Prisoner 237 was among the most difficult the interrogators at Bagram Air Base in Afghanistan had encountered.

He was a Saudi who had been captured in Pakistan and delivered by the CIA to the U.S. military prison north of Kabul. He spoke perfect English, and had studied aeronautical engineering at an aviation school in Arizona not far from where one of the September 11 hijackers had trained. He had then gone on to work with so-called charities that had clear links to Al Qaeda. He was, it seemed, a major catch.

Chris Mackey, a reserve sergeant who was the crew chief at Bagram, picked one of his best interrogators to go in and question the Saudi. But the prisoner detected her approaches with such nonchalance that after seven hours she was the one who emerged unnerved and reluctant to go back in.

So Mackey went in himself, and found a prisoner so undaunted by his situation that it was the detainee who asked the first question: “May I help you?” After another unproductive interrogation, Mackey crafted an elaborate plan. It involved convincing the detainee that he might be released in a prisoner exchange if he cooperated. To scare him toward the bait, the plan also included a ploy to make him think America was executing Al Qaeda operatives this prisoner knew.

Neither element was true, but the plan worked to perfection. When Mackey had a phony newspaper clip delivered to 237’s cell mentioning terrorist trials and executions, the prisoner was so stricken he turned to his side and vomited.

Then, just when it seemed 237 was on the verge of surrender, something happened. Mackey went back into the booth armed with new information that other detainees had seen 237 swear loyalty to Osama Bin Laden. Confronted with these claims, the prisoner shifted position in his chair, got a serene look on his face and refused to say another word.

Not long after, 237 was on a plane to the U.S. base in Guantanamo Bay, Cuba, where detainees from the war in Afghanistan are held. He may or may not have since revealed his secrets.

His case underscores one of the moral dilemmas in the war on terrorism: having failed to break 237, what if the interrogators had used harsher methods, maybe even torture? Would the prisoner have talked? Would the ends have justified the means?

•••

Americans didn’t have cause to give much thought to Army interrogation methods
before 9/11. But in the past three years, interrogation and incarceration have become global enterprises for the United States, which operates a constellation of military prisons stretching halfway around the world.

Bagram was among the first of these facilities to open, serving as the main point of entry for prisoners scooped up in Special Forces raids in the rugged territories along the border of Afghanistan and Pakistan. But Bagram is just one node in a network that also includes the detention camp at Guantanamo Bay, clandestine cia facilities whose locations are not publicly disclosed, and, of course, the notorious Abu Ghraib prison in Iraq.

The pictures of American soldiers abusing Iraqi prisoners have become devastating icons. But another image from that scandal jumped out at me: a sign that an officer at Abu Ghraib had posted over the entrance to the interrogators’ meeting room.

That night, panicked prisoners pressed against the edges of the cages . . .

The sign was labeled “Interrogation Rules of Engagement.” Listed down the left side were approved “approaches”—standard techniques that have been in the Army field manual for decades. But running down the right side were methods that could be employed with special permission. They included “Dietary Manip,” “Isolation for longer than 30 days,” “Presence of Mil Working Dogs,” “Sensory Deprivation” and “Stress Positions.”

I had spent hundreds of hours with the interrogation teams first assigned at Bagram. At one time or another, the Bagram crew had considered versions of the methods listed in the right column of that Abu Ghraib sign. They had rejected most as beyond the law, and adopted a few others in less potent form. But the trend line was clear—the longer they were at the prison, the harsher the methods they employed, and the better the information they got.

I had wondered where that trend line led after Mackey’s unit was relieved. The Abu Ghraib “Rules of Engagement” seemed an answer: the line continued, its trajectory maintained by the unit that replaced Mackey’s, as well as by those who subsequently were deployed to Iraq. The pictures drove the Abu Ghraib scandal, but in many ways it was the sign that traced the story’s arc.

After the September 11 attacks, it was evident that the high-tech satellites, listening posts and other intelligence platforms that the United States had spent billions of dollars developing during the Cold War were going to be of limited use against the nation’s new enemy. Success against Al Qaeda would depend on what spies call “human intelligence.” And with little chance of planting a spy inside Al Qaeda’s inner circle, much of that human intelligence would have to come from interrogations of captured terrorists.

