, this language can more plausibly be interpreted as illustrative rather than limiting,^[47]    i.e., that the treaty bans the development and deployment of any and all systems and components that perform the same function as systems "currently consisting of" radars, launchers, and missiles. Garthoff, who was responsible for proposing the change, certainly so understood it.^[48]    Had the adverb "currently" not been inserted, the parties might have been free to interpret the language as Sofaer does—to permit testing of systems or components that did not fall under the designation of radars, launchers, and missiles. Agreed Statement D thus merely clarifies and reinforces the treaty's intent to permit development (or "creation") of ground-based ABM systems and components based on new physical principles. The provision opens the possibility of deployment after discussion

insofar as the parties agree that such deployment could be made without endangering the treaty's central aim of preventing the deployment of territorial defenses. As Abram Chayes and Antonia Chayes have pointed out, to interpret an agreed understanding not as a clarification of treaty provisions but as a modification that significantly alters the scope and bearing of the text is to put more weight on such a statement than is warranted by its character and to contend, in effect, that the drafters were deliberately opening a loophole by which the parties could escape the intended limitations of the treaty—"making an absurdity out of the text of the treaty proper."^[49]   

As to the negotiating record, Garthoff points out that the U.S. delegation dropped its effort to get the Soviets to agree to ban the development of all conceivable ABM "devices." It was too broad a term, they decided, one that "could constrain and interfere with testing and development of multipurpose technologies." The delegation decided it would be preferable that the treaty "should refer only to ABM systems and components, including future systems or components based on other physical principles and capable of performing the functions of the then-current (1972) types of components of antiballistic missile systems." When the Soviets submitted a reformulation of U.S. proposals on future technology, the Soviet representative, Karpov, indicated that the new draft "took into account the wishes of the U.S. side" by using the word "components" to cover any and all ABM components. Graybeal, Garthoff notes, was asked to make sure that the new formulation was understood to cover future components using other physical principles. The negotiating record indicates that Karpov agreed with Graybeal's interpretation. Another Soviet negotiator subsequently agreed that the prohibition on various ABM systems other than those that were fixed and ground-based covered "the problem of future systems," contrary to Sofaer's reading of the Soviet understanding, which would have restricted the language of the treaty to future components based on known technologies and not to "unknown devices."^[50]   

Both subsequent statements and practice on both sides support Garthoff's interpretation. The reliance on the traditional interpretation in the annual ACDA arms control impact statements has already been noted. Although Reagan administration officials have sometimes construed a remark of Marshal Andrey Grechko (then the Soviet minister of defense) before the Presidium of the Supreme Soviet to mean that the Soviets themselves understood the treaty to allow space testing, Grechko's comment does not support such an interpretation. The treaty, he

said, "does not place any limitations on the conduct of research and experimental work directed toward solution of the problems of defense of the country against nuclear-missile strikes."^[51]    Grechko made no mention of testing in space and there is no reason to assume that what he had in mind by "experimental work" was anything other than what is permitted under the terms of the treaty. So far as is known, Soviet experimentation on ballistic missile defense has been configured in accordance with the traditional interpretation. After SDI was announced, Soviet officials referred to the traditional interpretation in criticizing the initiative as a treaty violation. On the U.S. side, it is noteworthy that six former secretaries of defense (Melvin Laird, Harold Brown, Clark Clifford, Robert S. McNamara, Elliot Richardson, and James Schlesinger), three of them Democrats, three of them Republicans, have joined in urging that the United States adhere to the traditional, restrictive interpretation.

The Treaty And SDI Experiments

As we have seen, some of the experiments planned for the SDI program would require revision or abrogation of the treaty. In the opinion of critics, "many of the demonstrations planned under the Strategic Defense Initiative … are almost certainly inconsistent with the terms of the Treaty."^[52]    Critics admit that some relevant considerations were not fully resolved in the negotiations. Despite numerous exchanges, the parties did not reach an agreed interpretation of "develop." The United States proposed that development be defined as the phase that follows research, understanding research to include conceptual design and laboratory testing, which precedes field testing. Development was understood to be associated with construction and testing of prototypes of a weapons system or its major components. The Soviets argued that the line dividing research from development should be placed where national technical means could identify specific systems as ABM-related.^[53]    "Testing in an ABM mode" is also not defined. The United States issued a unilateral statement explaining that, in its view, the term means testing against an RV with the flight trajectory of an ICBM, but the Soviets never agreed to the definition. Nevertheless, in the SCC in 1978 agreement was reached on criteria for defining testing in an ABM mode. (As previously noted, exact terms remain classified, but it appears that this agreement was similar to the United States' unilateral statement.) Similarly, because the two sides could find no easy formal definition for distinguishing

air defense from ballistic missile defenses, this issue was left unresolved. The issue of the difference between a "component" and an "adjunct" also arose in the negotiations. It was agreed that a telescope used to calibrate a radar would not have meaningful ABM capabilities. But, again, no precise distinction was drawn between a component and an adjunct, no authoritative U.S. statement was issued, and no agreed definition was subsequently reached in the SCC.

