by Sylvia Nasar
From the beginning, RAND’s work was a curious mix of narrowly focused engineering, cost-benefit studies, and blue-sky conjecture. A now-famous 1946 study, completed more than a decade before the launch of Sputnik in 1957, proved remarkably prescient. In “Preliminary Design of an Experimental World-Circling Spaceship,” RAND scientists argued that “the nation which first makes significant achievements in space travel will be acknowledged as the world leader in both military and scientific techniques. To visualize the impact on the world, one can imagine the consternation and admiration that would be felt here if the US were to discover suddenly that some other nation had already put up a successful satellite.”11
RAND’s civilian scientists soon made a mark on American defense policy. Poundstone reports that RAND played a leading role in the development of the ICBM; RAND convinced the Air Force to adopt in-flight refueling of jet bombers; it was responsible for the fail-safe protocol whereby bombers are kept in the air at all times and during a crisis head for targets in an enemy nation. Its worry that a psychotic individual in a position of power could trigger a nuclear war convinced the Air Force to adopt a safer button that required cooperation of several individuals to arm and detonate a nuclear warhead.
To be plucked from academe and initiated into the secret world of the military had become something of a rite of passage for the mathematical elite. In World War II, the very best had traveled into the New Mexico desert to Los Alamos to work on the A-bomb alongside von Neumann, and to Bletchley Park north of London to help Turing and his team break the Nazi code.12 Many others, less well known or simply younger, wound up at dozens of less famous sites working on weapon design, encryption, bomb targeting, and submarine chases.13
The recruitment of scientists by the military hadn’t stopped when the war ended, much to everyone’s surprise. Many of the mathematicians and scientists did not return to their quiet prewar routines but instead took on military research contracts, made frequent visits to the Pentagon and the Atomic Energy Commission, and, in a few cases, stayed on at Los Alamos and the other government weapons labs. For an elite cadre of applied mathematicians, computer engineers, political scientists, and economists RAND was the equivalent of Los Alamos.14
The problems the military asked the scientists to solve called for new theories and new techniques, which in turn attracted the top scientific talent on which RAND’s credibility depended. “We had so many practical problems that involved mathematicians and we didn’t have the right tools,” said Bruno Augenstein, a former RAND vice-president, years later. “So we had to invent or perfect the tools.”15 Mostly, according to Duncan Luce, a psychologist who was a consultant at RAND, “RAND capitalized on ideas that surfaced during the war.”16 These were scientific, or at least systematic, approaches to problems that had been previously considered the exclusive province of men of “experience.” They included such topics as logistics, submarine research, and air defense. Operations research, linear programming, dynamic programming, and systems analysis were all techniques that RAND brought to bear on the problem of “thinking the unthinkable.” Of all the new tools, game theory was far and away the most sophisticated.
The spirit of quantification, however, was contagious, and it was at RAND, more than anywhere else, that game theory in particular and mathematical modeling in general entered the mainstream of postwar thinking in economics. At that point, the military was the only government sponsor of pure research in the social sciences — a role later taken over by the National Science Foundation — and it bankrolled a great many ideas that turned out to have little true relevance for the military but a great deal for other endeavors. RAND attracted a younger generation of mathematically sophisticated economists who embraced the new methods and tools, including the computer, and attempted to turn economics from a branch of political philosophy into a precise, predictive science.
Take Kenneth Arrow, one of the early Nobel Laureates in economics. When Arrow came to RAND in 1948, he was an unknown youngster.17 His famous thesis, written in the as-yet-unfamiliar language of symbolic logic, was a product of a RAND assignment. The assignment was to demonstrate that it was okay to apply game theory, which is formulated in terms of individuals, to aggregations of many individuals, namely nations. Arrow was asked to write a memorandum showing how it could be done. As it turned out, the memorandum became Arrow’s dissertation, an attempt to restate the theories of British economist John Hicks in modern mathematical language. “That was it! It took about five days to write in September 1948,” he recalled. “When every attempt failed I thought of the impossibility theorem.”18 Arrow showed that it is logically impossible to add up the choices of individuals into an unambiguous social choice not just under a constitution based on the principle of majority rule, but under every conceivable constitution except dictatorship. Arrow’s theorem, along with his proof of the existence of a competitive equilibrium, which also owes something to Nash, earned him the Nobel Prize in 1972 and ushered in the use of sophisticated mathematics in economic theory.
