Policy Adopted in 1970s, Still in Effect Today Despite Danger of False Alerts
Top Defense Official in 1978 Called It “Dangerous, Wrong, and Ineffective”
Reagan Directive Called for Leaving Soviet Planners “with Strong Uncertainty” about U.S. Response to Warning
Washington, D.C., June 11, 2019 – “Launch-on-warning,” a feature of U.S. nuclear warfighting strategy since the late 1970s, has frequently faced intensive criticism because of the high risk of accidental launches and uncontrollable outcomes, including massive casualties, according to recently declassified records posted today by the nongovernmental National Security Archive.
Yet, successive presidential administrations have stood by a prompt-launch approach. The new documents, obtained through archival research and declassification requests, are combined here with an earlier National Security Archive e-book to further illuminate high-level thinking about a key aspect of nuclear war planning the public rarely hears about.
Two newly declassified highlights in the posting are White House adviser William Odom’s critique of launch-under-attack and President Ronald Reagan’s National Security Decision Directive 13, which provided criteria for nuclear war planning, including launch-on-warning as a way to keep Moscow "uncertain."
_____________________________________________
In the fall of 1979, as the Carter administration was revamping U.S. nuclear strategy, Lt. Colonel William Odom, an official on the White House national security staff, raised doubts about the recent inclusion of a launch-under-attack option in the nuclear war plan, the Single Integrated Operational Plan (SIOP). In a memorandum published today by the National Security Archive, Odom argued against rapid launching of Minuteman missiles in response to electronic warnings because warning systems were “just not good enough to let us know that our ICBMs are under attack.” Moreover, according to Odom, not enough thought had been given to the launch-under-attack targets: they would strike “empty Soviet silos.” Further, if the U.S. launched a follow-up nuclear attack, the nuclear explosions and radiation effects caused by a launch-under-attack strike could destroy incoming U.S. missiles through “fratricide.” Launch-under-attack as a SIOP option, Odom advised, should be cancelled.
Odom’s advice was disregarded but since then, presidents, presidential candidates, former officials, and defense policy experts have questioned the U.S.’s reliance on a high-alert, instant reaction posture for Minuteman missiles that provide a launch-on-warning capability. When he was a presidential candidate George W. Bush declared that “keeping so many weapons on high alert may create unacceptable risks of accidental or unauthorized launch.” His successor, Barack Obama, highly aware of the danger of nuclear weapons, explicitly called for reducing the role of launch-under-attack in U.S. nuclear planning. Yet, neither president changed the quick-reaction/launch-on-warning posture, perhaps to avoid a conflict with the Pentagon.
Today’s publication updates a National Security Web posting from April 2001, “Launch on Warning: The Development of U.S. Capabilities, 1959-1979.” The original documents are included in addition to more recently declassified items, , not only the Odom memorandum cited above, but records on policy development during the Carter and Reagan administrations when launch-on-warning and launch-under-attack postures were incorporated into the SIOP and official “nuclear weapons employment policy.”
Among the new documents are:
During the Carter administration and the years that followed, defense officials along with journalists, political scientists, and policy analysts debated the risks of launch-under-attack/on warning. Science adviser Richard Garwin saw it as a method to strengthen deterrence, while future Reagan administration official Fred C. Iklé decried it in the Washington Post. During the early 1980s, deeply critical assessments emerged sporadically as a result of hints that launch-on-warning or launch-on-attack had become embedded in Pentagon policy. The pros and cons attracted less attention, however, as a new détente emerged during Reagan’s second term.[1]
The end of the Cold War brought many changes, but not in the hair trigger alert status of the Minuteman ICBM force. Bruce G. Blair, a former Minuteman launch officer who was at the Office of Technology Assessment and the Brookings Institution during the 1980s and 1990s, emerged as the most persistent and well-informed critic/analyst of launch-on-warning. Drawing on his personal experience and wealth of knowledge, Blair has written major studies of command-and-control vulnerabilities and accidental nuclear war, especially the danger of the U.S.’s hair-trigger alert posture for Minutemen ICBMs and the risk of an accidental launch because of a warning system failure. One of his studies for the Office of Technology Assessment, addressing command and control problems and launch-on-warning, was so highly classified by the Pentagon that it has proven impossible to locate, at least so far.[2]
The risks that Blair and others associated with launch-on-warning were the logical consequence of U.S. nuclear war plans. As soon as the Soviet Union and then China began to develop a nuclear weapons complex, U.S. military planners defined the most crucial installations slated for rapid destruction as "time urgent" high-value targets; they included air defense, nuclear command centers, and missile and air bases. A capability to strike those targets as rapidly as possible, once warning information became available, became an enduring high priority for war planners.
The potential threat posed by Soviet nuclear forces prompted U.S. military commanders and intelligence agencies to look for signs that the Soviet leadership might be readying them for use in a surprise attack. If Washington received "strategic warning" of an impending Soviet attack, top commanders wanted the option of a preemptive strike (sometimes called taking the “initiative”) against Soviet strategic and command and control targets. Consistent with this, the first SIOP, approved by President Dwight Eisenhower in the fall of 1960, included choices for preemptive and retaliatory nuclear strikes.[3]
When the U.S. Air Force began to deploy intercontinental ballistic missiles during the late 1950s, they envisaged a strategic force that could be launched within minutes to deliver enormously destructive nuclear weapons. The "Minuteman" ICBM embodied that idea. While the possibility of rapid launch ICBMs supported ideas of preemptive attacks, preemption assumed strategic warning, specifically, intelligence indicating an imminent attack by an adversary, such as dispersal of nuclear forces and other signs of alerting and readiness activities. Whether such warning signs would be detected or properly interpreted is another matter.
If information became available that an attack was on the way – tactical warning intelligence – White House science advisers and Pentagon planners were reluctant to accept a strategy based on launching a retaliatory blow after absorbing a Soviet first strike. According to White House science adviser and MIT professor Jerome Wiesner, once electronic sensors could detect the launch phase of a Soviet ICBM attack, they could provide the "[warning] time necessary to ready our missiles so that they can be fired before they are destroyed."[4] What Wiesner was pointing to was the possibility of a launch-on-warning capability, a prospect that other U.S. government officials were beginning to recognize during the early 1960s.
Not all defense planners accepted the logic of launch-on-warning and some were skeptical of preemption. Apparently, Secretary of Defense Robert McNamara had strong objections to launch-on-warning, which he held long after he had left the Pentagon.[5] For some top officials, the development of Soviet ICBMs raised doubts about preemption. During a grim briefing by the National Security Council's Net Evaluation Subcommittee (NESC) President Kennedy made one of his last documented statements about nuclear strategy. Analyzing the consequences of U.S. and Soviet preemptive nuclear attacks on their respective societies, the NESC study introduced U.S. casualty figures---30 million--that were higher than Kennedy had heard before. With the devastating U.S. losses from Moscow's response to a preemptive strike, Kennedy observed that preemption was "not possible for us."[6] Despite Kennedy's misgivings, a preemptive strategic option remains embedded in the SIOP and nuclear war plans to this day.[7]
The deployment of the Ballistic Missile Early Warning System (BMEWS) in the early 1960s provided a rudimentary capability for launch-on-warning by giving command authorities fifteen minutes’ tactical warning of a missile attack. Also, in the works during the 1960s and deployed in the early 1970s was a satellite-based electronic warning system originally known as the Missile Defense Alert System (MIDAS) but later camouflaged behind the designation, Defense Support Program (DSP).[8]
With the deployment of DSP satellites, the possibility of launch-on-warning became increasingly imbedded in policy discussion, arms control negotiations, and in the training of Minuteman launch officers. Bruce Blair recalls that he “practiced LOW a hundred times during my Minuteman days 1972-74, which coincided with the U.S. DSP program becoming operational.” Although some officials looked favorably at the prospects of launch-on-warning, others saw great risk. One veteran official, Paul Nitze, warned that launch-on-warning would be "inexcusably dangerous" during a "time of intense crisis." What worried Nitze and others in particular was the danger of a false alarm, which was not a hypothetical problem. During the Cold War and after, both the United States and Russia received mistaken warnings of strategic attack, including the famous NORAD false warning incident on 9 November 1979.
