The Pentagon's Brain (16 page)

Hickey and Donnell were flown to the Pentagon, where they were scheduled to brief Harold Brown and Walt Rostow, the president’s national security advisor, on the Strategic Hamlet Program. The Pentagon was a world away from Saigon and from Cu Chi, and yet the anthropologists knew firsthand what an impact the Defense Department’s work was having on the villagers living
there. They made their way through security, into the mezzanine, past the food shops and the gift shops and the employee banks. They walked up stairs, down corridors, and into Harold Brown’s office in the E-Ring, not far from the secretary of defense and the Joint Chiefs of Staff. Brown’s office was spacious and well decorated, with large leather chairs and couches, and a view of the Potomac River.

Hickey recalled paraphrasing from their written ARPA report. “In the present war,” he said, “the Vietnamese peasant is likely to support the side that has control of the area in which he lives, and he is more favorably disposed to the side which offers him the possibility of a better life.” Hickey and Donnell told Brown and Rostow that Diem’s army was simply not holding up its end of the bargain. As a result, and despite the well-intended efforts of the Strategic Hamlet Program, local Vietnamese peasants were more likely to side with the Vietcong.

Then something strange happened. “As we began our first debriefing at the Pentagon with Harold Brown,” Hickey noted, “[he] swung his heavy chair around and looked out the window, leaving us to talk to the back of his chair.” Hickey and Donnell kept talking. Perhaps Brown was simply contemplating the severity of the situation.

“Farmers were unwilling to express enthusiasm for the program and appeared to harbor strong doubts that the sacrifices of labor and materials imposed on them could yield any commensurate satisfaction,” the anthropologists explained. If something wasn’t done, the entire Strategic Hamlet Program was at risk of collapse. Hickey and Donnell suggested that the Pentagon put pressure on Diem’s forces to pay the farmers a small amount of compensation, immediately.

Harold Brown did not respond. Throughout most of the meeting, he kept his back turned on the two men, and though now they had finished their briefing, Brown still didn’t turn around to face
them. National security advisor Walt Rostow, who had been paying attention, looked away. An aide walked into the room, and Hickey and Donnell were shown the door.

Escorted out of Harold Brown’s office, the two men were led down the corridor to where they were scheduled to brief Marine Corps lieutenant general Victor “Brute” Krulak, now serving as special assistant for counterinsurgency and special activities. Krulak was a hard-charging militarist. During World War II he had masterminded the invasion of Okinawa, the largest amphibious assault in the Pacific theater and the last battle of the war. In the Korean War, Krulak had pioneered the use of helicopters in battle. Krulak was not happy with what Hickey and Donnell had to say, and he was demonstrative in his disapproval. He told them that he wasn’t going to pay a bunch of Vietnamese peasants for their support. “He pounded his fist on the desk [and said] that ‘we’ were going to make the peasant do what’s necessary for the strategic hamlets to succeed,” Hickey recalled.

The anthropologists from RAND were shown the door. Their thirty-page report, originally prepared for ARPA as an unclassified report, was now given a classification of secret, which meant it could not be read by anyone without an appropriate government clearance. Harold Brown told RAND president Frank Collbohm about his dissatisfaction with what he saw as Hickey and Donnell’s overly pessimistic analysis of the Strategic Hamlet Program. The anthropologists’ findings were “too negative,” ARPA officials complained, and they prepared an official rebuttal to be attached to each copy distributed around the White House and the Pentagon.

Determined to repair any damage that Hickey and Donnell might have done, Collbohm sent a new set of RAND researchers to Saigon with specific instructions to reevaluate the Strategic Hamlet Program. This included Fulbright scholar Joe Carrier, who worked in cost analysis at RAND, and Vic Sturdevant, from systems analysis. With no previous knowledge of Southeast Asia, and
with no local language skills, the two men studied incidents in strategic hamlets initiated by the Vietcong over a nine-month period, from December 1962 to September 1963. Their findings were markedly different from Hickey and Donnell’s. In this new ARPA report on the Strategic Hamlet Program, Carrier and Sturdevant concluded that it would likely prove promising in the long run, if only the Defense Department would take a “more patient approach.”

Another RAND analyst dispatched to Vietnam to write a similarly themed report was George B. Young, an expert in missile design, aerodynamics, and nuclear propulsion. Young, who was Chinese American, became the first RAND employee assigned full-time to the Combat Development Test Center in Saigon. His analysis of the Strategic Hamlet Program was enthusiastic. Young said the villagers were committed to participating. In his ARPA report, called “Notes on Vietnam,” Young wrote about the fluid “delivery of intelligence” information that was taking place. Locals in the program had been taught to make written notes on any Vietcong activity they observed, Young reported. In turn, that information was taken to village elders, who wrote up reports for the Diem government. Soon, Young declared, the Vietcong forces would be “ground to a pulp.”

