The Doomsday Machine

by Daniel Ellsberg

  Harry Rowen told me what happened next. Soon after the directive went out from McNamara, Nitze happened to be at a meeting in McNamara’s office on another matter, along with Admiral Burke. At the end of the meeting, Burke asked Nitze to return with him to Burke’s office, in a different part of the Pentagon. When they arrived at the office, Burke sat down at his desk. Nitze saw immediately that he had in front of him a “burn copy” (the predecessor to the Xerox copy—a somewhat fuzzy duplicate on tan, flimsy paper) of my “Top Secret—Eyes Only” memo, which was intended for Nitze alone and wasn’t supposed to be copied.

  It was obvious that some commander or captain or rear admiral working for Nitze had seen my memo, “burned” it, and delivered the copy to Admiral Burke. He also had on the desk a copy of the ISA investigative report, along with McNamara’s directive to him.

  Burke started discussing my memo and the report, not bothering to explain what they were doing on his desk. “Burke was furious,” Nitze told Harry. Burke was famously given to rages, but this one, in front of an assistant secretary of defense, was surprising to Nitze. Burke made no attempt to deny the facts of the reports or to justify anything. The only thing he had to say, in a fury, was “What did you think you were doing, as a civilian, interfering with the operations of ships of the U.S. Navy?”

  The fact that this ship was in violation of one of our most important security treaties and was posing enormous diplomatic risks, that it was carrying nuclear weapons in violation of regulations on their whereabouts and in deliberate deception of the secretary of defense, that the special assistant to the secretary had been lied to by the Navy—none of these points was addressed by Burke, nor was he willing to hear about them. He maintained that it was absolutely unacceptable that the secretary of defense should presume to tell the Navy where to put its ships.

  Rowen got the impression that Nitze left the office quite shaken by Burke’s willingness to confront him in this way but determined to have the Navy brought into line. He himself wasn’t in a clear-cut command position with respect to Burke, except as he was accepted as a direct representative of the secretary. So everything depended on McNamara’s standing by his directive and backing Nitze on this issue. Rowen told me that Nitze went to McNamara and told him that this was of the highest urgency and that he should order Burke to comply with his directive and with the treaty.

  I asked Harry, “So, what’s happened?”

  “McNamara decided to withdraw the directive. He backed off. With all the fights he’s having with the services, he didn’t want to add this one.”

  I asked, “Does McNamara know he was lied to by the Navy?”

  “Yes.” Harry said, “That’s what made him furious in the first place. It’s what got him to send the directive.”

  But faced with Nitze’s account of Burke’s own rage, McNamara had to pick his fights, which included a struggle over the number of nuclear-powered carriers. In this case, I could guess, he would face the likelihood that the Navy would leak the dispute to a friendly committee in Congress, in distorted fashion, and make him defend himself from the charge he was unduly entering into operations by ordering around individual ships.

  I myself was faced with questions from the vice president of RAND, Richard Goldstein, when I returned to California. General LeMay had recently sat in on a meeting of the Air Force Advisory Board, which controlled the RAND budget. Goldstein called me into his office and said, “Dan, this is hard to believe, but we have a charge here from General LeMay—he’s been told by Admiral Burke—that you have been giving the Navy orders on how to operate a destroyer squadron. Is this possible?”

  I said, “What?!” It was true that most of the things I was doing in Washington would look madly presumptuous to most military officers, but I told him I had never done anything remotely like that. It took a second or two to guess what it must be referring to. The mention of Burke was the tip-off. I told Goldstein the whole story, and he passed it on. No one reprimanded me, though I learned that Burke had asked LeMay to have me fired from RAND.

  The San Joaquin County went back to Iwakuni from Okinawa, with its cargo of nuclear weapons. A couple of years later Nitze made another abortive attempt to move it, but it stayed until 1966, when Edwin Reischauer, our ambassador to Japan, learned of its presence—through a leak to his office—and demanded that it be removed. He threatened to resign if it wasn’t. In 1967 it finally moved back to Okinawa.

  a Long after drafting this account, I read John Rubel’s memoir Reflections on Fame and Some Famous Men (New Mexico, 2009), in which he confirms my memory of General LeMay’s locution in discussing presidential command authority. “The matter came up as I stood facing the general as he stood across from me at his desk. Somehow … I must have uttered the term, ‘command and control.’

  “LeMay expostulated contemptuously, ‘Command and control! Command and control! What’s that? It’s telling the fighting man what to do, that’s what it is. And that’s a job for the professional soldier. They talk about the president exercising command and control. What is the president?’ He spit out the ‘p’ in ‘president.’ ‘A politician.’ ‘What does a politician know about war?’ He dwelled on w-a-a-h-r. ‘Who needs the president if there’s a w-a-a-h-r? Nobody! All we need him for is to tell us that there’s a war. [Not that Curtis LeMay needed a president for that.] We are professional soldiers. We’ll take care of the rest.’ ” (pp. 65–66)

  CHAPTER 8

  “My” War Plan

  Meanwhile, throughout the spring of 1961 I was working on a project that Harry Rowen had passed on to me. He had been asked by his boss, Paul Nitze, to draft the new basic national security policy (BNSP) for the Department of Defense. President Eisenhower had initiated this series of annual Top Secret policy documents to serve as the civilian authority’s statement on the objectives and guidelines for all war planning within the Department of Defense.

