The Doomsday Machine

by Daniel Ellsberg

  The clearest memory I have of the game is that, when it ended, I was walking out of the building with Abe Chayes, the former Harvard law professor who was legal counsel of the State Department, when he turned to me in the dusk and said, “We’ve got to get out of Berlin.”

  I looked at him and didn’t say anything. He said, “You know, our position there is totally untenable. This game shows it. There’s no way we can defend that place.” In national security circles, that was the greatest heresy imaginable. I never heard anyone else say it, before or after. Yet in conventional military terms, there was no question: West Berlin was indefensible by NATO. It was not only in the middle of East Germany, surrounded by Soviet forces, but also the forces that surrounded it were the best in the Soviet Army. There were twenty-two Soviet divisions in the vicinity, mostly tank divisions with their latest model tanks, tremendously surpassing anything we could throw in there. And if it came to theater nuclear warfare, the Soviets had as many or more tactical nuclear weapons. There was no possibility of militarily confronting them effectively.

  If we had any plan at all for responding if the Soviets simply walked a division into West Berlin—in effect, arresting our small military garrison there, or capturing the remnants if they put up any fight—I never heard of it. When it came to that contingency, I believe we didn’t even want to think about it. The only challenge we envisioned was their repeating what they had done in 1948, blockading, using East Germans, in the first instance, to cut off all access to the city, but this time also blocking air access as well. The only way you could prevent them militarily from doing that with total effectiveness relied on the threat of initiating nuclear war.

  Of course there would be a series of steps leading up to that. Paul Nitze was in charge in the Pentagon of contingency planning for a Berlin crisis. The plans anticipated sending a small American unit to test the blockage at first, very small to begin with, a couple of platoons or a company along the access road. And if they were stopped, we’d send a battalion. If that was outnumbered by a blocking force, we might send a brigade or a regiment. But the proposals I’d seen all stopped there.

  At that point, we would be facing Eisenhower’s definition of “general war.” In mid-1961, the only plan the Pentagon actually had for a large conflict with Soviet forces, left over from Eisenhower, was an all-out nuclear attack on the Soviet Union. Roswell Gilpatric had told me that it was the intent of McNamara and himself, if it came to a crisis like that, to throw out the old plans and create new ones. But really, what was there to do? NATO planning simply wasn’t designed—by the nature of the alliance, and by the realities of the situation—for offensive operations into the Warsaw Pact territory.

  Kennedy and McNamara did, over time, introduce to NATO the notion of a “flexible response” starting with conventional defense against a large-scale Soviet offensive into West Germany. That might even include “demonstration” nuclear shots, one or two warheads against carefully selected targets, to warn the Soviets of the cataclysm that threatened just ahead and get them to back off. McNamara had indicated to me in the luncheon in his office that he would never recommend any such thing. And as I’ve said, much later he revealed that in fact he had advised both Presidents Kennedy and Johnson never, under any circumstances, to initiate the use of nuclear weapons. He said that they had agreed with him. So that was, privately, a bluff. But the recklessness of such a crazy “demonstration,” the likelihood of its simply triggering heavy nuclear attacks from both sides, seemed obvious to me. I could only hope the same was true for our allies.

  So there was no way for me to answer Chayes about the indefensibility of West Berlin without sounding to myself wildly reckless: as wild as our actual strategy appeared to be. There was, after all, just one way to hold on to West Berlin (without negotiating an agreement with the Soviets and East Germans!). It was what we had relied on since Khrushchev’s first ultimatum in 1958, and continued to do for the next generation. It was to threaten to carry out our actual Berlin planning. The planning we were threatening to carry out was best described by a skeptical Pentagon colleague: “We send in a series of increasingly larger probes. If they’re all stopped, we fire a [nuclear] warning shot. If that doesn’t work, we blow up the world.”

  That’s what we were doing. Chayes was saying, in effect, that this was not a good plan: not even, he implied, as a bluff. It was not sufficiently credible or reliable as a threat, and catastrophic if carried out. Holding on to West Berlin was not worth it. I wasn’t inclined to argue with that. Yet I wasn’t ready to come out and say that I agreed with what he had just said, even to myself. My memories of the Berlin blockade when I was seventeen, reinforced by my years in the Marine Corps in my twenties and my Cold War beliefs, were still too strong for that.

  At the same time, the idea of actually carrying out the threats we were relying on was total anathema to me. And incredibly, in August 1961, we were making those threats at a time when, according to our official estimates, we were inferior in power to the Soviet Union in strategic nuclear weapons. I couldn’t believe that anyone in the Pentagon or the administration could take the idea of actually launching a U.S. first strike seriously in face of that opposing reality.

  Yes, Andy Marshall had told me the year before—with no explanation—that “there isn’t going to be a missile gap,” and McNamara had said in February that there was no gap,84 in a backgrounder press briefing that he had thought was not for attribution. When his statement was published, he had offered to resign for embarrassing the president, who had just campaigned to eliminate a missile gap; Kennedy had brushed off the gaffe. But McNamara might have been wrong. (He had been persuaded of his judgment by Eisenhower’s departing secretary of defense, Thomas Gates.)