There is disagreement between those in charge of the program and its critics as to whether current and proposed experiments conform to the treaty. The SDIO director, Lieutenant General Abrahamson, admitted in 1985 that "at some point … we would have to depart from the treaty." He thought that would be necessary "about the turn of the next decade."^[54]    Critics have argued that already the experiment to test an airborne optical system (AOS—formerly named "Adjunct") is inconsistent with the treaty. Administration representatives argue that a device is not a component if it does not perform the function of the component on a "stand alone" basis, or, in other words, if the technology is not sufficiently mature to be incorporated into a workable ABM system. Thus, development and testing of assemblies and subassemblies that do not constitute full ABM components are permissible. The AOS is a case in point. It is mainly an adjunct to a terminal imaging radar. To be considered a component, it would have to perform all the functions of sensing and battle management. The critics argue that the history of negotiations suggests that sensors need not perform the full spectrum of ABM battle-management functions in order to be considered components. Most missile-defense systems have more than one sensor component. The early Nike-Zeus had, not one or two, but four separate types of radars—for target acquisition, decoy discrimination, target tracking, and interceptor tracking. The AOS, they point out, performs a role similar to that of the perimeter acquisition radar (PAR) in Sentinel/Safeguard; these radars were clearly considered components and therefore are subject to strict limitations. The United States has officially contended that the Soviet Henhouse radars counted as components even though they played less of a role in the ABM system than did the PAR.^[55]   

With current verification capabilities, it is often impossible to distinguish between a component and an adjunct. Detailed knowledge of performance capabilities would be required—knowledge that cannot be ascertained by national technical means. It might be important to know that a large infrared telescope was being tested in the AOS in conjunction with strategic missile testing; the aperture of the telescope would

also be useful information. National technical means cannot tell if the sensor hardware and computer software can perform an ABM mission. Meanwhile, development would be completed and the treaty circumvented. Current redesign of the AOS experiment has enhanced it so that it will include a laser range-finder and on-board computers, enabling the system to do virtually the entire range of ABM battle management.

By general admission, there are gray areas in the treaty—notably in the case of ATBMs and ASATs, but even in that of large phased-array radars (LPARs). Both the United States and U.S.S.R. "are deploying air defense systems that with some modifications could be upgraded to some anti-tactical missile capability" (the U.S. Patriot and improved Hawk and Soviet SA-10 and SA-X-12). ATBMs might be effective against "SLBMs which fly at a similar speed, flight trajectory, and reentry angle to MRBMs and IRMs,"^[56]    and in the SALT treaties, SLBMs are considered strategic missiles. The ABM Treaty does not prohibit development and testing of ASATs, but the technologies involved can also be applied in the development of space-based defenses. LPARs serve many functions. Distinguishing LPARs developed for one of these functions from those designed to be used in an ABM system is very difficult. Modernization of radars (involving replacement of old radars with PARs) might be inconsistent with the treaty, especially if this entails installations outside national territory. The United States' declared intention of modernizing the BMEWs facilities falls in this category. Because it is not clear whether the issue of modernizing a "grandfathered" early-warning system (i.e., previously constructed for a different purpose and ignored by the treaty) was settled during the negotiations, it is also unclear whether the BMEWs installed in Thule, Greenland, and Fylingdales, England, can be modernized without violating Article VII of the treaty.

The Boost Phase Surveillance and Tracking System (BSTS) experiments incorporate greatly enhanced infrared sensors that provide high resolution and precision for tracking missiles in their boost phase. But because MIRVed warheads are released, the system cannot give improved impact prediction or attack characterization unless it can also track the bus. BSTS could be a useful part of a layered ABM system, however. But some SDI opponents question these experiments because Article VI bans development and testing of space-based ABM systems or components. The technology, however, is not intrinsically ABM-related because the same technology could be used for early warning, long recognized as a legitimate goal for spaced-based technology, and thought by many also to have been "grandfathered into" the treaty.

The proposed Space Surveillance and Tracking System (SSTS) will use cryogenically cooled infrared sensors for midcourse detection and tracking. Initial versions are to be tested in space in the early 1990s. This system, too, would provide information that would be relayed to an interceptor. It may not be tested in an ABM mode; if tested against satellite targets, however, it might acquire an ABM capability prohibited in Article VI(a), which pledges the parties "not to give … radars … capabilities to counter strategic ballistic missiles."

