Foreign Relations of the United States, 1989–1992, Volume XXXI, START I, 1989–1991

21. Memorandum From Linton Brooks of the National Security Council Staff to the President’s Deputy Assistant for National Security Affairs (Gates)¹

Washington, May 18, 1989

SUBJECT

• Deputies Committee Meeting, Saturday, May 20, 1989, 10:00–11:00 a.m.—Security and Arms Control

This Deputies Meeting, which precedes a May 25 (tentative) NSC meeting on the same subject, will be your first opportunity to have the Deputies Committee deal with arms control. The nominal purpose of the meeting is to consider the arms control implications of the Defense Strategy Review. After several false starts, the Defense Strategy Review arms control section has been organized around the answers to six fundamental questions:

— Is our fundamental approach to START (deep reductions, shift to air breathing systems) correct?

— Are SLCMs in our interest?

— Is START verification good enough?

— Are our Defense and Space objectives correct?

— How low can we reduce ground and air forces and still meet NATO objectives?

— Are there alternatives to a global, verifiable ban on chemical weapons?

The paper responding to these six questions is at Tab A. It is quite long, in part because it attempts to deal with a number of subsidiary questions. There is consensus, which we endorse, that only the first four topics should be discussed at this Deputies meeting. We will plan on a Deputies meeting and an NSC meeting on CW in mid-June; at that time we will consider both the results of the defense strategy review on chemical weapons and the various arms control tasks set forth in NSR-14. We will deal with the conventional arms control in a similar fashion; although the paper is decent in this area, there is additional tasking from the Deputies Committee in NSR-14, and the subject, therefore, is not yet ready for senior-level discussion.

We believe that the remaining four questions offer more than ample scope for discussion, even though the paper, in part out of a concern over leaks, poses no specific issues for decision. We recommend you use the Deputies meeting (and we will recommend that Brent and the President use the NSC meeting) for an exceptionally frank, heart-to-heart discussion of where we are going in START and Defense and Space. The meeting should not focus on details of our position; there will be dozens of meetings over the next two years to do that. Rather, this meeting must answer the fundamental question: “Is this a trip we want to take?”

The Basic START Issues. With regard to the basic START position, the paper concludes:

— Present START reductions don’t imperil military sufficiency although deeper reductions might. Likely START reductions probably don’t save money. Our goal should be stability and reduced risk of war, not reductions, per se.²

— Plausible alterations of weapon-to-target ratios (by which people usually mean the ratio of hard target kill RVs to U.S. silos) won’t solve the ICBM vulnerability problem. Downloading and deMIRVing may be good ideas on grounds of flexibility, but mobility and strategic defense are better approaches to solving the ICBM vulnerability problem.³

— START allows us to keep an adequate SSBN force.⁴

— We are right to encourage a shift to air breathing systems and to accept looser constraints on such systems. We shouldn’t include air defenses in START.⁵

We recommend you structure discussion as follows:

— Summarize the above points and ask if there is agreement.

— Ask specifically if the Deputies can agree that the basic START levels (6000 warheads, 4900 ballistic missile warheads, and 1600 delivery vehicles) allow us to meet our essential military aim. (You should note, that while there is consensus on this point outside of the NSC staff, we believe Brent has reservations on the 1600 number.)⁶

— With respect to air breathing systems, force the Deputies to face up to the fact that we have chosen to preserve our flexibility at the expense of verifiability. To a first approximation, none of our air breathing limitations (other than basic counts of bombers) is verifiable. The Deputies need to face up to that they will each individually sooner or later have to testify that such a provision is in our interest.⁷

— Similarly make sure that the Deputies agree that even though we have no air defense and the B-2 may be unaffordable, we still want to encourage a long-term Soviet shift to air breathing systems.⁸

Sea-Launched Cruise Missiles. The paper says that some people think SLCMs are important and irreplaceable and others don’t; that everyone agrees we don’t know how to verify limits on them now; that there is disagreement on our prospects for finding an acceptable verification regime; and that we don’t have to make any decisions now. We think that everyone but State believes that a SLCM ban isn’t in our interest and is unverifiable.

SLCM is endlessly fascinating, but we’re not sure a discussion of it at this time leads anywhere. We suggest the following approach:

— Note that the paper sets for both sides of the issue on both SLCM utility and SLCM verification.

— Note that the Arms Control PCC has concluded that we need not alter our SLCM position during the initial round of negotiations.

— State that the review has identified this as an important issue where there is no consensus, but that we need not deal with it now.

— Seek agreement to note it and move on.⁹

We think the Deputies will probably agree, but you should recall that Brent personally believes SLCMs aren’t in our interest. Even if he’s correct, this isn’t a tar baby we need to grab now.

Verification. This agenda item is the most important of all. The paper:

— Describes the verification process;

— Mentions problems with verification of limits on mobile ICBMs (including disagreement over the significance of cheating in this area);¹⁰

— Mentions problems with verification of RV loading (noting disagreement over the extent of such problems);

— Says that we are accepting poor verifiability of ALCMs and bombers since (a) we want to protect our programs and (b) these are stabilizing systems;¹¹

— Says START won’t add requirements for unique NTM assets; and

— Says conclusions on verifiability are premature.

We think the paper is useful background, but misses the central issue: If the paper is correct in its assessment of the monitoring and verification problems we face, is this the road we want to walk down? If so, everybody should be prepared for the possibility that two years from now we will be testifying on a treaty that we will assert is in our interest even though not all provisions of that are fully verifiable.¹² If on the other hand, this is not a road we want to walk down, the President should either:¹³

— Direct a radical change in our START objectives to focus on an incomplete, but verifiable, set of limits;

— Direct massive new intrusive verification measures (which may or may not help and will be opposed by many on security grounds); or

— Plan on substantial new expenditures for collection systems (which may or may not actually solve any problems).

If “verification” means knowing how many systems or warheads the Soviets have or what those systems are armed with, it is almost certain that we will not be able to verify many of the START limits. The Deputies meeting offers us the first chance to force the new senior leadership to face up to that fact. Ultimately (but not at this meeting) we will have to decide what we mean by “effective” verification. We may find we have to accept a standard of “good enough,” even though “good enough” is quite low in traditional terms.

We suggest you avoid getting into a discussion of specifics and seek to focus on the basic question. You might suggest that the goal is not to solve problems but to make certain that we’re marching ahead with our eyes open.