Other giants of modern economics who did seminal work at RAND in the early 1950s included Paul A. Samuelson, probably the most influential economist of the twentieth century, and Herbert Simon, who pioneered the study of decisionmaking inside organizations.
RAND’s location was part of its allure. The corporation’s headquarters, in a once-sleepy beach colony, lies five miles to the south of the Santa Monica Mountains at the far end of the Malibu Crescent, just west of Los Angeles. In the early 1950s, Santa Monica looked the way Nash imagined that certain towns in Italy or France might look. Wide avenues were lined with pencil-thin palm trees. Cream-colored houses were topped with tiled roofs and encircled by shoulder-high walls. Seaside hotels and rest homes were across from a seaside promenade. The magentas and reds of the bougainvillea and hibiscus were improbably intense. The breeze, surprisingly cool, smelled of oleander and seawater. Some of the best work was done in beach chairs.
RAND itself was tucked out of sight of the ocean on Fourth and Broadway at the edge of Santa Monica’s slightly rundown business district. The 1920s bank building was a white stucco affair ornamented with Victorian flourishes. The building had recently housed the presses of the Santa Monica Evening Outlook; the newspaper had moved catty-corner to a former Chevy dealership when RAND moved in. By 1950, RAND was already spilling over into several annexes located over storefronts, including ones occupied by the Outlook and a bicycle shop. A year later, when Fortune magazine discreetly introduced RAND to the wider public, it described “bright walls shining through fog-sunny days and its wide, white-lighted windows shining on uninterruptedly through the night. The building is never closed, nor is it ever really open.”19
It was one of the most difficult buildings in the United States to get into, Fortune said. On Nash’s first day, members of RAND’s uniformed, armed police force stood guard in front of the building and in its lobby, scrutinizing him closely and memorizing his face.20 After that, for the rest of the summer and in subsequent years, the guards always greeted him with a cool, respectful “Hello, Dr. Nash.” There were no ID cards in those days. Inside were a series of locked doors, with offices clustered by types of security clearance needed to gain access to them. The math division occupied a group of small private offices in the middle of the first floor, upstairs from the electronics shop where von Neumann’s new computer, the Johnniac, stood.21 Nash got an office to himself, a small windowless cubicle whose walls didn’t quite extend to the ceiling, with a desk, blackboard, fan, and, of course, a safe.
RAND bristled with self-confidence, a sense of mission, an esprit de corps.22 Military uniforms signaled visitors from Washington. Executives from defense firms came for meetings. The consultants, mostly under thirty, carried briefcases, smoked pipes, and walked around looking self-important. Big shots like von Neumann and Herman Kahn had shouting matches in the hallways.23 There was a feeling around the place of “wanting to outrun the enemy,” as a former RAND vice-presiden
t later put it.24 Arrow, who was an army veteran from the Bronx, said, “We were all convinced that the mission was important though there was lots of room for intellectual vision.”25
RAND’s sense of mission was propelled largely by a single fact: Russia had the A-bomb. That shocking news had been delivered by President Truman the previous fall, a mere four years after Nagasaki and Hiroshima, and many years before Washington had expected it. The military had hard evidence, the president said in a speech on September 13, 1949, of a nuclear explosion deep inside the Soviet Union.26 Nobody in the scientific community, especially around Princeton, where von Neumann and Oppenheimer were engaged in an almost daily debate over the wisdom of pushing ahead with the Super, doubted that the Soviets were capable of developing nuclear weapons.27 The shock was that they had succeeded so quickly. Physicists and mathematicians, who were less convinced of Russia’s scientific and technological backwardness, had been warning the administration all along that predictions by senior government officials that America’s nuclear monopoly would persist another ten, fifteen, or twenty years were hopelessly naive, but the sense of being caught off guard was still very great.28 The news effectively ended the debate over the hydrogen bomb more or less immediately. By the time the president delivered the news of the Soviet explosion to the public, he had authorized a crash program at Los Alamos to design and manufacture an H-bomb.29