  Under Eisenhower each BNSP, usually no more than three or four pages, had embodied the “New Look” and “massive retaliation” doctrines of John Foster Dulles and chairman of the JCS Arthur W. Radford. These emphasized “main but not sole reliance” on nuclear weapons, as opposed to conventional, or non-nuclear, weapons. In fact, this emphasis was expressed in a trend by Eisenhower officials of describing nuclear weapons themselves as “conventional.” For some years as a senator, John Kennedy had been associated with a critique of massive retaliation and an espousal of what General Maxwell Taylor, chairman of the JCS, called a “strategy of flexible response.” Thus, following JFK’s election as president, it was understood in the Pentagon that a significant change in the policy directives guiding war planning was in order. It was assumed that this would take the form of a radically revised basic national security policy.

  Already in February, Bill Kaufmann of the RAND social science department had briefed Secretary of Defense McNamara on some proposals within the Air Staff aimed at moving away from what our RAND colleague Herman Kahn labeled a “spasm” concept of general war (or, as he often put it, referring to its all-out, nothing-held-back character, “a wargasm”). Kaufmann had pressed instead for developing a capability for sustained and controlled “war fighting,” focused mainly on military targets, with cities witheld from initial attack.

  In the eyes of advocates in the Air Force (which did not include LeMay), this was a strategy aimed as much at their service rivals in the Navy as it was at the Soviet Union. The sub-launched missiles on Polaris submarines, after all, were smaller than those on Air Force land-based ICBMs and much smaller than those on bombers. They were also, in the days before GPS systems, much less accurate than either of those vehicles. While the relatively invulnerable Polaris missiles were well suited to a deterrent policy aimed at targeting cities, they were less effective against hardened military targets like silo-based ICBMs (of which the Soviets had none in 1961, but were expected shortly to have hundreds to thousands). For that strategy, the Air Force bombers and missiles had an advantage. In fact th
e only role in which they had a relative advantage was as counterforce weapons—limiting damage to the United States in an all-out first strike by destroying Soviet land-based missiles before they were launched. Naturally, this was the preferred Air Force position, and after hearing Kaufmann’s briefing on the arguments, McNamara had seemed sympathetic.

  It would have been natural then for Harry to assign Kaufmann—who was also, like me, working as an ISA consultant in Washington at this time—to draft the general war section of the new BNSP. But somewhat to my surprise, Rowen asked me to draft that section, and instead assigned Kaufmann the task of drafting a section on limited nuclear war. I knew that my own views as to what the policy should be were closer to Rowen’s attitudes than Kaufmann’s on the importance of counterforce, and I presumed that was why I was given the job. This encouraged me to undertake the drafting as a process of refining my own views and making them as specific as possible, with the expectation that the end result would probably be acceptable to Rowen. That must have been his expectation too, because the only directive he had given me was, “Write what you think the guidance should be.” That’s what I proceeded to do.

  The Air Force concept of “war fighting” or “damage limiting” (limiting damage to the United States from Soviet offensive nuclear forces) involved prolonged and controlled counterforce attacks on a military target system in the Soviet Union and its satellites in Eastern Europe. That included precise attacks against hardened missile sites and command and control centers. Along with our land-based ICBMs, this concept called for increased numbers of high-performance bombers capable of penetrating Russian defenses—by either flying underneath the radar or higher than the range of air defense missiles—to deliver high-yield payloads more precisely than missiles could do. Thus it gave support to the Air Force proposals for the B-70 bomber program (later called the B-1).

  General Curtis LeMay, the former head of SAC, was by instinct hostile to ideas that downplayed attacks on cities. But he wanted the Air Force to have big, high-flying, fast bombers that could carry a heavy payload, and he was passionately devoted to the B-70 program. My friend Lieutenant Colonel Bob Lukeman said to me that he had heard a civilian in the Pentagon, on being told by a general officer that the first priority of the Air Force in 1961 was to get the B-70, ask what its second priority was. He was told, “There is no second priority.” Lukeman explained to me, “The Air Force has one priority at a time.”

  The counterforce strategy also implied a crash effort to improve the accuracy of our missiles. That was another objective then thought to disfavor the Navy’s mobile missile, the Polaris, in favor of land-based missiles controlled by the Air Force. Meanwhile, it called for large numbers of the land-based missiles to compensate for their current inaccuracy against small and hardened military targets. Before the year was out, LeMay and General Thomas Power, who had earlier replaced LeMay as the head of SAC, were calling for ten thousand of the prospective Minuteman land-based ICBMs.

  In making this pitch to McNamara, Kaufmann certainly gave him what our RAND USAF sponsors wanted in the way of a rationale for the B-70 and for land-based missiles controlled by the Air Force. Whether he truly believed in the pitch for the Air Force’s counterforce mission I never knew for sure. Others at RAND were unconvinced by the Kaufmann-USAF strategy. Yet it was not unwelcome to RAND management and analysts to find that Kaufmann’s briefing was met with enthusiasm both by Air Force high command (some of whose subordinates had actually inspired it) and by the new secretary of defense.