Defense & Space Talks. The paper raises the right question: Do we want to retain as the fundamental objectives of Defense and Space preserving our options to deploy SDI even if this risks blocking a START agreement? The paper sets forth alternate views on (a) the need to have a treaty right to deploy SDI, (b) the need for the broad interpretation, and (c) whether Defense and Space concessions are needed for START. The paper establishes consensus on the value of retaining the ABM Treaty for now and of seeking a cooperative transition to a world of strategic defenses.¹⁴

As with other elements of the review, the Defense and Space issues are complicated and detailed for the amount of time available to devote to them. Moreover, as we will be but three weeks from resumption of negotiations when the NSC considers the paper, we will be hard pressed to make significant changes in our negotiating position in an orderly fashion. Furthermore, most of the issues raised by the review (e.g. Defense and Space objectives versus seeking a START agreement) need to be borne in mind when making future decisions, but are not now ripe for decision. Thus, we suggest you seek consensus on the following hypotheses:

— The review raises the right issues about the Defense and Space Talks.

— However, given that we are not at endgame in either START or D&S, there is no need now to make decisions on how we would resolve fundamental tradeoffs.

— As the negotiations proceed, we will continue to review the issues raised in the paper.

— For the time being at least, the following objectives remain valid:¹⁵

a) to preserve the U.S. right to pursue the SDI program within the obligations of the ABM Treaty in order to determine the feasibility of Strategic Defenses;

b) to discuss with the Soviet Union a cooperative transition to effective strategic defenses; and,

c) to establish the right to deploy strategic defenses if we choose to do so, at a future time without further reference to the ABM Treaty¹⁶

— We can also reaffirm the consensus on the value of retaining the ABM Treaty for now and of seeking a cooperative transition to a world of strategic defenses.

If people are not willing to endorse these objectives, it will be important to have the bluntest possible discussion at the Deputies meeting, in preparation for an equally candid discussion at next week’s NSC meeting. It will not lead to sound policy if the agencies working on our Defense and Space position do not agree on what we are trying to do.

If time permits, you may also want to pursue a discussion of whether or not the United States needs to continue to protect the option to conduct tests under the broad interpretation of the ABM Treaty. There is no event in terms of either tests required by SDIO or negotiating dynamics that requires a decision on this now. Nonetheless, a candid discussion of this issue might be worthwhile given its high political visibility. If you hold such a discussion, we suggest framing it as follows:

— The advantages of sticking to the broad interpretation are that (a) we need not withdraw from the ABM Treaty to conduct end-to-end testing of boost-phase defenses, and (b) we retain significant bargaining leverage in Geneva.

— The advantage of moving away from this position would be to gain goodwill on the Hill and narrow a major difference with the Soviet Union.

Next Steps. In summarizing the discussion and preparing for the NSC, it would be useful to gain the consensus that the paper at Tab A, and the NSC discussions it will engender, are an adequate response to tasking of NSR-12. We have duplicated that tasking at Tab B; while several items aren’t covered in this paper, they have been dealt with in the longer backup paper. We believe it is time to declare victory on NSR-12 and quit. It would be useful if the Deputies came to a similar conclusion.

RECOMMENDATION

That you read the paper at Tab A and use the points outlined above in guiding discussions at the Deputies meeting on security and arms control.¹⁷

Will Tobey, Don Mahley, Judyt Mandel, Rich Davis, Bob Blackwill, Condi Rice, and Don Snider concur.

Tab A

Paper Prepared by the Ad Hoc Interagency Group on National Defense Strategy¹⁸

Washington, May 9, 1989

NSR–12, PART IV

(C) The Defense PCC directed that the response to Part IV of NSR–12 focus on six key sets of questions in the areas of START, Defense and Space, Conventional Forces, and Chemical Weapons. This report frames, and briefly discusses, the issues with respect to those questions.

START

(S) The Reagan Administration set out four fundamental criteria for a START agreement: deep reductions, greater stability, equality and effective verification. One of these—insistence on equal rights and limits under an agreement (consistent with the Jackson Amendment)—has not been brought into question. However, there are tensions among the other three criteria. For example, reductions to too low a level could jeopardize stability, rather than enhance it. There have also been criticisms that the general U.S. approach in START may not actually lead to enhanced stability—that it might not, for instance, lower the Soviet weapon-to-target ratio or that it might not allow us to deploy the forces necessary for deterrence. Finally, there are questions as to how to balance U.S. verification needs against our other security requirements. The following questions discuss these concerns.

Question #1: General Approach to START

What do deep START reductions, as currently envisioned, accomplish, and at what cost or savings? What will they contribute to: deterrence, stability, and constraining the threat? Should we be trying to reduce strategic offensive capability without at the same time significantly reducing Soviet weapon-to-target ratios (by downloading/de-MIRVing or other measures)? Will START restrict the U.S. to too low a number of SSBNs? Is channeling strategic forces away from ballistic missiles and toward air breathing systems in the U.S. security interest given the asymmetry in air defenses and our probable inability to establish limits on air defenses that are both verifiable and protect U.S. capability?

(S) When the START talks recessed in late 1988, the United States left on the table a negotiating position which sought deep reductions and tight constraints in ballistic missiles (and especially certain categories of ICBMs), with looser constraints on strategic bombers, bomber weapons, and cruise missiles.

Deep Reductions

(S/NF) Today the Soviets deploy about 11,000 intercontinental nuclear weapons. The Intelligence Community believes that all elements of Soviet intercontinental nuclear forces will be extensively modernized between now and the late 1990s. Without START, Soviet forces would probably grow to a level between 12,000 and 15,000 weapons in the next ten years, and the Soviets have the capability for expansion in the late 1990s to 16,000 or even 18,000 weapons if, for example, they decided to expand forces in response to a U.S. deployment of strategic defenses. Based on the current U.S. bomber/ALCM counting rules, a Soviet START force of 6,000 accountable warheads could legally contain in excess of 8,000 weapons. In a crisis or wartime situation, the Soviets might over the course of several months be able to deploy a few thousand additional weapons in their post-START force, even if a treaty does not permit de-MIRVing.

(S) In comparison, today the U.S. deploys 12,000 to 13,000 intercontinental nuclear weapons, and all legs of the strategic Triad are undergoing modernization. In the absence of START, this number is not projected to change significantly over the next decade. With START, the U.S. could legally deploy 9,000 to 10,000 strategic weapons, based on the U.S. bomber/ALCM counting rules.

