American psychiatrist Dorothy Lewis has spent 20 years examining violent criminals. Her work with murderers such as Ted Bundy, Mark Chapman and the racist killer Joseph Franklin has led her to conclude that violence is not a sin, but a symptom. It's the product of a damaged brain. Report by Malcolm Gladwell. Illustration by Toby Morison
On the morning of 18 November, Joseph Paul Franklin was led into courtroom 15 of the St Louis County courthouse in Clayton, Missouri. He was wearing a pair of black, high-top sneakers, an orange jump suit with short sleeves that showed off his prison biceps, and a pair of thick, black glasses. There were two guards behind him and two guards in front of him, and four more guards stationed around the courtroom, and, as he walked into the room - or rather shuffled, since his feet were manacled - Franklin turned to one of them and said, "Wassup" in a loud, Southern- accented voice. Then he sat down next to his attorney and stared straight ahead at the Judge, completely still except for his left leg, which he bounced up and down in an unceasing nervous motion.

Joseph Franklin, by his own admission, shot and paralysed Larry Flynt, the publisher of Hustler, outside a courthouse in Lawrenceville, Georgia, in August 1978, apparently because Flynt had printed photographs of a racially mixed couple. The following summer, he gunned down civil rights leader Vernon Jordan outside a Marriott hotel in Fort Wayne, Indiana, tearing a hole in Jordan's back the size of a fist. In the same four-year span in the late 1970s, as part of what he would later describe as a "mission" to rid America of blacks and Jews and whites who like blacks and Jews, Franklin also robbed several banks, bombed a synagogue in Tennessee, shot and killed an inter-racial couple walking down a Madison, Wisconsin street, killed a black man and a white woman coming out of a Pizza Hut in a suburb of Chattanooga, Tennessee, and is thought to have shot a black man in a Burger King in Falls Church, Virginia, as well as two 13- and 14-year- old boys as they walked across a railroad overpass in Cincinnati; and on and on, in a violent cross-country spree that spanned eight states, claimed close to 20 lives and, upon Franklin's arrest in 1980, earned him six consecutive life sentences.

Two years ago, while imprisoned in Marion Federal Penitentiary in Illinois, Franklin confessed to another crime. He was the one, he said, who lay in the bushes outside a synagogue in suburban St Louis in autumn 1977 and opened fire on a crowd of worshippers, killing 42-year-old Gerald Gordon. The state of Missouri then filed first-degree murder charges against him. He was moved from Marion to St Louis County jail, and from there, on a sunny November morning last autumn, he was brought before judge Robert Campbell of the St Louis County Circuit Court to determine whether he was fit to stand trial - to determine, in other words, whether embarking on a campaign to rid America of Jews and blacks is an act of evil or of illness.

The prosecution went first. On a television set at one side of the courtroom, they showed two videotapes, one of an interview with Franklin by a local news crew, and the other of Franklin's formal confession to St Louis police. In both, he is lucid and calm, patiently retracing how he planned and executed his attack on the synagogue.

He bought the gun, he said, in Dallas, answering a classified ad so the purchase couldn't be traced. He drove to the St Louis area and registered at a Holiday Inn. He looked through the Yellow Pages to find the names of synagogues. He filed the serial number off his rifle and bought a guitar case to carry it in. He bought a bicycle. He went to Radio Shack and bought a police scanner. He scouted out a spot near the synagogue where he could shoot without being seen. He parked his car in a nearby parking lot. He lay in wait in the bushes for several hours until there was a good-sized crowd. He fired five shots. He wiped the rifle clean, then dropped it on the ground. He rode the bicycle to the parking lot, climbed into his car, pulled out of the lot, checked the police scanner to see if he was being chased, then drove south, back toward Memphis.

In the interview with the news crew, Franklin answers every question, soberly and directly, in a flat, Southern-accented voice. He talks about his tattoos ("This one is the Grim Reaper. I got it in Dallas"), and respectfully disagrees with the media's description of racially motivated crimes as "hate crimes" since, as he points out, "every murder is committed out of hate". In his confession to the police, after he detailed every step of the synagogue attack, Franklin is asked if there is anything he'd like to say. He stares thoughtfully over the top of his glasses. There is a long silence. "I can't think of anything," he answers. Then he's asked if he feels any remorse. There is another silence. "I can't say that I do." He pauses again. "The only thing I'm sorry about is that it's not legal."