When it was first published in 2001, this collection demonstrated the limits of the declassified record. Few documents from military organizations, such as the Defense Department and the Strategic Air Command, had been declassified, although both played critically important roles in making launch-on-warning a capability. Since then, however, significant defense-related records have become available that illuminate the development of the launch-on-warning posture and the technological developments that underlay it. Exactly when strategic planners believed that a capability was actually at hand remains classified and it is likely, as Bruce Blair has suggested, that it was an evolutionary process.[9]
In 1970, Caltech President and former Secretary of the Air Force Harold Brown, and a member of the SALT I delegation, found launch-on-warning tactically useful: it could make a Soviet attack on U.S. Minuteman fields "a relatively risky and unattractive" proposition. Seeing launch-on-warning as a potential deterrent, when Brown became secretary of defense under President Jimmy Carter, he supported including a variant – launch-under-attack – in the SIOP. Initially, the option that Brown and others supported was an ICBM-only attack, but SAC later convinced him that achieving SIOP objectives required a larger-scale attack involving bombers and SLBMs – the entire triad.[10] Documents from the late 1970s and early 1980s that have recently become available shed light on the internal discussion of launch-on-warning and launch-under-attack and their significant role in the development of nuclear strategy.
Launch-under-attack is often used interchangeably with launch-on-warning. During the Cold War, the Joint Chiefs of Staff defined them identically: as a launch of forces between the detection of an attack and the arrival of the first warhead. According to Bruce Blair, pre-delegation instructions gave specific meaning to launch-under-attack because top military commanders with nuclear missions would have to delay action until confirmation of an attack was available, although they would not have waited for evidence of massive destruction. Under such circumstances, SAC defined the delayed reaction as a launch-under-attack.[11] In any event, the point of both launch-on-warning and launch-under-attack was to ensure that ICBMs would be launched rapidly enough to destroy time urgent targets specified in war plans.
Soviet nuclear strategy is largely undocumented and no definitive information is available on the role of launch-on-warning in Soviet policy. Available information, however, suggests that by the late 1960s and early 1970s, defense planners considered preemptive, retaliatory, and launch-on-warning options. Preemption had low feasibility because defense officials believed that U.S. nuclear forces were too widely dispersed to be destroyed by a first blow. Soviet defense officials saw launch after ride-out (“otvetnyy udar”) as an option by giving policymakers time for deciding on how to retaliate. Initially, top officials did not see launch-on-warning as a possibility because early warning systems were not effective enough to warrant a quick launch. By the early 1980s, however, launch-under-attack had greater plausibility because Moscow had acquired some capability, mainly by hardening some ICBMs, such as the SS-18. against the effects of nuclear detonations. With improvements in warning systems, Blair has argued, launch-on-warning became central to the Soviet and post-Soviet strategic posture. Indeed, Russian President Vladimir Putin’s October 2018 statements at the Valdai Discussion Club strongly suggest that Russia has a “counter-strike on warning” posture.[12]
Document 01
National Archives, Department of State Records, Record Group 59 (hereinafter RG 59), Decimal Files, 1955-59, 611.61/5-1958 (also available on Digital National Security Archive, U.S. Nuclear History I, 1958-5-19
This brief memo unambiguously conveys the notion that in the missile age, even civilian officials would take it for granted that launch-on-warning of attack would be possible and necessary. In late April 1958, Prime Minister Harold Macmillan proposed that President Eisenhower agree to Anglo-American talks for a "fully agreed and understood procedure" for making decisions to launch nuclear retaliation against a Soviet attack. With the U.S.'s major nuclear deployments in the United Kingdom and the close nature of Anglo-American relations, Macmillan sought U.S. agreement on consultations before making the most fundamental military decision of them all. British leaders had been pressing Washington for agreements on consultation since the November 1950 Korean War crisis but U.S. leaders, anxious to preserve freedom of action, would agree to only the most general commitments.[13]
In commentary on the problem of consultation, State Department official Robert Fearey broke down the issue into "four possible cases": The first scenario amounted to "launch under attack": when nuclear bombs and missiles are raining on British and U.S. territory, consultations would not be necessary or possible because of the urgent necessity to launch a retaliatory strike. "Launch-on-warning" characterizes the second scenario: with electronic sensors detecting a Soviet bomber-missile attack "there might be time" for consultations on whether warning information was accurate and whether missiles or bombers should be launched in retaliation. Only if the Soviets launched a non-nuclear attack or if Western intelligence had advance warning of a Soviet nuclear strike would there be time for consultations on nuclear weapons use. Although the Eisenhower administration accepted the importance of consultations between president and prime minister, in the June 1958 Murphy-Dean agreement, it reaffirmed the initial understanding that decisions to launch bombers or missiles had to be made "in the light of the circumstances at the time." In other words, consultation might not always be possible.
Document 02
Dwight D. Eisenhower Library, Anne Whitman File, Dwight D. Eisenhower Diaries, box 42, Staff Notes June 1-15 1959 (2); also available on DNSA, U.S. Nuclear History I, 1955-68.
The possibility and desirability of a launch-on-warning capability for the United States was a premise of a briefing given on 3 June 59 to President Eisenhower by MIT professor Jerome Wiesner, then a member of the President's Science Advisory Committee (PSAC) (He became White House science adviser for President Kennedy in 1961). While doubtful of the value of anti-ballistic missile systems, Wiesner saw advantage in an infra-red warning capability that would permit missile launch after receipt of a warning but his presentation did not approve MIDAS. Skeptical that MIDAS could overcome technical obstacles, the science advisers were far more interested in using high-altitude U-2 aircraft as a platform for an infra-red detection system.[14]
Document 03
National Archives, Record Group 59, Department of State Records, Policy Planning Staff Records, 1957-61, box 20, file: Owen, H. Chron (also available in DNSA, U.S. Nuclear History, 1955-68.
While Wiesner endorsed a launch-on-warning capability, other civilians had their doubts. The reference to launch-on-warning in this document appears in the context of the late 1950s-early 60s debate over the creation of a medium-range missile force for NATO that would enable non-nuclear powers like West Germany to participate in decisions on nuclear weapons use. In this memorandum, Policy Planning Staff director Gerard C. Smith cited a briefing by Supreme Allied Commander Europe (SACEUR) General Lauris Norstad where the latter argued that the NATO missile force had to be "ready to react two to five minutes after warning." Smith interpreted that statement as support for "fir[ing] after warning of impending attack and before Soviet missiles had landed." What troubled him was that it was inconsistent with Norstad's emphasis on the importance of a survivable missile force. Perhaps worried about the possibility of inaccurate warning, Smith questioned the need for "instant reaction."
Document 04
National Archives, Record Group 59, Department of State Records, Records of Department of State Participation in the Operations Coordinating Board and the National Security Council, 1947-196, Box 94, "NSC 5802 Memoranda," also available on DNSA, U.S. Nuclear History I, 1955-68.
