The Last Warrior

by Andrew F. Krepinevich

  Project RAND’s mission was to recommend “preferred instrumentalities and techniques” to the US Army Air Forces (after September 1947, the US Air Force).14 While the organization was to be highly independent, it also reported to the Air Staff—initially to the deputy chief of staff (DCS) for plans, General Lauris Norstad, and subsequently to the deputy chief of staff/research and development (DCS/R&D), Major General Curtis LeMay. In 1943, as commander of the Eighth Air Force’s 305th Bomb Group in Europe, LeMay had emerged as the US Army Air Forces’ preeminent combat leader in the European Theater. In 1945, as the commander of the Twenty-First Bomber Command in the Pacific, he oversaw the fire raids against Japanese cities and the atomic bombings of Hiroshima and Nagasaki.

  Project RAND grew rapidly. In March 1946 the Army Air Forces executed a $10 million contract that put RAND under Frank Collbohm’s direction in a separate area within the Douglas Aircraft plant at Santa Monica’s municipal airport, Clover Field.15 Collbolm, an aviation engineer, had helped design and test fly such trailblazing aircraft as the Douglas DC-3. He would head RAND until his retirement in 1967.

  In the early years both RAND’s divisions and personnel were heavily skewed toward engineering and the physical sciences. The first people hired into Project RAND were primarily engineers and mathematicians recruited from the aircraft industry. By late 1947 RAND’s hiring had expanded to include electronic engineers and physicists from the wartime Radiation Laboratory at MIT and such nuclear scientists as Arnold Kramish, who had been involved in the Manhattan Project that produced the atomic bomb.

  In May 1946 RAND published its first report, “Preliminary Design of an Experimental World-Circling Spaceship.” Requested by LeMay, the report discussed the design, performance, and potential uses of man-made, Earth-circling satellites. This report was followed in 1947 by a comparison of ramjets and rockets, including ballistic missiles, as offensive weapons.16 These studies presented remarkably prescient ideas at least a decade ahead of their time. But although RAND analysts were offering US Air Force leaders a view of where military technology was headed, the service, dominated by World War II bomber generals, initially showed little inclination to pursue either satellites or ballistic missiles.

  By 1947 it was becoming clear that RAND needed more than engineers and hard scientists. That fall, two years before Marshall joined the organization, RAND held a conference in New York to which a number of prominent economists and other social scientists were invited. The event was a first step toward enlisting experts from these disciplines to join the organization. The motivation was a growing realization within RAND’s top management that physical scientists were trained to shun making value judgments about the “preferred instrumentalities and techniques” for deterring or waging nuclear war.17 The conference, it was hoped, would provide an opportunity to elicit from social scientists whether and how they might be able to help RAND address the problems of judging the nontechnical aspects of a new capability’s military worth.18

  The conference succeeded. Frank Collbolm was sufficiently impressed that he decided to organize economics and social science divisions at RAND. In 1948 he brought in Charles Hitch to head the Economics Division and Hans Speier to lead the Social Science Division. Both men had been among the conference’s most impressive participants.19

  By early 1948 RAND had over two hundred staff members and was rapidly outgrowing its Douglas Aircraft facilities at Clover Field. In May Project RAND was incorporated in the state of California as a nonprofit research institution for the US Air Force, which itself was less than a year old. RAND relocated its headquarters (or “main campus,” as staffers called it) to facilities at Fourth Street and Broadway in downtown Santa Monica. The organization’s future was secured by an interest-free $1 million loan guaranteed by the Ford Foundation. By then RAND’s professional staff were mostly physicists, mathematicians, statisticians, aerodynamicists, and chemists but now included a sprinkling of economists and other social scientists.20 Young Andrew Marshall would soon be among them.

  The RAND Corporation may have been young, but by the time Marshall joined the organization’s Washington Office in January 1949 it was already entering its golden age as an incubator of US nuclear strategy. The think tank’s principal focus in the 1950s was, initially, to explore the profound problem of how the United States should deal with the threat of atomic weapons delivered by long-range bombers and, subsequently, the threat of thermonuclear weapons delivered by ballistic missiles. The destructive power of thermonuclear-tipped ballistic missiles was not only without precedent, it threatened the United States’ very existence. In the nuclear missile age, the oceans that had rendered the country beyond the immediate reach of hostile powers in Europe and Asia no longer guaranteed safety.