Even terminal-imaging radars would violate the treaty if they are designed for deployment in a mobile mode to make them more survivable. Such a development would violate the provision in Article V(i) banning the testing and development of mobile components. The testing of space-based imaging radars and lasers would "raise questions" about compliance, because these devices could arguably perform exactly the functions "currently" performed by ground-based radars.^[57]    The successful Homing Overlay Experiment (HOE) in June 1984 also raised questions because the payload was carried aboard a modified Minuteman IICBM. Article VI(a) stipulates agreement "not to give missiles … other than ABM interceptor missiles … capabilities to counter ballistic missiles or their elements in flight trajectory, and to test them in an ABM mode." In January 1984 the U.S.S.R. protested the use of Minuteman to test HOE in an aide memoir listing several dozen alleged U.S. violations. ACDA replied: "The test missile was observably different from Minuteman I, as were its performance characteristics. In any case, the Minuteman I is no longer deployable by the U.S." HOE experiments have used a single-intercept vehicle, but the program was originally designed to investigate the use of multiple-kill vehicles. The ACDA Bulletin noted that "the U.S. is not developing ABM interceptors with multiple warheads." Once ERIS is developed, however, pressure to deploy multiple warheads, which can significantly enhance kill prospects against a proliferated offense, will certainly be felt.^[58]   

Other experiments are also questionable. The supervelocity or advanced kinetic-energy launcher using an electromagnetic railgun could be considered a violation of the treaty's provision banning "rapid reload" because of its high rate of fire. Tests of the space-based kinetic-kill vehicle scheduled for early 1990s would be in blatant contradiction of Article VI(a), which forbids the development and testing of space-based ABM systems and components. The effort to develop a space-based laser via Talon Gold, LODE, and Alpha could also be inconsistent with the treaty. The Talon Gold project aims to attach a telescope to the laser

to test pointing and tracking technologies. It was supposed to have been tested on the space shuttle in 1987 and 1988 but has been delayed both by the Challenger disaster and to allow for the addition of a second telescope. In 1984 congressional budget cuts reduced the scope of the study. Now, a new, more capable system is to be developed. Some officials argue that it is only a "generic" experiment, one that is not capable of substituting for an ABM component. This, however, is its purpose, and it will be ABM-capable. The LODE project (Large Optical Demonstration Experiment) would raise questions by putting mirrors in space that could be considered components of a BMD system. The Alpha project to develop a hydrogen fluoride chemical laser does not violate the treaty so long as experiments are made in a laboratory, but field testing may constitute a treaty violation inasmuch as the ultimate objective is to mount the lasers on space-based satellites for use in an ABM mode. The same would be true for experiments with neutral-particle-beam weapons to be deployed in space.^[59]   

The SDIO view is that because there was no agreed definition of some elements of the treaty, such as the definition of a component, it is necessary to infer specific standards. The DOD has adopted four working principles:^[60]   

Under the DOD directive 5100.70 of January 9, 1973, the director of defense research and planning (DDRE) ensures that all military research programs comply with the treaty, on the advice of the Defense Department's general counsel. DOD Instruction S-5100.72 establishes general guidelines, and questions of interpretation are considered on a case-by-case basis. The DDRE undertakes to consult with general counsel, the office of International Security Policy in the Office of the Assistant Secretary of Defense, and the office of the Joint Chiefs. Military departments

and agencies must certify compliance quarterly and, moreover, establish internal procedures to monitor and ensure it.

The DOD divides SDI research into several categories. The first, conceptual design or laboratory testing, obviously precedes field testing and is not verifiable by national technical means. It is therefore not subject to treaty limits. The second, field testing of devices that are neither ABM components or prototypes, is also said to be permissible. The DOD holds that for a device to be limited by the treaty, "it must constitute an ABM system or component (an ABM interceptor missile, ABM launcher, or ABM radar) or be capable of substituting for such an ABM component." In the third category are experiments involving field testing of fixed land-based ABM components. These must take place at "current" or "additionally agreed" test ranges. For the United States the authorized sites are the Kwajalein Missile Range (on an atoll in the Marshalls chain) and the White Sands Missile Range in New Mexico.^[61]    According to the DOD, most technology research projects for SDI fall in either category one or two; none falls in category three. Many are "under roof" experiments "using devices incapable of achieving ABM performance levels." Although some are designed to be based in space, the space environment is simulated in a laboratory.^[62]   

Laser systems currently under development are being designed to fit within the DOD guidelines. In the case of the Mid-Infrared Advanced Chemical Laser (MIRACL), or of the Sea Lite Beam Detector subsystems from a former navy program, the effort is to determine in a ground experiment whether a laser and beam director can be successfully integrated. Whether separately or combined, however, these cannot substitute for an ABM component because "the power, optics, and laser frequency are not compatible with atmospheric propagation at ranges useful for ABM applications."^[63]    The experiments are planned against ground-based static targets; another experiment is being planned in which a moving target will be launched from a point close to the MIRACL. Because MIRACL is ground-based, it would comply with the treaty. Other experiments involving acquisition, tracking, and pointing will demonstrate technologies but will not be capable of achieving ABM performance levels. For the time being, the free-electron laser will not have enough power to serve as an ABM weapon. But even if at a later stage it does obtain this power, the ground-based FEL will not violate the treaty, although it may have to be tested at an acknowledged test range. The High Brightness Relay (HIBREL) project involves experiments

to demonstrate the feasibility of relay mirrors in space for beams from ground-based lasers. The mirrors will be able to handle light only from low-powered lasers and will use technologies that form just a part of what would ultimately be required for an ABM weapon; they, too, will not be able to achieve ABM performance levels. The neutral-particle-beam experiment will be tested in space to determine the technologies needed for midcourse discrimination or to detect nuclear material. The experiment will be conducted using nearby co-orbital targets at low average power; it will not be capable of "autonomously acquiring or tracking ballistic targets" and will therefore not have ABM capabilities, nor will the test be conducted in an ABM mode.^[64]   