  I was one of the doubters, and so was Harry Rowen. On the basis of the calculations I’d seen at RAND, the merits of counterforce attacks in limiting damage either to the target area or to the United States did not impress me. Assuming as we did in the spring of 1961 that the Soviets had sizable numbers of ICBMs, before long to be hardened and accompanied by sub-launched missiles, it seemed obvious to me that once a process of general nuclear war was under way, nothing could be relied on very much to limit damage to less than catastrophic levels. Thus, there was an incalculably vast premium for all nuclear powers on deterring, preventing, and avoiding a general nuclear war under any circumstances.

  However, the question remained: What did we do if deterrence failed? Like Kaufmann, I accepted the notion that if a general nuclear war did commence, our land-based missiles should be aimed at military targets, primarily Soviet missiles and air bases, for whatever prospect of limiting damage that might offer. But what seemed to me to offer relatively more promise than simply the “damage-limiting by controlled counterforce attacks” that Kaufmann was expounding were the other parts of the “coercive,” “no cities” strategy he had presented to McNamara, aimed at terminating the war as quickly as possible, before all weapons on both sides had been employed, and particularly before any, many, or all had been employed against urban targets. (It was a long time before I recognized the fatal contradictions between these two components of what was supposedly the same strategy, along with the effective infeasibility of either of them.)

  This latter goal required both deterring—if possible—an opponent from launching strikes against the cities of the United States and its allies, even if nuclear war was initiated by one side or the other, and at the same time inducing the opponent’s command authority to stop operations short of expending all his weapons.

  Both of these sub-goals called for three characteristics in our own planning and operations. First, it meant avoiding enemy cities altogether in our own initial strikes: what came to be known as a “no cities” approach. We would have to announce that intention long before hostilities began. Right away, this would be a marked departure from a policy of indicating beforehand and then carrying out our intent to destroy cities in all circumstances, a policy that removed any restraint on the enemy from targeting our own cities.

  Second, it required maintaining protected and controlled U.S. reserve forces under virtually all circumstances, thus preserving a threat capability in order to terminate the war. That might also deter enemy preparations to destroy our cities as an inevitable and automatic wartime strategy.

  Third, it called for preserving on both sides a command and control system capable of both controlling reserve forces and terminating operations. We would need a survivable command system capable of more than a simple “Go” decision; and we could not afford to deprive the Soviets62 of the same capability.†

  The first two of these points were part of Kaufmann’s briefing to McNamara in early 1961. They were rooted in RAND discussions and studies going back years before I arrived there, starting with Bernard Brodie, Andrew Marshall, Jim Digby, Charlie Hitch, and others, as Fred Kaplan has ably shown.63 Both notions appealed to me from my own background as soon as I encountered them at RAND: “no cities” from my earliest aversion to bombing civilians; the notion of inducing the Soviets, possibly, to do the same, and to end the war, by holding their own cities hostage to our reserve forces, was in line with my more recent analysis of coercion. That focus led me to emphasize, more than others did, the key importance of preserving command and control on both sides, not only our own. (That turned out to be an insuperable disability of the approach in the eyes of military planners on both sides.)

  Neither I nor Harry Rowen had at that time the illusion that any such planned measures had a great likelihood of achieving the desired effects either on enemy planning or on the course of hostilities. Still, the possibility of having some desirable effect—saving some, perhaps many cities on both sides, and ending a war, once started, short of mutual annihilation—seemed greater than with the Air Staff strategy, which focused exclusively on counterforce tactics and failed to provide any mechanism for either terminating the war or significantly limiting damage.

  This combination of damage-limiting and war-terminating goals implied that there were some choices to be made by the highest surviving U.S. authority even after general war hostilities had begun. Given the nature and urgency of such decisions—which might mean life
or death for our entire society—it was obviously desirable that the president himself, or at least someone having his full confidence, be physically capable of making such decisions after general war had begun. (Eisenhower’s wholly preplanned, truly “single” integrated operational plan didn’t really require this.) That required preserving the president or his representative physically as well as preserving reliable communication capabilities.

  At the same time, I saw a second, more important, advantage in demanding that a highest civilian authority be able to choose among options even in the midst of war.

  It provided a rationale—a “need to know”—for the president to inform himself and his civilian advisors before the war of the detailed nature of the proposed war planning.

  And there was a third advantage to my plan: Once it was acknowledged that the capability of highest-level authority (civilian or military) to command should be preserved during war, and once physical measures had been taken to achieve that with high reliability, there could no longer be compelling military objections against implementing lower-level physical controls over nuclear weapons—Permissive Action Links (PALs)—that would make it impossible for lower commanders to mistakenly or insubordinately initiate the use of nuclear weapons on their own.

  Finally, focusing critical attention on the existing, current plans, which called for the prompt destruction of Soviet and Chinese urban targets under all circumstances of general war, opened up the possibility of a strategic and moral critique of such plans. Up until now these plans had been regarded as beyond question—especially by civilians—because there was supposedly no alternative to them for purposes of deterrence or wartime operations.