(S) Our reduction objectives in START have two elements. First, reduction in overall strategic delivery vehicles and warheads, and second, reductions in the most destabilizing elements of the Soviet threat—ballistic missile warheads, warheads on heavy ICBMs, and throw-weight. The overall reductions envisioned in START halt and reverse the trend under SALT II toward increasing inventories of strategic nuclear weapons; however, overall reductions were not by themselves meant to carry the burden of improving crisis stability. The emphasis on deep reductions in destabilizing weapons was designed to do that. The effectiveness of our deep reduction provisions in improving stability is discussed in the next section.

(S) With U.S.-proposed bomber and ALCM counting rules, START will allow the U.S. to deploy sufficient weapons to accommodate U.S. targeting requirements. With appropriate modernization of U.S. strategic forces (which is not constrained by START), the U.S. can deploy strategic nuclear capability to support a strong deterrent policy and, if war occurs, will have the power to render further Soviet actions ineffectual.

(S) While START limits allow for military sufficiency, they take our military forces toward the margins of their ability to satisfy existing guidance for nuclear weapon employment, even with continued force modernization. Deeper reductions would require reconsideration of that guidance, and could affect allied perceptions of the credibility of the U.S. extended nuclear deterrent. Due to the larger number and greater hardness of targets in the Soviet Union, the Soviets may be more able than the U.S. to accept deeper cuts.

(S) As to the cost or savings of deep START reductions, it is not clear whether significant sums would be saved, as is sometimes assumed. Our experience with INF shows that we should not expect near-term savings. Over the longer term, savings are problematic and would depend upon several factors. For instance, the U.S. must continue to modernize and retain its remaining strategic nuclear forces at a high level of operational effectiveness, and the costs of monitoring must be considered.

Weapon-to-Target Ratios

(S) A second difficult question is whether the numerical limits and sublimits we have proposed adequately contribute to stability given their likely impact on Soviet weapon-to-target ratios. START would constrain the Soviet advantage in the overall weapon-to-target ratio, given the anticipated numbers of weapons and the pronounced target base asymmetry. Critics of START have tended to focus, however, on a specific weapon-to-target ratio of Soviet prompt, hard-target-kill weapons to U.S. ICBM silos. There is broad agreement that this is not a meaningful measure either of ICBM vulnerability or crisis stability. The ratio omits consideration of strategic defenses and ICBM mobility. These latter factors can enhance ICBM survivability and crisis stability, and—while not lowering the ratio of hard-target-kill weapons to U.S. ICBM silos—can make that ratio less relevant. That ratio also fails to reflect the role that other elements of our strategic forces play in deterrence.

(S) We have not identified a program-specific, post-START force, although we have constructed a number of theoretical possibilities for the purpose of assessing the military sufficiency of START proposals as they evolved. As currently envisioned, START would result in at least a fifty percent reduction in Soviet heavy ICBM weapons. Even with this reduction, however, U.S. ICBM silos will remain vulnerable. The Soviets are deploying a new variant of the SS-18 heavy ICBM and will deploy hard-target-capable SLBMs by the late 1990s. There are some options under consideration which could help U.S. ICBM survivability temporarily, but in the long run U.S. ICBM silos will not be survivable absent defenses. Enhancing crisis stability also depends on our decision regarding mobile ICBMs.

Mobile ICBMs

(S) Mobile ICBMs tend to be more survivable than those in fixed silos, although the degree of survivability of mobile systems depends upon the specifics of the mobile basing scheme, whether or not the mobile systems are dispersed, and the ability of the enemy to locate and hold at risk those systems. The Soviets currently deploy two types of mobile ICBMs.

(S) There is broad consensus that mobile ICBMs would be more stabilizing than silo-based ICBMs in a crisis, providing that actions necessary to ensure their survival are taken. There also is broad consensus that mobile ICBMs will pose serious monitoring, verification, and breakout problems. Much work has been done to date; however, whether a verification regime that meets our standards (with acceptable risks and impacts) can be identified and negotiated remains an open question (see pages 17–18 for a more detailed discussion of mobile ICBM verification). It also is unclear whether the U.S. can muster the sustained political support required to deploy a robust mobile ICBM force. A post-START world in which only the Soviets deploy mobile ICBMs would adversely affect the strategic balance. We would not have a survivable ICBM force (in the absence of effective defenses) to hedge against a failure in one of the other two legs of the U.S. strategic triad. It would also compound U.S. targeting, verification, and threat assessment problems.

Downloading/De-MIRVing

(S) Another potential approach—downloading our ICBMs—would lower the weapon-to-target ratio, but in the long run would not improve US ICBM silo survivability. There are, however, other reasons to download. Downloading (sometimes called de-MIRVing) refers to removing some reentry vehicles from existing MIRVed missiles like the Minuteman III, possibly modifying the attachment points for the reentry vehicles, and counting the missiles at the reduced warhead loading for purposes of satisfying the START numerical limits. Downloading would allow flexibility to deploy a greater number of silo-based ICBMs under START. Downloading also is attractive from the perspective of the operational flexibility it affords a post-START ICBM force. On the other hand, if not tightly constrained, downloading could undercut the purpose of placing strict limits on ballistic missile warheads in START. Downloading is a reversible process and poses both verification and break-out problems. For example, RVs could be replaced on Soviet missiles over a period of months, allowing the Soviets to deploy additional weapons. No U.S. decision has been made yet on how to handle downloading in START.

Number of SSBNs

(S) The report for Part II of NSR–12 concludes that the presently programed twenty-plus Trident SSBNs appear to meet U.S. security requirements. START places no direct restriction on SSBNs, but the size of the force would be indirectly limited by the number of accountable SLBM warheads under the 4900 sublimit. The U.S. approach allows for exempting from START accountability the warheads from as many as three Trident SSBNs in repair yards at any time. It also could allow (depending on the decision on downloading) for increasing the number of SSBNs by decreasing the number of accountable SLBM warheads deployed on them, and it could be revised to allow explicitly for altering the number of deployed launchers on an existing SSBN. It also allows deployment of future SSBNs with fewer than 24 launchers (the number on the Trident SSBN), and of future SLBMs with fewer than eight RVs (the number attributed to the Trident II SLBM). Within this array of options, the U.S. should be able to meet its security requirements with respect to SSBNs in a post-START environment. It should be noted that these provisions also have implications for the number of SSBNs that the Soviets could deploy.

Air Breathing Systems

(S) The U.S. approach in START has been to channel strategic forces away from ballistic missiles and toward air breathing systems. The U.S. determined that this was in its security interest despite the asymmetry in air defenses. However, it is worth reviewing why this approach was adopted in the first place.

