Hellhole

Annals of Human Rights

The United States holds tens of thousands of inmates in long-term solitary confinement. Is this torture?

by Atul Gawande March 30, 2009 New Yorker

Human beings are social creatures. We are social not just in the trivial sense that we like company, and not just in the obvious sense that we each depend on others. We are social in a more elemental way: simply to exist as a normal human being requires interaction with other people.

Children provide the clearest demonstration of this fact, although it was slow to be accepted. Well into the nineteen-fifties, psychologists were encouraging parents to give children less attention and affection, in order to encourage independence. Then Harry Harlow, a professor of psychology at the University of Wisconsin at Madison, produced a series of influential studies involving baby rhesus monkeys.

He happened upon the findings in the mid-fifties, when he decided to save money for his primate-research laboratory by breeding his own lab monkeys instead of importing them from India. Because he didn’t know how to raise infant monkeys, he cared for them the way hospitals of the era cared for human infants—in nurseries, with plenty of food, warm blankets, some toys, and in isolation from other infants to prevent the spread of infection. The monkeys grew up sturdy, disease-free, and larger than those from the wild. Yet they were also profoundly disturbed, given to staring blankly and rocking in place for long periods, circling their cages repetitively, and mutilating themselves.

At first, Harlow and his graduate students couldn’t figure out what the problem was. They considered factors such as diet, patterns of light exposure, even the antibiotics they used. Then, as Deborah Blum recounts in a fascinating biography of Harlow, “Love at Goon Park,” one of his researchers noticed how tightly the monkeys clung to their soft blankets. Harlow wondered whether what the monkeys were missing in their Isolettes was a mother. So, in an odd experiment, he gave them an artificial one.

In the studies, one artificial mother was a doll made of terry cloth; the other was made of wire. He placed a warming device inside the dolls to make them seem more comforting. The babies, Harlow discovered, largely ignored the wire mother. But they became deeply attached to the cloth mother. They caressed it. They slept curled up on it. They ran to it when frightened. They refused replacements: they wanted only “their” mother. If sharp spikes were made to randomly thrust out of the mother’s body when the rhesus babies held it, they waited patiently for the spikes to recede and returned to clutching it. No matter how tightly they clung to the surrogate mothers, however, the monkeys remained psychologically abnormal.

In a later study on the effect of total isolation from birth, the researchers found that the test monkeys, upon being released into a group of ordinary monkeys, “usually go into a state of emotional shock, characterized by . . . autistic self-clutching and rocking.” Harlow noted, “One of six monkeys isolated for three months refused to eat after release and died five days later.” After several weeks in the company of other monkeys, most of them adjusted—but not those who had been isolated for longer periods. “Twelve months of isolation almost obliterated the animals socially,” Harlow wrote. They became permanently withdrawn, and they lived as outcasts—regularly set upon, as if inviting abuse.

The research made Harlow famous (and infamous, too—revulsion at his work helped spur the animal-rights movement). Other psychologists produced evidence of similarly deep and sustained damage in neglected and orphaned children. Hospitals were made to open up their nurseries to parents. And it became widely accepted that children require nurturing human beings not just for food and protection but also for the normal functioning of their brains.

We have been hesitant to apply these lessons to adults. Adults, after all, are fully formed, independent beings, with internal strengths and knowledge to draw upon. We wouldn’t have anything like a child’s dependence on other people, right? Yet it seems that we do. We don’t have a lot of monkey experiments to call upon here. But mankind has produced tens of thousands of human ones, including in our prison system. And the picture that has emerged is profoundly unsettling.

Among our most benign experiments are those with people who voluntarily isolate themselves for extended periods. Long-distance solo sailors, for instance, commit themselves to months at sea. They face all manner of physical terrors: thrashing storms, fifty-foot waves, leaks, illness. Yet, for many, the single most overwhelming difficulty they report is the “soul-destroying loneliness,” as one sailor called it. Astronauts have to be screened for their ability to tolerate long stretches in tightly confined isolation, and they come to depend on radio and video communications for social contact.

The problem of isolation goes beyond ordinary loneliness, however. Consider what we’ve learned from hostages who have been held in solitary confinement—from the journalist Terry Anderson, for example, whose extraordinary memoir, “Den of Lions,” recounts his seven years as a hostage of Hezbollah in Lebanon.

Anderson was the chief Middle East correspondent for the Associated Press when, on March 16, 1985, three bearded men forced him from his car in Beirut at gunpoint. He was pushed into a Mercedes sedan, covered head to toe with a heavy blanket, and made to crouch head down in the footwell behind the front seat. His captors drove him to a garage, pulled him out of the car, put a hood over his head, and bound his wrists and ankles with tape. For half an hour, they grilled him for the names of other Americans in Beirut, but he gave no names and they did not beat him or press him further. They threw him in the trunk of the car, drove him to another building, and put him in what would be the first of a succession of cells across Lebanon. He was soon placed in what seemed to be a dusty closet, large enough for only a mattress. Blindfolded, he could make out the distant sounds of other hostages. (One was William Buckley, the C.I.A. station chief who was kidnapped and tortured repeatedly until he weakened and died.) Peering around his blindfold, Anderson could see a bare light bulb dangling from the ceiling. He received three unpalatable meals a day—usually a sandwich of bread and cheese, or cold rice with canned vegetables, or soup. He had a bottle to urinate in and was allotted one five- to ten-minute trip each day to a rotting bathroom to empty his bowels and wash with water at a dirty sink. Otherwise, the only reprieve from isolation came when the guards made short visits to bark at him for breaking a rule or to threaten him, sometimes with a gun at his temple.