Reservations about launch-on-warning appear in this analysis of the problem of defense against bomber and missile attack. Written during the period of the "missile gap" controversy, when actual Soviet intercontinental ballistic missiles (ICBM) capabilities were in doubt and worst-case analyses were routine, this study by the NSC Planning Board predicted that Soviet missiles would "constitute a great threat" to U.S. cities by the end of 1960. Question 3 (page 11) was especially apposite to the launch-on-warning problem: "should the United States revise ... its doctrine on response to attack and on response to warning of attack, in the light of decreased reaction time and in view of the increasing U.S. emphasis on retaliatory ballistic missile forces?"
The analysts were confident that the DEW (Distant Early Warning) line would provide sufficient warning of a bomber attack and "adequate time" for decisions and action, such as putting Strategic Air Command bombers in the air. A new warning system--the Ballistic Missile Early Warning System [BMEWs]--was in the works that could give U.S. authorities fifteen minutes to respond to an incoming ICBM attack (although the proposed MIDAS system would be able to provide more time), hardly enough time for decisionmakers to assess the situation, make a decision, and transmit it to commanders. Until BMEWs was available, only the Bomb Alarm System, then being deployed, could give definitive information on nuclear detonations.
According to the authors note, an important advantage of strategic bombers was that they could be recalled. By contrast, an unrecallable ICBM nuclear force made launch-on-warning of doubtful value: it was "questionable whether U.S. response doctrine will permit the launch of `irrecallable' ballistic missiles solely on the basis of information received from a warning system." The analysts doubted that BMEWs and any follow-on systems could provide "high confidence high early warning" and judged it "essential" to avoid launching unrecallable missiles based on a false warning (see paragraph 43).
Document 05
National Archives, Record Group 200, Papers of Robert S. McNamara, box 113, Reading File Nov. 1961 (also available in on DNSA, U.S. Nuclear History I, 1955-68
Interested in the status of U.S. warning and detection capabilities, Senator Stennis (D-Ms) sent McNamara a list of eight "requirements" to which McNamara responded with detailed information describing deployed and proposed systems. In the course of this assessment of various deployed and proposed systems--DEW Line, BMEWS, MIDAS, etc.--McNamara responded on page 17 to Stennis's request for information on whether a fifteen-minute warning time "would be sufficient for the warning to be transmitted, the command to be given and communicated, and our weapons actually launched before enemy missiles or bombs impact in our territory" (see page 17). McNamara confidently observed that fifteen minutes would be enough to assess warning intelligence, convene an emergency conference of the president and other National Command Authorities, and transmit an execution order to commanders, as well as launch "all SAC alert aircraft and Atlas E ICBMs and one third of the Atlas D ICBMs.[15]
Document 06
Library of Congress, Papers of General Curtis LeMay, Box 141, AFSC (AF Systems Command) 1962
McNamara may have been ambivalent about a launch-on-warning posture; according to one account, at one point during the Kennedy administration he stated that he strongly opposed it. Meeting with McNamara, General Bernard Schriever, then commander of Air Force Systems Command, cited a launch-on-warning capability to justify MIDAS. A witness to the meeting later recalled that McNamara was "furious" and told Schriever that "as long as he was secretary of Defense and Jack Kennedy was President, the United States would never launch on warning, even if that required a force of 10,000 Minuteman ICBMs [to assure the survivability of enough forces to retaliate]."[16]
To Schriever's dismay, in early August 1962, McNamara ruled against Air Force plans to deploy MIDAS satellites; from McNamara's perspective, MIDAS was too costly, it duplicated other warning systems, and the hardening of missile silos reduced the importance of early warning.[17] As this document shows, Schriever was firmly convinced that maximum warning information and he lobbied the Secretary of the Air Force to urge McNamara to reconsider. Despite Schriever's efforts, however, it took several years before more senior officials were convinced that MIDAS could work and to approve a development plan.
Document 07
Library of Congress, Papers of Curtis M. LeMay, box 153, 19-3 White House 1962 (also available on DNSA, U.S. Nuclear History I, 1955-68.
A capability for nearly instantaneous launch of strategic missiles, an important technical condition for launch-on-warning, came into play during the fall of 1962. At the height of the Cuban Missile Crisis, the Strategic Air Command began to deploy nuclear-armed Minuteman I missiles in silos located near Malmstrom Air Force Base in Montana. Secretary of the Air Force Zuckert reported to President Kennedy that the deployment was occurring under "unusual safety conditions" so that it would take hours to launch the missiles. Zuckert's confidence in safety procedures on the ground was misplaced; the missiles could actually be launched immediately, foreshadowing their normal alert status.[18] He also informed Kennedy that once the Minutemen in the first complex had been deployed in their "normal alert status," all "twenty missiles will be able to be launched in thirty seconds."
Document 08
National Archives, Record Group 200, Robert S. McNamara Papers, box 42, Defense Projects and Operations (also available on DNSA, U.S. Nuclear History I, 1955-68.
This memorandum elucidated the counterforce, or "damage limiting", assignment of the Minuteman ICBM force. Its chief targets were the 600 "time-urgent" Soviet bomber bases and missile sites, among others, that had to be destroyed before they could endanger U.S. allies or U.S. territory. The problem of "known failure"--that some percentage of ICBMs would fail to reach their target--made it necessary to assign an average of 1.67 missiles to assure that one "on-launch reliable" Minuteman hit its target. With U.S. reconnaissance satellites expected to locate more "time-urgent targets," Secretary of the Air Force Zuckert sought Robert McNamara's approval for a total force of 1200 Minutemen missiles to strike them. McNamara, however, had decided that 1000 Minutemen was "enough"; moreover, a new technology then still on the drawing-boards--multiple-independently targetable re-entry vehicles (MIRVs)--would make it possible to strike more targets with the same number of ICBMs.[19]
Document 09 NEW
Lyndon B. Johnson Presidential Library, Alain Ethoven Papers, box 5, Summary Memos and Comments, 1967
A reference to "launch-on-warning" by McNamara aide Alain Enthoven suggests that the concept was certainly in the air and not even an unusual one during the Johnson administration (whatever McNamara thought about it). Enthoven's paper was part of the Draft Presidential Memoranda (DPM) process that McNamara inaugurated to convey his policy views and budget recommendations to Presidents Kennedy and Johnson. Typically, the DPMs would produce comments and reclamas (requests for reconsideration) from the various armed services. In this one, Enthoven subjected Air Force and Joint Chiefs of Staff views on deterrence, "damage limiting," and other matters to detailed critical analysis. Noting that the Air Force believed that the U.S. had inadequate means to deter large-scale nuclear attack, Enthoven argued that U.S. nuclear forces gave the Soviet Union "a great deal to worry about", including "launch-on-warning tactics, ... Soviet fatalities due to long-term fall-out, epidemics, or secondary effects (e.g., starvation) because of bottle-necks in their economy."
Document 10 NEW
During 1968, SAC planners considered the possibility of "fire on warning" because they were wondering about the possibility and the impact of a Soviet "pindown" attack, whereby the Soviets would launch SLBMs at Minuteman sites and detonate the warheads in the air to stop the American missiles from launching. For the planners, "a fire on warning doctrine was the best military answer to the pindown threat," but it was "politically unacceptable," probably because it reduced presidential control over the decision-making process. Nevertheless, the planners anticipated that political objections might diminish and that the availability of new warning systems and more "streamlined national command authority procedures" might make fire on warning more generally acceptable. A "minimum reaction posture" for ICBMs would be also be necessary, presumably by reducing the time needed to launch them. Changes in safety rules and crew checklists would also be necessary to assure quick launch.
