The Doomsday Machine

by Daniel Ellsberg

  Any JCS staff papers to be referred to the secretary of defense or his office were to be retyped to eliminate any possible references to the JSCP. If there was an absolute need to refer to such plans in some oblique fashion, the directive continued, reference was to be made to “capabilities planning” (lowercase), which would, again, not suggest the existence of a specific plan or suggest that it was any kind of war plan at all.

  That phrase—no less than the official title, Joint Strategic Capabilities Plan—was a euphemism, a cover. It was meant to obscure from the secretary—and more important, from his deputy and assistant secretaries and their civilian staffers—that there existed a single highest-level annual operational plan for the conduct of general and limited war, the authoritative guidance for all lower-level operational war plans.

  All this was intended to preempt the JCS nightmare: that the secretary or a civilian working for him might see this acronym in a document, might ask what it meant, and then ask to see the plan. This could open the possibility of civilians working for the president actually reviewing the plan and demanding changes. A vague reference to “issues arising in capabilities planning,” which the JCS directive prescribed, gave such officials no handle to ask for a specific document, or a hint that there was an overriding piece of paper that would be worth their while to read.

  As a result, almost no civilian, including the secretary of defense, was aware that a piece of paper of the character of the JSCP existed. That, of course, extended to its critical “Annex C”—the SAC war plan, which laid out in some detail the nature of our general (nuclear) war operations. The JSCP stated, “In the event of general war, Annex C would be executed.”

  Reading that statement by itself, knowing what Annex C was, one might naturally infer that that provision virtually defined “general war,” in operational terms. It was when the president would direct that the SAC war plan, attached, should be executed against our principal adversary, the Soviet Union. But when might he do that?

  Obviously, such a fateful decision would depend on circumstances and the president’s judgment, and could be left unspecified. At the same time, it would have seemed natural to give some sense of the various circumstances contemplated. Obviously, the arrival or unmistakable imminence of a Soviet surprise nuclear attack on the United States or its forces would be one such circumstance. This was the scenario that RAND strategic analysts focused on almost exclusively.

  Those with knowledge of NATO planning and commitments (almost any planner working in the Pentagon) would know that a massive Soviet non-nuclear attack that threatened to overwhelm NATO forces and to occupy Western Europe would be a compelling occasion for the president to launch the SAC war plan. (Polls throughout the Cold War showed, surprisingly, that most Americans—in contrast to most citizens of Western Europe—were not aware that the United States had made such an official commitment to our NATO allies.) Were there any other circumstances that justified the deployment of SAC and other tactical nuclear forces?

  In fact, an explicit definition of “general war” did appear in the JSCP. This was perhaps the most sensitive piece of information in the entire document, and the main reason for protecting it from the eyes of civilian authority. I’ll never forget the moment when—thanks to Lieutenant Colonel Bob Lukeman of the Air Staff—I was given an opportunity in a basement room of the Pentagon to read this holy of holies, and I finally came upon the definition: “General War is defined as armed conflict with the Soviet Union.”

  To properly understand the hair-raising import of this proposition, one had to read it in the context of two other key assertions in the JSCP: “In general war, Annex C will be executed”56; and “in general war, a war in which the armed forces of the USSR and of the U.S. are overtly engaged, the basic military objective of the U.S. Armed Forces is the defeat of the Sino-Soviet Bloc.”

  The meaning of “armed conflict,” in this case the key trigger to unleashing the full fury of the SAC war plan against both the Soviet bloc and China, was subject to some narrow controversy in military circles. It was generally accepted that a platoon or company-level skirmish with Russian forces in the Berlin corridors or on the borders of East Germany—which might not represent a deliberate decision by any Soviet leader—was not to be regarded as “armed conflict” for the purposes of this definition in the JSCP. But what about a brigade-level conflict (two battalions) or a division or two (as could quickly erupt in a hot Berlin crisis)? That would undoubtedly meet the conditions for the definition, and what was directed to follow from it.

  And the definition was not confined to Europe. It implied that any conflict pitting U.S. forces against any more than several battalions of Soviet troops anywhere in the world—Iran, Korea, the Middle East, Indochina—would lead to instant U.S. strategic attacks on every city and command center in the Soviet Union and China. It was hard to imagine that such a plan could actually be carried out. Yet according to what I had already come to discover in the Pacific, and what turned out to apply worldwide, no alternative plans existed for a war involving Soviet forces on a level beyond a division or two except for the general war plan. And that lack was by the directive of President Eisenhower, who had decreed that there should be no plans for “limited war” with the Soviet Union, whether nuclear or conventional, under any circumstances, anywhere in the world.

  This reflected Eisenhower’s military judgment that no war between any significant forces of the United States and the Soviet Union could remain limited more than momentarily. Therefore if such a conflict were pending, the United States should immediately go to an all-out nuclear first strike rather than allow the Soviets to do so.

