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

63. Memorandum From the Legal Adviser of the Department of State (Sofaer) to the Special Advisor to the President and the Secretary of State on Arms Control Matters (Nitze)¹

Washington, October 3, 1985

SUBJECT

• Extent of U.S. Flexibility Under the ABM Treaty to Develop, Test and Deploy SDI Components

The differing positions of OSD² and ACDA³ may obscure significant areas of agreement. Both agencies agree on the following propositions:

— Compliance must be based on objective assessments of capabilities that support a single standard for both sides, not on subjective judgments as to intent that could lead to a double standard.

— The ABM Treaty permits research short of field testing of a prototype ABM system or component, which is the research being conducted under the initial phase of the SDI program.

— The ABM Treaty restricts defenses against strategic ballistic missiles, but not defenses against cruise missiles or non-strategic ballistic missiles (provided they are not given ABM capability).

— Fixed land-based ABM systems may be developed and tested, subject to the restrictions of Articles IV and V(2); and fixed land-based ABM launchers, interceptors and radars may be deployed, subject to the limits of Articles III and V(2).

The issue that you asked me to address is whether the ABM Treaty prohibits the development, testing, or deployment of SDI systems or components that have ABM capability and which are “based on other physical principles” than those current when the Treaty was ratified in 1972.⁴ ACDA contends that the Treaty unambiguously prohibits development, testing, and deployment of such systems or components; in addition, it argues that the negotiating record and subsequent U.S. statements establish that such development, testing, or deployment would violate the Treaty. OSD contends that the text of the Treaty is ambiguous with respect to systems or components based on new physical principles. OSD apparently concludes that the negotiating history demonstrates that development, testing, and deployment of such systems or components are not controlled by the Treaty, except for the requirement in Agreed Statement D that the parties discuss specific limitations on such systems and components.

This memorandum analyzes three textual arguments regarding U.S. flexibility under the ABM Treaty to develop, test, and deploy an SDI system: ACDA’s and OSD’s readings, as well as a third interpretation that construes the Treaty to allow testing and development, but not deployment, of an SDI system or components. In the near future, I will be sending you a memorandum analyzing the negotiating history of Agreed Statement D and the post-negotiation statements of the parties.

Textual Analysis

The central feature of the Treaty is the commitment not to deploy a territorial ABM system. Article I(2) provides that: “Each party undertakes not to deploy ABM systems for a defense . . .” of its territory. Article III prohibits deployment of ABM systems or components, except for certain fixed land-based systems and components. Article II(1) defines “an ABM system,” “for the purpose of this Treaty,” in functional terms (“a system to counter strategic ballistic missiles or their elements in flight trajectory”) and immediately thereafter refers to contemporaneous technology: “currently consisting of” ABM interceptor missiles, launchers, and radars, which are defined in terms of both function and mode of testing.

1. ACDA’s Interpretation

ACDA’s textual interpretation rests on two lines of argument. First, ACDA argues that, because Article II(1) defines ABM system functionally and then goes on to state that it is “currently consisting of” interceptor missiles, launchers, and radars, the definition necessarily was meant to include all systems, present and future, that have the requisite capability. “If the term ABM system only applied to systems utilizing technology current as of 1972, there would have been no reason to use the phrase ‘currently consisting of.’” (Graham Memo, 9/20/85, at 3.) ACDA argues that, a fortiori, the commitment not to deploy (contained in Articles I and III) applies to all ABM systems as functionally defined in Article II and to their components, whether or not specifically listed in Article II(1). ACDA also relies on the agreement in Article V(1) “not to develop, test, or deploy ABM systems or components which are sea-based, air-based, space-based, or mobile land-based.” ACDA contends that the functional definition of an ABM system in Article II(1) applies to Article V(1) as well, and that the U.S. thereby has agreed not to develop, test, or deploy any system or component that could serve an ABM function, based in the sea, air or space, or mobile land-based.

This argument does not establish unambiguously that all future systems are covered by Articles II(1) and V(1). Article II(1) may, as OSD contends, use the words “currently consisting of” simply to describe what systems are covered by the other provisions of the Treaty, in which case Article V(l) could be said to relate only to current systems and components. Indeed, Article II(2) supports this view. That paragraph provides as follows:

The ABM system components listed in paragraph 1 of this Article include those which are:

(a) operational;

(b) under construction;

(c) undergoing testing;

(d) undergoing overhaul, repair or conversion; or

(e) mothballed.