  There were no experts in nuclear strategy when Marshall joined RAND. Nobel Prize winners were not necessarily any better than graduate students when it came to determining how nuclear weapons would change the conduct of war. Consequently RAND meetings and working group sessions accepted no hierarchy.21 Individuals were only as good as the ideas they brought to the table.

  It was this combination of talent and intellectual freedom that enabled RAND to shape how senior American political and military leaders viewed strategy in the nuclear age. As political scientist and military strategist Bernard Brodie later recalled from his days at RAND that virtually “all the basic ideas and philosophies about nuclear weapons and their use” were generated by civilians at such institutions as RAND, working independently of the military.22 Worth adding is that while RAND strategists had a remarkably extensive impact on American nuclear strategy, “the group of real strategists at RAND probably never numbered more than about 25 people.”23

  Upon his arrival at RAND’s Washington office in January 1949 to work in the Social Science Division, the twenty-seven-year-old Andrew Marshall knew very little about the organization he had joined. In fact, his main reason for choosing RAND over the other job offer he had considered in Chicago had been pecuniary.24 He thought the higher salary he got from RAND would shorten the time it would take him to save enough money to pursue a PhD in statistics.

  There was little in Marshall’s background or his arrival at RAND to suggest that he would become a leading member of RAND’s foremost strategic thinkers. Goldhamer, after all, had hired him simply to do statistical analysis of data on mental illness among the draft-age male cohort. Yet in the course of little more than a decade, that is precisely what happened.

  Marshall’s development as a strategist was unquestionably influenced by RAND colleagues and mentors such as Goldhamer and Hitch. But how much of his thinking was shaped by these prominent individuals is impossible to say. His early interests in fields as diverse as mathematics, military history, human evolution, and literature; his deep innate intellectual curiosity; and his strong preference for empirical data over abstract models and theories all predated his arrival at RAND. What is indisputable, however, is that once Marshall joined RAND, he was quickly captivated by the real-world problems of the nuclear age. Given his involvement in these issues, it should come as no surprise, then, that he would later recall the RAND of the 1950s as a “fantastic place” with “wonderful people.”25

  The strategic implications of the rapidly evolving long-term competition between the United States and the Soviet Union would occupy Marshall’s professional life for the next four decades. Yet, upon arriving at RAND he was set to work on something far more mundane: helping Goldhamer to determine whether the incidence of mental illness or psychosis in the US population had increased during the preceding century. In both world wars, the US military had rejected substantial numbers of individuals for mental or psychological reasons. Was the incidence of mental illness in the US population growing, as many presumed? RAND’s formal interest in this question stemmed from the scale of US military manpower mobilization that had been required to defeat Nazi Germany and Imperial Japan in the Second World War. In the four years after the Japanese at
tack on Pearl Harbor, over 16 million Americans would serve in the US military out of a population of roughly 137 million.26 In 1945 the number of active duty US military personnel had peaked at just over 12 million; but by 1947 less than 1.6 million remained in uniform.27 Many senior defense officials believed that another major conflict could not be ruled out and that if it did occur, a European war against Stalin’s massive Soviet Army might require the United States to mobilize manpower on a similar scale. If so, then a growing incidence of mental illness could significantly constrain the country’s military potential.

  The prevailing wisdom at the time held that the rate of mental illness in the US population had been growing since the 1840s due to the increasing urbanization of the nation’s population combined with the “killing” pace, insecurity, and intense individualism associated with life in American cities. Goldhamer, with Marshall’s help, aimed to discover whether data would confirm this view, and, if this hypothesis was confirmed, whether the trend had been accelerating in the first half of the twentieth century. If the conventional wisdom proved correct, the growing incidence of mental illness among draft-age men could significantly undermine American national defense in the event of a major-power conflict on the scale of World War II.