He missed people terribly, especially his fiancée and his family. He was despondent and depressed. Then, with time, he began to feel something more. He felt himself disintegrating. It was as if his brain were grinding down. A month into his confinement, he recalled in his memoir, “The mind is a blank. Jesus, I always thought I was smart. Where are all the things I learned, the books I read, the poems I memorized? There’s nothing there, just a formless, gray-black misery. My mind’s gone dead. God, help me.”

He was stiff from lying in bed day and night, yet tired all the time. He dozed off and on constantly, sleeping twelve hours a day. He craved activity of almost any kind. He would watch the daylight wax and wane on the ceiling, or roaches creep slowly up the wall. He had a Bible and tried to read, but he often found that he lacked the concentration to do so. He observed himself becoming neurotically possessive about his little space, at times putting his life in jeopardy by flying into a rage if a guard happened to step on his bed. He brooded incessantly, thinking back on all the mistakes he’d made in life, his regrets, his offenses against God and family.

His captors moved him every few months. For unpredictable stretches of time, he was granted the salvation of a companion—sometimes he shared a cell with as many as four other hostages—and he noticed that his thinking recovered rapidly when this occurred. He could read and concentrate longer, avoid hallucinations, and better control his emotions. “I would rather have had the worst companion than no companion at all,” he noted.

In September, 1986, after several months of sharing a cell with another hostage, Anderson was, for no apparent reason, returned to solitary confinement, this time in a six-by-six-foot cell, with no windows, and light from only a flickering fluorescent lamp in an outside corridor. The guards refused to say how long he would be there. After a few weeks, he felt his mind slipping away again.

“I find myself trembling sometimes for no reason,” he wrote. “I’m afraid I’m beginning to lose my mind, to lose control completely.”

One day, three years into his ordeal, he snapped. He walked over to a wall and began beating his forehead against it, dozens of times. His head was smashed and bleeding before the guards were able to stop him.

Some hostages fared worse. Anderson told the story of Frank Reed, a fifty-four-year-old American private-school director who was taken hostage and held in solitary confinement for four months before being put in with Anderson. By then, Reed had become severely withdrawn. He lay motionless for hours facing a wall, semi-catatonic. He could not follow the guards’ simplest instructions. This invited abuse from them, in much the same way that once isolated rhesus monkeys seemed to invite abuse from the colony. Released after three and a half years, Reed ultimately required admission to a psychiatric hospital.

“It’s an awful thing, solitary,” John McCain wrote of his five and a half years as a prisoner of war in Vietnam—more than two years of it spent in isolation in a fifteen-by-fifteen-foot cell, unable to communicate with other P.O.W.s except by tap code, secreted notes, or by speaking into an enamel cup pressed against the wall. “It crushes your spirit and weakens your resistance more effectively than any other form of mistreatment.” And this comes from a man who was beaten regularly; denied adequate medical treatment for two broken arms, a broken leg, and chronic dysentery; and tortured to the point of having an arm broken again. A U.S. military study of almost a hundred and fifty naval aviators returned from imprisonment in Vietnam, many of whom were treated even worse than McCain, reported that they found social isolation to be as torturous and agonizing as any physical abuse they suffered.

And what happened to them was physical. EEG studies going back to the nineteen-sixties have shown diffuse slowing of brain waves in prisoners after a week or more of solitary confinement. In 1992, fifty-seven prisoners of war, released after an average of six months in detention camps in the former Yugoslavia, were examined using EEG-like tests. The recordings revealed brain abnormalities months afterward; the most severe were found in prisoners who had endured either head trauma sufficient to render them unconscious or, yes, solitary confinement. Without sustained social interaction, the human brain may become as impaired as one that has incurred a traumatic injury.

On December 4, 1991, Terry Anderson was released from captivity. He had been the last and the longest-held American hostage in Lebanon. I spoke to Keron Fletcher, a former British military psychiatrist who had been on the receiving team for Anderson and many other hostages, and followed them for years afterward. Initially, Fletcher said, everyone experiences the pure elation of being able to see and talk to people again, especially family and friends. They can’t get enough of other people, and talk almost non-stop for hours. They are optimistic and hopeful. But, afterward, normal sleeping and eating patterns prove difficult to reëstablish. Some have lost their sense of time. For weeks, they have trouble managing the sensations and emotional complexities of their freedom.

For the first few months after his release, Anderson said when I reached him by phone recently, “it was just kind of a fog.” He had done many television interviews at the time. “And if you look at me in the pictures? Look at my eyes. You can tell. I look drugged.”

Most hostages survived their ordeal, Fletcher said, although relationships, marriages, and careers were often lost. Some found, as John McCain did, that the experience even strengthened them. Yet none saw solitary confinement as anything less than torture. This presents us with an awkward question: If prolonged isolation is—as research and experience have confirmed for decades—so objectively horrifying, so intrinsically cruel, how did we end up with a prison system that may subject more of our own citizens to it than any other country in history has?