This enumeration is limited to existing types of ABM system components. Components are included within the Article II(1) definition even if they are not yet fully operational (e.g., because they are being tested or constructed). But the paragraph omits devices as to which only research is under way, as well as devices that have not yet been envisioned or designed. Had the negotiators intended to bring future technologies within the purview of Article II(1), they easily could have added “undergoing research” or “developed in the future” to the litany in Article II(2).

Moreover, as OSD contends, ACDA’s reading would render Agreed Statement D superfluous. If Article III deals with all land-based systems or components based on present or future physical principles, and if Article V(l) absolutely prohibits development, testing, and deployment of all systems, present or future, with other basing modes, then nothing is left for coverage by Agreed Statement D. OSD argues that it is in Agreed Statement D that the parties “treated separately” the problem of future systems and components, and its language lends facial support to OSD’s position:

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 [establishing the Standing Consultative Commission] and agreement in accordance with Article XIV of the Treaty [providing for amendment].

ACDA responds to this argument as follows:

Article III of the ABM Treaty, along with the ABM Protocol, limits the deployment of an ABM system to a numerically limited number of fixed land-based ABM missiles, launchers, and radars situated in one geographically limited deployment area. Since the Article specifies the type of ABM components that can be deployed, any other types—including those based on other physical principles—cannot be deployed without amendment of the Treaty. The purpose of Agreed Statement D is simply to clarify this point. Deployment of 100 large, fixed land-based lasers in place of 100 fixed land-based ABM interceptor missiles would create a different balance than that established by Article III. (Graham Memo, 9/20/85, at 3.)

Treaties are sometimes read to contain statements intended “to clarify” a point covered by other provisions. But such redundancy tends to undermine the viability of a proposed construction—especially where no clarification is needed. A reading that renders a provision redundant or superfluous cannot be said to make the treaty unambiguous. ACDA might respond that redundancy is to be expected in an “Agreed Statement,” which appears along with other “Common Understandings.” This observation lacks force here, however, because Agreed Statement D—unlike Agreed Statements A, B, and C—does not contain the word “understand”. Rather, it is written with the operative phrase “the Parties agree,” which is closer to the language in the Treaty itself that “the Parties undertake.” Similarly, the Statement’s phrase—“In order to insure fulfillment of the obligation not to deploy”—is close to the analogous phrases in Article VI (“To enhance assurance of the effectiveness of the limitations”), and Articles IX, XII (1), and XIII. In short, Agreed Statement D looks more like a substantive obligation than like merely another shared interpretation.

Even if “ABM system” in Article II(1) should be read to mean more than just the particular systems in existence when the Treaty was signed, it does not follow that Article II(1) should be read to include all future systems or components. The phrase can be read to go beyond those systems currently consisting of interceptors, launchers, and radars, but to fall short of all future systems. In particular, the definition could be understood to include all systems other than the current ones that could be developed through the application of existing physical principles, but not those systems that could be developed only by resort to other physical principles.

ACDA’s second argument also rests in part on its construction of Article II(1), but is a more comprehensive analysis based on the Treaty’s other provisions and purposes. ACDA views Article III(1) as prohibiting any deployment of fixed, land-based ABM systems except in accordance with that paragraph’s specific limitations. It reads Agreed Statement D as permitting the development and testing of fixed land-based ABM systems and components, based on other physical principles, but precluding their deployment until after discussion and agreement on specific limitations. This interpretation presents Agreed Statement D as the Treaty’s means of ensuring implementation of the fundamental purpose of Article III, which (according to ACDA’s view) is to prevent the deployment of any number of ABM systems, of any type, in any place, except for the fixed land-based systems in the places and numbers specifically authorized.

Though ACDA’s interpretation is strengthened by this attempt to harmonize Article III(1) and Agreed Statement D, it nevertheless remains only one possible construction of the relevant provisions. ACDA’s reading is not necessary to ensure implementation of the Treaty’s fundamental purpose. In Article I(2), the Parties agree not to deploy ABM systems for territorial defense, except as provided in Article III. Article III also limits deployment, and Agreed Statement D was specifically adopted “to insure fulfillment of the obligation not to deploy ABM systems and their components except as provided in Article III . . .” This central purpose of the Treaty—to limit deployment—is satisfied by any reading that prevents the deployment of future systems and components; it does not also require a prohibition on their development and testing. Nothing in the structure or stated purposes of the Treaty requires an absolute prohibition against the development and testing of all systems or components, other than those described by Article III.