Getting prisoners to talk was a dark art of the intelligence world about which very little had been written. After spending several months in the early spring of 2002 persuading officials at the Pentagon to grant me some access to the interrogators in Afghanistan, I traveled to Bagram in May 2002.

The interrogators, part of a unit known as Task Force 202, had recently moved from a larger but far more vulnerable facility at Kandahar, in southwestern Afghanistan. There, the prisoners were kept in open-air cages, and the interrogations took place in dusty tents that baked during the day but were so cold at night that bottles of water froze.

The conditions were significantly better at Bagram. The Army had appropriated a cavernous building that served as a machine shop during the Soviet occupation in the 1970s and turned it into a prison. There, the interrogations took place in six cement-floored offices along a second-floor balcony. The “booths,” as interrogators called them, were protected from the sun and wind. But they were still spartan, with nothing inside but a simple table and two or three chairs.

Given the job description of an Army interrogator, I expected to encounter menacing—or at least intimidating—personalities, people drawn to the task because of some aptitude or affinity for personal confrontation. Instead, the interrogators I met were completely against type.

Among them was a blonde female who was 21 but looked no more than 17, who wanted to become a schoolteacher when she got out of the Army. Another was a 28-year-old reservist whose crop of black hair made him resemble Hawkeye Pierce. He had worked in public relations in New York. Senior interrogator Mackey, also a reservist, had been called away from his career as an international tax expert in London.

Many in the unit had liberal arts degrees, and had joined the service either to pay for school or to learn a foreign language. Their politics skewed more liberal than is typical in the military, so much so that the lone Republican in the bunch, having grown up with a poster of William F. Buckley in his bedroom—called his colleagues “NPR people.”

Army interrogators are trained at Fort Huachuca, Ariz. During a four-month course, the recruits study the Geneva Conventions, learn “order of battle” data on enemy armies and spend hundreds of hours practicing the 16 “approaches”—the psychological levers they use to get prisoners to talk.

Most approaches are designed to exploit a prisoner’s mindset. If he seems angry with his former colleagues, a “hate of comrades” approach might get him to reveal their coordinates. If a photo of a relative is found in his pocket, “love of family” seeks to convince him that the quickest route to being reunited is cooperation. The most popular approach among the students at Huachuca—as well as the screenwriters in Hollywood—is the “fear up.” Its aim is to heighten a prisoner’s anxiety by shouting, tossing chairs and pounding fists.

The rules surrounding interrogation are strict but ambiguous. The guiding principles are the Geneva Conventions, which ban “physical or mental torture,” and outlaw “outrages upon personal dignity, in particular humiliating and degrading treatment.” But what constitutes torture, or an outrage? There are no clear answers.

The Army’s interrogation field manual is more explicit. It bans “the use of force, mental torture, threats, insults or exposure to unpleasant and inhumane treatment of any kind.” But in the same breath, it notes that the ban on the use of force “is not to be confused with psychological ploys, verbal trickery or other nonviolent and non-coercive ruses.”

The first few months at Bagram were marked by frustration and futility, as the interrogators struggled to adapt to an enemy they were ill-prepared to face. Even in the late 1990s, the Army’s training was largely built around a Cold War model that envisioned a clash of conventional forces. The objective was tactical information— where was the prisoner’s unit and how adequate were its supplies? Moreover, the Army assumed that uniform insignia would provide an easy means of sorting prisoners, and that most prisoners would cooperate when asked a direct question. Instruction focused more on knowing what to ask than on how to break a prisoner’s will.

But in Afghanistan, interrogators encountered prisoners who not only wore no uniforms but also sometimes claimed to have no last name. There was no “order of battle” for Al Qaeda, and the direct approach got nowhere with militants who saw their cause as a holy mission. Many prisoners would feign cooperation, even praise the United States, and then spend the next several hours talking in circles, claiming to have forgotten every significant name, place and date.

It was a wall of deception and it took months before the interrogators began to find gaps. But gradually they did.