George Tanham returned to the CDTC in Saigon in 1963, now under a long-term ARPA contract. Much had changed since Tanham’s first trip, at Harold Brown’s behest, in the summer of 1961. In his “Trip Report: Vietnam, 1963,” Tanham showed great optimism about how things were shaping up in Vietnam. An Air Force officer from the Combat Development and Test Center took Tanham in an airplane ride over the strategic hamlet regions, just outside Saigon—some of the very same hamlets that Gerald Hickey and John Donnell had written so pessimistically about in their report, the one that caused Harold Brown to turn his back on them. Tanham marveled at the little villages down below. He said
he could see the bamboo huts, the barbed-wire fences, even the distinct perimeter ditches, and that it all looked wonderful. In Tanham’s estimation, the Defense Department could look ahead to “successfully concluding the war in two or three years or even less.” He included in his report an interview with an officer from the U.S. Air Force who said that the Air Force was “proud of its contribution to the war in Vietnam” and that it planned to “leave behind helicopters and airplanes when it left, ideally sometime in 1964.” Things were looking very positive, Tanham wrote. He quoted a high-ranking general as telling him, “Given a little luck we can wind this one up in a year.”

I
n October 1962, a quiet forty-seven-year-old civilian scientist from Missouri arrived at the Pentagon to begin a new job with the Advanced Research Projects Agency. His work would change the world. By 2015, 3 billion of the 7 billion people on the planet would regularly use technology conceived of by him. The man, J. C. R. Licklider, invented the concept of the Internet, which was originally called the ARPANET.

Licklider did not arrive at the Pentagon with the intent of creating the Internet. He was hired to research and develop command and control systems, most of which were related to nuclear weapons at the time. The idea that a bright red telephone, like the one installed in Herb York’s bedroom in the first week of the Kennedy presidency, was the only way for heads of state to communicate the dreaded “go or no-go” decision in a potential nuclear launch scenario was absurd. In the world of push-button warfare, fractions of seconds mattered. World leaders could not afford the extra seconds it would take to dial a 1962 telephone.

The mandate to update the command and control system,
which would become known as C2, came from the president. Within months of taking office, Kennedy ordered Congress to allocate funds to rapidly modernize the U.S. military command and control system, specifically to make it “more flexible, more selective, more deliberate, better protected, and under ultimate civilian authority at all times.” The directive for “new equipment and facilities” was sent to the Pentagon, where it was tasked to ARPA. Harold Brown recruited J. C. R. Licklider for the job.

Licklider was a trained psychologist with a rare specialization in psychoacoustics, the scientific study of sound perception. Psychoacoustics concerns itself with questions such as, when a person across a room claps his hands, how does the brain know where that sound is coming from? It involves elements of both psychology and physiology, because sound arrives at the ear as a mechanical sound wave, but it is also a perceptual event. People hear differently in different situations, and those “conditions have consequences,” Licklider liked to say. During World War II, while working at Harvard University’s Psycho-Acoustic Laboratory, Licklider conducted experiments with military pilots in all kinds of flight scenarios, with the goal of developing better communication systems for the military. Aircraft were not yet pressurized, and at altitudes of 35,000 feet, cockpit temperatures descended below freezing, which profoundly affected how pilots heard sound and how they responded through speech. Licklider conducted hundreds of experiments with B-17 and B-24 bomber pilots, analyzed data, and published papers on his findings. By war’s end, he was considered one of the world’s authorities on the human auditory nervous system.

After the war, Licklider left Harvard for the Lincoln Laboratory at MIT, where he became interested in how computers could help people communicate better. Engineers at the Lincoln Laboratory were working on an IBM-based computer system for the Air Force called the Semi-Automatic Ground Environment, or SAGE, which was being built to serve as the backbone of the North
American Air Defense Command (NORAD) air defense system. SAGE was the first computer to integrate radar with computer technologies, and to perform three key functions simultaneously: receive, interpret, and respond. The SAGE system received information from tracking radar; it interpreted data as it came in; and in response, it pointed America’s defensive missile systems at incoming threats. It was a gargantuan machine, so large that technicians walked inside it to work on it. SAGE system operators were among the first computer users in the world required to multitask. While sitting at a console, they watched display monitors, typed on keyboards, and flipped switches as new information constantly flowed into the SAGE system through telephone lines.