Recently, I met a man who had spent more than five years in isolation at a prison in the Boston suburb of Walpole, Massachusetts, not far from my home. Bobby Dellelo was, to say the least, no Terry Anderson or John McCain. Brought up in the run-down neighborhoods of Boston’s West End, in the nineteen-forties, he was caught burglarizing a shoe store at the age of ten. At thirteen, he recalls, he was nabbed while robbing a Jordan Marsh department store. (He and his friends learned to hide out in stores at closing time, steal their merchandise, and then break out during the night.) The remainder of his childhood was spent mostly in the state reform school. That was where he learned how to fight, how to hot-wire a car with a piece of foil, how to pick locks, and how to make a zip gun using a snapped-off automobile radio antenna, which, in those days, was just thick enough to barrel a .22-calibre bullet. Released upon turning eighteen, Dellelo returned to stealing. Usually, he stole from office buildings at night. But some of the people he hung out with did stickups, and, together with one of them, he held up a liquor store in Dorchester.

“What a disaster that thing was,” he recalls, laughing. They put the store’s owner and the customers in a walk-in refrigerator at gunpoint, took their wallets, and went to rob the register. But more customers came in. So they robbed them and put them in the refrigerator, too. Then still more customers arrived, the refrigerator got full, and the whole thing turned into a circus. Dellelo and his partner finally escaped. But one of the customers identified him to the police. By the time he was caught, Dellelo had been fingered for robbing the Commander Hotel in Cambridge as well. He served a year for the first conviction and two and a half years for the second.

Three months after his release, in 1963, at the age of twenty, he and a friend tried to rob the Kopelman jewelry store, in downtown Boston. But an alarm went off before they got their hands on anything. They separated and ran. The friend shot and killed an off-duty policeman while trying to escape, then killed himself. Dellelo was convicted of first-degree murder and sentenced to life in prison. He ended up serving forty years. Five years and one month were spent in isolation.

The criteria for the isolation of prisoners vary by state but typically include not only violent infractions but also violation of prison rules or association with gang members. The imposition of long-term isolation—which can be for months or years—is ultimately at the discretion of prison administrators. One former prisoner I spoke to, for example, recalled being put in solitary confinement for petty annoyances like refusing to get out of the shower quickly enough. Bobby Dellelo was put there for escaping.

It was an elaborate scheme. He had a partner, who picked the lock to a supervisor’s office and got hold of the information manual for the microwave-detection system that patrolled a grassy no man’s land between the prison and the road. They studied the manual long enough to learn how to circumvent the system and returned it. On Halloween Sunday, 1993, they had friends stage a fight in the prison yard. With all the guards in the towers looking at the fight through binoculars, the two men tipped a picnic table up against a twelve-foot wall and climbed it like a ladder. Beyond it, they scaled a sixteen-foot fence. To get over the razor wire on top, they used a Z-shaped tool they’d improvised from locker handles. They dropped down into the no man’s land and followed an invisible path that they’d calculated the microwave system would not detect. No alarm sounded. They went over one more fence, walked around a parking lot, picked their way through some woods, and emerged onto a four-lane road. After a short walk to a convenience store, they called a taxi from a telephone booth and rolled away before anyone knew they were gone.

They lasted twenty-four days on the outside. Eventually, somebody ratted them out, and the police captured them on the day before Thanksgiving, at the house of a friend in Cambridge. The prison administration gave Dellelo five years in the Departmental Disciplinary Unit of the Walpole prison, its hundred-and-twenty-four-cell super-maximum segregation unit.

Wearing ankle bracelets, handcuffs, and a belly chain, Dellelo was marched into a thirteen-by-eight-foot off-white cell. A four-inch-thick concrete bed slab jutted out from the wall opposite the door. A smaller slab protruding from a side wall provided a desk. A cylindrical concrete block in the floor served as a seat. On the remaining wall was a toilet and a metal sink. He was given four sheets, four towels, a blanket, a bedroll, a toothbrush, toilet paper, a tall clear plastic cup, a bar of soap, seven white T-shirts, seven pairs of boxer shorts, seven pairs of socks, plastic slippers, a pad of paper, and a ballpoint pen. A speaker with a microphone was mounted on the door. Cells used for solitary confinement are often windowless, but this one had a ribbonlike window that was seven inches wide and five feet tall. The electrically controlled door was solid steel, with a seven-inch-by-twenty-eight-inch aperture and two wickets—little door slots, one at ankle height and one at waist height, for shackling him whenever he was let out and for passing him meal trays.

As in other supermaxes—facilities designed to isolate prisoners from social contact—Dellelo was confined to his cell for at least twenty-three hours a day and permitted out only for a shower or for recreation in an outdoor cage that he estimated to be fifty feet long and five feet wide, known as “the dog kennel.” He could talk to other prisoners through the steel door of his cell, and during recreation if a prisoner was in an adjacent cage. He made a kind of fishing line for passing notes to adjacent cells by unwinding the elastic from his boxer shorts, though it was contraband and would be confiscated. Prisoners could receive mail and as many as ten reading items. They were allowed one phone call the first month and could earn up to four calls and four visits per month if they followed the rules, but there could be no physical contact with anyone, except when guards forcibly restrained them. Some supermaxes even use food as punishment, serving the prisoners nutra-loaf, an unpalatable food brick that contains just enough nutrition for survival. Dellelo was spared this. The rules also permitted him to have a radio after thirty days, and, after sixty days, a thirteen-inch black-and-white television.