"What's not legal?"

Franklin answers as if he's just been asked the time of day: "Killing Jews."

After a break for lunch, the defence calls Dorothy Otnow Lewis, a psychiatrist at Bellevue Hospital in New York City and a professor at New York University Medical School. Over the past 20 years, Lewis has examined, by her own rough estimate, somewhere between 150 and 200 murderers. She examined, for the defence, Mark David Chapman, who shot John Lennon. Once, in a Florida prison, she sat for hours and talked with mass-murderer Ted Bundy. It was the day before his execution, and, when they had finished, Bundy bent down and kissed her cheek. "Bundy thought I was the only person who didn't want something from him," Lewis says.

Usually, Lewis works with Jonathan Pincus, a neurologist at Georgetown University in Washington. Lewis does the psychiatric examination, Pincus the neurological examination. But Franklin put his foot down. He could tolerate a Jewish woman on his defence team, but not a Jewish man. Lewis testified alone.

Lewis is a petite woman in her late fifties, with short, dark hair and large, liquid eyes. She was wearing a green blazer and a black skirt with a gold necklace, and she was so dwarfed by the witness stand that from the back of the courtroom only her head was visible. Under direct examination, she said that she had spoken with Franklin twice, once for five hours and once for an hour, and her conclusion was that he is a paranoid schizophrenic, a psychotic whose thinking is delusional and confused, a man wholly unfit to stand trial. She talked of the brutal physical abuse he had suffered as a child. She mentioned that he has scars on his scalp, suggestive of some kind of brain injury. She talked about his hallucinations, his obsession with being castrated, his grandiosity, his belief that he was once Jewish in an earlier life, his bizarre statements and beliefs.

At times, Lewis seemed nervous, her voice barely audible, but perhaps that was because Franklin was staring at her directly and unblinkingly, his leg bouncing faster and faster under the table. After an hour, Lewis stepped down. She paused in front of Franklin and, ever the psychiatrist, suggested that, when everything was over, they should talk. Then she walked slowly through the courtroom, past the defence table and the guards, and out of the door.

Later, on the plane home to New York City, Lewis worried that she had come across as some kind of caricature: liberal, Jewish, New York City psychiatrist comes to Middle America to tell the locals to feel sorry for a murderer. But she insisted that she wasn't making an ideological point of Franklin. She was saying that she did not feel Franklin's brain works the way brains are supposed to work, that he had identifiable biological and psychiatric problems that diminished his responsibility for his actions.

"I just don't believe people are born evil," she said. "To my mind, that is mindless. Forensic psychiatrists tend to buy into the notion of evil. But I felt that that's no explanation. The deed itself is bizarre, grotesque. But it's not evil. To my mind, evil bespeaks conscious control over something. Serial murderers are not in that category. They are driven by forces beyond their control."

The plane was in the air now. By some happy set of circumstances, Lewis had been bumped up to first class. She was sipping champagne. Her shoes were off. "You know, when I went to see him, he shuffled in and sniffed me right here" - she touched the nape of her neck and flared her nostrils in mimicry of Franklin's gesture. "He said to his attorney, `you know, if you weren't here, I'd make a play for her'." She had talked to him for five hours and he had come on to her, just like that, this guy who hated Jews so much he lay in the bushes and shot at them with a rifle. She shivered at the memory. "He said he wanted some pussy."

When Dorothy Lewis graduated from Yale Medical School in 1963, neurology, the study of the brain, and psychiatry, the study of behaviour and personality, were entirely separate fields. This was still the Freudian era. There was little attempt made to search for organic causes of criminality. When she left medical school, and began working with juvenile delinquents in New Haven, Connecticut, the theory was that these boys were robust, healthy: a delinquent was simply an ordinary kid who had been led astray by a troubled home life, by parents who were too irresponsible or addicted to drugs and alcohol to provide proper discipline.

Lewis accepted this dogma. The problem was that, when she began working with delinquents, they didn't seem like that at all. They didn't lack for discipline. If anything, she felt, they were being disciplined too much. And they weren't robust, rowdy teens. They seemed to be damaged and impaired.