Document 11
National Archives, Nixon Presidential Materials Project, National Security Council Files (hereinafter Nixon NSF), box 840, Sentinel ABM System Vol. II, 4/1/69; also published in DNSA, U.S. Nuclear History II, 1969-1976.
Discussions of launch-on-warning also surfaced in Congress and the White House early in the Nixon administration. Lawrence Lynn, an NSC defense analyst, prepared a briefing paper for national security adviser Henry Kissinger's use in discussion with "prominent news columnists." Apparently, leading opponents of Anti-Ballistic Missiles (ABMs), including Sen. Albert Gore Sr. (D-TN) had suggested a "launch-on-warning" option as a method for preserving the ICBM force from attack. The White House, however, wanted to shoot down "firing on warning" as "dangerous and irresponsible" because early warning sensors had such a high rate of false reports. Thus, the "possibility of a disastrous mistakes" would be a "very real one." Drawing on classified information, Lynn reported that existing warning systems, BMEWs and Over-the-Horizon Radar (OTH)[20], had significant false reports rate; for example, 50 percent of initial OTH reports were false. Lynn demonstrated, however, why some would find launch-on-warning to be workable: the "new early warning satellite [647 project] may produce one false alarm per year."
Document 12
Nixon NSF, box 710, USSR Vol V. 10/69, also published in DNSA, Nuclear History II, 1969-1976.
References to launch-on-warning emerged during a conversation at an Institute for Strategic Studies meeting, between Georgy Arbatov, a Soviet American specialist who headed the Institute of USA and Canada Studies, and Helmut Sonnenfeldt, a European and Soviet affairs specialist, who had joined Kissinger's NSC Staff from the State Department's Bureau of Intelligence and Research. Reflecting the problems in U.S.-Soviet relations during the period, the conversation turned to strategic nuclear issues, including U.S. uncertainties about the SS-9, the Soviet ICBM that appeared to threaten U.S. Minuteman silos. To Sonnenfeldt's surprise, Arbatov observed that there was little to "worry" over because "neither side would wait if it received warning of an attack but instead ... would simply empty out its silos by launching a counter-strike at once." Sonnenfeldt objected, noting the danger to "strategic stability" of a launch-on-warning posture. Sonnenfeldt also questioned whether Arbatov's statement reflected "existing doctrine."
Document 13
ACDA FOIA release to National Security Archive; also published in DNSA, Nuclear History II, 1969-1976.
Some months after the Arbatov-Sonnenfeldt discussion, during the spring of 1970, the problem of launch-on-warning received more attention during the Strategic Arms Limitations Talks (SALT), summarized in this ACDA report. During an April 1970 discussion of a possible ban on multiple reentry vehicles, chief Soviet negotiator Vladimir Semenov raised the problem of launch-on-warning when he noted that new warning systems would enable governments to launch missiles and to empty silos before "the enemy [strikes] a blow at them." A few weeks later, Gerard Smith, the head of the U.S. SALT delegation, showed his concern about launch-on-warning when he asked whether governments should plan to fire missiles "solely on the possibly fallible reading of signals from ... early-warning systems." Such a posture would be "very dangerous and would increase the risks of unwanted war." The discussion did not go much further, although it became evident that General Ogarkov, the top military official on the Soviet SALT delegation, was resentful that Smith had taken the discussion further and told U.S. General Royal Allison that "as a military man, [he] should know the answer" to Smith's question, suggesting later that U.S. military manuals assumed a launch-on-warning posture.
Smith wanted to influence the discussion further. While acknowledging that Soviet uncertainty about a U.S. launch-on-warning posture could have "some deterrent value" and even some provide some "bargaining leverage" in the SALT talks, on balance Smith believed there would be more "risk and danger" if the Soviets had a "mistaken view" of U.S. policy. Thus, on 19 May, he cited a slightly equivocal statement of "hope" by Secretary of Defense Melvin Laird that "that kind of strategy would never be adopted by any Administration or by any Congress."
Document 14
RG 59, Policy Planning Council Miscellaneous Records, 1959-72, box 299, SALT 1970 October 1-13; also published in DNSA, Nuclear History II, 1969-1976.
The apparent growing vulnerability to a surprise attack of U.S. land-based ICBMs worried the hawkish Cold War veteran Paul H. Nitze, the author of NSC-68 and a former Deputy Secretary of Defense and Secretary of the Navy and a member of the U.S. SALT delegation. During a conversation with a like-minded State Department official Seymour Weiss, Nitze worried that even with a SALT agreement, Moscow might be in a position to install multiple independently targetable reentry vehicles (MIRVs) on the SS-9 ICBMs, thus giving Moscow a "first strike capability against US land-based missiles." Nitze saw several alternatives to address this vulnerability: 1) developing a first strike capability, 2) a launch-on-warning doctrine, or 3) abandoning land-based missiles and "move entirely to sea" by relying on submarine-launched ballistic missiles.
Nitze argued that launch-on-warning was "always contrary to US strategic policy" and would be "inexcusably dangerous" during a "time of intense crisis." Nevertheless, he acknowledged that Washington could "be forced" into a launch-on-warning posture if Minuteman vulnerability "seems at least theoretically possible."
Document 15
Donation from ACDA; also published in DNSA, Nuclear History II, 1969-1976.
While Paul Nitze and Gerard Smith were alert to the risks, a launch-on-warning posture was already embedded in U.S. nuclear planning. Bruce Blair, who was a Minuteman launch officer at the time, recalls that he "was postured for LOW [launch-on-warning] during the early 1970s, and the whole force and command system were geared to this timing", that is, to rapid response.[21]
This excerpt from a meeting of the General Advisory Committee on Arms Control and Disarmament [GAC], a group of prominent civilian experts on arms control and nuclear forces, shows that military officials showed little doubt about the value of launch-on-warning. Focusing on the possible vulnerability of the U.S. Minuteman force to an attack by Soviet ICBMs, GAC heard testimony on, and discussed, Soviet ICBM forces, plans for hardening missile silos, and the possible role of anti-ballistic missile defenses. Toward the end of the session, one of the Committee members, Kermit Gordon, then president of the Brookings Institution, asked whether there was a "plausible scenario" for a simultaneous Soviet first strike against U.S. Minuteman and bomber forces.
The discussion that Gordon's query prompted was less than straightforward but a launch-on-warning option flowed from the discussion of different scenario in which the Soviets targeted ICBMs and SLBMs on Minuteman silos and bomber bases respectively. According to Caltech President (and future Secretary of Defense) Harold Brown, once the Soviets launched their ICBMs, they would risk a [U.S.] "launch on warning." Commander James Martin then observed that "there's about 20 minutes in there when the President might decide to launch on warning."
Document 16
RG 59, Subject-Numeric Files, 1970-1973, Def 12 USSR; also published in DNSA, Nuclear History II, 1969-1976
Prepared only a few days after the GAC discussion, this significant document shows that State Department intelligence officials recognized that a capability for launch-on-warning existed. They gave it relatively uncritical support, although one of the officials, Frank Perez, observed that he was "not advocating [its] adoption."
Both Perez and his superior officer, Leonard Weiss, constructed the discussion of launch-on-warning around the logic of deterrence: because the Soviets could detect a U.S. capability to get the Minutemen "off the ground in time," that could deter them from the "possibility of undertaking a first strike." Even if Moscow struck first, launch-on-warning would enable the United States to inflict "intolerable damage" on the Soviet Union.
