Foreign Relations of the United States, 1981–1988, Volume XLIV, Part 1, National Security Policy, 1985–1988

8. Memorandum for the Record¹

West Point, New York, December 20, 1984

SUBJECT

• An Arms-Control Context for SDI

The assumptions linking SDI and arms-reduction negotiations seem to be these:

(1) The President wishes to make significant gains in negotiating reductions in offensive arms.

(2) The President wishes to continue with the SDI as a strong program.

(3) The Russians will not negotiate seriously on offensive arms unless we are willing to negotiate seriously on SDI.

It is the third assumption which seems to make the first two logically incompatible, and to place us in a bimodal situation with regard to SDI: preserve it, and undermine the negotiations; or consider it fair territory for trading, and thus lose it. The problem, then, is whether there is a middle ground.

WHAT IS THE SDI?

I believe the key to this dilemma lies in the definition of the SDI. Is the SDI

• (a) A long-range R&D program to determine whether nationwide defense is possible, and how it could be done?
• Or (b) A goal-oriented development program specifically charged to produce (that is, design and deploy) a nationwide defense, either as soon as possible or within some fixed timespan?

If the correct answer is (b), then the dilemma above is real, for we cannot in good faith profess that we have anything to bargain with unless we intend to give up the whole thing. If, on the other hand, the correct answer is (a), then there are a good many things which we could realistically discuss with the Russians, things which would not materially damage the SDI or retard its success, but which would provide the Russians with real incentive to bargain (given their manifest concern over SDI).

Definition (a) is in fact consonant with the President’s expressed desire on strategic defense: “. . . a long-term research and development program to begin to achieve our ultimate goal of eliminating the threat posed by strategic nuclear missiles.” The significance and wisdom of this formulation lie in the fact that it recognizes the actual state of technology: We don’t know today what sort of system will accomplish the objective of nationwide defense, nor even whether it can actually be accomplished.

An important turning point in the logic is a recognition of this fact, which is a strong consensus among real experts in the field. It was said plainly in the report of the Fletcher Commission,² and it was said in the year-long DOD study of BMD chaired by Johnny Foster (which concluded about the time of the Fletcher Commission). If you accept the fact that we do not know yet about the feasibility of nationwide defense, then the logical orientation of the SDI is (a). In either case, you must begin with R&D to provide the data for analysis and evaluation of alternatives. The real difference between (a) and (b) is thus that in (b) we have already decided that some defense which satisfies our criterion will be found and that we will deploy it. The general argument in favor of (b) is that, having decided now to go ahead full steam, we can speed up the process by doing, in parallel with the R&D to get the answers, all the advanced development and testing technology will allow. Much of that work will be wasted because it will pertain to systems that don’t show up in the final solution, but we will move toward deployment faster. In fact, this argument is not correct. Parallel demos of undeveloped technology nearly always prove disastrous to the programs they belong to. The demos change the direction of work because they overrun their costs and suck up the money from the R&D. Having thus destroyed the potentially more promising options, they leave one with a commitment to a second-rate solution, engineered into a system too soon; thus we get the poor version of the poor option.

We must pause a moment in the argument to make an important distinction about demonstrations and testing. The goal enunciated by the President is nationwide missile defense. A different goal, enunciated by prominent SDI supporters from Senator Garn to Henry Kissinger, has been reinforcement of deterrence by reducing vulnerability of our military forces. This latter job is something we can do with existing technology—interceptor missiles with nuclear warheads—and we can test our systems without violating the ABM Treaty. Because of strictly numerical limitations, we can’t deploy a viable system without violating or revising the treaty. I shall discuss the relevance of this near-term technology at the end of the memo. The important distinction is that when I speak of big demonstrations now I am speaking of demos of immature technology, proposals for things which, after considerable further improvement, may be part of solving the problem of nationwide defense, but whose pursuit now actually constitutes a danger to the success of the SDI.

Thus far I have argued that R&D is the indispensable core of the SDI in the near future, by which I mean at least five to ten years. I have also suggested that (a) is thus the logical SDI strategy. I will now make several points about disadvantages of (b) which I claim settle the case. I will then show how choosing (a) gives the U.S. a plausible negotiating position to satisfy both the President’s wishes. Finally, I will address the question of what we should be doing in the near term.

Disadvantages of SDI Definition (b)

I shall enumerate five severe disadvantages of this approach.

(1) Arms Control Dilemma As explained above, approach (b) reduces us to the bimodal position of bargaining with the whole SDI or refusing to discuss it. If the interpretation of the Russian position is correct, we cannot keep SDI and achieve arms reductions.