For starters, they got better at catching lies. They would make prisoners describe how they had arrived in Afghanistan, and what they were doing there, in excruciating detail, then make them retell the story back to front, homing in on inconsistencies.

Later, the unit learned to attack groups of prisoners by ethnicity and nationality—splitting up all the Algerians, for example, and then questioning them one by one, using information from one against the others, until a group of prisoners that had once been united in silence turned into a squabbling, finger-pointing mass. The “echoes,” as the interrogators called themselves, also drove wedges between ethnic groups, convincing the Arabs that the Afghans were selling them out, and so on.

One of the biggest breaks came midway through the war when a Special Forces team returned to Bagram with a pile of documents. Among them was an Al Qaeda training manual on resisting interrogation.

The interrogators were stunned. It spelled out all of the maddening tactics they had faced for months. It coached detainees to withhold key information until their former colleagues had time to adjust their plans, to claim to have forgotten all names and places, to use the Islamic calendar for all references to dates—anything to slow or confuse captors.

The manual practically taunted the interrogators, saying prisoners had little to fear in U.S. custody, that Americans were weak and disinclined to use the harsh methods employed by Middle East countries. Indeed, it urged prisoners to bait American interrogators into physical confrontations, saying bruises or broken bones witnessed by the Red Cross could create an international outcry.

The interrogators were infuriated by the text, but studied it closely and learned to spot patterns of deception. They also took advantage of one of the manual’s inadvertent disclosures—that even if Arab prisoners were not frightened of Americans, they were terrified of being sent back to their home countries, especially Egypt, Jordan and Saudi Arabia, which were notorious for their brutal treatment of Islamic radicals.

The interrogators began holding brainstorming sessions on how to exploit this and other fears, and concocted elaborate ruses, including one designed to make detainees believe they were being placed on a plane to Egypt.

The execution of the plan took a full week. It began with military police planting rumors that an important visitor was coming, and culminated in a staged event on the main prison floor. One of the interrogators—who was born in the Middle East and spoke Arabic with a flawless Gulf region accent—posed as an Egyptian colonel who had come to select prisoners to take home with him. As the prisoners were lined up and brought forward, a soldier slapped duct tape across their chests and scrawled “Egypt” in felt pen on those the “colonel” designated.

That night, panicked prisoners pressed against the edges of the cages, begging for one last chance to come clean and be taken off the fight manifest. The interrogators took their information, and loaded them on a plane the next morning anyway. Only when the plane touched down in Cuba did the detainees realize they had been duped.

In spy movies, one villain (usually an evil genius) knows all, and conveniently spills the pertinent information in a quick two-minute stretch. Real espionage doesn’t work that way. Interrogators spend most of their time collecting tiny bits of truth, fragments of information. Ideally, the information is confirmed by other sources—another prisoner, an intercepted phone call, maybe a photo from a satellite. But often interrogators have nothing but intuition to tell them whether what they got was solid.

During their eight months in Afghanistan, the interrogators of Task Force 202 collected a great deal of valuable intelligence. They documented the escape routes of fleeing members of Al Qaeda; the operations of charities that were really fronts for terrorist groups; the hierarchy of the terrorist network itself. The information they provided led to follow-on raids in Afghanistan and the roll-up of terrorist cells elsewhere in the world.

In some cases, their work is probably paying dividends still. One of the most delicate operations involved Mackey’s recruitment of a 17-year-old prisoner who had been captured with Al Qaeda operatives in a CIA sting. The street-smart teen spoke multiple languages and was fascinated with American culture.

COVERED?: Because Al Qaeda doesn't comply with the rules of war, its 'soldiers' may not be protected by the Geneva Conventions, says a Stanford expert on internation law.

Over a period of weeks, Mackey put him through tests to see whether he could be trusted as an informant, and rewarded his successes by letting him join the interrogators for meals and movies, including the raunchy comedy American Pie.

Eventually, Mackey came to think the prisoner could be more than a mere snitch. So he arranged a meeting between the teen and an agent from a Western intelligence service. A short time later, the boy boarded a plane that would take him back to his former Al Qaeda haunts to work as a spy.