SDIO also defends other experiments as treaty-compliant. The BSTS experiment will be designed to "determine if sufficiently sensitive tracking and signature data can be collected on-orbit against the earth's background." The BSTS will have only enough capacity to function as an early-warning system. It will collect ballistic-missile-plume data but will lack the real-time data-processing capacity needed to pass the data to a boost-phase interceptor. (Other capabilities will also be limited—by nature and resources.) The AOS experiment will test the feasibility of using an airborne infrared sensor to support a ground-based radar; it will involve only a single, passive sensor, and because of sensor and platform limitations it cannot substitute for an ABM component. The SBKKV project is designed to enable space-based kinetic interception by chemically propelled rockets, but the demonstration hardware for any space-based experiment will not be an ABM component, cannot substitute for an ABM component, and will not be tested in an ABM mode. (There will be no intercepts of strategic ballistic missiles or their elements in flight-trajectory in space, but intercepts of certain orbital targets simulating anti-satellite weapons can be compatible.) Otherwise, the planned experiments are said to raise no questions of compliance. The ERIS development concerns a ground-based ABM missile which will be provided with a single nonnuclear warhead; its terminal-imaging radar will be housed in a fixed site.^[65]   

But it is not clear that the development and testing in space of the space-based kinetic-kill vehicle system, now designated the space-based interceptor (SBI), would be allowable. It would use technology known at the time the treaty was adopted and incorporated into existing ground-based ABMs. Development and testing of this technology for purposes of a space-based ABM could therefore fall under the prohibitions in the provision on testing and development. Because the Reagan administration

claimed an exception to the testing provisions only for technologies based on "other physical principles"—other, that is, than those incorporated into ABM systems when the treaty was signed—it would seem that SBI could not legally be tested, even under the "broad" interpretation. The only argument in favor of testing is that the adaptation of the technology for space-based purposes represents the use of "other physical principles." In 1987 and 1988 the Reagan administration was reported to be considering whether to adopt the view that while the ERIS ABM missile is based on traditional technology, the SBI is a "future" system, even though it relies on previously known kinetic principles. Plans have been made to conduct two suborbital flight tests of the SBI against a ballistic-missile target at Kwajalein Atoll between 1991 and 1992, and possibly earlier. Because these tests are suborbital, the SDIO asserts that they conform to the traditional interpretation of the treaty. The issue of whether or not these tests are legal has reportedly divided the experts in the administration, those in the Defense Department leaning toward approval, those in the State Department and ACDA toward opposition. Other critics have charged—correctly in our view—that all such development and testing are clearly prohibited.^[66]   

The Disagreement At Reykjavík

Despite these efforts to argue that SDI research conforms with a restrictive interpretation of the ABM Treaty, the issue of what research is permitted under the treaty became the sticking point of the negotiations in Reykjavík. What happened at Reykjavík is in some dispute, but it appears that the two sides came close to a much broader agreement on arms reduction than had been anticipated by the United States. At most, U.S. negotiators had hoped to reach agreement with the Soviets on the removal of longer-range intermediate weapons from the European theater. When Gorbachev proposed a more sweeping agreement—which would have included the halving of strategic weapons in five years—followed by a total elimination of strategic offensive weapons, Reagan's response was encouraging (although the president appears to have proposed the elimination of strategic ballistic missiles rather than of all strategic weapons). The Soviet leader's proposal was associated with three others. One was that all longer-range intermediate weapons—the Soviet SS-20s and the U.S. Pershing Is and IIs and ground-launched cruise missiles—be removed from the European theater. Both sides would be free to retain up to one hundred of such launchers, but the

Soviets would keep theirs beyond the Urals and the United States would house its stock in the continental U.S. The second proposal was that both sides reaffirm the ABM Treaty by making a commitment to refrain from deploying defensive systems beyond what the treaty allowed for at least fifteen more years. During this time research on new systems might continue, but development and testing of space-based systems and components would be confined to the laboratory. The U.S. side, which had offered a seven-and-a-half-year commitment, compromised on ten years, which was acceptable to the Soviets, but would not accept the Soviet proposal for restrictions on development and testing. The third Soviet proposal was that all nuclear testing be banned, which the United States sought to defer pending further investigation of problems of verification.