"I was studying for my boards in psychiatry, and in order to do a good job you wanted to do a careful medical history and a careful mental status exam. I discovered that many of these kids had had serious accidents, injuries or illnesses that seemed to have affected the central nervous system and that hadn't been identified previously."

In 1976, she was given a grant by the state of Connecticut to study juvenile delinquents. Immediately, she went to see Pincus, then a young professor of neurology at Yale. They had worked together once before. "Dorothy came along and said she wanted to do this project with me," Pincus says. "She wanted to look at violence. She had this hunch that there was something physically wrong with these kids. I said, `That's ridiculous. Everyone knows violence has nothing to do with the brain'."

But Lewis would have none of it. "She said, `We should do it, anyway'. I said, `I don't have the time'. She said, `Jonathan, I can pay you'. So I would go up on Sunday, and I would examine three or four youths, just give them a standard neurological examination." It was at that point, after he had seen the kids for himself, that Pincus, too, became convinced that the prevailing wisdom about juvenile delinquents - and by extension about adult criminals as well - was wrong. Lewis was right. "Almost all the violent ones," Pincus remembers, shaking his head, "were damaged."

Over the past 20 years, Lewis and Pincus have testified for the defence in dozens of criminal cases, usually in an attempt to save a murderer from the death penalty. Together, they have published a series of studies on murderers and delinquents, painstakingly outlining the medical and psychiatric histories of the very violent. Lewis and Pincus believe that child abuse and neurological impairment - the two conditions that, for example, appear to surface in the history of Joseph Franklin - are experiences central to understanding the most dangerous and violent of criminals. Criminals, they believe, aren't people who have merely been abused as children, since lots of people who have been abused never end up harming anyone else. Nor are they simply people with brain injuries, since most people with brain injuries aren't criminals, either. They believe that criminals are, overwhelmingly, people with both problems - with abusive childhoods and brain injuries, that somehow these factors in combination create such terrifying synergy that normal neurological functioning is profoundly altered.

The virtue of this theory is that it sidesteps all the topics that so cripple contemporary discussions of violence - genetics, biological determinism and, of course, race. It's really a return to the old liberal idea that environment counts and that it is possible to do something significant about crime by changing the material conditions of people's lives. Only this time, the maddening imprecision of the old idea (what exactly was it about, say, bad housing that supposedly led to violent crime?) is gone. Lewis and Pincus and the neurologists and psychiatrists working in the same field think they can actually pinpoint what it is that helps to turn some people into criminals - right down to the parts of the brain that are warped and disabled by abuse and injury.

Lewis works out of a tiny set of windowless offices on the 21st floor of the new wing of Bellevue Hospital on Manhattan's East Side. The rooms are decorated in institutional colours - grey carpeting and bright orange trim - and her desk is stacked high with boxes of medical and court records from the cases she has worked on, as well as dozens of videotapes of interviews with murderers she has conducted over the years. She talks about some of her old cases - especially some of her Death Row patients - as if they had just happened, going over and over details, worrying, sometimes, if she made the absolutely correct diagnosis.

The fact that everyone else has given up on these men (and they are almost always men) seems to be just what continues to attract her. Years ago, when Lewis was in college, she found herself sitting next to the well-known Harvard theologian, Paul Tillich, on the train from New York to Cambridge. "When you read about witches being burned at the stake," Tillich asked her, in the midst of a long and wide-ranging conversation, "do you identify with the witch or the people looking on?" Tillich said he himself identified with the crowd. Not Lewis. She identified with the witch.

The cortex is the thick carpet of grey matter on the upper part of every human brain, and the function of the cortex - and, in particular, those parts of the cortex beneath the forehead known as the frontal lobes - is to modulate the impulses that surge up from within the brain. The cortex and the frontal lobes are supposed to provide judgement, to organise behaviour and decision-making, to learn and adhere to rules of everyday life. In other words, they are responsible for making us human, and it is the central argument of the criminality school to which Lewis and Pincus belong that what distinguishes many violent criminals from the rest of us is that something has happened inside their brain to throw the functioning of the cortex and the frontal lobes out of whack. One of the first things that Lewis and Pincus do, then, when they evaluate a murderer, is to check for signs of frontal lobe impairment. This, the neurological exam, is Pincus's territory.