According to Perez, a launch-on-warning capability would depend on the availability of "unambiguous warning," which could be provided by systems that were becoming available. The 440-L Over-the-Horizon system and the 647 early warning satellite, also known as the Defense Support Program, could detect mass missile launches.[22] Perimeter Acquisition Radars (PAR), a type of phased array radar, could provide "absolute certainty as to the size of the attack and ... where [it] originated and to where it was directed," for example, whether Minuteman fields were a target.
While Perez was overoptimistic about the extent to which new warning systems like PAR could quickly provide actionable information, he believed that they would give a President a choice other than "rid[ing] out the attack and then respond[ing] with what residual [forces] remained." Instead, the president could "respond to a Soviet attack based on his assessment of the situation." For example, in response to a Soviet attack, Perez suggested that escalation and mass civilian casualties could be avoided with a controlled response of some 200-300 Minutemen against high-value Soviet military targets away from urban-industrial centers. Whether an attack by 250 Minuteman could actually limit escalation looks questionable in retrospect.
Document 17
RG 59, Records of Undersecretary of State John Irwin, 1969-73, box 5, SALT Jan-June 1971; also published in DNSA, Nuclear History II, 1969-1976
One of the problems that generated interest in a launch-on-warning capability was the alleged vulnerability of the U.S. Minuteman force to Soviet attack. This summary of a meeting of the National Security Council's Defense Program Review Committee [DPRC], chaired by national security adviser Henry Kissinger, suggests interagency agreement that the Soviets had the wherewithal to destroy Minutemen but differences over when the threat would materialize. Taking the most cautious, "worst case" approach, the Defense Department estimated a threat by the mid-1970s. The "intelligence community", presumably the CIA, however, did not see a vulnerability problem at least until later in the decade depending on when the Soviets could deploy accurate MIRVs on the SS-9.
As some argued at the time, a "vulnerable" Minuteman force might not be a serious liability when more survivable U.S. SLBMs could threaten Soviet cities. Nevertheless, for some on the DPRC, the vulnerability problem posed important political questions, for example, "how would US political leadership react in a crisis if a significant portion of US force was considered vulnerable"? One possible implication was that if national authorities saw a danger of a Soviet preemptive move against U.S. missile silos, they might raise alert levels for possible recourse to launching Minutemen on warning. Raising alert levels, of course, could increase anxiety levels at the Kremlin heightening the risks of nuclear war. Only future declassification releases, however, may elucidate the DPRC's later discussions of the broader implications of the survivability problem.
Document 18
Nixon NSF, box 808, Brown, Harold; also published in DNSA, Nuclear History II, 1969-1976
As a member of the SALT delegation, Caltech President Harold Brown was a member of the SALT delegation would pass on his thinking to Henry Kissinger. For Brown, curbing the arms race by limiting ABMs was highly important, even more than limiting large Soviet ICBMs such as the SS-9. If the Soviets deployed MIRVs on the SS-9 they could pose a threat to U.S. Minuteman silos but, Brown believed, a launch on "unambiguous" warning capability, if not a doctrine, would make such an attack "a relatively risky and unattractive" proposition.
Attentive to the danger of false warning, Brown conceded that launch-on-warning was not a "sure tactic," but it could be "relatively easily ... achieved during the mid-70s." Nevertheless, he did not necessarily support a launch-on-warning "doctrine," perhaps to give civilian authorities more flexibility in a crisis. Like Perez, Brown believed that Minutemen launched on warning could hit military targets, although he may have had Soviet bomber bases in mind.
In his comments, K. Wayne Smith, the NSC's director for program analysis during the early 1970s, found value in Brown's argument on the importance of controls over ABM radars, but questioned the merits of launch-on-warning. For Smith, it was a problematic deterrent, because of the danger that, during an international crisis, it could increase the risks of war by encouraging one side or the other to act precipitously.
Document 19 NEW
U.S. National Archives, Record Group 383, Records of U.S. Arms Control and Disarmament Agency, Bureau of Intelligence, Verification, and Information Management, 1968-81: box 1 of 1.
This report assessed the technical practicality of launch-on-warning (LOW) as a means to assure Minuteman survivability in the face of an apparent Soviet strategic threat. With satellite-based warning systems providing nearly 30 minutes of warning of a Soviet ICBM attack, the National Command Authorities (NC) could order "the Minuteman [to] be launched within that 30-minute interval" and "the accuracy of the Soviet ICBMs will be unimportant because the Minuteman silos will be empty."
Whether the command-and-control system could function within the 30-minute time limit, whether the warning signals were accurate, and whether the Soviets might be able to defeat LOW by using submarine-launched ballistic missiles to "pin-down" Minuteman ICBMs were problems that the McDonnell Douglas analysts reviewed. For example, on false warning, the analysts maintained that the "probability of a false alarm by any one of the warning systems is negligible; the probability that two warning systems measuring entirely different phenomena would report correlated false alarms at the same time is infinitesimal." (Future analysts were not so certain.) On "pin-down," the analysts believed that the Soviets would not have the capability for several years and that it was possible to make "pin-down" impractical, e.g., by making Minutemen more resistant to radiation ("hardening").
In retrospect, one of the most problematic elements of the McDonnell Douglas analysis are the facts presented about the time required to make, transmit, and implement a decision to launch Minuteman missiles. According to the analysts, once the NCA made a launch decision, "significant delays" would occur, of which "the most important and the most surprising" was the "11 minutes required by the launch control crew to receive, decode, authenticate and execute the launch command." The analysts observed that "at one time," six minutes was the time required, bur 11 minutes was the "interval" established by "Air Force doctrine to insure that no crew attempts to launch before all crews have completed their prelaunch functions." Another four minutes was required to decode messages and transmit the launch command.
Bruce Blair, who was a Minuteman launch officer in the early 1970s, finds that the McConnell Douglas analysts were mistaken; it would take three minutes to launch the Minuteman at the most. The eleven-minute time line may have been the standard during the late 1960s but not in the years that followed.[23]
Document 20
FOIA request to Department of Defense; also published in DNSA, Nuclear History II, 1969-1976
A launch-on-warning capability depended on warning information, quick-reaction nuclear forces, but also a command-and-control apparatus that could assess strategic intelligence, make appropriate decisions, and rapidly convey them to military commanders. Pages 345-347 from chapter XXVI of this Institute for Defense Analyses history describe the Defense Support Program (DSP) satellites and their role in providing "usable warning time" as well as qualitatively better information so National Command Authorities would be better able to assess an attack and chose one of the SIOP options. Chapter XXX describes the central features of the National Military Command System as it stood during the early 1970s, including procedures for transmitting and implementing orders for the execution of the SIOP.
As important as the command-and control system was, confidence in its reliability was not high and reports on its failings were "continuous" during this period (and beyond). Thus, whether the NCA could properly assess warning information, much less make a decision to launch-on-warning and successfully transmit it to commanders in the field, would be problematic. For example, in 1970, even though the Defense System Program had already been successfully tested, the Blue Ribbon Defense Panel reported on the difficulty of providing warning information to the president: "it is possible that no President could be sure ... that an attack was in progress or that retaliation was justified," unless confirmation of nuclear detonations was already available (p. 408).
Document 21 NEW
FOIA release by U.S. Air Force
This report shows how Carl H. Builder and his RAND colleagues looked at launch-on-warning when they considered future roles for ICBMs. Strongly interested in identifying the possibility for special roles for the Minuteman force, e.g., for limited strategic operations or counterforce missions, the RAND analysts assumed that the value of ICBMs as a deterrent depended on their survivability when under attack. While they believed that worries about a preemptive counterforce attack were exaggerated, they saw no good choices for assuring Minuteman survivability.