  But even if those military judgments were challenged—as they were, repeatedly, by Army Chief of Staff General Maxwell Taylor—Eisenhower believed that any alternative approach was unacceptable from a fiscal point of view. Under the influence of conservative economic advisors, he was convinced that preparation to fight even a limited number of Soviet divisions on the ground (as Taylor proposed), with or without the use of tactical nuclear weapons, would compel an increase in defense spending that would cause inflation, precipitating a depression and “national bankruptcy.”

  The budgetary battle between the services had come to be fought, oddly, over the definition of “general war.” All the services accepted that “general war,” in the nuclear era, implied all-out nuclear war with the Soviet Union, in which the Strategic Air Command would play a predominant role. The Navy, with its carrier aircraft and submarines, would be second in importance, with Army operations problematic and subordinate at best. For purposes of planning—the structure of forces, deployments, and operations—and, above all (from the services’ point of view), determining the size and division of the budget, the vital question was when, among the wide range of possible circumstances in the world threatening to U.S. interests, such an apocalyptic response might be invoked.

  Army leaders like Taylor, and initially those in the Navy as well, wanted to define “general war” as narrowly as possible, leaving a broad range of conflict situations to be planned for, budgeted for, and addressed if they arose, without necessarily involving an attack by SAC on the Soviet Union or China. They argued, with a good deal of plausibility, that since such an attack involved a high risk, if not a certainty, of devastating retaliation against the United States, it should be reserved for only the most extreme, exigent contingencies.

  One definition they proposed was that “general war” was an armed conflict with the USSR and the United States as the principal protagonists “with the national survival of both deemed to be at issue.” Air Force Intelligence countered, according to my notes, that USAF “does not accept implication that there could be armed conflict between the US and SU in which the national survival of both was not at issue.” As early as 1956 Eisenhower sided with the Air Force on this, against Taylor, asserting that the qualifying phrase at the end should be omitted from the definition, which would simply be “armed conflict
with the USSR.”

  The Army and Navy didn’t give up, though they continued to be overruled. In my notes of the Army and Navy view as of October 30, 1959, general war should be defined as “governmentally directed overt armed conflict between nations with the objective of complete subjugation or destruction of the national entity of the enemy,” with other forms of armed conflict, including between U.S. and Soviet forces, characterized by “limitations on locale, weapons, forces, participants, or objectives.”

  To the layperson, this might appear sensible enough. But what Eisenhower, the Air Force, and successive chairmen of the JCS detected behind these innocent-sounding definitions was a charter for the Army to go to their allies in Congress to seek capabilities for fighting even multiple Soviet divisions in a limited, non-nuclear, and non-general war. That was precisely what the budget-obsessed President Eisenhower and the Army’s service rivals feared and wanted to avoid. My friend Colonel Ernie Cragg in Air Plans was pointing out in dueling memos with the Army as late as January 21, 1961 (the day after Kennedy’s inauguration):

  Adoption of the “view” that limited wars between the US and the USSR are possible is an “invitation” to attack. It also could open Pandora’s box with respect to forces for limited war at the expense of general war forces.… It would allow the Army and Navy to increase their “requirement” for forces for limited war to almost unlimited levels.

  The latter point, frequently echoed by Eisenhower, seemed especially compelling since, for years, both American and NATO intelligence had produced enormously inflated estimates of Soviet ground strength. For example, they ignored that a Russian division was less than half the size of an American division. Moreover, the supposed number of Soviet divisions was grotesquely overstated. Most of the oft-quoted figure of “175 Soviet divisions” referred to units that existed only on paper, subject to wartime mobilization, or to units that were grossly undermanned and under-equipped, and many that consisted of nothing more than a headquarters staff. Still, matching even the twenty crack Soviet armored divisions deployed in East Germany would legitimize large budgetary Army requests. These, if granted, would come at the direct cost of Air Force and Navy budgets.

  One major reason for the JCS to keep any dispute over numbers from the attention of the secretary of defense was a fear that he would decide them in a way that would unfavorably influence the budget for the Air Force, or for one or another service. Even though Secretary of Defense Gates increasingly insisted on having a say in operational matters, in practice he became aware of only those problems that the chiefs unanimously agreed to submit to him. This had to represent a definite judgment on each of their parts that they had more to gain by such a submission than they did by bargaining among themselves. So while many important problems were never brought to the defense secretary’s attention, one matter that was, however, was the definition of “general war” to be used in “capabilities planning.” In June 1960, by my notes, Secretary Gates confirmed the definition: “war with the USSR.”

  Since there was to be only one plan for fighting Russians anywhere in the world under any circumstances—including, along with SAC, Polaris submarines, and theater forces—Eisenhower endorsed in 1959 the coordination of a single strategic plan at SAC headquarters in Omaha. Annex C of the JSCP came to be, by December 1960, the Single Integrated Operational Plan (SIOP).

  By 1960, the planners of the SIOP, the Joint Strategic Target Planning Staff, had gathered all the general-war target lists of the various commands, including SAC, NATO, and PACOM, into one coordinated target list to allocate weapons more efficiently to targets all over the world. A major argument for centering the targeting effort at SAC headquarters was the allegedly unique capabilities of SAC’s computers. In reality, these were still at such a rudimentary state of development that much or most of the computation had still to be done by hand, with the aid of calculating machines.

