The Imagineers of War: The Untold Story of DARPA, the Pentagon Agency That Changed the World

by Sharon Weinberger

Another proposal examined under Star was to have a continuous stream of air flowing in front of the president’s speaker stand. It was thought this airflow might slightly deflect bullets or other projectiles, at least enough to protect the president from a direct hit. Simple calculations showed that the airstream would have a minimal effect on anything except tomatoes, and even then the ARPA-funded analysts predicted the thrower would be able to correct his or her aim by the second or third throw.

“During my five years at ARPA, I spent $100 million of the taxpayers’ hard earned money,” Stark later wrote. “I’d like to have it all back.”

Back in 1964, Godel had laid out a proposal to stop the flow of Vietcong fighters and arms into South Vietnam by “sealing” the border using “novel” technologies, a nearly monumental task given the country’s combination of mountainous and jungle terrain.

The resulting proposal called for deforesting 80 to 90 percent of a crucial 180-mile portion of South Vietnam’s border with Laos that formed a part of the Ho Chi Minh Trail. A handwritten list of proposed technologies was indeed novel, if in some cases horrifying. The barrier would require 100,000 “throw-away” shotguns, 250,000 rocket pistols, one million tetrahedrons (ground-based spikes, also known as caltrops), two million mines disguised as rocks, twenty thousand bomblets loaded with chemical defoliants, and an unidentified amount of “insect attractants” (as opposed to insect repellant).

the purpose of the raid was not simply to bomb the Vietcong but rather to spark an out-of-control forest fire using a combination of defoliants and incendiary bombs that would eliminate ground cover for insurgents.

The JASONs devised a wholly unique and arguably improved border system—a virtual barrier made up of thousands of air-dropped ground sensors linked to a computer system that would cue strike aircraft to a suspected infiltration.

By linking sensors and aircraft to computers in real time, the JASON proposal transformed Godel’s idea of a barrier into something far more technologically advanced: the world’s first electronic battlefield. What the barrier failed to do, by all accounts, was to have any appreciable effect on the course of the war.

An American firm, called Simulmatics Corporation, had sent Slote to Vietnam in 1966 to help the Pentagon understand the growing insurgency. Slote believed the Rorschach test, popular at the time among psychotherapists to diagnose personality traits, could be used to understand the reasons behind growing resentment of the United States and the South Vietnamese government.

Even the antigovernment Buddhist monk Slote interviewed was more cooperative. “You know, I’ve never seen one, except on a child,” the monk replied in astonishment, when Slote asked him if a particular ink blot resembled a vagina.

The chief proponent of this new line of work was Seymour Deitchman, an emerging member of Secretary of Defense McNamara’s technocratic elite and a longtime nemesis of William Godel’s. Deitchman was convinced it was an engineer’s slide rule, not a soldier’s intuition, that would determine who won or lost battles. More important, he believed that people could be studied and their actions predicted the way engineers measure and track the flight of a ballistic missile.

General Electric in August 1965, for example, wrote to ARPA suggesting the company be given a “continuing open ended type contract” that would allow it to apply its technology to counterinsurgency. Its first proposal was a “mass polygraph for internal village security.”

Deitchman, however, was resolved to clean up AGILE’s problems and bring in grounded science. One of his first decisions was to get rid of Herman Kahn, the portly nuclear war theorist.

That included, in one report, a proposal to build a moat around Saigon to protect it from the Vietcong.

Deitchman got rid of Kahn, and when Kahn threatened to complain to McNamara, Deitchman called him on his bluff. “Go ahead,” Deitchman said.

One of Rand’s most significant ARPA-funded projects, and the one that would become its most famous wartime social science work, was the Viet Cong Motivation and Morale Project, which sought to understand support for the the communist insurgency. Two Rand analysts, Joe Zasloff and John Donnell, were sent over to supervise the interviews, which were conducted with captured Vietcong, as well as those who had surrendered under amnesty. The initial analysis of the interviews was rather dismal for the prospects of U.S. intervention in Vietnam. The study produced insights that did not mesh with the official line, because it showed that the Vietnamese were joining the Vietcong out of genuine political conviction, rather than being forced to join,

Yet he found that growing opposition to American involvement in Vietnam, particularly on college campuses, was making it almost impossible for ARPA to find good academics to work for the Defense Department on Southeast Asia issues.