(S) Ballistic missiles (and particularly certain types of ICBMs) have been seen as the greater threat to stability. Ballistic missiles in general are characterized by short flight times, high assurance of penetration to target, and the fact that they cannot be recalled once launched. Today, heavily MIRVed ICBMs possess features which make them especially well suited for use in a pre-emptive attack: high alert rates, good prompt-attack accuracy, and the requisite communications capability.

(S) Air-breathing systems (heavy bombers and cruise missiles), on the other hand, do not pose as great a threat to crisis stability. Generating the bomber force (which has much lower alert rates than ICBMs) could provide considerable strategic warning, and bomber flight times are measured in hours, not minutes. Moreover, bombers can be recalled. And while stealth technology increases the ability of a bomber to penetrate air defenses, a bomber incorporating stealth technology still would likely have detectable force generation and time-of-flight characteristics similar to its non-stealth counterparts. It has been considerations of this sort which led the U.S. to emphasize tighter constraints and deeper reductions on ballistic missiles than on air-breathing systems. Partly for the same reasons, the U.S. has emphasized program and operational flexibility for its air-breathing systems over verifiable limits on Soviet air-breathing systems.

(S) There are compelling reasons to continue this approach, despite the pronounced asymmetry in U.S.-Soviet air defenses and the fact that the Soviets are believed to be moving in the direction of a heavy bomber force which will have a greater role than before, with more weapons and greater force diversity. Today our bombers would deliver approximately 35 percent of our strategic nuclear weapons. With appropriate bomber counting rules, such as those in the current approach, and given the currently proposed ballistic missile warhead limits, we would require approximately 50 percent of our post-START strategic nuclear weapons to be bomber delivered to cover the Soviet targets. A further delay or cutback in the B–2 program, however, might require reconsideration of our ALCM/Heavy bomber counting rules to meet post-START requirements. Our success in meeting targeting objectives is somewhat attenuated by the elaborate Soviet air defense network, but continued modernization of our bomber forces allows us to cope with that problem.

(S) The Soviets could also exploit this feature of a START treaty, but it would take them a good many years to effectively match the U.S. in bombers. And, as was discussed earlier, we have wanted them to shift the emphasis in their force structure from ballistic missiles to bomber delivery weapons which we consider less destabilizing.

(S) As for the judgment reached in the preceding discussion that we should continue emphasizing looser constraints on air-breathing systems, despite the pronounced asymmetry in U.S. and Soviet air defenses, several considerations have influenced this assessment. The U.S. traditionally has relied more heavily (and with greater confidence) on its bomber leg of strategic forces than has the U.S.S.R., and has lived with the asymmetry in air defenses. As pointed out earlier we depend heavily upon bomber weapons for target coverage. We have kept our bombers at a higher state of alert than the Soviets, and generally have remained ahead of the Soviets in technologies related to effectiveness of the bomber force. The Soviets can, of course, also exploit START’s looser constraints on air-breathing systems, but it would take them a good many years to effectively match the U.S. in bombers. And (unlike the situation in ICBMs), if the Soviets attempt to break out of a START treaty by uploading bomber weapons, we could match or exceed them in this category over a period of time.

(S) U.S. decisions in the past to pursue modest air defenses were arrived at in large part because of the absence of effective defenses against ballistic missiles. If the U.S. were to decide in the future to deploy advanced ballistic missile defenses, that would require a fundamental review of whether to also deploy extensive air defenses or to seek significant constraints on air defenses. There are several considerations which lead us to assess the prospects of effective and acceptable constraints on air defenses as being poor.

(S) First, verifying compliance with limits on air defenses would be severely challenging. The complexity of an air defense infrastructure, the mobility (and multiple uses) of air defense systems, and the sheer numbers of activities involved raise the question of whether we could set verification standards with respect to strategic air defenses which would be as demanding as those for systems dealt with in START.

(S) Second, systems for strategic air defense (such as interceptor aircraft or airborne surveillance radars) have the inherent capability to be shifted rapidly to non-strategic missions. It would be difficult to construct constraints which limited dedicated air defenses without adversely affecting their readiness for non-strategic uses. An analogy can be drawn to attempting to constrain airborne refueling capability—an area we have carefully avoided in START.

(S) Finally, air defense of North America is a mission which we share by treaty with Canada. Any proposal to negotiate restraints on air defenses would, at a minimum, have to be carefully coordinated with Canada. And introducing air defenses into START could give the Soviets a stronger case for drawing one of our allies into the negotiations, an outcome which we have carefully avoided for strategic offensive forces.

Conclusion

(S) No specific decisions are required at this time, as a result of the discussion on Question #1. This intent was to familiarize the reader with major issues relating to our general approach to START. Decisions will come as part of other reviews.

Question #2: SLCM

Would the elimination of nuclear-armed, land-attack SLCMs enhance or degrade security and stability?

(S) U.S. and Allied concern over Soviet nuclear-armed, land-attack SLCMs stems from their potential for use in a precursor attack on critical leadership and C3I targets, as well as their potential for attack on a range of industrial and military targets. Allied concerns focus on the latter. Arguing that nuclear-armed SLCMs do not play a central role in U.S. strategic forces, some have suggested that a ban on such systems would be in the U.S. interest.

(S) The Soviet Union possesses other assets that could be used for precursor attacks, in particular SLBMs launched from SSBNs close to U.S. coasts though these are easier to detect than SLCMs once launched. In the future, conventionally-armed SLCMs may have sufficient accuracy to hold some of the same targets at risk.

(S) Thus far, U.S. attempts to find an effective verification regime for nuclear-armed SLCMs have been unsuccessful. Verifying a ban or limits on nuclear SLCMs—with the requirement to distinguish between nuclear and conventional SLCMs—remains very difficult at best. Verification schemes would also have to be designed so as not to disrupt naval operational plans or undermine our policy to “neither confirm nor deny” the presence or absence of nuclear weapons on ships. These elements make verification of SLCM limits in START particularly difficult.

(S) The United States deploys nuclear-armed SLCMs to perform a number of important missions. First, nuclear-armed Tomahawk SLCMs provide a continuously-deployed, world-wide nuclear deterrent presence that is both visible and free to move in international waters. Because SLCMs are spread over a large number and variety of platforms (some of them submarines), they complicate any Soviet effort to destroy U.S. nuclear retaliatory capability, and thereby enhance stability. Second, these systems help deter Soviet attacks on our naval forces by Soviet land-based naval bombers armed with nuclear anti-ship missiles, by threatening nuclear retaliation against naval and air bases on Soviet territory. Threatening to retaliate with nuclear weapons against Soviet surface ships alone is an inadequate deterrent, for the Soviets would accept such a trade, given the far greater U.S. dependence on naval forces for security. Deployment of nuclear-armed, land-attack SLCMs underscores U.S. policy that we will not permit the Soviets to confine a nuclear war to the sea. Third, a portion of the deployed nuclear-armed SLCM force will be dedicated to the nuclear reserve force; because they are highly survivable (especially on submarines), highly flexible, and carry a single warhead, they are ideally suited for this role. The remainder of the nuclear-armed SLCM force will be available to theater commanders, and may ultimately play a vital role in support of flexible response options in post-INF Europe.