  Again, there was an intense concern for minimizing the “interference” or “fratricide” of vehicles aimed at targets in close proximity. There was also a desire by Eisenhower to reduce “duplication” of efforts by different commands. In the actual planning, both concerns were totally frustrated; the latter because each command and service was determined to cover important targets by its own forces. By some counts, over eighty weapons were dedicated to Moscow; other counts put this at a hundred and eighty. Meanwhile, the prevention of “interference,” as in the Pacific, remained a delusional objective.

  As with the CINCPAC plan I had read earlier, the coordination involved in this higher-level plan was so complex that there was room for only one real strategy. The price of bringing all the theater and component service plans into harmony with each other, into one plan, was the total elimination of any flexibility in carrying it out. So much planning was involved in producing this one scenario that there was simply no staff or computer time available to produce an alternative. As with the CINCPAC planners I met earlier, the SIOP planners themselves were appalled at the confusion and chaos that might ensue if any alternative was proposed.

  Following the guidance of the JSCP, the planners at SAC headquarters set out to weld all the warheads in the U.S. arsenal into one hydra-headed monster that would arrive on its targets as near simultaneously as possible, preferably before any Soviet warheads had launched.

  On strips like Kunsan or Kadena and on aircraft carriers surrounding the Sino-Soviet bloc (as it was still described in 1961, though China and the Soviets had actually split apart a couple of years before that), more than a thousand tactical fighter-bombers were armed with H-bombs in range of Russia and China. Each of them could devastate a city with one bomb. For a larger metropolitan area, it might take two. Yet until this time, SAC planners had regarded these tactical theater forces as so vulnerable, unreliable, and insignificant a factor in all-out nuclear war that they had not even bothered to include them in calculating the outcome of attacks in a general war.

  In 1961 there were about seventeen hundred SAC bombers, including over six hundred B-52s and a thousand B-47s. In the bomb bays of the SAC planes were thermonuclear bombs much larger than those I had seen in Okinawa. Many were from five to twenty-five megatons in yield. Each twenty-five-megaton bomb—with 1,250 times the yield of the fission bomb that destroyed Nagasaki—was the equivalent of twenty-five million tons of TNT, or over twelve times the total bomb tonnage we dropped in World War II. Within the arsenal there were some five hundred bombs with an explosive power of twenty-five megatons. Each of these warheads had more firepower than all the bombs and shells exploded in all the wars of human history.

  These intercontinental bombers and missiles had come to be stationed almost entirely in the continental United States, though they might be deployed to forward bases overseas in a crisis. A small force of B-52s was constantly airborne. Many of the rest were on alert. I had seen a classified film of an incredible maneuver in which a column of B-58s—smaller than B-52s but still intercontinental heavy bombers—taxied down a runway and then took off simultaneously, rather than one at a time. The point—as at Kadena and elsewhere—was to get in the air and away from the field as fast as possible, on warning of an imminent attack, before an enemy missile might arrive. In the time it would normally have taken for a single plane to take off, a squadron of planes would be airborne, on their way to their pre-assigned targets.

  In the film, these heavy bombers, each as big as an airliner, sped up in tandem as they raced down the airstrip, one behind the other so close that if one had slackened its pace for an instant, the plane behind, with its full fuel load and its multiple thermonuclear weapons, would have rammed into its tail. Then they lifted together, like a flock of birds startled by a gunshot. It was an astonishing sight: beautiful and terrible at once.

  On carriers, smaller tactical bombers would be boosted on takeoff by a catapult, a kind of large slingshot. But since the general nuclear war plan, as I knew, called for takeoff around the world of as many U.S. planes and missiles as were ready
at the time of the Execute order—as near simultaneously as possible—the preparations contemplated one overall, inflexible global attack, as if the entire destructive arsenal of the United States were launched by a single catapult—a slingshot made for Goliath.

  The preplanned targets for the whole force included, along with military sites, every city in the Soviet Union and China. There was at least one warhead allocated for every city of 25,000 people or more in the Soviet Union. The “military” targets (many of them in or near cities, and many only tendentiously described as military) were by far the great majority, since all the cities could be totally destroyed by a small fraction of the attacking vehicles.

  In 1960–61 it was in reality quite possible—though USAF and CIA “missile gap” estimates implied otherwise—that not a single nuclear warhead would land on U.S. territory after such an American first strike. Worldwide fallout in the stratosphere from our own strikes would certainly kill Americans, but over so long a time, with radioactivity decaying in the atmosphere on the way over and deaths from cancer long delayed, that the increase in mortality in any one year might not be statistically perceptible. But our Western European allies in NATO would be quickly annihilated twice over: first from the mobile Soviet medium-range missiles and tactical bombers targeted on them, which our first strike couldn’t find and destroy reliably, and second from the close-in fallout from our own nuclear strikes on Soviet bloc territory.