“This is going to be a piece of cake,” Dellelo recalls thinking when the door closed behind him. Whereas many American supermax prisoners—and most P.O.W.s and hostages—have no idea when they might get out, he knew exactly how long he was going to be there. He drew a calendar on his pad of paper to start counting down the days. He would get a radio and a TV. He could read. No one was going to bother him. And, as his elaborate escape plan showed, he could be patient. “This is their sophisticated security?” he said to himself. “They don’t know what they’re doing.”

After a few months without regular social contact, however, his experience proved no different from that of the P.O.W.s or hostages, or the majority of isolated prisoners whom researchers have studied: he started to lose his mind. He talked to himself. He paced back and forth compulsively, shuffling along the same six-foot path for hours on end. Soon, he was having panic attacks, screaming for help. He hallucinated that the colors on the walls were changing. He became enraged by routine noises—the sound of doors opening as the guards made their hourly checks, the sounds of inmates in nearby cells. After a year or so, he was hearing voices on the television talking directly to him. He put the television under his bed, and rarely took it out again.

One of the paradoxes of solitary confinement is that, as starved as people become for companionship, the experience typically leaves them unfit for social interaction. Once, Dellelo was allowed to have an in-person meeting with his lawyer, and he simply couldn’t handle it. After so many months in which his primary human contact had been an occasional phone call or brief conversations with an inmate down the tier, shouted through steel doors at the top of their lungs, he found himself unable to carry on a face-to-face conversation. He had trouble following both words and hand gestures and couldn’t generate them himself. When he realized this, he succumbed to a full-blown panic attack.

Craig Haney, a psychology professor at the University of California at Santa Cruz, received rare permission to study a hundred randomly selected inmates at California’s Pelican Bay supermax, and noted a number of phenomena. First, after months or years of complete isolation, many prisoners “begin to lose the ability to initiate behavior of any kind—to organize their own lives around activity and purpose,” he writes. “Chronic apathy, lethargy, depression, and despair often result. . . . In extreme cases, prisoners may literally stop behaving,” becoming essentially catatonic.

Second, almost ninety per cent of these prisoners had difficulties with “irrational anger,” compared with just three per cent of the general population.* Haney attributed this to the extreme restriction, the totality of control, and the extended absence of any opportunity for happiness or joy. Many prisoners in solitary become consumed with revenge fantasies.

“There were some guards in D.D.U. who were decent guys,” Dellelo told me. They didn’t trash his room when he was let out for a shower, or try to trip him when escorting him in chains, or write him up for contraband if he kept food or a salt packet from a meal in his cell. “But some of them were evil, evil pricks.” One correctional officer became a particular obsession. Dellelo spent hours imagining cutting his head off and rolling it down the tier. “I mean, I know this is insane thinking,” he says now. Even at the time, he added, “I had a fear in the background—like how much of this am I going to be able to let go? How much is this going to affect who I am?”

He was right to worry. Everyone’s identity is socially created: it’s through your relationships that you understand yourself as a mother or a father, a teacher or an accountant, a hero or a villain. But, after years of isolation, many prisoners change in another way that Haney observed. They begin to see themselves primarily as combatants in the world, people whose identity is rooted in thwarting prison control.

As a matter of self-preservation, this may not be a bad thing. According to the Navy P.O.W. researchers, the instinct to fight back against the enemy constituted the most important coping mechanism for the prisoners they studied. Resistance was often their sole means of maintaining a sense of purpose, and so their sanity. Yet resistance is precisely what we wish to destroy in our supermax prisoners. As Haney observed in a review of research findings, prisoners in solitary confinement must be able to withstand the experience in order to be allowed to return to the highly social world of mainline prison or free society. Perversely, then, the prisoners who can’t handle profound isolation are the ones who are forced to remain in it. “And those who have adapted,” Haney writes, “are prime candidates for release to a social world to which they may be incapable of ever fully readjusting.”

Dellelo eventually found a way to resist that would not prolong his ordeal. He fought his battle through the courts, filing motion after motion in an effort to get his conviction overturned. He became so good at submitting his claims that he obtained a paralegal certificate along the way. And, after forty years in prison, and more than five years in solitary, he got his first-degree-homicide conviction reduced to manslaughter. On November 19, 2003, he was freed.

Bobby Dellelo is sixty-seven years old now. He lives on Social Security in a Cambridge efficiency apartment that is about four times larger than his cell. He still seems to be adjusting to the world outside. He lives alone. To the extent that he is out in society, it is, in large measure, as a combatant. He works for prisoners’ rights at the American Friends Service Committee. He also does occasional work assisting prisoners with their legal cases. Sitting at his kitchen table, he showed me how to pick a padlock—you know, just in case I ever find myself in trouble.