(S) None of these requirements would be eliminated under a nuclear SLCM ban. However, there are differing views as to whether other systems could perform them effectively. One such view is that nuclear SLCMs’ contribution to deterring attack by Soviet land-based naval bombers may be relatively small. This view also holds that SLBMs and mobile ICBMs may be at least as useful for a nuclear reserve role as SLCMs, which would be on numerous ships with other important wartime missions. Assigning SLBMs and mobile ICBMs to the nuclear reserve force role now planned for SLCMs, however, would require an expansion of those forces beyond current plans and currently-envisioned START constraints.

(S) While some Allies—especially littoral states—would support a ban on nuclear-armed SLCMs, others would be concerned about shifting NATO nuclear emphasis more to land-based systems, thus intensifying the controversy surrounding modernization of NATO theater nuclear forces.

(S) In conclusion, nuclear-armed SLCMs make an important contribution to U.S. security. Their deployment on a large number of relatively survivable platforms also helps to enhance strategic stability. There are differences of view on the verifiability and effect on U.S. security of a ban on nuclear-armed SLCMs. Most believe that an effectively verifiable ban or limits on nuclear-armed SLCMs do not appear possible, and that—even if effective verification were possible—it would not be in the U.S. interest to ban or limit such systems. Another view is that it is premature to exclude a ban or limits on nuclear-armed SLCMs. In this view, it is held that continued study of SLCM verification may eventually produce a regime that meets our most serious concerns, and elimination or constraint of the Soviet nuclear SLCM threat may be in the U.S. net benefit.

Conclusion

(S) The SLCM issue does not have to be resolved at the present time. Verification issues relating to SLCMs will be addressed in due course.

Question #3: START Verification

Given present uncertainties about the verifiability of our current START position, would an agreement along its lines be acceptable? What would we be willing to pay to improve verifiability in terms of: additional resources for NTM; negotiating leverage for additional cooperative verification measures; intrusiveness, as the measures will be applied to us; increased risk of no START agreement?

Objectives

(S) The U.S. has stated consistently that effective verification must be the cornerstone of any future arms control agreement. The objectives of an effective verification regime for START must be: to ensure confidence in the agreement; to help deter Soviet violation of the Treaty by increasing both the likelihood of U.S. detection and the difficulty to the Soviets of committing an undetected violation; and to permit timely detection of violations of military or political significance so that appropriate steps can be taken to protect U.S. and Allied security. An assessment of the acceptability of a START verification regime is a function of treaty provisions and definitions, monitoring judgments, the record of past Soviet compliance, our assessment of Soviet incentives and disposition to cheat, assessments of military and political risks, and the availability of safeguards and countermeasures. No arms control agreement will allow perfect verification, regardless of the effectiveness of our NTM assets or the on-site inspection regimes. In developing the currently proposed on-site verification provisions, the U.S. has attempted to balance the requirements for high-confidence monitoring against the necessity to safeguard vital U.S. national security assets.

Monitoring

(S) Monitoring is the process of collecting, analyzing and reporting information gained on Soviet activities relative to the agreement. Monitoring judgments are a measure of our ability to monitor Soviet activities related to the terms of the agreement.

Military Risk

(S) An assessment of military risk requires a judgment on the extent to which Soviet cheating would alter the strategic balance. The effect of uncertainty about Soviet compliance at levels less than would be considered to be militarily significant would also need to be considered to the extent that such violations undermine the viability of the treaty. In conducting these assessments, potential cheating scenarios must be examined for their likelihood, their military (and political) implications, and our ability to detect cheating on a timely basis.

Safeguards and Countermeasures

(S) An assessment of the availability of safeguards and countermeasures is a judgment of the ease and speed with which it would be possible to deny the benefits gained from Soviet violations through diplomatic and military countermeasures; the military or political difficulties which might inhibit effective response; and the constraints which might be imposed by the need to protect sensitive intelligence sources and methods.

[1 paragraph (10 lines) not declassified]

[2 paragraphs (23 lines) not declassified]

An alternative view is that non-deployed missiles in greater numbers than the U.S. will require under the Treaty provide the Soviets no meaningful increase in target coverage and thus no incentive to cheat. Whether changes to verification provisions or numerical limits can ameliorate these risks is open to question, both because such measures may not be identifiable, or because such measures may prove unacceptable to the U.S.

[1 paragraph (12 lines) not declassified]

(S/NF) Another view is that current provisions are sufficient, either because the Soviets would not exceed RV limits without a series of flight tests which would be detectable by NTM or because U.S. requirements for operational flexibility supersede monitoring concerns.

(S) An assessment on the acceptability of a START verification regime cannot be completed until the U.S. has formulated provisions for distinguishing between modernization of a missile and creating a “new type” of missile, and has identified provisions for counting RVs on new types of missiles. Because the U.S. START position may provide for more stringent verification provisions for new versus existing types of missiles, this is a significant gap which must be filled. Finally, some are concerned with current provisions which define and limit throw-weight for new types of missiles.

Heavy Bombers

[7½ lines not declassified]. Additionally, they believe proposals to allow heavy bombers to visit bases of other types greatly degrades any capability to monitor basing restriction. Another view is that basing restrictions are useful only as a confidence building measure.

(S) The policy community understands that in a breakout situation, the Soviets could quickly convert additional aircraft to carry ALCMs; however the U.S. has been willing to accept that possibility, not only to protect U.S. options, but also because bombers and ALCMs are primarily second strike, less destabilizing systems, and because the U.S. retains a similar capability. Moreover, some believe the uncertainty of random, short-notice, on-site inspections provides additional deterrence and detection of possible reconfiguration efforts. With respect to the distinguishability of conventional and nuclear ALCMs, the U.S. emphasis is on protecting U.S. options in both types of programs since the U.S. enjoys a technological advantage in both.

NTM

(S) Since the requirements of START verification are tied directly to U.S. national security needs, expenditure of resources for NTM assets will be justified by the requirements of both. In a period of constrained resources, NTM assets designed to collect necessary information for general intelligence purposes and for Indications & Warning purposes will also assist in the monitoring of central limits, albeit with some resultant competition for target coverage. Difficult decisions on the selection of new overhead assets, therefore, will not be further complicated by the need to procure assets only for verification purposes.

