Marshall did not choose the quantitative Game Theory approach as a clever tactic to suit his audience. Marshall, in this sense, was as much a part of this audience as anyone. The approach represented the way that he truly did think through problems. Marshall and Goldhamer only briefly acknowledged once that it “might be worthwhile investigating whether there are idiosyncratic Soviet dispositions that can be expected to increase the deterrent effect of various measures (or to suggest quite new ones).” In the main, they rested the study on the premise that in “trying to estimate possible U.S. courses of action, the Soviets would presumably attempt to construct their own version of the U.S. utility matrix”—in effect, on the premise that the Kremlin thinks about the bomb in much the same way as the RAND Corporation.
One upshot of the Marshall-Goldhamer study was that, because of its methodology and seemingly scientific approach, a broad segment of the RAND strategic community began to entertain counterforce and withholding strategies more seriously than at any time since Jim Digby’s aborted study four years earlier. Still, there were major doubts on the part of many that the U.S. lacked the intelligence data to find and destroy Soviet military targets. The only reassurance on this point that Marshall and Goldhamer offered in their study was the single sentence: “We believe that the United States has relatively good knowledge of the location of operational bases in the Soviet Union.” Ever discreet, Marshall made no reference to all the types of data that Joe Loftus had told him about—the World War II information, the aerial photographs from balloons and the U-2 aircraft, the COMINT intercepts, the spies and, in the near future, reconnaissance satellites.
Marshall and Loftus were frustrated by the skepticism, but did little to quash it. The information was just too sensitive to spread around indiscriminately. Their frustration heightened, however, in 1959–1960, during the biggest group project in RAND history, the Strategic Offensive Forces, or SOF, study. General Tommy White, Air Force Chief of Staff, asked RAND to undertake the project. He had read all the papers and monographs on the nature of deterrence, but he wanted analysts to start thinking about what we should do if deterrence failed, what were the objectives of war under those circumstances, and how do those objectives fit in with the weapons that the Air Force was planning to build in the 1960s. (White asked RAND to do this before the Marshall-Goldhamer study was issued.)
Tommy White was a peculiar Chief of Staff, more intellectual than most of his predecessors. The son of a Presbyterian minister, fluent in seven languages, White worked his way through military ranks as an attaché in Russia and China and then spent ten years inside the Pentagon—not as a commander in the field—before becoming Chief. He was genuinely interested in ideas, and had come to look toward RAND as a source of many.
The SOF study was supervised by Ed Barlow, who was Director of RAND Projects. Nearly everyone worked on some aspect of the study—the cost-effectiveness of bombers versus missiles, examinations of ballistic-missile defense and air defense, studies of Air Force research and development efforts, the vulnerability of SAC, different approaches to targeting, every topic one could possibly imagine that might concern the modern Air Force.
The man picked to analyze counterforce strategies was Ed Oliver, an engineer and technical adviser to the Gaither Committee of 1957. Oliver produced an anti-counterforce report arguing that however desirable such a strategy might be, it was unfeasible, given intelligence technology. Marshall and Loftus approached Oliver shortly after he completed his first draft and questioned his assumptions, his calculations and especially the degree of access he had to data on the most modern reconnaissance technology. Oliver delved into the matter again, and came up two weeks later with the confession that a fuller study should be conducted.
Marshall and Loftus wondered whether Ed Oliver was up to the task. Both felt that somebody had to do an analysis so compelling that skepticism at RAND, and perhaps in the Air Staff, would be overcome. Marshall and Loftus started to engage in intensive conversations with Bill Kaufmann in the social science division, a man they both liked from the time he came to RAND in 1956 and who seemed intuitively sympathetic to the counterforce idea.
Kaufmann had spent his first few years at RAND mainly as a student of sorts. He had done a few studies on his own, but they were “in the essay tradition.” Providing a political scientist’s perspective, he had served as one of several assistants on a few major projects. But he was still on the sidelines of RAND’s strategic community, still learning about weapons and technology and systems analysis, still absorbing this new methodology and new focus on national security. In this course of re-education, he was still trying to find the answer to the problem—how to convince the allies of America’s commitment to their defense without either committing suicide or spreading nuclear weapons in the process—that he had faced in that “diplomatic exercise” game on NATO during his early RAND days.
Kaufmann and Loftus got along quite chummily. When they talked about strategy, counterforce frequently came up. Kaufmann had heard about counterforce, but like most of the others at RAND, he had casually dismissed the matter from his mind, believing that the concept was impractical. Loftus set out to dispel this skepticism by telling Kaufmann of the highly secret techniques by which the intelligence agencies could know the location of the important counterforce targets. Andy Marshall instructed Kaufmann about the work that Jim Digby had partially completed in his counterforce study of 1955, and convinced him that going after the Soviet military targets was operationally feasible.
The more he thought about it, the more Kaufmann began to realize that counterforce went a long distance toward solving the NATO problem, that counterforce in fact fit quite nicely into the scheme of limited war that he himself had outlined in his Princeton essays of 1954–1956. Massive retaliation was an inadequate policy once the Soviets had amassed their own nuclear arsenal, because destroying their cities would provoke the Soviets to destroy our cities. The RAND game on NATO concluded that we should give the allies their own nuclear weapons because the Europeans could not reasonably trust the United States to blow up Russian cities in retaliation to a Soviet attack on Europe, thus sacrificing New York for Paris, Chicago for Bonn.
But what if we didn’t blow up Soviet cities at all? What if we executed a counterforce attack in retaliation to a Soviet invasion of Western Europe? Not only would that strike at the very heart of the Soviet military power, but it might also serve as a more credible deterrent to such an invasion in the first place. If we did not strike at Soviet cities, perhaps the Soviets would strike back at something other than American cities. And if, therefore, American cities were not placed so much at risk in the act of defending Western Europe, then America’s commitment to the NATO alliance could once more be made credible, which was to Kaufmann the objective of preeminent importance.
In the Princeton essays, Kaufmann had argued the case for restraint of force in future limited ground wars with the Soviet Union. The continuing competition with Communism was going to have its violent moments in the coming decades, and it was important to meet the challenges without provoking nuclear holocaust. Limited wars should be “managed,” Kaufmann had written in 1955, in such a way as to send “messages” to “the other belligerent,” which would “have a good chance of inducing him to accept limitations of geography, weapons and possibly time. The scope and method of the initial attack will tend to define the minimum limits of the ensuing conflict and the possibilities of controlling it.”