But it was impossible to talk to him about his time in isolation without seeing that it was fundamentally no different from the isolation that Terry Anderson and John McCain had endured. Whether in Walpole or Beirut or Hanoi, all human beings experience isolation as torture.

The main argument for using long-term isolation in prisons is that it provides discipline and prevents violence. When inmates refuse to follow the rules—when they escape, deal drugs, or attack other inmates and corrections officers—wardens must be able to punish and contain the misconduct. Presumably, less stringent measures haven’t worked, or the behavior would not have occurred. And it’s legitimate to incapacitate violent aggressors for the safety of others. So, advocates say, isolation is a necessary evil, and those who don’t recognize this are dangerously naïve.

The argument makes intuitive sense. If the worst of the worst are removed from the general prison population and put in isolation, you’d expect there to be markedly fewer inmate shankings and attacks on corrections officers. But the evidence doesn’t bear this out. Perhaps the most careful inquiry into whether supermax prisons decrease violence and disorder was a 2003 analysis examining the experience in three states—Arizona, Illinois, and Minnesota—following the opening of their supermax prisons. The study found that levels of inmate-on-inmate violence were unchanged, and that levels of inmate-on-staff violence changed unpredictably, rising in Arizona, falling in Illinois, and holding steady in Minnesota.

Prison violence, it turns out, is not simply an issue of a few belligerents. In the past thirty years, the United States has quadrupled its incarceration rate but not its prison space. Work and education programs have been cancelled, out of a belief that the pursuit of rehabilitation is pointless. The result has been unprecedented overcrowding, along with unprecedented idleness—a nice formula for violence. Remove a few prisoners to solitary confinement, and the violence doesn’t change. So you remove some more, and still nothing happens. Before long, you find yourself in the position we are in today. The United States now has five per cent of the world’s population, twenty-five per cent of its prisoners, and probably the vast majority of prisoners who are in long-term solitary confinement.

It wasn’t always like this. The wide-scale use of isolation is, almost exclusively, a phenomenon of the past twenty years. In 1890, the United States Supreme Court came close to declaring the punishment to be unconstitutional. Writing for the majority in the case of a Colorado murderer who had been held in isolation for a month, Justice Samuel Miller noted that experience had revealed “serious objections” to solitary confinement:

A considerable number of the prisoners fell, after even a short confinement, into a semi-fatuous condition, from which it was next to impossible to arouse them, and others became violently insane; others, still, committed suicide; while those who stood the ordeal better were not generally reformed, and in most cases did not recover suffcient mental activity to be of any subsequent service to the community.

Prolonged isolation was used sparingly, if at all, by most American prisons for almost a century. Our first supermax—our first institution specifically designed for mass solitary confinement—was not established until 1983, in Marion, Illinois. In 1995, a federal court reviewing California’s first supermax admitted that the conditions “hover on the edge of what is humanly tolerable for those with normal resilience.” But it did not rule them to be unconstitutionally cruel or unusual, except in cases of mental illness. The prison’s supermax conditions, the court stated, did not pose “a sufficiently high risk to all inmates of incurring a serious mental illness.” In other words, there could be no legal objection to its routine use, given that the isolation didn’t make everyone crazy. The ruling seemed to fit the public mood. By the end of the nineteen-nineties, some sixty supermax institutions had opened across the country. And new solitary-confinement units were established within nearly all of our ordinary maximum-security prisons.

The number of prisoners in these facilities has since risen to extraordinary levels. America now holds at least twenty-five thousand inmates in isolation in supermax prisons. An additional fifty to eighty thousand are kept in restrictive segregation units, many of them in isolation, too, although the government does not release these figures. By 1999, the practice had grown to the point that Arizona, Colorado, Maine, Nebraska, Nevada, Rhode Island, and Virginia kept between five and eight per cent of their prison population in isolation, and, by 2003, New York had joined them as well. Mississippi alone held eighteen hundred prisoners in supermax—twelve per cent of its prisoners over all. At the same time, other states had just a tiny fraction of their inmates in solitary confinement. In 1999, for example, Indiana had eighty-five supermax beds; Georgia had only ten. Neither of these two states can be described as being soft on crime.

Advocates of solitary confinement are left with a single argument for subjecting thousands of people to years of isolation: What else are we supposed to do? How else are we to deal with the violent, the disruptive, the prisoners who are just too dangerous to be housed with others?

As it happens, only a subset of prisoners currently locked away for long periods of isolation would be considered truly dangerous. Many are escapees or suspected gang members; many others are in solitary for nonviolent breaches of prison rules. Still, there are some highly dangerous and violent prisoners who pose a serious challenge to prison discipline and safety. In August, I met a man named Robert Felton, who had spent fourteen and a half years in isolation in the Illinois state correctional system. He is now thirty-six years old. He grew up in the predominantly black housing projects of Danville, Illinois, and had been a force of mayhem from the time he was a child.

His crimes were mainly impulsive, rather than planned. The first time he was arrested was at the age of eleven, when he and a relative broke into a house to steal some Atari video games. A year later, he was sent to state reform school after he and a friend broke into an abandoned building and made off with paint cans, irons, and other property that they hardly knew what to do with. In reform school, he got into fights and screamed obscenities at the staff. When the staff tried to discipline him by taking away his recreation or his television privileges, his behavior worsened. He tore a pillar out of the ceiling, a sink and mirrors off the wall, doors off their hinges. He was put in a special cell, stripped of nearly everything. When he began attacking counsellors, the authorities transferred him to the maximum-security juvenile facility at Joliet, where he continued to misbehave.