For all their gains, the interrogators always knew their success was incomplete. Few prisoners ever broke in the classic, cinematic sense of total surrender. There was almost always another layer of deception the interrogators hadn’t penetrated.

From the beginning, they had discussed, and sometimes tried, harsher methods. One interrogator was reprimanded early in the war for putting a prisoner in a stress position—making him kneel with his arms extended for 20 or 30 minutes at a time. At one point, MPs suggested using dogs to frighten detainees, an idea that was quickly dismissed. But the method that prompted the most internal debate—and came to be embraced like no other—was sleep deprivation.

As interrogators honed their methods, it became clear that the best information was elicited toward the end of lengthy sessions, particularly those that dragged on 10 hours or more. Tired prisoners were simply more prone to slip.

Keeping prisoners awake worked, but was it right? Initially, the interrogators and the officers in charge of the unit decided the answer was no, that there was no way to reconcile depriving prisoners of sleep with the principles of the Geneva Conventions—which don’t address the matter explicitly but generally disallow punishment for failure to cooperate.

But the interrogators kept returning to the idea, looking for a loophole. Mackey said he came to approach it like a tax problem: here’s the outcome the client wants, how can the tax code be interpreted in such a way to justify it?

In late spring 2002, Mackey devised an answer. He instituted a new rule that a prisoner could be kept awake and in the booth for as long as an interrogator could last. There could be no tag-teaming or substitutions. If the interrogator took a break, the prisoner got one, too. The same rule applied with meals. Technically, the method was sleep deprivation, but it was considered defensible because the interrogator was being deprived of as much sleep as the prisoner.

The interrogators called the new technique “Monstering” and employed it to powerful effect more than a dozen times in the coming months. The reservist who looked like Hawkeye Pierce proved particularly tireless, once going 30 straight hours in the booth before breaking a prisoner who had been caught with a stash of ricin, a poison. He emerged from the booth, wrote a report that identified the prisoner’s Al Qaeda and Taliban handlers, and collapsed on his cot.

The Bagram crew and their immediate officers were largely left on their own to sort out how far they could go within the constraints of the Geneva Conventions. (Though the Bush administration had declared that combatants in Afghanistan were not entitled to full prisoner-of-war status, the interrogators were instructed to behave as if the Geneva Conventions applied.)

What they didn’t know was that a similar debate was simultaneously playing out in Washington at the highest levels of government. Lawyers from the Pentagon, the CIA, the Justice Dept. and the White House were crafting memos defending harsh methods such as the use of dogs and stripping of detainees, and generally making the case that physical coercion might, in certain circumstances, be legally defensible.

The lawyers saw far more latitude than the soldiers. Looking back through the lens of Abu Ghraib, the debates that took place among the interrogators at Bagram in early 2002 seem enlightened.

What was an ending point for Mackey and his unit was a starting point for the teams of interrogators that followed. Military investigations have shown that their replacements, the 519th Military Intelligence Battalion from Fort Bragg, N.C., relaxed many of the rules. By December 2002, according to an investigative report by Major General George R. Fay, the interrogators at Bagram “were removing clothing, isolating people for long periods of time, using stress positions, exploiting fear of dogs and implementing sleep and light deprivation.”

The 519th was running the facility at Bagram when two detainees died there in December 2002. And one of the officers overseeing interrogations during that period, Capt. Carolyn A. Wood, was later deployed to Abu Ghraib, where she crafted the “Interrogation Rules of Engagement” sign.

The reason not to engage in torture is not that it is ineffective. Most interrogators will tell you that torture or physical coercion produces only bad information—that a prisoner under duress will say anything to end the pain. That may be true in the extreme, but there are other points along the continuum, and the experiences of Task Force 202 suggest that harsher methods do work.

The rationale for prohibiting torture is more fundamental: it is dehumanizing, it knocks the interrogator off the moral high ground, and when such methods are exposed, they breed more enemies. One of the many tragedies of Abu Ghraib is that the images of abuse there will inflame anti-American sentiment in the Muslim world for years to come.

Interrogations are critical to winning the war on terror. But the abuses at Abu Ghraib have shown that there is more than one way to lose.

RETURN TO TOP