  It was true that the lowest estimate, aside from the purportedly aberrant one by the Army and Navy, showed no significant gap—fifty Soviet ICBMs to our forty—but the majority opinion in the intelligence community was estimating a missile gap in favor of the Soviets as late as the June 7 NIE, a few days after the Vienna summit.

  * * *

  In the last week of September 1961, Alain Enthoven, now the assistant secretary of defense for systems analysis, and Harry Rowen in ISA informed me of a new national intelligence estimate. It was astonishing. It essentially confirmed what the Army and Navy estimators had been saying for two years in their footnote dissents to the NIEs: that the Soviets had “only a few” ICBMs. The number observed was actually four.

  “Observed”—that was the big secret. Neither Alain nor Harry told me, initially, just what type of information the new intelligence report reflected, but within days, discussions in the Pentagon had revealed it to me inadvertently. This was not just an “estimate,” based on inferences about production capabilities, or Soviet “requirements,” or ambiguous electronic intelligence. Four missiles had actually been seen, and photographed, at one site at Plesetsk by our most highly secret intelligence system at the time, the Corona satellite program. (The cover name for the program was Discoverer. It had replaced the U-2 spy plane program, which had been equally secret until the Soviets shot down a CIA U-2 over Russia and captured its pilot Gary Powers in 1960.) No other missile launchers had been seen elsewhere in the Soviet Union—except for a couple of prototype launch sites at the missile-test complex at Tyuratam—after what was finally nearly complete coverage of possible missile sites.

  The fact that this was “hard” intelligence based on actual photos was what is now called sensitive compartmented information (SCI), higher than Top Secret. Access to it required a Keyhole (K) clearance, higher than Top Secret, which I didn’t have at the time. The existence of clearances higher than Top Secret was in those years itself a well-kept secret, along with the nature of the information each of them covered and the actual information itself. It was extremely unusual for anyone holding such a clearance to give any hint of these secrets to someone who didn’t have the special clearance.

  The penalty for a security breach of th
at nature was to be dropped immediately, within minutes of the discovery of the indiscretion, from the computer listings of those with access to the special clearances. That meant exclusion from the list of those who counted in national security discussions within the government—those who had access to this information and could talk freely among themselves. That sanction helped keep those secrets very, very well. Leaks to the press were nonexistent, either about the clearances, the intelligence means, or the contents of the information. Breaches of discipline, either deliberate or inadvertent, even to close colleagues who hadn’t been specially cleared, simply didn’t occur, with few exceptions.

  I happened to benefit from several such exceptional breaches. Talking with Colonel Ernie Cragg, of the plans division of the Air Staff, one late night in the Pentagon cafeteria, I asked him something about the basis for the new missile estimates. He started to answer, then broke off, looked at me, and asked, “Are you cleared for T and K?”

  I said no, and Cragg clammed up, evidently realizing he’d already said more than he should have.

  Cragg’s question was breach number one. As I was briefed later, when I did get such clearances, if he were in doubt as to whether he was dealing with someone who was entitled to this information, he should never have mentioned to that person the code letters revealing the existence of these clearances. If he really wanted to discuss these matters, he should have excused himself, gone to a Pentagon phone to call a special number, identified himself by a code, and asked the officer at the other end, “Is Daniel Ellsberg cleared for T or K?” If the answer, based on a computer search in the control office he was calling, was no, he would come back and change the subject.

  If the answer was yes, he would come back and tell me that I had checked out and invite me to go to a phone to check his clearance out using the same process. For a uniformed colonel in Air Force Plans that I knew personally, that might not have seemed necessary. But in theory, he could have been bluffing, having heard the initials “T” and “K” or perhaps even having found out their nature, tricking me into a discussion to which he was not entitled.

  That possibility was the basic need for this rigmarole, and why only the first letters of the code words “Talent” (for U-2 photography) and “Keyhole” (for the reconnaissance satellite program and photos) were to be mentioned in a public place, where they might be overheard. Elaborate as it sounds, this two-phone-call routine was something I practiced many times in later years before talking with someone whose access was not known to me. Procedures like this—and the sanction of being summarily cut off from access, involvement, and advancement by violating them—kept a vast amount of information relevant to government decision-making (“higher than Top Secret,” SCI) secret from the public, Congress, and most of the government, along with foreigners and enemies, for long periods of time; they were proof against leaks for decades and generations, even when information was known to hundreds or thousands of individuals cleared for it.

  The cliché that “everything leaks; it all comes out in the New York Times eventually” is emphatically not true, above all for sensitive compartmented information. It’s a cover story, designed both to hide and sustain the effectiveness of the overall secrecy system. (Edward Snowden was the first ever to expose a large amount of SCI, including massively unconstitutional and criminal dragnet surveillance of American citizens and others in the world without probable cause for suspicion. Many thousands of NSA employees had known for a decade of that mass surveillance and its criminality. Not one other had disclosed it. Snowden is currently in exile, probably for life.)