The U.S. negotiators responded with two successive alternatives. The first committed each side "to confine itself to research, development and testing, which is permitted by the ABM treaty" for five years, during which time "strategic nuclear arsenals" would be reduced by 50 percent. Provided further reductions of offensive weapons took place, toward a goal of eliminating "all offensive ballistic missiles" by the end of a ten-year period, the same restrictions would hold. At the end of this ten-year period either side would be free to deploy additional defenses. When the Soviets declined this proposal, the United States tabled a second, pledging both sides not to exercise their right to withdraw from the treaty for ten years—adding, however, that the treaty "is of unlimited duration" (a change designed to meet the Soviet objection that the first draft would have changed the duration of the treaty to ten years). In addition, the draft added a comma after the word "testing," thus obviating the possible implication in the first draft that only the research, development, and testing specified in the treaty and further defined in the agreed statements and SCC clarifications would be permitted. In effect, the second draft left the United States free to adopt the "broad" interpretation of the treaty. In the event, however, this draft was also rejected by the Soviets, prompting Reagan to gather up his papers and end the negotiations.

The Soviet side was accused by Shultz and later by the president himself of trying to "kill" SDI by requiring that all development and testing be confined to the "four walls" of a laboratory—a point underscored by Foreign Minister Eduard Shevardnadze when he used his hands to describe the walls of a laboratory. In a television address from Moscow, Gorbachev repeated his assertion that the ABM Treaty banned development and testing of space defenses except in the laboratory, but he phrased it in ways that seemed somewhat less restrictive than U.S. representations.

"We firmly stated," he observed, "that it was necessary to comply with the ABM Treaty of 1972 without any time limit. Moreover, in order to strengthen the regime of this Treaty, we proposed … mutual commitments … not to use our right to withdraw from the treaty for at least 10 years, until we finally eliminated the nuclear weapons." He said that the Soviet emphasis on the restrictions was intended to "strengthen the treaty," which seemed to bear out the U.S. contention that he was trying to add to or amend the treaty, but he made clear that the intention was to prohibit the testing of "all space elements of antiballistic missile defense" except in laboratories. "We did not demand or require," Gorbachev insisted, "that he stop work in this area, but with the understanding that all provisions of the ABM treaty be observed." Several times Gorbachev repeated that the terms of the treaty required that development and testing be confined to the laboratory. In one formulation, however, he introduced a possibly significant specification by observing that "testing of all space elements of anti-ballistic missile defense in space [is] prohibited except research and testing in laboratories."^[67]   

By introducing the term "space elements," the Soviets may have been deliberately attempting to broaden the scope of the treaty so as to eliminate any distinction between components and adjuncts, or they may simply have been intending to use a synonym for components. Conceivably, they may also have been signaling a willingness to discuss which types of experiments might be allowable on grounds that they are not elements of anti-ballistic-missile defense, or not exclusively that. If, for example, the U.S. side had asked if it was permissible to test improved sensors for early warning, even though such sensors could also be used in conjunction with ballistic-missile defense, it is conceivable that the Soviets would have agreed that such development could continue. That latitude would also have to apply, it appears, to systems designed to develop ASATs or to test ATBMs, the two acknowledged loopholes in the treaty.

Strangely, the U.S. delegation did not request clarification of the Soviet position. This may have been a result of exhaustion, for the issue came up at the end of an unexpectedly detailed and comprehensive set of negotiations. Conceivably, however, it may have reflected the United States' unwillingness to enter into any sort of compromise that would have restricted its testing and development. Perhaps the president and his advisors thought that, especially in the second five-year period, testing would have to go beyond the restrictive interpretation, which would certainly have been disallowed under the Soviet interpretation.

It is possible, but unlikely, that this was simply a missed opportunity,

a failure on the part of both sides to recognize that they were not communicating their actual views. The Soviet position did not necessarily mean that the United States would be prevented from undertaking virtually all of the experiments on tracking, pointing, and acquisition planned for the SDI (because some of these could be justified as efforts to improve early warning) or for many other experiments that could be defined and configured as efforts to develop ASATs and to improve tactical air defense and anti-ballistic-missile defense. Because the Soviets were ready to drop their proposal for a comprehensive ban on nuclear testing, testing to develop the X-ray laser could also have continued as long as it did not involve the testing of the system in space. Even if the United States had accepted the Soviet formulation, much—arguably most—of what was planned for the SDI over the next ten years could apparently have been done in compliance with the treaty. It is also not clear why the Soviets introduced new language in speaking of "space elements." To find out what the Soviets intended, the United States could have proposed acceptance of the restrictive definition or, in other words, agreed that the treaty prohibits development and testing of ABM systems and components other than those that are ground-based. Given that the Soviets have an active program to develop ground-based lasers, they would presumably not object to such an interpretation and, indeed, may well have focused attention on space-based elements in order to distinguish them from the ground-based versions of new technologies they are developing.

Although it is intriguing to speculate about the possibility that the dispute might have been resolved if both sides had sought clarification, it would probably be a mistake to suppose that nothing more was at fault than fatigue and miscommunication. The same issue had been discussed in detail at the Geneva summit, but no final agreement could be achieved. The U.S. delegation went to Reykjavík without any expectation of an imminent breakthrough. The failure to reach an agreement was probably caused by more fundamental factors, not the least of which is that many of the Soviet proposals had not been anticipated, and thus not thoroughly studied, by any authoritative arm of the administration.