He begins by taking a medical history. He asks about car accidents and falls from trees and sports injuries and physical abuse and problems at birth and any kind of blows to the head that might have caused damage to the frontal lobes. He tests reflexes, and asks about headaches and limps and whether the left side is the same as their right side.

"I measure the head circumference. If it's more than two standard deviations below the normal brain circumference, there may be some degree of mental retardation, and, if it's more than two standard deviations above, there may be hydrocephalus," Pincus told me. "I also check gross motor co-ordination - ask them to spread their fingers and hold their hands apart, and I look for choreiform movement, discontinuous little jerky movements of the fingers and arms."

We were in Pincus's small, cluttered office at Georgetown University Medical Center in Washington, DC and Pincus, properly professional in a grey check suit and a dark-red waistcoat, held out his hand to demonstrate. "Then I ask them to skip, to hop," he went on, jumping out of his chair to demonstrate, hopping up and down in a small space on the floor between papers and books.

Pincus is well over six-feet, with the long-limbed gracefulness of an athlete, and he plays the part of neurologist to perfection: calm, in command, with a distinguished sprinkle of white hair. At the same time, though, he has a look of mischief in his eyes, a streak of irreverence that allows him to jump up and down in his office before perfect strangers. It's an odd combination, like Sigmund Freud being played by Walter Matthau.

At one point, Pincus held up a finger 45 degrees to my left and moved it slowly to the right. "Now we're checking for frontal functions. A person should be able to look at the examiner's finger and follow it smoothly with their eyes. If they can only follow it jerkily, the frontal eye fields are not working properly. Then there's upward gaze." He asked me to point my eyes at the ceiling. "The eye should go up five millimetres and a person should also be able to direct his gaze laterally and maintain it for 20 seconds. If he can't, that's motor impersistence." Ideally, Pincus will try and back these tests up with EEG, which measures electrical patterns in the brain or an MRI scan, to see if he can spot scarring or lesions in any of the frontal regions that might contribute to impairment.

Pincus is also interested in measuring judgement. He tries to pick up evidence of the inability to cope with complexity, the lack of connection between experience and decision-making so characteristic of cortical dysfunction. At this point, Pincus walked behind me, reached over the top of my head, and tapped in a steady rhythm on the bridge of my nose. I blinked once, then stopped. That, he tells me, is normal.

"When you tap somebody on the bridge of the nose, it's reasonable for a person to blink a couple of times, because there is a threat from the outside," Pincus says. "When it's clear there is no threat, they should be able to accommodate that. But, if the subject blinks more than three times, that's `insufficiency of suppression', which may show frontal lobe dysfunction. The inability to accommodate means you can't adapt to a new situation. There's a certain rigidity there."

Arthur Shawcross, the Rochester serial killer, who had a cyst in one temporal lobe and scarring on both frontal lobes (from, among other things, being hit on the head with a sledgehammer and a discus and falling on his head from the top of a 40-foot ladder) used to walk in absolutely straight lines, splashing through puddles instead of walking around them, and tearing his trousers on a barbed wire fence rather than using a gate a few feet away.

That's the kind of behaviour Pincus is trying to pick up. "I give them a word-fluency test. I ask them to name as many words that they can think of that begin with the word F. Normal is 14, plus or minus five. Anyone who does less than nine is abnormal."

This is not an intelligence test. People with frontal lobe damage might do just as well as anyone else if asked, say, to list the kinds of products they might buy in a supermarket. Under these rules, "most people can think of at least 16 products in a minute and rattle them off", Pincus says. But that's a structured test, with familiar objects. The thing that people with frontal lobe damage can't do is cope with situations where there are no rules, where they have to improvise, where they make unfamiliar associations.

"Very often they get stuck on one word - they'll say `four', `fourteen', `forty-four'. They'll use the same word again and again. `Farm' and then `farming'. Or, as one fellow in a prison once said to me, `fuck', `fucker', `fucking'. They don't have the ability to come up with something else."