2. OSD’s Interpretation

OSD makes no claim that its interpretation is the only plausible one of the Treaty’s language, and it therefore acknowledges that resort to the negotiating history is appropriate. Substantial doubt exists, however, as to whether one of OSD’s claims is plausible enough to require reference to the negotiating record. The Treaty on its face casts grave doubt on OSD’s contention that the Parties are permitted to deploy systems or components based on other physical principles after discussion but without agreement on specific limitations. The ABM Treaty is of unlimited duration (Article XV(1)), and its most fundamental objective is to prevent deployment of systems that can serve the function defined in Article II(1). To read Articles II(1) and V(1) to apply only to current systems or components, and to read Agreed Statement D to require only discussion prior to deployment of future systems or components, would tend to undermine the Treaty’s essential purpose.

Agreed Statement D states, moreover, that its purpose is “to insure fulfillment of the objective not to deploy ABM systems and their components except as provided in Article III,” and it requires both discussion of “specific limitations” for future systems or components under Article XIII “and agreement in accordance with Article XIV of the Treaty,” upon the creation of any such systems or components. The text suggests a general prohibition against deployment of such systems or components, as well as a commitment to discuss and to agree on specific limitations such as those contained in Articles III, IV, and VI.

Nevertheless, OSD’s theory is not absolutely precluded by the text of the Treaty: one could reasonably claim that parties are often reluctant to enter into legally binding obligations with respect to hypothetical, future problems. The Parties may have assumed that the same factors that led them to agree on limits for current land-based systems and components, as well as to prohibit all otherwise-based current systems and components, would ultimately have led to agreement on limiting the deployment of future systems and components. The Treaty contemplates, moreover, that agreement about fundamental issues might not be possible, and each Party retains the power to terminate in order to protect its supreme interests.

3. An Alternative Interpretation

The problems identified with respect to both ACDA’s and OSD’s interpretation suggest a third construction that is at least as plausible as the other two. Neither ACDA nor OSD adequately deals with the different degree of concern reflected in the Treaty with respect to deployment, on the one hand, and with respect to activities short of deployment (development and testing), on the other. A powerful argument can be made, based on Articles I(2) and III(1) and Agreed Statement D, that the Treaty prohibits deployment of any ABM system or component, current or future, other than as provided in Article III, but that it permits the “creation” of systems or components based on other physical principles and would require discussion of and agreement on “specific limitations” on such systems or components before deployment.

This interpretation is consistent with the textual argument based on the Treaty’s express and fundamental purpose: to prevent development of systems or components other than as specified by Article III. It adopts OSD’s approach to Article V(1), in which each Party undertakes “not to develop, test, or deploy ABM systems or components which are sea-based, air-based, space-based, or mobile land-based.” The theory rests, therefore, on the proposition that Article II(1), and therefore Article V(I), relate only to “current” systems and components (with the word “current” encompassing any system or component based on physical principles known at the time the treaty was signed).

The Treaty in fact consistently uses a definition of “ABM systems or their components” that is at odds with ACDA’s expansive view of Article II(1). To begin with, Article II(2) includes in its description of covered systems or components only those that were already developed and undergoing testing prior to the Treaty’s adoption. Article III(l) uses the phrase “ABM systems” to describe systems consisting of missiles, launchers, and radars, the components of “current” systems. Likewise, Article IV, which makes clear that Article III does not apply to systems and components used for development or testing at agreed ranges, limits to fifteen the number of “launchers” at those ranges, thereby suggesting that it is addressed only to current technology. Even Article V(2) is written with current technology in mind, as it controls the development, testing, or deployment of launchers of multiple ABM interceptor “missiles” and automatic or rapid reload “launchers.” In short, the Treaty is permeated with references to, and concern about, current systems or components. One can therefore argue plausibly that the Treaty also refers in Article V(1) to current systems and components, prohibiting the development, testing, or deployment of such current systems or components “which are sea-based, air-based, space-based, or mobile land-based.”

In sum, both ACDA’s and OSD’s constructions of the Treaty present textual difficulties. Both are logically defensible, although OSD’s claim that deployment of future systems is permitted borders on the untenable. The view advanced here—that the Parties agreed not to deploy future systems or components, but that they permitted development and testing of future systems or components—is also a reasonable interpretation. Reference to the negotiating record is therefore necessary to determine which of these three views best reflects the Parties’ understandings.

1. Source: Department of State, Ambassador Nitze’s Personal Files 1953, 1972–1989, Lot 90D397, October 1985. Secret. All brackets are in the original.
2. See Document 56.
3. See Document 55.
4. Another important issue that may remain disputed is the scope of permissible research on, and impermissible testing of, traditional ABM interceptors, launchers or radars in the employment modes proscribed by Article V(1). [Footnote is in the original.]