Builder and his co-authors considered whether a "launch-under-attack-assessment" would be useful for giving the Minuteman force enough time to be used in a crisis. Highlighting "attack assessment" instead of "warning," their definition of attack assessment demanded more authoritative information than warning from DSP satellites. Moreover, for the decision-making process to be credible, two problems had to be solved. One was "the attack assessment thresholds for considering launch commitment," for example, weighing the consequences of launching or not. The other was the "level of confidence in assessment information for launch decision." Was it necessary to have "confirmed reports" that Soviet warheads had detonated on U.S. soil? Also necessary for credible launch under attack assessment" were targets other than (empty) missile silos.
According to the report, once the President had made a decision, Minutemen could reach high altitudes "seven minutes" after a president's launch decision. Although the report mentioned possible vulnerabilities of ICBM silos to SLBM strikes, during this period Soviet SLBMs were not accurate enough to pose such a threat.
To support a launch-under-attack assessment, the authors drew on a logic similar to that employed by INR's Frank Perez: "we believe that the technical capabilities to launch ICBMs on attack assessment should be developed for their deterrence value--so that no adversary would dare assume that the U.S. could not launch the force out from any attempted disarming attack." Nevertheless, the authors argued against openly declaring the policy; the idea of launch-on-warning was so controversial that "it would be rigorously opposed as both dangerous and unstable (an accident could theoretically precipitate a nuclear war)." They also argued that the survivability of U.S. ICBMs was not important enough to require a decision to launch-under-attack. Implicitly, the danger of nuclear war was too terrible to allow the "assurance of ICBM retaliatory capabilities [to] rest upon such an awesome commitment."
Document 22
Gerald R. Ford Library, National Security Council Meetings Files, Box 2 1975-12-22; also published in DNSA, Nuclear History II, 1969-1976.
During a briefing by Director of Central Intelligence William Colby on Soviet forces, Secretary of State Henry Kissinger, JCS Chairman George Brown, Secretary of Defense Donald Rumsfeld (in his first incarnation in that role), and their colleagues discussed the possibility and problems of launch-on-warning (see pages 8-9). Looking at a worst-case scenario--a Soviet ICBM attack on U.S. Minuteman silos--Kissinger argued that it would be highly difficult for Soviet leaders to accomplish such an attack. Not only could the United States respond by launching SLBMs and bombers, it could also launch ICBMs on warning with the Minuteman force alone producing 80 million Soviet casualties. When ACDA Director Fred Ikle mentioned the dangers of a launch-on-warning posture--"accident prone" and "dangerous"-Kissinger suggested that it was already an available option by noting that command-and-control arrangements could be fixed to ensure that missiles were never launched without "presidential authority." Ikle remained a critic of launch-on-warning.[24]
Kissinger and top Pentagon officials were more interested in preserving the ambiguity of the U.S. posture so that the Soviets could not know with any certainty whether the United States had a "launch on warning policy." To complicate Soviet planning, US policymakers wanted to keep Moscow guessing. Further, as National Security Adviser Brent Scowcroft suggested, it was "not to our disadvantage if we appear irrational to the Soviets in this regard." The implication was that U.S. ambiguity or even irrationality could encourage Soviet diplomatic caution.[25]
Document 23 NEW
MDR Request to Defense Department; also published in DNSA, Nuclear History II, 1969-1976.
Wallace Henderson's memorandum gave general support to an Air Force program to demonstrate launch-on-warning by launching a Minuteman from Vandenberg Air Force Base in response to warning information from a Defense Support Program satellite. The point would be to "add credibility" to Defense Department statements that the U.S. had a "capability" for launch-on-warning while also mitigating concerns about "the low survivability of Minuteman in the face of mounting Soviet ICBM throw weight and accuracy." Henderson acknowledged a "negative" - that a demonstration "could revive emotional responses in Congress to discussion of "Launch on Warning."
According to the cover memorandum, Deputy Secretary of Defense Robert Ellsworth cancelled the plans, perhaps concluding that a demonstration was too controversial.
Document 24 NEW
Jimmy Carter Presidential Library, National Security Files, NLC-6-79-4-20-3
Odom brought to Brzezinski's attention an INR report on recent intelligence that the Soviets had acquired a "fast reaction" ICBM launch capability that made launch-on-warning feasible. According to the report, the quick-launch feature encouraged the Soviets to see their ICBMs as their "prime retaliatory force." Odom observed that the INR report represented an "upgrading" of an intelligence estimate that had been made a year earlier. "The least comforting implication is the [limited] readiness warning time we might have."
Document 25 NEW
Jimmy Carter Presidential Library, National Security Files, NLC-1-5-2-29-2
In this report to President Carter, the declassification reviewers released an important tidbit that was excised from the INR paper (see Document 22): "According to State, recently acquired evidence indicates that as much as 70 percent of the current Soviet ICBM force could be ready to launch in about three minutes, even from a normal readiness posture." Before, only a small portion of the ICBM force was at "the highest level of preparedness with gyroscopes continuously running." Brzezinski reported that the new information was "consistent with the USSR's continuing efforts to improve its early warning capabilities."
U.S. intelligence information may well have been correct. According to Pavel Podvig, an expert on Soviet nuclear forces, ICBMs could be launched in 2.5 to 3 minutes after having withstood a nuclear strike, with 90 seconds needed for a pause, after the strike, and 60 seconds for the 'launch cycle." However, the Soviets may not have deployed ICBMs with spinning gyroscopes until the early 1980s.[26]
Document 26 NEW
(Program Analysts & Evaluation) to Secretary of Defense, "Interim Report on the PD-18 Study Modernization of the ICBM Force -- ACTION MEMORANDUM," 26 June 1978, with "Introduction and Summary," "Paper # 2 Launch-Under-Attack," and "Paper #3 "The Strategic Options Matrix," attached, Top Secret, excised copy MDR request to Defense Department
During 1977-1978 the Carter administration launched a major review of nuclear strategy and forces, not only so the president would have alternatives to the massive strike options in the SIOP but also to strengthen deterrence and to mitigate the perceived vulnerability of the ICBM force. One of the studies of ICBM modernization reviewed the feasibility of a launch-under-attack option. Apparently finding that the U.S. had an "operational" launch-on-warning capability, the study found advantages and disadvantages in LUA. The key advantage was that a "credible LUA capability may enhance deterrence," by producing "great uncertainty" in the minds of Soviet leaders about the possibility of a first strike against the U.S.
If deterrence failed, the study questioned the value of "LUA as a tactic to maintain an effective ICBM force." The conclusion was that an "operational LUA capability is not a viable alternative for modernizing the ICBM force." The short amount of time was one consideration; another problem raised constitutional questions: "confidence in this tactic would require a system for transferring authority to a properly briefed and survivable figure quickly in the event communications with the President failed." The need for expert knowledge was another problem: "The President and his potential successors would have to be thoroughly familiar with the attack assessment systems and be prepared to make a decision in very little time." And such a requirement was wrong: "the President could decide to retaliate in a few minutes [but] he shouldn't, as a matter of policy, have to do so." The Defense analysts did not want to see the president forced into a decision (or "jammed" in a crisis.) [27]
Document 27 NEW
MDR request to Defense Department, under appeal at ISCAP
This massively excised critique of launch-on-warning as "dangerous, wrong, and ineffective" may have been influenced by the criticism of launch-under-attack found in Document 26. Some of the content can be gleaned by reading between the lines. For example, the second-to-last paragraph might address the problem of forcing the president to make a quick decision in a crisis.