Pincus has found that when he examines murderers, these kinds of neurological problems come up again and again. Recently, Lewis and Pincus compiled a list of all the verifiable brain injuries suffered by 15 randomly selected Death Row inmates. Here are the injuries listed for the first two murderers in the study:

1. Three years: beaten almost to death by father (multiple facial scars); early childhood: thrown into sink onto head (palpable scar); late adolescence: one episode of loss of consciousness while boxing.

2. Childhood: beaten on head with two-by-fours by parents; childhood: fell into pit, unconscious for several hours; 17 years: car accident with injury to right eye; 18 years: fell from roof, apparently because of a blackout.

After Pincus is finished, Lewis looks for evidence of child abuse. She does this because the second, critical argument of the Lewis-Pincus hypothesis is that child abuse causes profound and pathological changes in the structure of the brain as surely as injury does. Bruce Perry, a psychiatrist at Baylor College of Medicine, for example, has done brain scans of children who have been severely neglected and found that entire structures of their cortex never properly developed; as a result, these cortical regions were something like 20 or 30 per cent smaller than normal. According to Perry, in someone abused or neglected, the section of the brain involved in attachment, in making emotional bonds, would actually look different. The wiring wouldn't be as dense or as complex. "It's like emotional retardation. They are literally lacking some brain organisation that allows them to actually make strong connections to other human beings."

Abuse also disrupts the brain's stress response system - with profound results. When something traumatic happens - a car accident, a fight, a piece of shocking news - the brain responds by releasing several waves of hormones, the last of which, cortisol, is supposed to bring everything back to normal. The problem is that cortisol is toxic. If someone is exposed to too much stress over too long a time, all that cortisol begins to eat away at the organs of the brain known as the hippocampus, which serve as the brain's archivists - organising and shaping memories, putting them in context, placing them in space and time, tying together visual memory with sound and smell. J Douglas Bremner, a psychiatrist at Yale, has actually measured this damage by taking brain scans of adults who suffered severe sexual abuse as children and comparing them to the brains of healthy adults. In those who had been abused, Bremner found, the hippocampus was on average 12 per cent smaller.

Abuse also seems to affect the relationship between the left hemisphere of the brain, which plays a large role in logic and language, and the right hemisphere, which is thought to play a disproportionate role in creativity and depression. Martin Teicher, a psychiatrist at Harvard University, recently gave an EEG - the scan that measures electrical activity in the brain - to 115 children admitted to a psychiatric facility with a history of some kind of abuse. Not only did the rate of abnormal EEGs among the group turn out to be twice as high as a non-abused control group, but in every case the abnormality was entirely on the left.

Teicher argues that in some of his EEG and MRI analysis of the left and right hemispheric imbalance, what he is really describing is the neurological basis for the kind of polarisation so often found in psychiatrically disturbed patients, the mood swings, the sharply contrasting temperaments. Instead of having two integrated hemispheres, these patients have brains, in some sense, divided down the middle. "What you get is a kind of erraticness," says Frank Putnam, who heads the Developmental Trauma Unit at the National Institute of Mental Health in Maryland. "These kinds of people can be very different in one situation compared with another. There is the sense that they don't have a larger moral compass."

In one of the classic studies in the field of child abuse, Mary Main, a psychologist at UC Berkeley, observed a group of 20 toddlers over three months, half of whom had been subjected to serious physical abuse and half of whom had not. Main was interested in how the toddlers responded to a classmate in distress. What she found is that the healthy almost always responded to a crying or unhappy peer with concern and sadness, or, alternately, showed interest and made some attempt to provide comfort. But the abused toddlers never showed any concern. At the most, they showed interest. The majority either got fearful and distressed themselves, or lashed out with threats, anger and physical assaults.

Here is Main's description of "Martin", an abused boy of 32 months, who seems entirely blind - emotionally retarded in the way that Perry describes - to what it takes to relate to another human being:

"Martin... tried to take the hand of the crying other child, and, when she resisted, he slapped her on the arm with his open hand. He then turned away from her to look at the ground and began vocalising very strongly, `Cut it out! CUT IT OUT!,' each time saying it a little faster and louder. He patted her, but when she became disturbed by his patting, he retreated, `hissing at her and baring his teeth'. He then began patting her on the back again, his patting became beating, and he continued beating her despite her screams."