  Ironically, the second breach was by an unlikely person, a normally very tight-lipped colleague who had long been known at RAND to have “intelligence clearances,” whatever that meant. After Cragg’s slip, I asked my friend, who was in D.C. consulting, the meaning of “T” and “K,” and he actually told me.

  In retrospect, it’s amazing, even perplexing, that he did so, which was not only against rules that were almost never violated but was highly out of character for him. Moreover, he said that I should make an effort to get those clearances, along with SI clearance (for special intelligence, a cover term for signals intelligence, comprising communications intercepts and other electronic signals). The three together gave one what was called “all-source access,” the output of communications and reconnaissance intelligence.

  Those who had (only) SI, T, and K in addition to Top Secret clearance were told, and almost all believed, that with their “all-source access” they had all the existing clearances. That was another cover story. There were in fact many clearances higher than these.

  The existence of special access programs (SAPs) known as “operational” clearances about special programs—including, say, the actual operations and decision-making process concerning the U-2 or its successors or the family of reconnaissance satellites or covert operations—was unknown to those who had “only” all-source intelligence. I got a dozen of these clearances when I was special assistant to the assistant secretary in 1964–65. For example, Ideal (I) was clearance for information about the operations of the U-2 program and the decision-making in connection with its uses and priorities. The existence of this clearance, and what it covered, would be unknown to the much larger number of people who had only Talent clearance to view the U-2 photography.

  The final critical clue I got about how much to trust the new estimate of Soviet missile capabilities was that Harry Rowen (who now, like Alain Enthoven, had all-source clearances plus many more in his position in the Pentagon, as I later learned) described to me a conversation he’d had in Carl Kaysen’s White House office with Carl, Alain, and some CIA officials. They had been passing around, Harry said, actual photographs from the Corona satellite of the four Soviet ICBMs and the nonexistence of such missiles at other suspected sites. One of them had said, Harry told me, laughing, “These pictures are worth a billion dollars.” And someone else had answered, “That’s about what they cost.”

  Harry’s telling me that there were now photographs of all the suspected sites, and the one real one, was the third breach. It was the big secret that I wasn’t cleared to know. Along with Cragg’s question to me about T and K and my friends’s explanation of the clearances to me, that clicked. The version of the new NIE I’d seen was only Top Secret. It didn’t tell, or even hint, what the new evidence was that led to the new assurance of the astonishing pronouncement on the Soviet ICBM force, or the lack of one. The new NIE would not be available to my colleagues back in Santa Monica. But even if they’d read it, as I now had, they wouldn’t have known enough of the evidence on which it was based to know whether to believe it. Now I did.

  I’ve gone into all this to emphasize that the credibility of this new estimate—fantastic, inherently incredible to anyone who had been relying on Air Force estimates or even CIA estimates (anything but Army and Navy estimates)—depended on knowledge of a kind of information that most people in the national security field, inside and outside the government, had no inkling existed. From the internal leaks—“unauthorized disclosures”—to me within the bureaucracy, I did believe it, even though it totally contradicted the fundamental basis for my concerns and work for the past several years.

  It wasn’t just a matter of numbers, though that alone invalidated virtually all the classified analyses and studies I’d read and participated in for years. Since it seemed clear that the Soviets could have produced and deployed many, many more missiles in the three years since their first ICBM test, it put in question—it virtually demolished—the fundamental premise that the Soviets were pursuing a program of world conquest like Hitler’s.

  As the Air Force chief of intelligence had put it in his dissent to the low figures in the June estimate, that pursuit of world domination would have given them an enormous incentive to acquire at the earliest possible moment the capability to disarm their chief obstacle to this aim, the United States and its SAC. His assumption of Soviet aims was shared, as f
ar as I knew, by all my RAND colleagues and with everyone I’d encountered in the Pentagon:

  The Assistant Chief of Staff, Intelligence, USAF, believes that Soviet determination85 to achieve world domination has fostered recognition of the fact that the ultimate elimination of the US, as the chief obstacle to the achievement of their objective, cannot be accomplished without a clear preponderance of military capability.

  If that was their intention, they really would have had to seek this capability before 1963. The 1959–62 period was their only opportunity to have such a disarming capability with missiles, either for blackmail purposes or an actual attack. After that, we were programmed to have increasing numbers of Atlas and Minuteman missiles in hard silos and Polaris sub-launched missiles. Even moderate confidence of disarming us so thoroughly as to escape catastrophic damage from our response would elude them indefinitely.

  Four missiles in 1960–61 was strategically equivalent to zero, in terms of such an aim. They could have hit Washington and SAC headquarters, but that would neither have disarmed nor paralyzed SAC’s ability to annihilate them in response. The Soviets could hit a city or two, striking first. Suicidally. They had no second-strike missile capability at all against the continental United States.