It was unthinkable that such destructive power would be unleashed. Therefore RAND insisted that it was necessary to ponder the possibility.30 The rational life was worshiped to an almost absurd degree. RAND was full of men and women committed to the idea that systematic thought and quantification were the key to the most complex problems. Facts, preferably detached from emotion, convention, and preconception, reigned supreme. If reducing complex political and military choices, including the problem of nuclear war, to mathematical formulae could produce light, why then the same approach must be good for more mundane matters. RAND scientists tried to tell their wives that the decision whether to buy or not to buy a washing machine was an “optimization problem.”31
RAND was privy to the military’s most highly guarded secrets at a time when the nation was growing increasingly nervous about the safeguarding of those secrets to the point of paranoia. From the summer of 1950 on, RAND would be increasingly affected by the growing alarm over Russian access to American military secrets.32 It began with the Fuchs trial in the winter of 1950.33 Fuchs was a German émigré scientist who had fled to Britain during the war and eventually wound up working with von Neumann and Edward Teller at Los Alamos. A clandestine member of the British Communist Party, Fuchs subsequently confessed in January 1950 to passing atomic secrets to the Russians and was tried and convicted in London that February. Senator Joseph McCarthy had embarked that same month on his anticommunist campaign, accusing the federal government of security breaches.34 Four years later, in April of 1954, Robert Oppenheimer, the former head of the Manhattan Project, the director of the Institute for Advanced Study, and the most famous scientist in America, was declared a security risk by Eisenhower and stripped of his security clearances in the full glare of national publicity.35 The ostensible reason was Oppenheimer’s youthful left-wing associations, but the real reason, as von Neumann and most scientists testified at the time, was Oppenheimer’s refusal to support the development of the H-bomb.
The fact that McCarthy himself ultimately became a target of censure would do little to dispel the atmosphere of paranoia and intimidation at RAND, which lived on Air Force and AEC money and had projects on the H-bomb and ICBMs.36 Most of what the mathematicians worked on was not in fact classified, but that didn’t matter. RAND, which harbored a collection of oddballs like Richard Bellman (a former Princeton mathematician who had all kinds of communist associations, mostly accidental, including a chance encounter with a cousin of Julius and Ethel Rosenberg), would become particularly careful about minding its Ps and Qs.37
Everybody needed a top-secret clearance. People who arrived without a temporary security clearance were banished to “quarantine” or “preclearance” and weren’t permitted to sit with everybody else. Nash’s secret clearance was granted on October 25, 1950.38 His recollection that he had a top-secret clearance — which a large contingent-in the math division did have — is probably faulty. Nash also recalls that he applied for a Q clearance in 1952.39 Any consultant to the math division who worked on Atomic Energy Commission contracts was required to have a Q clearance because of access to documents related to the construction and use of nuclear weapons. But despite a November 10, 1952, postcard to his parents telling them that he had applied for a higher clearance at RAND, Nash now says it was never approved — meaning that his role at RAND was largely confined to highly theoretical excercises as opposed to applications of game theory concepts to actual questions of nuclear strategy — the province of men like von Neumann, Herman Kahn, and Thomas Schelling.40
Everyone had a safe in his office for storing classified documents, and everyone was warned about taking documents out of the building or talking out of school.41 Papers had to be put in the safes at the end of every day. There were spot checks. There was a public address system and there were parts of the building that were off-limits to people who didn’t have a Q clearance.
By 1953, soon after Eisenhower issued a new set of security guidelines, security consciousness, in the sense of not overlooking anyone who might be thought remotely unreliable, grew.42 The Eisenhower guidelines broadened the grounds for denying a clearance or stripping someone of an existing clearance. Without a doubt, fear about potential leaks brought to a boil many simmering antagonisms against individuals and groups who posed little or no actual threat to security. Almost any sign of nonconformity, political or personal, came to be considered a potential security breach. The notion, for example, that homosexuals were unreliable, because of either poor judgment or vulnerability to blackmail, was first codified in the Eisenhower guidelines.