Felton wasn’t a sociopath. He made friends easily. He was close to his family, and missed them deeply. He took no pleasure in hurting others. Psychiatric evaluations turned up little more than attention-deficit disorder. But he had a terrible temper, a tendency to escalate rather than to defuse confrontations, and, by the time he was released, just before turning eighteen, he had achieved only a ninth-grade education.

Within months of returning home, he was arrested again. He had walked into a Danville sports bar and ordered a beer. The barman took his ten-dollar bill.

“Then he says, ‘Naw, man, you can’t get no beer. You’re underage,’ ” Felton recounts. “I says, ‘Well, give me my ten dollars back.’ He says, ‘You ain’t getting shit. Get the hell out of here.’ ”

Felton stood his ground. The bartender had a pocket knife on the counter. “And, when he went for it, I went for it,” Felton told me. “When I grabbed the knife first, I turned around and spinned on him. I said, ‘You think you’re gonna cut me, man? You gotta be fucked up.’ ”

The barman had put the ten-dollar bill in a Royal Crown bag behind the counter. Felton grabbed the bag and ran out the back door. He forgot his car keys on the counter, though. So he went back to get the keys—“the stupid keys,” he now says ruefully—and in the fight that ensued he left the barman severely injured and bleeding. The police caught Felton fleeing in his car. He was convicted of armed robbery, aggravated unlawful restraint, and aggravated battery, and served fifteen years in prison.

He was eventually sent to the Stateville Correctional Center, a maximum-security facility in Joliet. Inside the overflowing prison, he got into vicious fights over insults and the like. About three months into his term, during a shakedown following the murder of an inmate, prison officials turned up a makeshift knife in his cell. (He denies that it was his.) They gave him a year in isolation. He was a danger, and he had to be taught a lesson. But it was a lesson that he seemed incapable of learning.

Felton’s Stateville isolation cell had gray walls, a solid steel door, no window, no clock, and a light that was kept on twenty-four hours a day. As soon as he was shut in, he became claustrophobic and had a panic attack. Like Dellelo, Anderson, and McCain, he was soon pacing back and forth, talking to himself, studying the insects crawling around his cell, reliving past events from childhood, sleeping for as much as sixteen hours a day. But, unlike them, he lacked the inner resources to cope with his situation.

Many prisoners find survival in physical exercise, prayer, or plans for escape. Many carry out elaborate mental exercises, building entire houses in their heads, board by board, nail by nail, from the ground up, or memorizing team rosters for a baseball season. McCain recreated in his mind movies he’d seen. Anderson reconstructed complete novels from memory. Yuri Nosenko, a K.G.B. defector whom the C.I.A. wrongly accused of being a double agent and held for three years in total isolation (no reading material, no news, no human contact except with interrogators) in a closet-size concrete cell near Williamsburg, Virginia, made chess sets from threads and a calendar from lint (only to have them discovered and swept away).

But Felton would just yell, “Guard! Guard! Guard! Guard! Guard!,” or bang his cup on the toilet, for hours. He could spend whole days hallucinating that he was in another world, that he was a child at home in Danville, playing in the streets, having conversations with imaginary people. Small cruelties that others somehow bore in quiet fury—getting no meal tray, for example—sent him into a rage. Despite being restrained with handcuffs, ankle shackles, and a belly chain whenever he was taken out, he managed to assault the staff at least three times. He threw his food through the door slot. He set his cell on fire by tearing his mattress apart, wrapping the stuffing in a sheet, popping his light bulb, and using the exposed wires to set the whole thing ablaze. He did this so many times that the walls of his cell were black with soot.

After each offense, prison officials extended his sentence in isolation. Still, he wouldn’t stop. He began flooding his cell, by stuffing the door crack with socks, plugging the toilet, and flushing until the water was a couple of feet deep. Then he’d pull out the socks and the whole wing would flood with wastewater.

“Flooding the cell was the last option for me,” Felton told me. “It was when I had nothing else I could do. You know, they took everything out of my cell, and all I had left was toilet water. I’d sit there and I’d say, ‘Well, let me see what I can do with this toilet water.’ ”

Felton was not allowed out again for fourteen and a half years. He spent almost his entire prison term, from 1990 to 2005, in isolation. In March, 1998, he was among the first inmates to be moved to Tamms, a new, high-tech supermax facility in southern Illinois.

“At Tamms, man, it was like a lab,” he says. Contact even with guards was tightly reduced. Cutoff valves meant that he couldn’t flood his cell. He had little ability to force a response—negative or positive—from a human being. And, with that gone, he began to deteriorate further. He ceased showering, changing his clothes, brushing his teeth. His teeth rotted and ten had to be pulled. He began throwing his feces around his cell. He became psychotic.