This does not mean that abuse, in and of itself, results in violence. There are plenty of abused children who never become violent. The same is true for neurological impairment. It creates a vulnerability to certain kinds of anger or to social situations that might end in violence. But it's hardly an explanation for violence. What Lewis and Pincus argue, however, is that what matters is neurological impairment and child abuse, that the layering of frontal lobe lesions on top of the damage wrought by cortisol somehow makes everything disproportionately worse.

Several years ago, Lewis did a follow-up study of the 95 juveniles she and Pincus had first worked with in Connecticut in the late 1970s. She broke the group down into several groups: Group One, consisting of those who had neither psychiatric nor neurological vulnerabilities nor an abusive childhood; Group Two, consisting of those with vulnerabilities but no abuse at home; Group Three, consisting of those with abuse but no vulnerabilities; and Group Four, consisting of those with extensive vulnerabilities and abuse.

Seven years later, as adults, those in Group One had subsequently been convicted of an average of just over two criminal offences, none of which were violent, resulting in essentially no jail time. Group Two, the neurologically impaired kids, had been convicted of an average of 10 offences, two of which were violent, resulting in just under a year of jail time. Group Three, the abused kids, had 11.9 offences, 2.1 violent, and 562 days in jail. But Group Four - those with both - were in another league entirely. In the intervening seven years, they have been convicted of, on average, 16.8 crimes, 5.4 of which were violent, resulting in a total of 1,214 days in prison.

These statistics have broad implications for how we think about criminality. What if, for example, as a result of his brutal childhood, Joseph Franklin has a lesion on his frontal lobes, an atrophied hippocampus, a maldeveloped left hemisphere, a lack of synaptic complexity in the pre-cortical limbic area, and a profound left-right hemisphere split? What if his remorselessness was just the grown-up version of the little boy, Martin, whose ability to understand and relate to others was so retarded that he kept on hitting and hitting, even after the screams began? He might still be sane, in the strict legal definition. But that kind of medical diagnosis suggests that his ability to live by the rules of civilised society, and to appreciate and act on the distinctions between right and wrong, is quite different from someone with a happy childhood and a normal brain.

The organic perspective on violence promises enormous insights about how to rehabilitate prisoners and prevent violence and restore to health those turned into criminals by no fault of their own. But there is no doubt that there is something a little unsettling about using medical evidence to explain certain kinds of criminal behaviour. It may be a more flexible and sophisticated way of looking at criminality, but something is clearly lost in the translation. The moral force of the old standard, after all, was in its inflexibility and lack of sophistication. Murder was murder, and the allowances made for aggravated circumstances were kept to a minimum. Is a moral standard still a moral standard when it is freighted with exceptions and exemptions and physiological equivocation?

When Lewis went to see Bundy in Starke, Florida, on the day before his execution, she asked him why he had allowed her - out of all the people lining up outside his door - to see him. He answered, "Because everyone else I've talked with these past days wants to know what I did. You are the only one who wants to know why I did it." It's impossible to know what the supremely manipulative Bundy meant by this: whether he genuinely appreciated Lewis, or whether he simply regarded her as his last conquest. What's clear is that, over the handful of times they met in Bundy's last months, the two reached a kind of understanding, and by the end she sensed a certain breakthrough.

"The day before he was executed, he asked me to turn off the tape recorder. He said he wanted to tell me things that he didn't want recorded, so I didn't. It was very confidential." To this day, Lewis has never told anyone what Bundy said. There is something almost admirable about this. But there is also something strange about extending the physician's privilege to a killer like Bundy, about turning the murderer so completely into a patient. It is not that the premise is false, that murderers can't also be patients. It's just that once you make that leap, once you make the criminal into an object of medical scrutiny, the crime itself inevitably gets pushed aside and normalised.

The difference between a crime of evil and a crime of illness is the difference between a sin and a symptom. And symptoms do not intrude in the relationship between the murderer and the rest of us: they don't force us to stop and observe the distinctions between right and wrong, between the speakable and the unspeakable, the way sins do. It was at the end of that final conversation that Bundy reached down and kissed Lewis on the cheek. But this is not all that happened. Lewis then reached up, put her arms around Bundy's neck, and kissed him back

Malcolm Gladwell. A longer version of this article first appeared in the `New Yorker'