(2) Domestic Risks The SDI has already attracted an entirely disproportionate amount of public criticism and comment. Because of domestic politics, going ahead with (b) now means that we risk losing the entire SDI. Maintaining an R&D program as requested by the President largely defuses further criticism, since even most liberal critics admit that absolute defense would be good if it were possible, and object to various problems of transition. But announcing that we know now that it will work is doubly dangerous. First, this contention can be disproved to the satisfaction of any good scientist, and in debate to any intelligent congressman. Second, the Administration opens itself to a variety of criticisms based on the unstable nature of partial defenses and transitions, arguments which have been made thus far only as possibilities: in short, a highly effective rallying-point will have been created for enemies of the program. Particularly in its currently-constituted form, the SDI is not strong enough technically or programmatically to deal with either of these problems; the combination of them is very likely to permit the enemies of strategic defense to wreck the entire effort.

(3) Poor Product For reasons explained in the section above, demos of immature technologies are not merely wasteful, they actually lead to a poor solution to the proposed task. In short, if we get any nationwide defense at all, we will have elected to develop one which comprises a poorly integrated complex of poorly selected ideas. It won’t work well, and it will rapidly become dangerously obsolete. All this also adds to the disadvantage above.

(4) Ally problems Our allies have expressed sizeable fears about the success and implications of SDI. Although these fears are not well-founded, they are real. If SDI is publicly elevated to demonstration-and-test status, and if any of these demos fail, the allies will become truculent in their attitude toward SDI. Thus, definition (b) places great pressure on all the demos to succeed, distorting the technical plan (degrading or delaying the objective) and needlessly risking the credibility of the entire SDI. Giving alliance politics leverage as a criterion in managing the SDI is another way of gambling with its success.

(5) Real Soviet Responses If we understand the Russians correctly, they are seriously worried about the SDI. So far this is fine. But if we proceed with a large program of tests and demos which may not lead anywhere, and which will certainly not lead anywhere for some time, what will be their response? Since the ABM Treaty, the U.S. has had no deployed missile defense. The Soviets have a limited BMD deployment, twice upgraded. They could proliferate this system easily. Perhaps more important, the Soviets also have an interceptor system, the SA–12, which represents a powerful, current military capability. It has performance to spare in killing Pershing II, and it probably could have (with additional acquisition radar) disturbingly good performance against SLBMs. They are building SA–12 as anti-Pershing weapons through a loophole in the Treaty. If the Soviet response to SDI(b) is to proliferate SA–12 as an anti-SLBM, they will have a real defense capability while we have one on paper. Our lead-time to deployment of any similar system is longer than five years.

Advantages of SDI Definition (a)

Choosing (a) minimizes the problems (2), (4) and (5) and eliminates (3). The pertinent question is, what does it do for (1)? It allows us to adopt the following bargaining position with the Russians:

1. We will not violate the ABM Treaty. We are embarked on long-term R&D, and so are you. If we reach a positive answer sometime in the future, and decide that a system based upon that answer is good enough to be worth deploying, we will give the required notice before withdrawing from the Treaty. We believe that you would do the same, which is why the withdrawal clause was mutually agreed to. Meanwhile, our R&D efforts will respect the limits of the Treaty. We would like to know the character and extent of your R&D programs, since ours has been announced and yours remains a very large but completely secret undertaking.

2. We will not violate the treaty banning weapons of mass destruction in space.

3. We are willing to discuss limitations on large-scale demonstration of SDI-relevant technologies, particularly demonstrations in space.

4. We are willing to discuss limitations of space battle stations for BMD.

The first two items will not interfere materially with our R&D program and can be volunteered unilaterally. The effect will be to establish our bona fide intent with the Russians, with our allies, and with Congress. The ABM Treaty continues to be valuable to us as a tripwire to Soviet BMD breakout or SA–12 deployments as SLBM defense, at least until we have a near-term capability of our own.

The third item comprises a large class of elaborate space demos almost all of which are premature or can be tested other ways. Specifics are complex for a memo of this length, but there are only a few limited prerogatives which we would have to retain in order for our R&D of the next ten years or so to lead us to the answers we need. Presumably the new treaty, like the ABM Treaty, would have a renewal time or withdrawal clause in case we reach a stage where big tests in space are essential. Thus, there are a large number of things here which will appear to be substantive to the Russians but whose loss would not injure SDI, and which we can bargain with in good faith, one item at a time.

The fourth item represents an apparently large restriction which, if handled properly, is none at all; thus we can appear to be offering a lot. Teller’s dictum says that satellite battle stations are all cheaper to shoot down than they are to put in place. Some people may argue with this dictum, but the real experts (Johnny Foster, for example) agree with it. We should be careful to preserve special cases, however: satellites which do not themselves house weapons systems but which may be key parts of BMD systems. Two examples are: relay mirrors for ground-based lasers and high-altitude satellites containing decoy discrimination techniques, such as particle beams.