It is unclear how many prisoners in solitary confinement become psychotic. Stuart Grassian, a Boston psychiatrist, has interviewed more than two hundred prisoners in solitary confinement. In one in-depth study, prepared for a legal challenge of prisoner-isolation practices, he concluded that about a third developed acute psychosis with hallucinations. The markers of vulnerability that he observed in his interviews were signs of cognitive dysfunction—a history of seizures, serious mental illness, mental retardation, illiteracy, or, as in Felton’s case, a diagnosis such as attention-deficit hyperactivity disorder, signalling difficulty with impulse control. In the prisoners Grassian saw, about a third had these vulnerabilities, and these were the prisoners whom solitary confinement had made psychotic. They were simply not cognitively equipped to endure it without mental breakdowns.

A psychiatrist tried giving Felton anti-psychotic medication. Mostly, it made him sleep—sometimes twenty-four hours at a stretch, he said. Twice he attempted suicide. The first time, he hanged himself in a noose made from a sheet. The second time, he took a single staple from a legal newspaper and managed to slash the radial artery in his left wrist with it. In both instances, he was taken to a local emergency room for a few hours, patched up, and sent back to prison.

Is there an alternative? Consider what other countries do. Britain, for example, has had its share of serial killers, homicidal rapists, and prisoners who have taken hostages and repeatedly assaulted staff. The British also fought a seemingly unending war in Northern Ireland, which brought them hundreds of Irish Republican Army prisoners committed to violent resistance. The authorities resorted to a harshly punitive approach to control, including, in the mid-seventies, extensive use of solitary confinement. But the violence in prisons remained unchanged, the costs were phenomenal (in the United States, they reach more than fifty thousand dollars a year per inmate), and the public outcry became intolerable. British authorities therefore looked for another approach.

Beginning in the nineteen-eighties, they gradually adopted a strategy that focussed on preventing prison violence rather than on delivering an ever more brutal series of punishments for it. The approach starts with the simple observation that prisoners who are unmanageable in one setting often behave perfectly reasonably in another. This suggested that violence might, to a critical extent, be a function of the conditions of incarceration. The British noticed that problem prisoners were usually people for whom avoiding humiliation and saving face were fundamental and instinctive. When conditions maximized humiliation and confrontation, every interaction escalated into a trial of strength. Violence became a predictable consequence.

So the British decided to give their most dangerous prisoners more control, rather than less. They reduced isolation and offered them opportunities for work, education, and special programming to increase social ties and skills. The prisoners were housed in small, stable units of fewer than ten people in individual cells, to avoid conditions of social chaos and unpredictability. In these reformed “Close Supervision Centres,” prisoners could receive mental-health treatment and earn rights for more exercise, more phone calls, “contact visits,” and even access to cooking facilities. They were allowed to air grievances. And the government set up an independent body of inspectors to track the results and enable adjustments based on the data.

The results have been impressive. The use of long-term isolation in England is now negligible. In all of England, there are now fewer prisoners in “extreme custody” than there are in the state of Maine. And the other countries of Europe have, with a similar focus on small units and violence prevention, achieved a similar outcome.

In this country, in June of 2006, a bipartisan national task force, the Commission on Safety and Abuse in America’s Prisons, released its recommendations after a yearlong investigation. It called for ending long-term isolation of prisoners. Beyond about ten days, the report noted, practically no benefits can be found and the harm is clear—not just for inmates but for the public as well. Most prisoners in long-term isolation are returned to society, after all. And evidence from a number of studies has shown that supermax conditions—in which prisoners have virtually no social interactions and are given no programmatic support—make it highly likely that they will commit more crimes when they are released. Instead, the report said, we should follow the preventive approaches used in European countries.

The recommendations went nowhere, of course. Whatever the evidence in its favor, people simply did not believe in the treatment.

I spoke to a state-prison commissioner who wished to remain unidentified. He was a veteran of the system, having been either a prison warden or a commissioner in several states across the country for more than twenty years. He has publicly defended the use of long-term isolation everywhere that he has worked. Nonetheless, he said, he would remove most prisoners from long-term isolation units if he could and provide programming for the mental illnesses that many of them have.

“Prolonged isolation is not going to serve anyone’s best interest,” he told me. He still thought that prisons needed the option of isolation. “A bad violation should, I think, land you there for about ninety days, but it should not go beyond that.”

He is apparently not alone among prison officials. Over the years, he has come to know commissioners in nearly every state in the country. “I believe that today you’ll probably find that two-thirds or three-fourths of the heads of correctional agencies will largely share the position that I articulated with you,” he said.

Commissioners are not powerless. They could eliminate prolonged isolation with the stroke of a pen. So, I asked, why haven’t they? He told me what happened when he tried to move just one prisoner out of isolation. Legislators called for him to be fired and threatened to withhold basic funding. Corrections officers called members of the crime victim’s family and told them that he’d gone soft on crime. Hostile stories appeared in the tabloids. It is pointless for commissioners to act unilaterally, he said, without a change in public opinion.