In 1966, ARPA gave Simulmatics a wide-ranging contract for social science work in Vietnam, and soon the first research teams started showing up in Saigon. It was to be one of the most disastrous contracts ever let under AGILE.

Most dismal, perhaps, was a “sorcerer’s project,” which had enlisted Vietnamese sorcerers—essentially local magicians—to sway villagers against the Vietcong. It failed because, as Hoc put it without a hint of irony, “the sorcerers did not say what they were supposed to say.”

He ended with a request that was never fulfilled, because his original letter was lying in the National Archives more than forty years later: “Burn this after reading.”

ARPA’s counterinsurgency work, though largely unsuccessful, rolled forward out of classic bureaucratic inertia. It had become central to ARPA—its third-largest program after missile defense and nuclear test detection—so to admit failure would have been to resign from a core mission.

Glenard Lipscomb,

In the end, it was Congress that put an end to ARPA’s support for the social sciences.

In 1969, the Democratic senator Mike Mansfield, an ardent critic of the Vietnam War, pushed through what became known as the Mansfield Amendment, which prohibited the Defense Department from funding research that “lacked a direct or apparent relationship to a specific military function.” The amendment struck at the heart of ARPA’s social science funding, ending much of the agency’s work in Southeast Asia and the Middle East.

Over the next few years, almost all of the governments that ARPA worked with on counterinsurgency collapsed. Only Thailand managed to avoid a total political implosion: a 1973 revolution toppled the military-led government, but the country never fell to a communist insurgency. In 1974, a communist junta took power in Ethiopia. The following year, South Vietnam fell to a conventional invasion by the North, and the country was united under a communist government; sectarian divisions in Lebanon ignited a civil war

“As to the reason you are here,” the fortune-teller told the two Americans, “it will be like scissors cutting water.”

AGILE—and counterinsurgency in Vietnam more generally—failed in large part because it was supporting a government that was incapable of providing the security the population wanted, and no amount of American forces could change that.

On October 22, 1964, half a mile underground in a salt mine in Mississippi, the United States set off a nuclear device about a third as powerful as the bomb that leveled Hiroshima. Above where the nuclear device was detonated, someone had placed a Confederate battle flag next to a sign that read, “The South Will Rise Again.” It was an unexpectedly apt description for what happened next: shock waves from the nuclear explosion, which scientists had expected to be contained inside the underground cavity, sent tremors through the earth.

It took the government two years of pumping in fresh air to remove the gas and lower the interior temperature to a still scorching three hundred degrees Fahrenheit. Then, on December 3, 1966, ARPA set off another nuclear explosion in the same salt mine.

In 1965, ARPA established the Advanced Sensors Office, specifically to market technology—derived from its work in nuclear test detection—to the CIA and intelligence community.

In the days of the Corona spy satellite, ARPA had been an unwanted interloper from the CIA’s perspective. Similarly, ARPA’s involvement in nuclear test detection intruded on what the intelligence community believed was its territory. And, of course, there was William Godel and his work in Vietnam, which had also been viewed with suspicion by the CIA.

The Advanced Sensors Office was established in 1965 with a modest budget of just under $5 million, and its first program was Project Pandora, a top secret research program into mind control.

The Moscow Viral Study, as it was called, was the cover story for the American government’s top secret investigation into the effects of microwave radiation on humans. The Soviets, it turned out, were bombarding the embassy in Moscow with low-level microwaves. The “Moscow Signal,” as officials in Washington called the radiation, was too low to do any obvious harm to the people in the building.

Cesaro was one of ARPA’s earliest employees, hired back in 1958, and also perhaps its most notorious, with a reputation for being creative, aggressive, and obnoxiously rude. No one quite knew what Cesaro was doing, and that was the way he liked it.

With the passage of time, the government’s concerns about microwave-induced mind control might sound like something born of the worst sort of Cold War paranoia—the sort of thing easily parodied as a tin-foil-hat conspiracy—but set in the landscape of the 1960s, it seemed a plausible concern.