Thus, the latter two items do contain sensitive details which must be handled carefully by experts, but they offer up what appears to be considerable ground for bargaining. None of what we are actually offering should damage the attainment of SDI’s goals.

A potential disadvantage of SDI(a) will no doubt be raised, so I should comment on it here. There is a contention that a program without big demonstrations can never get sufficiently big appropriations, or maintain its technological momentum for long. As far as the appropriations go, I have already argued that the demos themselves represent a threat to the health, if not the life, of the program. The early demos constitute more of a danger than an assistance. Note that big chemical laser proponents have tried for years to get funding for big space demos. Not only have they not succeeded, but the chemical laser program has demonstrated negligible progress during the period. As far as technological momentum, it is true that some big projects are necessary, but we can build plenty of big projects in laboratories or on the ground (big ground-based lasers, for example). Such projects can serve as technological centerpieces for Congress, if such are really deemed necessary.

Near-Term Proposals

I recommend that we consider an additional position for discussion with the Russians:

5. We will consider bans on testing and deployment of space-based ASATs, but not of ground-based or air-based ASATs.

And an additional course of action, which might lead to discussion with the Russians:

6. We should revitalize our conventional hardsite BMD development program, and prototype a working system. When we have an adequate benchmark, we should consider the relative benefits of re-negotiating the ABM Treaty limits, in terms of numbers (of sites, interceptors and radars) and basing (mobility and deception), for “conventional” hardsite defense of ICBM silos.

Position 5 permits us to refine and upgrade our current ASAT and to develop and upgrade ground-based laser ASATs. It prevents the Soviets from threatening our high-altitude satellites with anything but a large ground-based laser, a technology in which we have a substantial lead. In short, there is no real reason why we would prefer space-based ASATs over the ones we have now in development. This agreement would permit the Soviets to keep the ASATs we can’t verify anyhow, but stop them from developing ones we can actually verify.

Position 6 essentially argues that we close the gap on the Soviets’ existing advantage in real military capability for missile defense. We can actually construct a defense that will provide sufficient ICBM survivability, but it will require changing several limits in the ABM Treaty—changing a few numbers, but leaving it qualitatively the same. The Soviets have a big lead-time-to-deployment advantage (4–7 years), because we never prototyped our system. More important, when SDI started up, conventional hardsite—the only BMD system with demonstrated military effectiveness—was cancelled.

Such a system could address the principal concern of Dr. Kissinger and Senator Garn, and do it sooner than any result from the nationwide defense program is likely to be ready. Everyone agrees that terminal hardsite defense of ICBMs is stabilizing. The technology to do that job, to eliminate the unstable vulnerability, is actually available. Critics will be concerned about meddling with the Treaty, but not nearly so concerned as if we embarked on SDI(b), in which we would effectively be announcing an intention to withdraw from the treaty, and for an objective those same critics regard as destabilizing rather than stabilizing.

There is absolutely no reason why we should permit the Soviets to maintain and improve their capability while we throw ours away, just because we hope to get something better. There is, however, good reason for concern about the Soviet capability, because of the linkage between offense and defense. In the interim between now and the success of the SDI effort—an interval whose length is uncertain—reductions in offensive arms will act to the benefit of the USSR unless we can have the option for an equivalent defense, to deter or match Soviet BMD breakout. At the very least, we must be able to evaluate the capability of their defense by more reliable means than paper studies.

Whether or not we pursue actually deploying hardsite, we should revive our activities in “conventional” hardsite BMD, prototype a system, keep modernizing it, and test pen-aids against it to ensure we can get through Russian defenses. We should not allow the Russians a large unilateral advantage in defensive capability in the near term. While we are looking for the better thing, let us keep the good one.

Thomas H. Johnson
Director, Science Research Laboratory

1. Source: Reagan Library, Tyrus Cobb Files, Subject File, SDI 84–85. No classification marking. Drafted by Johnson. Poindexter sent the memorandum to Matlock on December 28 under cover of an undated handwritten note: “Jack, this is an interesting paper—lot of good thoughts. As you stated there are some points with which I disagree but I do think substance is basically correct. JP” (Reagan Library, Jack Matlock Files, Chronological Files, 1980–1986, Matlock Chron December 1984)
2. Documentation pertaining to the Fletcher Commission and the Department of Defense study of Ballistic Missile Defense chaired by Foster is scheduled for publication in Foreign Relations, 1981–1988, vol. XLIII, National Security Policy, 1981–1984.