Document 28 NEW
Jimmy Carter Presidential Library, Zbigniew Brzezinski Collection, Subject File, box 35, Presidential Directive 59- 9/78-4/79
This synopsis of the Carter administration's targeting study, directed by Defense Department official Leon Sloss, reiterated some objections to LUA, such as the risk of accidental war and the danger of forcing the president to "make a 'use or lose' choice if there are other options available." It also included a telling criticism of the role of ICBM deployments in targeting policy: "the set of targets for our ICBM force is not in itself planned to meet any specific objective." Moreover, then-current ICBM targeting would kill so many people that it would "likely invite retaliation against US urban/industrial assets." To reduce the escalatory risks, the study recommended a specialized launch-under-attack option using ICBMs only that would target a "broad set of nuclear and non-nuclear targets and command and control," while trying to "minimize collateral damage to population consistent with achievement of the attack objective." Whether attacking a "broad set" of targets would produce fewer fatalities looks doubtful in retrospect because many Soviet military-related targets were near population centers.
Document 29 NEW
MDR release by Defense Department
Secretary of Defense Harold Brown had reviewed this targeting review study and directed officials to begin implementing some of the recommendations promptly. One of the recommendations, on page 9, concerned launch-under-attack. Along the lines of the targeting review study, the recommendation found that "LUA cannot be a substitute for measures to reduce ICBM vulnerability; rather it is an interim measure designed to strengthen deterrence." While an ICBM-only, launch-under-attack option was to be ready for implementation during 1981-1982, it was not to be incorporated into war plans "until it adequately supports the building blocks that are developed to support SIOP planning," (with each block representing a discrete option interrelated with others).
Document 30 NEW
Jimmy Carter Presidential Library, Brzezinski Collection, Subject File, box 42, Missile Warning Incidents 11/79-8/80
To Odom's dismay, the latest version of the SIOP included what he termed an "unwise" launch-under-attack option. First, the U.S.'s tactical warning system was "just not good enough to let us know that our ICBMs are under attack" (although a few weeks later, the famous NORAD false alert incident demonstrated the importance of the DSP and other warning systems for disconfirming mistaken information about Soviet incoming missiles). Second, "little thought has been given to. the rationale of the target set." The SIOP launch-under-attack option "is targeted at empty Soviet silos and some conventional military targets," (in striking contrast to the "broad set" of targets recommended by the nuclear targeting review). That could provoke a "major Soviet response." Odom also argued that "if the LUA is executed and a few minutes later there is a decision to execute one of the MAOs [Major Attack Option] large numbers of our RVs would be destroyed through fratricide." That is, the blasts and radiation caused by the LUA strikes on Soviet targets would damage incoming U.S. reentry vehicles, "fratricide."
Odom recommended that Zbigniew Brzezinski ask Secretary of Defense Brown to revoke the LUA option from the SIOP, but it appears unlikely that the memo went any further than his boss's desk.
Document 31
excised copy released on appeal by Air Combat Command
This study confirms that U.S. nuclear planners incorporated the Minuteman-only, launch-under-attack option into SIOP-5D, as of October 1 1979. In keeping with efforts during the 1970s to breakdown the SIOP into more discrete attack options, war planners initially conceived of LUA as a Selective Attack Option (SAO) because Minuteman missiles only would be committed to this option. During the 1980s, launch-under-attack was expanded to include bombers and submarine-launched ballistic missiles, thus turning it into a Major Attack Option.[28]
As a sign that launch-on-warning was becoming routinized in operational planning but that a requirement for definitive information on detonations would not be integral to planning, the Joint Chiefs of Staff began to include "launch under attack" in their dictionary of military terms. The Chiefs explained it as "execution by National Command Authorities of Single Integrated Operational Plan Forces subsequent to tactical warning of strategic nuclear attack against the United States and prior to first impact."[29]
Document 32
Jimmy Carter Presidential Library, Brzezinski Collection, box 23, Meetings-Muskie/Brown 7/80-9/81. Initially published in National Security Archive Electronic Briefing Book 390, Jimmy Carter's Controversial Nuclear Targeting Directive PD-59 Declassified
Jimmy Carter's Presidential Directive 59 on "Nuclear Weapons Employment Policy" sought a nuclear strategy that enhanced deterrence, but if deterrence should fail it sought to "terminate the war on acceptable terms that are as favorable as practical." While there was "no plausible definition" of victory in nuclear war, the U.S. had to be "capable of fighting successfully." PD 59 did not directly endorse launch-on warning - it was not U.S. "policy" - nor did it mention launch-under-attack Nevertheless, launch-on-warning was a capability that was to be available in the form of "pre-planned options" for attacks on "vulnerable" ICBMs. What senior officials had in mind was probably something like a Minuteman-only LUA option.
Document 33
FOIA appeal to the Department of Defense, Initially Published in National Security Archive Electronic Briefing Book No. 403, Declassified Pentagon History Provides Hair-Raising Scenarios of U.S. Vulnerabilities to Nuclear Attack through 1970s, 19 November 2012
With launch-on-warning options available in PD 59, defense officials in the new Reagan administration analyzed the capabilities of U.S. command-control-communications systems for implementing various launch-under-attack scenarios. A mid-1981 report found that in light of C3 vulnerabilities the prospects for a successful launch-under-attack were poor, either during a surprise attack or if U.S. strategic forces were on high alert ("fully generated"). For example, in the second scenario, "no combination of systems and procedure was fast enough to complete the process of warning assessment, decision-making, and emergency action message dissemination in the time available (3.5 minutes) between the first submarine-launched ballistic missile breakwater and attacks on command, control, and communications systems." Thus, the probability was "extremely low" that command-control-and-communications systems would provide "assured execution" of a launch-under-attack option. Trying to solve this problem was an important motive behind the Reagan administration's push for improvements in the C3 system during the years that followed,
Document 34 NEW
Release by Interagency Security Classification Appeals Panel, 2018
Following the logic of PD 59, Reagan's NSDD 13 sought to enhance deterrence partly through a military posture that "makes Soviet assessments of war outcomes, under any contingency, so uncertain and dangerous as to remove any incentive for initiating attack." If, however, deterrence should fail, Reagan mandated "pre-planned options" and "immediate options" for attack. Like PD 59, NSDD 13 supported a capability to "wage war successfully," but implied the possibility of winning nuclear war by seeking to "deny the Soviet Union a military victory at any level of conflict." Moreover, like PD 59, NSDD 13 assumed that command-control-communications-intelligence (C3I) systems had to be "sufficiently survivable and enduring to support" U.S. target plans, although that was a nearly Sisyphean task.
In some contrast to Carter policy, which authorized launch-on-warning options, NSDD 13 stipulated that it was U.S. policy not "to rely on" launch-on-warning in an "irrevocable manner." That appeared to proscribe ICBM launch-on-warning, if not bomber launch, but it did not rule out launch-under-attack as incorporated in the SIOP. In any event, the point was to "leave Soviet planners with strong uncertainty as to how we might actually respond to such warning." Moreover, upon receipt of warning information, the U.S. "must be prepared to launch our recallable bomber forces upon warning that a Soviet nuclear attack has been initiated."[30]
[1]. Richard Garwin, “Launch Under Attack to Redress Minuteman Vulnerability?,” International Security 4 (1979/80): 117-139; Fred C. Iklé, "The Growing Risk of War by Accident,” Washington Post, 24 June 1980; Charles Mohr, “A Scary Debate Over ‘Launch Under Attack,’” New York Times, 18 July 1982; Hendrick Smith, “Colonel Stirs Question On MX-Firing Doctrine,” The New York Times, 8 April 1983; William Broad, “Pershings Stir Accidental War Fears,” New York Times, 12 December 1983; John Steinbrunner, “Launch Under Attack,” The Scientific American, 250 (1984): 33-41.