Cesaro found another area where he could market ARPA’s secret sensor technology: Vietnam. Instead of zapping people with microwaves, however, Cesaro was trying to use ARPA’s sensors to hunt them down and kill them.

Reliability problems plagued the QH-50. Trying to get the weapons to accurately hit anything also proved futile.

Among other misdeeds, he was sent to prison for impregnating possibly dozens of unsuspecting patients with his own sperm, rather than that of screened anonymous donors as they were expecting.

The legacy of ARPA’s work on armed drones would not become clear until thirty years later, when, in the weeks following September 11, 2001, the air force took a QH-50 and put a Hellfire missile on it. The test was a failure.

Vietnam began to take over the Pentagon figuratively and literally. ARPA’s offices were moved out of the Pentagon to make way for analysts working on the Vietnam War, and the agency moved into leased office space on Wilson Boulevard in Rosslyn, Virginia.

To protect its bureaucracy, the Pentagon began designating offices as field agencies to immunize them from reductions. On March 23 of that year, ARPA officially became the Defense Advanced Research Projects Agency, which meant that ARPA would hence be DARPA. On its own, the new name had no significance for the agency’s direction, but it was a symbolic defeat. Lukasik despised the name change, insisting that the acronym remain “ARPA,” which it did, until the end of his directorship. “This was not a minor point of civil disobedience,” Lukasik said. “I said freely that DARPA sounded like a dog food.”

Lukasik’s predecessor, Eberhardt Rechtin, had focused on quietly getting rid of projects that were subject to congressional scrutiny, like a $1 million “mechanical elephant.” Operated by a man sitting inside the machine using hydroelectric controls, the four-legged “Cybernetic Anthropomorphous Machine” was designed to navigate the jungles of Vietnam, hauling equipment for soldiers. Rechtin called it a “damned fool” project bound to land ARPA in hot water.

In 1970, ARPA proposed a new approach in Iran, which it called “high level systems analysis,” advising senior ministry officials on how to buy military weapons. The idea was to teach the staff the basics of systems analysis, which meant, for example, not just comparing the cost of an American missile with that of a British one but calculating the actual “kill per engagement” to come up with a total cost comparison.

“The trouble was that Tony Cordesman was a pain in the ass and he got everybody mad,”

What other analysis was being masked is unknown, and even fifty years later Cordesman declined to discuss any details of his work for ARPA, claiming that he believed it to still be classified.

When the navy demonstrated its F-14, the pilot put on an even better show: he did a roll, circled around, and made a loop right over the shah’s head. After the performance, the shah said to Currie, “You know, I’ve always viewed Iran as kind of an island in the middle of an ocean.” Currie marveled at the shah’s nonsensical statement. The F-14 was a fighter built for aircraft carriers, something that Iran did not possess. It was more expensive and had a shorter range than the F-15, but it was clear the Iranian leader’s mind was made up. Iran became the only foreign country to ever buy the F-14, a decision based on a thirty-minute air show.

Both men were eventually released unharmed, though Woods, the ARPA researcher, later groused that the government denied him per diem for the days he was held, because the hijackers had provided him with room and board.

Cordesman was brilliant but arrogant, producing thick analytical reports faster than most people could read them. “He projected an air of superiority over lesser beings, particularly Iranians who felt insecure to begin with,” recalled Precht.

Charles Herzfeld’s argument to Congress in the late 1960s had been that ARPA was doing valuable work that no one else was doing, and sometimes he was right. In 1971, the Beirut office was asked to look into the little-known threat of “improvised explosive devices,” the technical name for homemade bombs. More than forty years before IEDs entered the popular lexicon and became the leading killer of American and coalition troops in Iraq and Afghanistan, ARPA hired a contractor to study them and prepare a comprehensive report,

After spending nearly $20 billion, the director of that agency, called the Joint Improvised Explosive Device Defeat Organization, admitted in 2010 that the best method the Pentagon had for detecting bombs was still a dog. The 1971 ARPA report, in the meantime, sat in a box in the National Archives and Records Administration in College Park, Maryland.

“I took a really bad dish and I poured in some crap and some shit, and it became the Tactical Technology Office,” Lukasik joked.