  The Goldhamer-Marshall collaboration produced two RAND papers in 1949. Marshall’s contribution was a statistical analysis of the available data. Both papers were subsequently published together under the title Psychosis and Civilization.28 The first paper overturned the conventional wisdom, conclusively demonstrating that there had been no increase in the frequency of mental illness in the United States over the previous one hundred years. This was the good news. But there was also some bad news. The second paper estimated that individuals who survived to age forty-five would have one chance in twenty of being stricken by a serious mental illness, and one chance in ten if they lived to age sixty-five. The greater incidence of psychosis was not driven by urban living but by the growing number (and percentage) of Americans living well into old age. Mental health in the United States remained no better—and no worse—in the 1940s than it had been during the previous century; there were simply more mentally ill people now than there had been before.

  Through the course of their initial collaboration, Marshall and Goldhamer became close friends. The young Marshall found Goldhamer, who had once been the junior chess champion of Canada, both bright and intellectually stimulating. When the two men spent most of September and October 1949 in Santa Monica monitoring the statistical analyses and data processing for their mental illness studies, Goldhamer introduced Marshall to a variation of chess known as Kriegspiel (literally German for “war play”), which had become a popular recreational activity among RAND staffers. Kriegspiel is chess with incomplete information: each player can see the positions of his own pieces on the board, but not those of his opponent. A referee with complete information adjudicates each move as “legal” (that is to say, within the rules) or not and provides information on checks or captures. In the late 1940s John Williams, who headed RAND’s mathematics division, was the center of a lively Kriegspiel activity at lunchtime and in the evenings. During his fall 1949 trip to Santa Monica, Marshall spent many evenings playing Kriegspiel at Williams’s Pacific Palisades house just north of Santa Monica.

  Not long after joining RAND Marshall was granted a Top Secret security clearance. Granting such clearances to RAND employees was commonplace at the time. LeMay had been emphatic that members of RAND’s staff were to be kept current on both the Air Force’s plans and its intelligence information—requirements that necessitated high security clearances. Since Project RAND’s aim was to provide the Air Staff with unbiased advice, LeMay also specified that no one on the Air Staff would tell RAND researchers what to do or—just as important—what not to do.29 Even more remarkable, the Air Staff was instructed to provide Project RAND with information on its other contract programs to avoid unnecessary duplication. These policies continued into the 1960s and made Project RAND an unusual enterprise from the outset. Finally, as Marshall could attest, RAND’s top managers were both willing and able to pay top dollar to attract the best talent. When it came to intellectual horsepower, they preferred quality over quantity, hiring some very bright and unusual minds to address specific problems. Decades later, after being appointed the Pentagon’s director of net assessment, Marshall would apply much the same approach to his outside research program.

  While in Washington Marshall took advantage of the region’s universities to continue his education in statistics. He began taking courses from Solomon Kullback at George Washington University. Kullback, an American cryptanalyst and mathematician, had been one of the first three employees hired by the US Army’s Signal Intelligence Service in the 1930s. Before and during World War II he had worked on deciphering both Japanese and German codes, including Japan’s RED cipher machine messages. When Marshall arrived in Washington, Kullback was teaching evening classes in advanced statistics. Marshall availed himself of the opportunity, earning “As” in Kullback’s classes.

  A year before Marshall joined RAND, an economics division had been established under Charles Hitch. Hitch had come to RAND with an impressive résumé. Nearly two decades earlier he had entered Oxford as a Rhodes scholar. Three years later, in 1935, when he sat for his oral examination for a master’s degree in economics, the Oxford faculty members on his committee, rather than questioning him, just tipped their academic caps to him in a show of respect and awarded him his degree. Shortly thereafter Hitch became the first American Rhodes scholar to be appointed an Oxford don. He also became general editor of the Oxford Economic Papers, which still publishes refereed papers in economic theory, applied economics, econometrics, economic development, economic history, and the history of economic thought.

  During World War II Hitch gained practical experience in economic matters. While on leave from Oxford he served on the first Lend-Lease mission in London; the War Production Board; and in the Office of Strategic Services, where he helped evaluate the effects of strategic bombing on Germany. He ended his wartime service as chief of the Stabilization Controls Division of the Office of War Mobilization and Reconversion. When he joined RAND in 1948 Hitch was a visiting professor at Yale, the University of California–Los Angeles (UCLA), and the University of São Paulo in Brazil. He traveled frequently to speak before audiences throughout continental Europe and England.