Cooperative Measures

(S) The U.S. has made significant advances in gaining Soviet acceptance of wide-ranging intrusive on-site inspection provisions and other cooperative measures which play a major role in the verifiability of the START agreement. Given their strong desire to retain mobile ICBMs, the Soviets have already accepted major portions of our proposed mobile verification regime. One view is that our negotiating leverage rests in the proposed ban on mobile ICBMs and we should retain that card until an effective verification regime is identified and agreed. Another view is that the U.S. proposed ban on mobiles has had little effect on the Soviet approach to mobile ICBM verification. There is agreement, however, in attempting to establish an effective verification regime, that the risk to the security of U.S. forces and weapon programs posed by additional verification provisions should not outweigh the benefits offered by additional verification measures.

The Risk of No Start Treaty

(S) The U.S. has consistently maintained that failure to secure a START treaty is preferable to acceptance of a treaty which cannot be effectively verified. Despite the internal and international pressures which may be brought to bear in the near term, this position should remain unassailable.

Conclusion

(S) With regard to verifiability, it is premature to render a judgment at this time.

DEFENSE AND SPACE TALKS (DST)

Question #4

In light of the President’s decision on SDI, should the United States retain the fundamental objectives of DST—to preserve the right to pursue SDI within the ABM Treaty; to discuss a cooperative transition; and to establish a future right to deploy defenses without further reference to the ABM Treaty? Should it do so even if this risks blocking a START agreement?

DST Objectives

(S) The U.S. goal in the Defense and Space Talks (DST) has been to help set the stage for a possible future decision to deploy defenses. DST objectives have been: a) to preserve the U.S. right to pursue the SDI program within the obligations of the ABM Treaty in order to determine the technical feasibility of advanced strategic defenses; b) to discuss with the Soviet Union a cooperative transition to effective strategic defenses; and c) to establish the right to deploy strategic defenses if we choose to do so, at a future time without further reference to the ABM Treaty. However, there are two views on whether some of the DST objectives are more important to the SDI program and future U.S. deployment options than others.

(S) DST Objectives Have Equal Value—One view is that no useful distinction between objectives can be made in terms of their importance to any SDI program structured to support relatively near-term deployment decisions and options. According to this view, the DST objectives were developed to complement in equal measure an SDI program structured to support a deployment decision in the early to mid-1990s and initial deployments in the mid- to late-1990s.

(S) DST Objectives Not Equal—Another view is that for purposes of evaluating whether to retain these objectives, a useful analytical distinction can be drawn between those objectives that are likely to be enduring and those that respond to changing circumstances. Enduring objectives include “a”—except, according to this view, preserving broad interpretation testing rights during the nonwithdrawal period (see discussion below)—and “b” above and maintaining ABM Treaty constraints on Soviet ballistic missile defenses until the United States is ready to commit itself to deployment. They have guided our position since negotiations began, enjoy a wide political consensus, are independent of specific Soviet demands, and remain valid for any SDI program with the goal of developing, if feasible, effective defenses which would permit deterrence to be based increasingly on defenses, with the ultimate goal of a defense-dominant deterrent.

(S) According to this view, other objectives, including “c” above and Soviet acquiescence to the right to test under the broad interpretation, evolved and currently are explicitly linked to specific Soviet demands (i.e., the nonwithdrawal period and observance of the ABM Treaty). They evolved in response to changing conditions—programmatic requirements of the SDI program, Soviet demands, and domestic political considerations. The issue, then, is whether these objectives remain valid in all circumstances.

President’s SDI Decision

(S) [The President provided guidance for the Department of Defense regarding the SDI program. The new program is similar to the current SDI program, but includes greater emphasis on a modification of the space-based interceptor concept (e.g., Brilliant Pebbles). With a funding profile of $4.6 billion in FY 90 and $33 billion over the FYDP, this program structure would maintain the possibility of deploying a layered defense in the late 1990s, following a deployment decision in the mid-1990s. In effect, the President has elected to pursue an SDI program that is essentially the same as that pursued by the previous administration.]¹⁹

Preserve Right to Pursue SDI Within ABM Treaty

(S) Assuming Soviet compliance, it is in the strategic and political interest of the United States to maintain ABM Treaty constraints until it is ready to commit itself to deployment of defenses. Consequently, one objective of DST is to preserve the right to pursue SDI within the ABM Treaty. In this context, there are two views regarding the legal rights which DST should preserve during the nonwithdrawal period.

(S) Preserve Broad Interpretation—One view holds that it will continue to be in the interest of the United States to preserve its full legal rights under the ABM Treaty (broad interpretation) and to resist any Soviet efforts to impose a more restrictive legal obligation. This will enable the SDI program to maximize the foundation of technical information upon which development and deployment decisions will be based. Were the United States to agree to a more restrictive legal obligation, such as the narrow interpretation, it would not be able to conduct the system integration tests necessary to support a fully informed deployment decision in the mid-1990s. With the increased SDI program emphasis on Brilliant Pebbles, it is possible that the narrow interpretation also would adversely affect testing planned for the early 1990s. Under these circumstances the U.S. would be forced either to: a) make a decision on deployment before all testing necessary to establish high confidence in the feasibility of an effective defense has been conducted; or b) withdraw from the ABM Treaty in order to complete the testing prior to making a decision on deployment.

(S) Further, the Soviets have attempted to impose an even more restrictive legal obligation than the narrow interpretation. They have suggested in DST that the sides agree either to “strictly abide by the ABM Treaty,” without defining what that means, or agree to a list of permitted and prohibited space-testing activities which specifies precise technical limits to the testing of certain devices. While the Soviets appear to have distanced themselves somewhat from the “list” approach, it does give some insight into what they mean when they demand that the sides “strictly abide by the ABM Treaty.” Besides being largely unverifiable, in many cases the permitted performance parameters of devices identified in the “list” proposal are more restrictive than the narrow interpretation and would prohibit even some space-based tests which the United States plans to conduct under the narrow interpretation.