Like the decade itself, RAND had a split personality. Its style was informal. It tolerated quirky people. It was in some ways more democratic than a university. Almost everyone, including von Neumann, was called by his or her first name, except by the guards, never Doctor or Professor or Sir. Graduate students rubbed shoulders with full professors in a way unimaginable in most academic departments. RAND’s president, a former Douglas Aircraft executive, was a spit-and-polish man who was almost never seen in a suit and tie. All but one or two of the mathematicians, including Nash, came to work in short-sleeved shirts. Appearances were so casual that one mathematician, who found it all very déclassé, felt obliged to rebel by wearing a three-piece suit and a tie to the office every day.43
Practical jokes were as much a part of the RAND culture as pipes and crewcuts. Mathematicians and physicists mixed rubber bands into the pipe tobacco, substituted dog biscuits for cookies, and tilted desks so pencils rolled onto the floor.44 Wit was greatly appreciated. When John Williams, the head of RAND’s mathematics department, wrote a primer on game theory, published as a RAND study, it was illustrated with funny little cartoon figures and full of jokey examples starring John Nash, Alex Mood, Lloyd Shapley, John Milnor, and other members of the math department.45
The mathematicians were, as usual, the freest spirits.46 They had no set hours. If they wanted to come into their offices at 3:00 A.M., fine. Shapley, who had come back from Princeton for the summer and continued to insist on the sanctity of his sleep cycle, was rarely seen before midafternoon. Another man, an electrical engineer named Hastings, typically slept in the “shop” next to his beloved computer. Lunches were long, much to the annoyance of RAND’s engineers, who prided themselves on sticking to a more respectable routine. The mathematicians mostly took their bag lunches to a conference room and pulled out chessboards. They invariably played Kriegspiel, usually in total silence, occasionally punctuated by a wrathful outburst from Shapley, who frequently lost his temper over an umpire’s or opponent’s error. Even though the gam
es typically lasted well into the afternoon, they were rarely finished and finally reluctantly abandoned midgame. Poker and bridge groups met after hours.
There were no afternoon teas, formal seminars, or faculty meetings at RAND. Unlike the physicists and engineers, the mathematicians usually worked alone. The idea was that they would work on their own ideas but would help solve the myriad problems encountered by researchers, picking up problems to solve as the spirit moved them.47 People would drift into each other’s offices or, more frequently, simply stop to chat in the corridors near the coffee stations. The grids and courtyards of RAND’s permanent headquarters — to which the mathematics group moved in 1953, the year before Nash’s final summer at RAND — were designed, by John Williams, as it happens, “to maximize chance meetings.”48 Through such encounters new research was “announced” and mathematicians got hooked on problems that colleagues in other departments wanted solved. Most of the work wasn’t reported formally, and even when it was published as RAND memoranda, there was no formal approval process. A consultant would simply go to the math department secretaries, hand over a handwritten paper, and a day or two later a RAND memorandum would appear.49 Published reports for outside circulation didn’t go through a much more rigorous vetting process.
This copacetic atmosphere was mostly Williams’s doing.50 Witty and charming, weighing close to three hundred pounds, expensively suited, Williams looked like a businessman always about to reach into his pocket to pull out a wad of twenties. An astronomer from Arizona who had spent a couple of years in Princeton attending lectures in Fine Hall, playing poker, and developing an enthusiasm for the theory of games, Williams had been a dollar-a-year man in Washington during the war and became RAND’s fifth employee afterward. Williams hated flying. He loved fast cars. At one point, he spent an entire year outfitting his chocolate-brown Jaguar with a powerful Cadillac engine. It had taken substantial RAND resources (RAND had a repair shop) and considerable bravado to install the thing. Cadillac and Jaguar mechanics had both dismissed the idea as impractical, but Williams had prevailed. He disproved the mechanics’ conventional wisdom in late-night, 125–mile-an-hour drives along the Pacific Coast Highway.