This past year, both the Republican and the Democratic Presidential candidates came out firmly for banning torture and closing the facility in Guantánamo Bay, where hundreds of prisoners have been held in years-long isolation. Neither Barack Obama nor John McCain, however, addressed the question of whether prolonged solitary confinement is torture. For a Presidential candidate, no less than for the prison commissioner, this would have been political suicide. The simple truth is that public sentiment in America is the reason that solitary confinement has exploded in this country, even as other Western nations have taken steps to reduce it. This is the dark side of American exceptionalism. With little concern or demurral, we have consigned tens of thousands of our own citizens to conditions that horrified our highest court a century ago. Our willingness to discard these standards for American prisoners made it easy to discard the Geneva Conventions prohibiting similar treatment of foreign prisoners of war, to the detriment of America’s moral stature in the world. In much the same way that a previous generation of Americans countenanced legalized segregation, ours has countenanced legalized torture. And there is no clearer manifestation of this than our routine use of solitary confinement—on our own people, in our own communities, in a supermax prison, for example, that is a thirty-minute drive from my door.

Robert Felton drifted in and out of acute psychosis for much of his solitary confinement. Eventually, however, he found an unexpected resource. One day, while he was at Tamms, he was given a new defense lawyer, and, whatever expertise this lawyer provided, the more important thing was genuine human contact. He visited regularly, and sent Felton books. Although some were rejected by the authorities and Felton was restricted to a few at a time, he devoured those he was permitted. “I liked political books,” he says. “ ‘From Beirut to Jerusalem,’ Winston Churchill, Noam Chomsky.”

That small amount of contact was a lifeline. Felton corresponded with the lawyer about what he was reading. The lawyer helped him get his G.E.D. and a paralegal certificate through a correspondence course, and he taught Felton how to advocate for himself. Felton began writing letters to politicians and prison officials explaining the misery of his situation, opposing supermax isolation, and asking for a chance to return to the general prison population. (The Illinois Department of Corrections would not comment on Felton’s case, but a spokesman stated that “Tamms houses the most disruptive, violent, and problematic inmates.”) Felton was persuasive enough that Senator Paul Simon, of Illinois, wrote him back and, one day, even visited him. Simon asked the director of the State Department of Corrections, Donald Snyder, Jr., to give consideration to Felton’s objections. But Snyder didn’t budge. If there was anyone whom Felton fantasized about taking revenge upon, it was Snyder. Felton continued to file request after request. But the answer was always no.

On July 12, 2005, at the age of thirty-three, Felton was finally released. He hadn’t socialized with another person since entering Tamms, at the age of twenty-five. Before his release, he was given one month in the general prison population to get used to people. It wasn’t enough. Upon returning to society, he found that he had trouble in crowds. At a party of well-wishers, the volume of social stimulation overwhelmed him and he panicked, headed for a bathroom, and locked himself in. He stayed at his mother’s house and kept mostly to himself.

For the first year, he had to wear an ankle bracelet and was allowed to leave home only for work. His first job was at a Papa John’s restaurant, delivering pizzas. He next found work at the Model Star Laundry Service, doing pressing. This was a steady job, and he began to settle down. He fell in love with a waitress named Brittany. They moved into a three-room house that her grandmother lent them, and got engaged. Brittany became pregnant.

This is not a story with a happy ending. Felton lost his job with the laundry service. He went to work for a tree-cutting business; a few months later, it went under. Meanwhile, he and Brittany had had a second child. She had found work as a certified nursing assistant, but her income wasn’t nearly enough. So he took a job forty miles away, at Plastipak, the plastics manufacturer, where he made seven-fifty an hour inspecting Gatorade bottles and Crisco containers as they came out of the stamping machines. Then his twenty-year-old Firebird died. The bus he had to take ran erratically, and he was fired for repeated tardiness.

When I visited Felton in Danville last August, he and Brittany were upbeat about their prospects. She was working extra shifts at a nursing home, and he was taking care of their children, ages one and two. He had also applied to a six-month training program for heating and air-conditioning technicians.

“I could make twenty dollars an hour after graduation,” he said.

“He’s a good man,” Brittany told me, taking his arm and giving him a kiss.

But he was out of work. They were chronically short of money. It was hard to be optimistic about Felton’s prospects. And, indeed, six weeks after we met, he was arrested for breaking into a car dealership and stealing a Dodge Charger. He pleaded guilty and, in January, began serving a seven-year sentence.

Before I left town—when there was still a glimmer of hope for him—we went out for lunch at his favorite place, a Mexican restaurant called La Potosina. Over enchiladas and Cokes, we talked about his family, Danville, the economy, and, of course, his time in prison. The strangest story had turned up in the news, he said. Donald Snyder, Jr., the state prison director who had refused to let him out of solitary confinement, had been arrested, convicted, and sentenced to two years in prison for taking fifty thousand dollars in payoffs from lobbyists.

“Two years in prison,” Felton marvelled. “He could end up right where I used to be.”

I asked him, “If he wrote to you, asking if you would release him from solitary, what would you do?”

Felton didn’t hesitate for a second. “If he wrote to me to let him out, I’d let him out,” he said.

This surprised me. I expected anger, vindictiveness, a desire for retribution. “You’d let him out?” I said.

“I’d let him out,” he said, and he put his fork down to make the point. “I wouldn’t wish solitary confinement on anybody. Not even him.” ♦

*Correction, April 6, 2009: Three per cent of the general population had difficulties with “irrational anger,” not three per cent of prisoners in the general population, as originally stated.