(S) Narrow Interpretation May Be Sufficient—The other view agrees that the broad interpretation is fully justified from a legal standpoint, but questions the premise that a DST regime that does not implement the broad interpretation during the nonwithdrawal period would seriously undermine the ability of the United States to establish a sufficient degree of technical confidence in defense to support a fully informed decision in the mid-1990s on whether to deploy defenses. The real issue is what level of confidence in defense system feasibility prior to a deployment decision would be acceptable programmatically and politically. For example, the SDI program outlined in the 1989 Report to Congress—which was similar to President Bush’s new program—indicated that a certain level of confidence could have been achieved by continuing until FY 1997 under the narrow interpretation. It is not clear when the restructured program—in particular the Brilliant Pebbles element—would require broad interpretation testing (although it may require the broad interpretation in the early 1990s) or if it could benefit from such testing, since the issue of whether KKV are “based on other physical principles” remains to be resolved.

(S) For tactical purposes, it may be in our interest in DST to continue to seek Soviet acquiescence to the broad interpretation, but we should not assume that either Congress or the Soviets will eventually agree to it. The Soviets have clearly rejected the broad interpretation, although they appear to have changed their position from one of being more restrictive than the narrow interpretation to one approximating the narrow interpretation. In various explanations of the meaning of their current proposal, first tabled prior to the Washington Summit, to observe the ABM Treaty as signed in 1972, they appear to have indicated a willingness to step up to the narrow interpretation. For example, in the DST plenary of November 1, 1988, the Soviets stated for the first time, their understanding that Agreed Statement D “permits the development and testing of only those ABM systems and components based on other physical principles which are fixed, land-based.” In addition, in discussing their “list” proposal with U.S. negotiators, the Soviets stated that space-based KKV testing could be conducted for any purpose. Some have concluded that this indicates that the “list” proposal could permit some space-based KKV testing in an ABM mode that is banned by the narrow interpretation.

Cooperative Transition

(S) All agree that the DST objective of discussing a possible cooperative transition is valid. If the United States were to decide to deploy a large-scale ballistic missile defense system, there would be strategic and political advantages in conducting the transition to defenses in a cooperative fashion with the Soviet Union. This would enhance stability during the transition by increasing the predictability of potential deployments, and perhaps, by regulating the nature and rate of deployments. Representing Soviet approval of the SDI program and the concept of mutual defenses, agreement on a cooperative transition also could bestow political benefits that could reduce domestic and allied opposition to SDI. An additional view is that an uncooperative transition to defenses could be strategically and politically disadvantageous to the United States and its allies.

Right to Deploy Without Further Reference to ABM Treaty

(S) All agree that Article XV of the ABM Treaty provides the United States with the legal right to withdraw from the Treaty (following six months prior notice) and thereafter to deploy ballistic missile defenses beyond the limits prescribed if it “decides that extraordinary event related to the subject matter of [the] Treaty have jeopardized its supreme national interests.” This article was intended to provide a broad right of withdrawal to complement rights under customary international law, providing parties maximum flexibility in future situations.

(S) The terms of Article XV are deliberately general and, therefore, do not provide a specific legal standard which must be met in order to justify a decision by the President to withdraw from the Treaty to deploy effective ballistic missile defenses. Nevertheless, as a practical matter, the President’s flexibility would depend on the degree to which he could make a credible case that he faced “extraordinary events,” that these events are “related to the subject matter of the Treaty,” and that the “supreme national interests of the United States” have been “jeopardized.” Depending on the circumstances existing at the time of such a decision, (especially if the Soviets were in compliance with the ABM Treaty and a START agreement), it might be difficult to derive a politically sustainable rationale for withdrawal under Article XV. There are two views on the significance of this situation.

(S) Removing Article XV an Important DST Objective—One view is that to avoid this potential political hurdle, it is in the interest of the United States to establish the right to deploy defenses in the future, should it wish to do so, without further reference to Article XV of the ABM Treaty.

(S) Removing Article XV Not an Important DST Objective—The other view is that this divisive debate will not be mitigated by achievement of this DST objective. Those who oppose the withdrawal from the Treaty would oppose the exercise of the right to deploy, regardless of whether the Soviet agreed to the right to deploy without further reference to Article XV; they also might question whether it is in our interests to achieve this DST objective which would create a certain risk that the Soviets could gain a military advantage—if they were to deploy and the United States did not.

DST Objectives v. START Agreement

(S) There also are two views concerning the question of whether the United States should continue to pursue the fundamental objectives of DST, even if it risks blocking a START agreement.

(S) No Concessions on DST Objectives—One view is that at this point in the START and Defense and Space negotiations, this question need not be answered now and, in fact, may never need to be answered. In the first place, a START agreement is not imminent. Until START negotiations enter the end game, pressure to reach an accommodation with the Soviets in DST is not likely to increase appreciably. In the meantime, SDI testing will yield valuable information on the potential effectiveness of a ballistic missile defense which may have an important bearing on the Administration’s decision on this question. Ultimately, if the Soviets remain adamant in their START/DST linkage, the decision will hinge on an assessment of the importance to U.S. national security of START relative to the deployment of effective ballistic missile defenses.

(S) Second, it is not clear that a START agreement would require the United States to make concessions on one or more of its DST objectives (just as they were not required to achieve an INF agreement). If the United States were perceived by the Soviet Union to have achieved a favorable consensus on President Bush’s SDI program decision, the Soviets might eventually view cooperation with the United States—with the object of influencing the course of the SDI program—to be an approach that would better serve Soviet interests than taking an obstructionist approach on DST and/or START.

(S) Finally, as a practical matter, if a START agreement is reached in return for U.S. concessions on its DST objectives that reduce the viability of SDI, such an exchange could create a political climate in the U.S. Senate that would make it more difficult to achieve the 2/3 vote necessary to ratify a START agreement.

(S) DST Concessions May be Necessary—The other view is that because U.S. agreement to a nonwithdrawal period may be a necessary quid pro quo for Soviet agreement to START, and because the Soviets may require limitations on testing during the nonwithdrawal period, we will need to decide whether to insist on the broad interpretation, even if it risks blocking a START agreement. Assuming that a START agreement could be concluded in the next two years—well before the projected SDI program deployment decision point—we may need to reach an accommodation in the next year. Specifically, in the context of postponing SDI deployments in order to achieve START reductions, the following questions need to be answered: a) how long we can live within the narrow interpretation; b) when—and for what purpose—we will need the broad interpretation; and c) the duration of the nonwithdrawal period.

(S) Second, it is not clear that we could get a START agreement without adjustment to our DST position. The Soviets view the linkage between offensive reductions and defensive deployments as much more organic than the artificial linkage between INF and defenses, and may not delink the two without assurances regarding SDI. The United States already has lost much bargaining leverage as a result of the domestic debate over ABM Treaty interpretation. As long as the Soviets continue to believe that Congress will constrain SDI at least as effectively as a D&S agreement, it will see no benefit to agreement to the broad interpretation or to the right to deploy without further reference to the ABM Treaty.