The president's fundamental concern centered on perceptions. He did not want to be seen as trading away SDI in return for an agreement on the reduction of offensive weapons. He did not want to disappoint his domestic supporters, who had come to see SDI as a unilateral U.S. move, based on superior technological ability, to ensure against any

possibility of a Soviet preemptive strike and who would not have trusted the Soviets to adhere to any arms-control agreement—even one negotiated by Reagan. The president himself had come to see SDI as his legacy to the nation's defense and as an essential step away from the threat of nuclear catastrophe. He tried to persuade the Soviet leader at Geneva that strategic defenses would not pose an offensive threat and reiterated his willingness to share the system with the Soviets. (His own offer to share the technology was both clumsy and vulnerable to charges that it could hardly take place: (1) the United States would not be willing to share its most advanced technology with the U.S.S.R., and (2) knowing how to use the system means knowing how to defeat it. But Reagan may have had something else in mind, including the possibility that once the system were in place, both powers might at least use the data from sensors to identify third-party threats, or that the interceptors might be managed jointly or perhaps by some multinational entity.)

At Reykjavík, Reagan went out of his way to try to persuade Gorbachev that the only way to achieve total elimination of strategic weapons was to allow both sides to deploy defenses as an "insurance policy" against cheating or third-party threats. Presumably, as Fletcher and other proponents of strategic defenses (including Teller) have pointed out, such a defense need not be space-based, and some versions could be much less elaborate than a comprehensive layered shield, because they would need to defend only against a much smaller threat.^[68]    In any case, Reagan made clear that he was committed to the SDI and was unwilling to agree to any proposal that would prevent it from reaching a point at which a decision could be made ten years hence to deploy a system. Development would certainly require some forms of testing in space, though the experiments could be configured to conform to ABM Treaty restrictions—at least in the sense that the elements tested would not achieve ABM power levels or be tested against strategic launchers or RVs and could therefore not be used without further prototype development as BMD weapons. In effect, the official U.S. position has been that development and testing of the potential elements of a space-based defense are permissible, so long as they do not reach the point at which field testing of prototypes takes place. This seems to have been the president's own position. But he was also obviously aware of the danger that if an agreement were reached in his administration to reduce nuclear weapons and to restrict SDI research, forces in Congress would therefore be unwilling to fund it at a significant level.

There was apparently also a disagreement, or at least a difference of

emphasis, among the president's counselors. Weinberger made it clear, before the Geneva summit and after Reykjavík, that he was prepared to have no arms-control treaties if the alternative meant constraints on the modernization of U.S. strategic forces and the development of strategic defenses. Because of this position, he and Perle may well have exaggerated the intransigence of the Soviets. Perle went so far as to suggest the Soviets were concerned with the offensive capabilities of strategic defenses because their own research is so advanced that they have discovered offensive uses of which the United States is unaware. Shultz, however, appeared to have been genuinely disappointed by the failure of the Reykjavík summit. He went to Geneva after Reykjavík in the hope of breaking the deadlock, but to no avail.

Reagan seemed to be immobilized between the contradictory views of Weinberger and Shultz. After Reykjavík, in early 1987, Weinberger sought to capitalize on the breakdown of negotiations by proposing that the "legally correct interpretation" be formally adopted as the basis for SDI experimentation and that the president endorse early deployment in the January 1987 State of the Union address. But the new national security advisor Frank C. Carlucci reportedly succeeded in dissuading the president from endorsing early deployment by pointing out that no decision had so far been reached on the issue within the administration. On February 3 a meeting was held with the key policymakers at which the president said he was ready to adopt the new interpretation but not to call for early deployment. Shultz, joined by Sofaer and Nitze, recommended that the president defer a decision pending a further review. Accordingly, Reagan approved National Security Decision Directive 261, which called for new studies on the interpretation. The minutes of this meeting were leaked to the press. Apparently, the participants discussed the likely Soviet reaction to an announcement that the new interpretation was to be adopted. The president is quoted as saying: "Don't ask the Soviets. Tell them," and "Why don't we just go ahead on the assumption that this is what we're doing and it's right?"^[69]   

At the Washington summit in December 1987, controversy over the testing provisions continued to pose an obstacle to a START agreement, but this time the leaders agreed to set the issue to one side. At first, Foreign Minister Shevardnadze insisted that the United States must agree to abide by the ABM Treaty as "signed and ratified." Secretary Shultz reportedly persuaded him that this formula was an inappropriate allusion to the debate over the interpretation of the treaty between the administration and members of the Senate.^[70]    A last-minute compromise was

struck which enabled each side to claim that the other had accepted its position. In a prime specimen of diplomatic double-talk, the two leaders instructed their delegates in Geneva "to work out an agreement that would commit the sides to observe the ABM Treaty, as signed in 1972, while conducting their research, development and testing as required, which are permitted by the ABM Treaty, and not to withdraw from the treaty for a specified length of time."^[71]    In other words, the impasse over the kinds of testing that would be permitted was left unresolved, presumably to await further negotiation.