  In 1949 Russell Nichols, Hitch’s only representative in RAND’s Washington office, had become impressed by Marshall and approached him to point out the opportunities that the new department might hold for a young man of his talents. Then, in the spring of 1950, Hitch invited Marshall to spend the summer in Santa Monica with a team of RAND staff members that was exploring how the Air Force might best use atomic weapons to target the Soviet economy in the event of war. When Marshall arrived in California in late May 1949, he was still working with Goldhamer on the mental illness project, but he soon became immersed in the targeting project. In the end he concluded that destroying the Soviet economy with the Strategic Air Command’s then modest stockpile of a few hundred atomic bombs would be difficult given that the Soviet Union was an enormous country with widely dispersed industrial facilities.

  After the summer study Marshall returned to Washington and continued working with Goldhamer on the mental illness project. In the fall, however, Hitch, who would become one of Marshall’s most influential mentors, invited Marshall to relocate to California and join the Economics Department. Hitch’s brilliance, combined with his willingness to challenge the conventional wisdom, predilection for multidisciplinary research, and respect for empirical facts strongly resonated with the young Marshall’s own intellectual inclinations, and by November he had accepted Hitch’s offer and moved to California.

  Hitch and Marshall also became close outside their professional interactions. Although most of the division heads at RAND during the 1950s were not inclined to do a great deal of entertaining, Hitch
was a notable exception, regularly hosting parties and dinners at his home. After Marshall’s marriage, he and his wife Mary became regular guests at the Hitches’ social events. The close friendship between the two couples did not end with Hitch’s move to the Pentagon in 1961 to be Defense Secretary Robert McNamara’s comptroller. Even after Hitch had left the Pentagon in 1965 to head the University of California, the Hitches and the Marshalls often spent the Christmas holidays together in small cabins in the Sequoia National Forest.

  Another RAND colleague Marshall became very close to during the 1950s was a young physicist by the name of Herman Kahn. Like Marshall, Kahn had bounced around the academic world. After World War II, he had begun a PhD program at the California Institute of Technology (Caltech). He eventually dropped out, settling for a master of science degree. Kahn then briefly sold real estate until recruited into RAND’s Physics Division by a friend, Samuel Cohen.* Although he never acquired a doctorate, Kahn subsequently worked closely with three titans of science—nuclear physicists Edward Teller and Hans Bethe, and mathematician and polymath John von Neumann—on the hydrogen bomb. Later Kahn’s outspoken views on nuclear war would earn him a reputation as RAND’s l’enfant terrible, and he would gain fame as a strategist from his books on nuclear strategy, which included On Thermonuclear War and Thinking the Unthinkable.

  Kahn shared Marshall’s intense curiosity about how the world actually works. Both men had started self-educating themselves at an early age in fields other than those in which they later earned formal degrees. Given their wide range of interests, intelligence, and skepticism regarding what others accepted as “conventional wisdom,” the two men soon became close friends. During the early 1950s, they were inseparable whenever both of them were in Santa Monica. Marshall shared his thinking about economics and anthropology with Kahn, and Kahn reciprocated with insights into nuclear bomb design. By 1951 Kahn was regularly commuting to the Lawrence Livermore National Laboratory to work on the hydrogen bomb. Toward the end of the year, he began telling Marshall about the design breakthrough based on the concept of a two-stage device in which radiation confinement from the first-stage (or “primary”) fission “trigger” would compress the fuel in the second stage (or “secondary”) to produce a fusion reaction similar to what occurs in the sun. Turning the concept into a working thermonuclear device required extensive calculations of such things as the width of the channel through which radiation would flow from the primary to the secondary. During this period Kahn and Marshall were also collaborating on Monte Carlo statistical methods, which relied on random sampling of a large number of cases. To help Teller’s assistant Frederic de Hoffmann with such calculations, Kahn and Marshall ended up spending late nights at UCLA running Monte Carlo calculations on a new computer to estimate H-bomb radiation flows.30