Licklider was inspired by the SAGE system. To him, it exemplified how computers could do more than just collect data and perform calculations. He imagined a time in the future when man and machine might interact and problem-solve to an even greater degree. He wrote a paper outlining this concept, called “Man-Computer Symbiosis,” in which he described a partnership between humans and “the electronic members of the partnership,” the computers. Licklider envisioned a day when a computer would serve as a human’s “assistant.” The machine would “answer questions, perform simulation modeling, graphically display results, and extrapolate solutions for new situations from past experience.” Like John von Neumann, Licklider saw similarities between the computer and the brain, and he saw a symbiotic relationship between man and machine, one in which man’s burdens, or “rote work,” could be eased by the machine. Humans could then devote their time to making important decisions, Licklider said.

Licklider believed that computers could one day change the world for the better. He envisioned “home computer consoles,” with people sitting in front of them, learning just about anything they wanted to. He wrote a book,
Libraries of the Future,
in which he described a world where library resources would be available to
remote users through a single database. This was radical thinking in 1960 yet is almost taken for granted today by the billions of people who have the library of the Internet at their fingertips twenty-four hours a day. Computers would make man a better-informed being, Licklider wrote, and one day, “in not too many years, human brains and computing machines will be coupled… [and] the resulting partnership will think as no human brain has ever thought.”

It was exactly this kind of revolutionary thinking that interested the Advanced Research Projects Agency and why the work of J. C. R. Licklider caught ARPA’s attention. Computing power needed to be leveraged beyond its present capabilities in order to advance command and control systems, and J. C. R. Licklider was the man for the job. ARPA director Jack Ruina telephoned Licklider and asked him to come to Washington and give a series of seminars on computers to Defense Department officials. Then he offered Licklider a job. When Licklider arrived at the Pentagon just a few months later for his first day of work, the sign on his door read “Advanced Research Projects Agency, Command and Control Research, J. C. R. Licklider, Director.” It was a small office, in both physical size and relative importance. At the time, it was impossible to imagine just how colossal a program command and control would become. In 1962, it was just an idea.

When Licklider arrived at the Pentagon in the fall of 1962, the Department of Defense purchased more computers than any other organization in the world, and ARPA had just entered the world of advanced computer research. The agency inherited four computers from the Air Force, old dinosaurs called Q-32 machines. Each was the size of a small house. These were the computers that the SAGE program had run on at the MIT Lincoln Laboratory, starting in 1954; there was no way the Pentagon was going to throw them away. The Q-32s, built by Systems Development Corporation, a subdivision of RAND, had been incredibly expensive to construct,
each costing $6 million (roughly $50 million in 2015). ARPA had inherited them, and Licklider was given the job of making sure they got used.

Fifteen days after Licklider’s arrival at the Pentagon, the most harrowing of conflicts set the world on a razor’s edge. Photographs taken by a U-2 spy plane revealed that the Soviets had covertly placed nuclear-armed missiles in Cuba, ninety miles off the coast of Florida. President Kennedy demanded that the missiles be removed, but Premier Nikita Khrushchev refused. For thirteen days, starting on October 16, the United States and the Soviet Union played a game of nuclear chicken. At the height of the crisis, on October 24, the United States set up a military blockade off the island and a standoff in the ocean ensued. By all accounts, this thirteen-day period was the closest the world has ever come to nuclear war, before or since. The president raised the defense condition to DEFCON 2 for the first and only time in history. And yet new information from ARPA’s history has recently come to light that paints an even more dramatic Cuban Missile Crisis than was previously understood.

“Guess how many nuclear missiles were detonated during the Cuban Missile Crisis?” asks Paul Kozemchak, special assistant to DARPA director Arati Prabhakar, during an interview for this book. Kozemchak is a thirty-year veteran of DARPA, which makes him the longest-serving employee in its history. “I can tell you that the answer is not ‘none,’” said Kozemchak. “The answer is ‘several.’” In this case, “several” refers to four.

By the time of the Cuban Missile Crisis, Eisenhower’s test ban had failed, and the United States and the Soviet Union had both returned to nuclear weapons testing. Twice during the height of the Cuban Missile Crisis, on October 20 and October 26, 1962, the United States detonated two nuclear weapons—code-named Checkmate and Bluegill Triple Prime—in space. These tests, which sought to advance knowledge in ARPA’s pursuit of the Christofilos
effect, are on the record and are known. What is not known outside Defense Department circles is that in response, on October 22 and October 28, 1962, the Soviets also detonated two nuclear weapons in space, also in pursuit of the Christofilos effect. In recently declassified film footage of an emergency meeting at the White House, Secretary of Defense McNamara can be heard discussing one of these two Soviet nuclear bomb tests with the president and his closest advisors. “The Soviets fired three eleven-hundred-mile missiles yesterday at Kapustin Yar,” McNamara tells them, one of which contained a 300-kiloton nuclear warhead. “They were testing elements of an antimissile system in a nuclear burst environment.”