The Soviet Position

In the various meetings, the Soviets evidently did not make clear exactly why SDI was of such concern to them. Some Soviet statements indicated that they saw SDI as an offensive rather than a defensive program. Perhaps this explains their references to "space-strike" weapons. Some U.S. observers have observed that, in fact, these space-based weapons could be used for offensive purposes, i.e., for striking targets on the ground and in the air and for setting major fires.^[72]    These fears have been derided, however, by those who point out that there are far more efficient and inexpensive ways of mounting offensive strikes and that laser weapons cannot reliably penetrate the clouds to strike ground targets. Fires cannot readily be set in urban areas by the use of lasers, and in any case, the resulting exhaustion of fuel supplies for offensive purposes would render space-based defensive weapons useless for their primary purpose. Setting fires, moreover, would take time, and, by then, the Soviets would have ample warning of an attack and be able to launch their retaliatory strike.

The Soviet fear of SDI is more likely to be based on the general notion that it represents a major acceleration in the intensity of the technological arms race, which would only impose further strains on an economy already seriously handicapped by extravagant military expenditures. Furthermore, such a comprehensive program is virtually certain to produce technological advances that could be applied to nuclear and conventional weapons systems. Better sensors, computers, laser weapons, electromagnetic railguns, etc., could significantly increase the effectiveness of U.S. offensive weapons, forcing the Soviets to make substantial investments in similar modernization just to keep abreast. Their relative technological inferiority in most of the relevant technologies would put them at a disadvantage in this competition. The Soviets would

have to do more than develop countermeasures to the SDI; the applicability of any research findings to offense would force them to undertake similar programs.

Another Soviet objection is that SDI could give the U.S. a first-strike option. The stated concern of the Soviets, since Premier Andropov first responded to Reagan's "Star Wars" speech, is that SDI would enable the United States to launch a first strike and then to absorb a "ragged retaliatory strike." Although at most a remote possibility, the thought has occurred even to SDI supporters. Weinberger has expressed horror at the prospect that the Soviets might obtain a significant defensive capability, because of the very scenario described by Andropov. "I can't imagine a more destabilizing factor for the world," Weinberger declared, "than if the Soviets should acquire a thoroughly reliable defense against these missiles before we do."^[73]    He has also said that he "cannot imagine that the Soviets, if they had a monopoly position, would do anything other than try to blackmail the rest of the world."^[74]    Why should the Soviets think the United States would behave any differently?

The Soviets may also reasonably fear that BMD weapons would find a ready use as ASATs. Satellites are not very difficult targets, so the products of SDI research will probably be more useful in this capacity than in defending against a massive missile attack. In a crisis, such weapons could be used as part of an offensive effort to destroy or blind enemy sensor and communication satellites. In this limited sense, it is plausible to think of them as space-strike weapons.

For all these reasons, the Soviets seem to have decided that they had to make SDI a central issue in the arms-control negotiations. If they were to agree to reductions in offensive arms without constraining the U.S. effort to design and deploy defenses, they would risk putting themselves in a strategically inferior position. Even if the U.S. program were only partially successful, the United States might decide to erect defenses (at the end of the ten-year extension of the ABM Treaty), challenging the parity they now enjoy because of constraints on defensive deployments. The Soviets claim that their position would make it unnecessary for either side to deploy defenses. They propose instead that both sides agree to eliminate all nuclear weapons. In that way, neither would face a nuclear threat and neither would therefore require defenses. The U.S. position is that a total elimination of nuclear weapons is an unrealistic expectation. The United States, at any rate, could not safely assume that the U.S.S.R. would not stockpile some undeclared nuclear weapons; even a few such weapons would be a formidable

threat against a power with none. Defensive systems are therefore essential as "insurance" against cheating and third-party threats in conjunction with radical reductions in offensive arms.

Breaking The Impasse

Is there a way out of the impasse? One such way would be for both sides to agree on a more precise definition of allowable development and testing. Much of the work the SDIO has scheduled falls well within the customary, or restrictive, reading of the treaty. Some is in a gray area. The complexity of the proposed research makes it obvious, as various studies have shown in detail, that these technologies cannot possibly become candidates for deployment for many years—ten to fifteen at the very least. Restrictions on the testing of prototypes would therefore have no effect whatever on the pace of research. Restrictions on the testing in space of DEW satellites, however, or on satellites carrying beam generators with high power levels, would have some cramping effect. But, generally, much of the research on new physical principles can be done in the laboratory or in forms acceptable under the treaty. An agreement to abide by the restrictive interpretation for ten years would therefore not significantly delay the effort. In addition, it would assure the Soviets that ten years hence the U.S. effort could not possibly have reached the point where deployment could be implemented in a short period of time. Both sides are bound to continue their modernization of offensive weapons; provided SDI is not funded at extraordinary levels or surrounded by unrealistic rhetoric, the Soviets need not fear the likelihood of major breakthroughs that their own research and development could not possibly match.

In sum, the impasse can be broken if both sides finally come to realize that none of the systems being investigated under SDI is remotely likely to pose a threat to the current basis of deterrence for at least several decades. Certainly, there is no justification for supposing that the ABM Treaty must be violated within the next ten years to take account of new technological realities.