(S) Finally, with a well-planned strategy, the United States could achieve both a START agreement and adjust its position in DST in a manner that would neither reduce the viability of SDI nor jeopardize START ratification. In addition, deep reductions in START might reduce the requirements for a future deployment of U.S. defenses, such that some U.S. concessions in DST might not reduce the viability of future U.S. ballistic missile defenses. In view of the fact that the Congress as a whole, and the Senate in particular, have remained skeptical of the necessity of the broad interpretation, U.S. acceptance of less than the broad interpretation—i.e., the narrow interpretation—in exchange for a START agreement may meet with approval.

(S) The issues related to Question #4 on DST will be presented for review and decision in the context of NSR–14.

[Omitted here are discussions not related to START.]

Tab B

Paper Prepared in the National Security Council²⁰

Washington, undated

EXTRACT FROM NSR–12

IV. The Role of Arms Control in Promoting U.S. Defense Objectives and Strategy: This report should examine the degree to which arms control can and should be used to advance and support U.S. defense objectives and strategy. The following questions should be addressed:

General Issues

— To what extent can arms control enhance national security and promote strategic stability? (C)

— What is the potential value of arms control in promoting such U.S. defense objectives as: (1) enhancing the predictability of the threat; (2) reducing or bounding actual capabilities of adversaries; (3) constraining the diffusion of threatening technologies; (4) restructuring forces in stabilizing or otherwise desirable ways; (5) enhancing force survivability; and (6) channeling the long-term military competition in directions that ease U.S. security concerns and/or that take advantage of U.S. competitive strengths without imprudently hampering emerging technologies? (C)

— If arms control is to be used to limit or reduce both the strategic and conventional threat, assess the contribution or disadvantages of: (1) seeking deep reductions in Soviet forces; (2) encouraging Soviet force restructuring; (3) reducing Soviet weapon-to-target ratios; and (4) seeking to eliminate certain specific threats such as, for example, depressed trajectory missiles? (C)

— Can other options for negotiated threat reduction be identified which differ from traditional approaches? In addition to negotiated agreements, are there less formal arms control approaches (e.g., parallel, informal restraints) that can contribute to threat reduction? In addition to traditional arms control restraints, such as limits on forces and force levels, are there additional devices (e.g., operational constraints, confidence-building and openness measures, etc.) that could serve U.S. defense interests? (C)

Specific Issues. The arms control review should either affirm or suggest options for modifying the premises underlying our approach to current or prospective negotiations. It should not, as part of this effort, address particular details of U.S. negotiating positions. The review should examine among other things:

1. START/Defense and Space

— Is the existing U.S. approach to START consistent with the defense objectives and force structure analysis discussed in Sections II and III of this review? (C)

— At what point do negotiated reductions in U.S. strategic force structure cease to be consistent with U.S. security requirements? What kind of force limitations should be sought, and what avoided, for their impact on the ability of the U.S. to develop and deploy systems that exploit our long-term competitive advantages? (C)

— What is the security relationship and dynamic, if any, between START and the Conventional Forces in Europe (CFE) talks? (C)

— Is the existing U.S. approach to the D&S negotiations consistent with the defense objectives and force structure analysis discussed in Sections II and III of this review? (C)

— To what extent are continued constraints on ballistic missile defense in the U.S. interest? (C)

2. Other Current or Prospective Arms Control Issues. The review should address security considerations that bear on:

— Further negotiated reductions in theater nuclear systems in Europe;

— Further restrictions on nuclear testing;

— A global ban on chemical weapons;

— Conventional arms reductions in Europe; and

— Restraining the proliferation of destabilizing technologies (nuclear and chemical/biological weapons and ballistic missiles). (C)

1. Source: George H.W. Bush Library, Bush Presidential Records, National Security Council, H-Files—NSR Files, OA/ID 90010–006, NSC/DC 027—May 20, 1989—NSC/DC Meeting on Security and Arms Control. Secret. Sent for action. Sent through Kanter. No minutes for the May 20 Deputies Committee meeting were found.
2. Gates placed an asterisk beside this point, underlined “Our goal should be stability and reduced risk of war, not reductions,” drew an additional line under “reductions,” placed two vertical lines in the right-hand margin and wrote “abandon obj. of elim of nuc. weapons” in the left-hand margin.
3. Gates placed an asterisk beside this point, underlined “mobility” and “strategic defense are,” and wrote “not taking down SS-18s or de-MIRVing/T.Weight/MIRV limits not way to deal w. ICBM survival” in the right-hand margin.
4. Gates underlined “START allows us to keep an adequate SSBN force.”
5. Gates placed an asterisk beside this point, underlined “to air breathing systems,” and drew a downward diagonal arrow pointing to this paragraph in the left-hand margin.
6. Gates placed a checkmark beside this point, circled “1600,” and drew two short vertical lines in the left-hand margin beside the first sentence.
7. Gates drew two short vertical lines in the right-hand margin beside this sentence, and wrote “discounting rule” beside them. He also drew a vertical line in the left-hand margin beside the last two sentences of the point, and wrote: “? CMs—range, nuc vs. conv., bombers—US 10 cms, Sov. capable”
8. Gates underlined “encourage a long-term Soviet shift to air breathing systems” and drew two vertical lines in the left-hand margin beside this point.
9. Gates drew a bracket in the left-hand margin beside the four points in this section.
10. Gates drew a vertical line in the left-hand margin beside this point.
11. Gates drew a vertical line in the left-hand margin beside this point, and wrote in the left-hand margin beside the first four points of this section: “W/o serious JCS assess. of mil-risk of cheating scenarios—no ratif as long as verif. prob.”
12. Gates underlined “we will be testifying on a treaty that we will assert is in our interest even though not all provisions of that are fully verifiable.”
13. Gates wrote “chg. definition of verification” in the right-hand margin beside this paragraph, and drew two short vertical lines in the left-hand margin.
14. Gates wrote in the right-hand margin below this paragraph: “cost, access, doability.”
15. Gates underlined “the following objectives remain valid.”
16. Gates drew a bracket in the left-hand margin beside points A-C, placed a checkmark to the left of the bracket, and wrote: “reaffirm.”
17. Gates did not indicate his preference.
18. Secret; Noforn; Wnintel. Brackets are in the original.
19. This description of the President’s decision is based on unofficial sources. It will be revised, as appropriate, when an official source becomes available. [Footnote is in the original.]
20. Confidential.