It is hard to determine what is more shocking, that this information, which was made public by Russian scientists in the early 1990s, is not generally known, or that four nuclear weapons were detonated in space, in a DEFCON 2 environment, during the Cuban Missile Crisis. Firing off nuclear weapons in the middle of a nuclear standoff is tempting fate. The BMEWS system, at J-Site in Thule, could easily have misidentified the Soviet missile launches as a nuclear first strike. “The danger of the situation simply getting out of control, from developments or accidents or incidents that neither side—leaders on either side—were even aware of, much less in control of, could have led to war,” says the former CIA officer Dr. Raymond Garthoff, an expert in Soviet missile launches.

The information about the Soviet high-altitude nuclear tests remained classified until after the Berlin Wall came down. The Soviet nuclear weapon detonated on October 28, 1962, over Zhezqazghan in Kazakhstan at an altitude of ninety-three miles had a consequential effect. According to Russian scientists, “the nuclear detonation caused an electromagnetic pulse [EMP] that covered all of Kazakhstan,” including “electrical cables buried underground.”

The Cuban Missile Crisis made clear that command and control systems not only needed to be upgraded but also needed to be reimagined. It was J. C. R. Licklider who first challenged his
ARPA colleagues to rethink old ideas about what computers could do beyond mathematical tasks like payroll and accounting. Licklider proposed the development of a vast multiuser system, a “network” of computers that could collect information across multiple platforms—from radar and satellites to intelligence reports, communication cables, even weather reports—and to integrate them. What was needed, said Licklider, was a partnership between man and machine, and between the military and the rest of the world.

Of his ARPA bosses, Licklider wrote, “I kept trying to convince them of my philosophy that what the military needs is what the business man needs, is what the scientist needs.” Six months after arriving at ARPA, he sent out a memo calling this network the “Intergalactic Computer Network.” At the time, different computers spoke different programming languages, something Licklider saw as a hurdle that needed to be immediately overcome. It was an extreme problem, he wrote, one “discussed by science fiction writers: How do you get communications started among totally uncorrelated sapient beings?” Finding the answer would take decades, but it began at ARPA in 1962.

J. C. R. Licklider is sometimes called modern computing’s “Johnny Appleseed” for planting the first seeds of the digital revolution. What is not generally known about Licklider is that he ran a second office at the Pentagon called the Behavioral Sciences Program, an office that would eventually take on much more Orwellian tasks related to surveillance programs. This office grew out of a study originally commissioned by Herb York, titled “Toward a Technology of Human Behavior for Defense Use.” This study examined how computers, or “man-machine systems,” could best be used in conflict zones. The results, today, are far-reaching.

In its Behavioral Sciences Program, ARPA wanted to “build a bridge from psychology into the other social sciences” using computers, according to an early ARPA report. Because Licklider was
trained as a psychologist, ARPA director Jack Ruina believed he was the right man for this job, too.

One task of the Behavioral Sciences Program was to imagine a future world where computers could be used by the Defense Department as teaching tools. This was visionary thinking in 1962, when computers still took up entire rooms and cost millions of dollars to build and operate. “Computer assisted teaching systems and computer assisted gaming and simulation studies are examples of work chosen [for] human performance research believed to be defense relevant,” read an internal ARPA report. Training President Diem’s South Vietnamese army was a solid example. ARPA sought ways in which to teach Vietnamese recruits to be better soldiers and more efficient administrators so they could defeat communism. This was arduous, labor-intensive work. Language and culture barriers added an extra layer of toil. One idea behind the Behavioral Sciences Program was that computers could one day shoulder the burden of this kind of work.

The Behavioral Sciences Program initiated a number of projects. These were programs that had a public face but also had highly classified components. ARPA secretly opened a second Combat Development Test Center, this one on the outskirts of Bangkok, five hundred miles to the northwest of Saigon. Like its Vietnamese counterpart, this new CDTC would also research and develop techniques and gadgets but with a focus on longer-term counterinsurgency goals, including Licklider’s plans for computer-assisted teaching, gaming, and simulation studies. Congress was not told about the new Combat Development Test Center in Bangkok, nor was the House Committee on Appropriations, though the Defense Department was legally required to notify it before constructing new facilities.