[2]. For Bruce Blair's writings see, in particular, The Logic of Accidental Nuclear War (Washington, D.C., Brookings Institution, 1993), and Global Zero Alert for Nuclear Forces (Washington, D.C., Brookings Institution, 1995). For the Office of Technology Assessment study, see Tim Carrington, “The Ultimate Secret: A Pentagon Report Its Author Can't See – 'Nuclear Decapitation' Study Warns of Communication Being Destroyed in Crisis,” Wall Street Journal, 18 February 1986. See also Bruce Blair, The End of Nuclear Warfighting: Moving to a Deterrence-Only Posture: An Alternative U.S. Nuclear Posture Review, with Jessica Sleight and Emma Claire Foley (Princeton, N.J. and Washington, D.C.: Program on Science and Global Security, Princeton University, and Global Zero, 2019).
[3]. For accounts of early U.S. nuclear planning and the first SIOP, see Henry S. Rowen, "Formulating Strategic Doctrine," U.S. Commission on the Organization of the Government for the Conduct of Foreign Policy, Appendices, Vol. 4 (Washington, D.C., Government Printing Office, 1975), 217-34; David A. Rosenberg, "The Origins of Overkill: Nuclear Weapons and American Strategy, 1945-1960," Steven E. Miller, editor, Strategy and Nuclear Deterrence: An International Security Reader (Princeton, Princeton University Press, 1984), 113-82; and Scott D. Sagan, "SIOP-62: the Nuclear War Plan Briefing to President Kennedy," International Security 12 (Summer 1987): 22-51.
[4]. Blair, The Logic of Accidental Nuclear War, 170; Jerome Wiesner, "Warning and Defense in the Missile Age," 3 June 1959 (see document 2, below).
[5]. Robert S. McNamara, “No Second Use -Until,” The New York Times, 2 February 1983. For McNamara, Blair, and others in the mid-1980s debate, see "Science and the Citizen," The Scientific American 255 (1986): 74-76.
[6]. U.S. State Department, FRUS 1961-1963, 8:499-502.
[7]. Blair, Logic of Accidental Nuclear War, 170.
[8]. For a valuable and comprehensive account of the history of MIDAS and DSP, see Jeffrey Richelson, America's Space Sentinels: DSP Satellites and National Security (Lawrence, KS, University of Kansas Press, 1999).
[9]. Blair, Logic of Accidental Nuclear War, 168.
[10]. Ibid, 186-187.
[11]. Ibid, 187; information from Bruce Blair, 14 May 2019.
[12]. Pavel Podvig, “Does Russia Have a Launch-on-Warning Posture? The Soviet Union Didn’t,” Russian Strategic Nuclear Forces, 29 April 2019;
Blair, Logic of Accidental Nuclear War, 195-216; William Burr and Svetlana Savranskaya, eds., “Previously Classified Interviews with Former Soviet Officials Reveal U.S. Strategic Intelligence Failure Over Decades,” National Security Archive Electronic Briefing Book 285, 11 September 2009; “Putin Meets with Members of the Valdai Discussion Club, Full Transcript of the Plenary Session of the 15th Annual Meeting,” 18 October 2018.
See also Simon Saradzhyan, “Putin’s Remarks on Use of Nuclear Weapons Are Confusing, But Unlikely to Constitute a Shift in Nuclear Posture,” Belfer Center for Science and International Affairs, Russia Matters, 28 November 2018.
[13]. For background, see the major study by Stephen Twigge and Len Scott, Planning Armageddon: Britain, the United States, and the Command of Western Nuclear Forces, 1954-64 (Amsterdam, Harwood Academic Publishers, 2000), 36-37, 117-119.
[14]. For more details on PSAC thinking about MIDAS, see Richelson, America's Space Sentinels, 12-13, 17
[15]. Jacob Neufeld, Ballistic Missiles in the United States Air Force, 1945-1960 (Washington, D.C., Office of Air Force History, 1990), 213-14
[16]. Richelson, America's Space Sentinels, p. 256, n. 37, citing an interview with Jack Ruina. If Schriever ever made an explicit case for launch-on-warning in writing it remains classified or unknown to this researcher
[17]. For McNamara's decision and later developments, see Richelson, America's Space Sentinels, 28-48.
[18]. According to Scott Sagan, "SAC and Air Force contractor personnel appear to have improvised their own safety procedures in a manner that seriously compromised Minuteman nuclear safety." See Sagan, The Limits of Safety: Organizations, Accidents, and Nuclear Weapons (Princeton, Princeton University Press, 1992), 81-91.
[19]. For McNamara's decisions on the Minuteman force, see Stephen I. Schwartz et al., Atomic Audit: The Costs and Consequences of U.S. Nuclear Weapons Since 1940 (Washington, D.C., Brookings Institution, 1998),185-86, and Desmond Ball, Politics and Force Levels: The Strategic Missile Program of the Kennedy Administration (Berkeley: University of California Press, 1980).
[20]. Known also as Forward Scatter Radar, Over-the-Horizon (OTH) radar used high-frequency radio transmitters and receivers that were placed on either side of the Soviet Union and China. It would bounce continuous signals between the ionosphere and the earth until the signal reached the correct receiver. The system would detect a missile launch when it disrupted the stream of signals.
[21]. Bruce Blair communication with editor, 22 February 2001.
[22]. For the deployment of the DSP satellites during the late 1960s and early 1970s, see Richelson, America's Space Sentinels, 44-69.
[23]. Bruce Blair communication with the editor, 18 June 2019.
[24] Blair, Logic of Accidental Nuclear War, 342, note 40.
[25]. This resonates with a politico-military stratagem – the "Madman theory" – that has been associated with the Nixon administration: the notion that disproportionate threats and unpredictable irrationality could successfully coerce adversaries into compliance with U.S. goals. See William Burr and Jeffrey Kimball, Nixon's Nuclear Specter: The Secret Alert of 1969, Madman Diplomacy, and the Vietnam War (Lawrence: University of Kansas Press, 1998), 76-86.
[26]. E-mail from Pavel Podvig (editor of blog, Russian Strategic Nuclear Forces), 22 May 2019, drawing on his notes from the Vitalii Kataev papers (Hoover Institution Library and Archives, Stanford CA) and personal recollections.
[27]. For “jamming” the President in a crisis, see Bruce Blair, “Strengthening Checks on Presidential Nuclear Launch Authority,” Arms Control Today, January-February 2018, at 8.
[28]. For decisions on LUA, see Blair, Logic of Accidental Nuclear Warfare, 186.
[29]. Ibid., 168. For the downgrading of information confirming nuclear detonations in U.S. strategic planning during the 1980s, see ibid, 192.
[30]. For launch-on-warning policy as it developed during the Reagan administration, see Lee Butler, Uncommon Cause: A Life at Odds with Convention, Volume II: The Transformative Years (Denver, Outskirts Press, 2016), 10-12. For launch-on-warning not ruled out, see Michael Getler, “Joint Chiefs Back Plan for 100 MX,” Washington Post, 22 April 1982.