Nevertheless, in the absence of an arms-control agreement neither side is likely to rush toward a breakout from the ABM Treaty. Given that the technology of effective defense is so remote, the expense of early deployment cannot be justified. Early deployment of U.S. ground-based defenses is a possibility, but this could be done within the limits of the ABM Treaty. The Soviets have no incentive to abrogate the treaty unless

SDI testing blatantly disregards treaty limits. If either side decides to abrogate the treaty, however, the efforts to achieve negotiated strategic arms control are likely to collapse. Neither side would feel comfortable dismantling offensive weapons. Both would seek to develop force structures that would make the job of the defense more difficult, as both sides have already done. This could take the form of improving offensive penetration aids, shortening booster-burnout times, coating missiles, etc., and of efforts to develop defensive weapons, ground-based as well as space-based.

It is highly doubtful that either side would benefit from such a competition. As we have seen in chapter 3, history has shown that innovations by one side are easily matched by the other either by the development of a similar system or countersystem. In this case, another, more novel response might be attempted; the Soviets could decide to prevent the full deployment of a space-based defense. Striking at a satellite is not a clear violation of the international law that prohibits transgressing another state's territorial integrity because there is no established principle governing the deployment of a defensive satellite in what could be said to be an extension of the air space of another state. The state that did so could claim to have acted in self-defense, particularly if it warned that it viewed such satellites above its territory as hostile or invasive. Any state, furthermore, could disable satellites in ways that do not constitute outright attacks, such as blinding them or interfering with their mirrors by orbiting debris—a step that would not produce the equivalent of a smoking gun.

Another difficulty is that a shield in space would not protect against threats to go beneath it by attacking with depressed-trajectory weapons, tactical weapons, and air-breathing systems. The development of long-range, low-altitude stealthy cruise missiles is currently under active consideration in the United States. A separate initiative would be needed to produce an air defense and tactical BMD able to meet this kind of threat—including defense against the new stealth-type missile which, according to its promoters, is able to penetrate any known type of defense. Even if such an initiative were successful, it would not entirely remove the threat of a major nuclear attack. Even in the extremely unlikely event that all dangers from missiles and bombs could be averted, nuclear weapons could still be delivered to ports in the holds of cargo vessels, to major urban centers secreted in civilian aircraft, and otherwise smuggled in as suitcase bombs. Although such weapons would probably not have the devastating effect of a full-scale ICBM attack,

they could nevertheless cause damage on a scale unprecedented in the history of warfare. A determined or desperate adversary would always have the means to threaten such an attack. Such delivery systems would not be adequate for attacking military targets, but they would be entirely adequate as threats against civil targets.

Similarly, there is little merit in the contention that a partially effective defense is better than none. Advocates of SDI argue that, at the very least, a defensive shield would create so much uncertainty in the mind of a potential attacker as to rule out any possibility of a preemptive attack. If the goal is to instill uncertainty, however, it can be achieved far more easily (and more reliably and cheaply) by increasing deployments at sea, on mobile launch platforms, and so forth, which unlike SDI, do not risk destabilizing the weapons balance. An enormously expensive program like SDI (or the follow-on deployment) is not needed if the only purpose is to instill uncertainty in the mind of the adversary.

The very fact of a breakdown in arms-control negotiations, coupled with the resumption of the defensive as well as the offensive arms race, could well promote insecurity on both sides. In such a situation, the side that thinks its position is worsening could be tempted to take actions that would prevent a further decline before it is too late, or it could take actions that could divert the other side from its program. Thus, a U.S. military buildup could be countered by a Soviet use of proxy forces against U.S. interests. The Soviet Union's ability to make trouble for the United States and its allies in the Third World is now well appreciated. Although the U.S.S.R. has resisted efforts to link arms control with détente in these areas, it would have an incentive to turn up the heat should the United States seek to threaten to gain a strategic advantage by developing weapons that the Soviets perceive as a direct threat to their security. Such a policy of provoking trouble for the United States may carry costs that the Soviets would not be willing to bear, but the United States has no interest in encouraging Soviet efforts to destabilize the international system.

In view of the potential for instability, a further development of the arms-control regime seems highly preferable. Such a regime would aim at continual steps to reduce nuclear arms to whatever level could be considered minimally acceptable. If defenses are to be deployed, they could be installed on the basis of agreed understandings, worked out on the model of the ABM Treaty, or by amending or renegotiating the treaty. As confidence grows, a cooperative deployment of defensive systems may become conceivable in which, for example, both sides would

receive warning of any attack from sensor satellites. A multinational or transnational effort could then conceivably be mounted to administer certain peace-keeping devices, more or less in accordance with Reagan's belief that a shield could be shared with the Soviets. Unfortunately, these possibilities are undermined by the fact that so much new research—whether performed under military auspices or not—poses insuperable temptations to develop systems that confer temporary advantages but render cooperation with the other side more difficult. In the case of SDI, unilateral deployment would also jeopardize the unity of the Western alliance, for reasons we explore next.