Is the Criminal Law Based on One Huge Mistake?

Punishment and Violence: Is the Criminal Law Based on One Huge Mistake?

James Gilligan

FOR the past three millennia, since the time of the first law-givers–Hammurabi and Moses, Drakon and Solon, Plato and Aristotle, Cicero and Justinian–humanity has been engaged in a massive exercise in social research. We have been conducting a great social experiment to test the hypothesis that we could prevent violence by defining it as a crime (or war crime), and then punishing those who commit it with more violence of our own (which we define as justice). Three thousand years is long enough to test any hypothesis, and the results of this experiment have been in for a long time now. This approach to violence, which I will call the moral and legal approach, far from solving the problem of violence, or even diminishing the threat that it poses to our continued survival, has, instead, been followed by a continuing and ever-accelerating escalation of the scale of human violence–to the point that the century we have just survived has been the bloodiest in all of human history. Worse yet, we have now achieved, by deliberate effort, the technological ability to kill everyone on earth, thus becoming the first species in evolutionary history to be in danger of bringing about its own extinction–unless we increase our ability to prevent violence far more effectively than we have by means of the ways of thinking, and the strategies based on them, that we have employed for the past three millennia.

Nietzsche said that the history of the world is the ultimate refutation of the notion that there is moral order in the universe. I am simply paraphrasing him by observing that history is the ultimate refutation of the theory that punishment will prevent or deter violence. On the contrary, punishment is the most powerful stimulus of violence that we have yet discovered. In order to understand why this is true, we will need to understand the psychology of punishment. Because the etymology of words is one of the royal roads to understanding the collective or cultural unconscious, it is relevant here to note that the etymology of “punishment” tells us what its underlying meaning is. The word derives from the Greek poine, and its Latin derivative, poena, which mean revenge. Indeed, in its capitalized form, Poine was the Greek goddess of revenge. Poine and poena are also the roots of our word “pain,” as well as of penalty, penal (system), penitentiary, and penance. Hence, punishment is the deliberate infliction of pain on a person for the sake of attaining revenge. And penitentiaries, or prisons, are institutions whose purpose is to inflict pain on people for the sake of revenge (a task at which they are all too successful). But as one of the children whom Piaget interviewed for his studies of moral development recognized, the trouble with revenge is that it is endless; the moment one person gets revenge, the person on whom revenge was taken is motivated to return the favor, resulting in an endless vicious cycle (1932).

That may help us to understand why the moral and legal approach to preventing violence has been so unsuccessful–in fact, not merely unsuccessful, but counter-productive, leading to more violence than there was before the moral and legal approach was invented. On the other hand, if that approach has not worked, what reason do we have to think that any other approach would be more successful? One point of this article will be to suggest some answers to that question.

The main cognitive handicap that we impose on ourselves by defining violence as a moral and legal problem is that that point of view is incapable of informing us as to what causes violence and how we could prevent it. The only questions the moral and legal way of thinking can ask are: “How evil (or heroic) was this particular act of violence, and how much punishment (or reward) does the perpetrator of it deserve?” But even if it were possible to gain the knowledge that would be necessary to answer those questions (which it is not), the answers would not help us in the least to understand either what causes violence or how we could prevent it. Those are empirical questions, not moral or legal ones. It is only by approaching violence from the point of view of empirical disciplines, as a problem in public health and preventive medicine, including social and preventive psychiatry and psychology, that we can acquire answers to those questions–by engaging in clinical, experimental and epidemiological research on violence.

The main assumption on which the moral and legal approach has been based up to now is that punishment will deter, inhibit, or prevent violence. But that is an empirical hypothesis, not a moral or legal one. There are several different types of social and psychological research that have provided data regarding this question. The first type I will review is research that I have been conducting over the past thirty years with violent people (both criminals and psychiatric patients) in prisons, jails, youth detention facilities, and prison mental hospitals for the “criminally insane,” in the course of directing or evaluating mental health services focused on preventing violence (both homicide and suicide) in those institutions and in the community following the patients’ release.

From the time I first began working clinically with violent people in a prison mental health service, I realized that I had never met a group of people who had been punished as severely, from so early on, as this group had. The first prison inmate whom I ever saw in psychotherapy, a man who had committed several brutal muggings and armed robberies, described how his father had beaten everyone in his family–himself, his mother, his siblings. I thought at first that he was merely a “con artist” who was trying to enlist my sympathy and give himself an excuse for his antisocial behavior. Then I discovered that his father was also in prison–for the crime of murdering his own daughter (my patient’s sister). So I learned that whether or not he was untrustworthy in many other respects, on this subject he was not merely telling the truth, he was describing something that had been proven “beyond a reasonable doubt” in a court of law.

Another prisoner’s body was covered by scars from scalding water his mother had thrown on him repeatedly during his childhood to discipline him. A third inmate’s parents had punished him by locking him into an empty icebox for hours; although he did not die, he was there long enough that he suffered brain damage from anoxia. Another man, who had committed a terrible rape and murder, came into my office in the prison mental health service after he was convicted for his crime, and I noticed that he had small scars, about the size of a dime, on his wrists and ankles. I asked him what had caused them, and he said, “That is where my mother shot me.” I asked him what he meant, and he described how his mother, whenever she wanted to punish him, would not merely spank him. She would take out her pistol and shoot him–not where it would kill him, but where it would “teach him a lesson.” And of course it did teach him a lesson–that that is how you treat other people. (Not surprisingly, this student, like many students, surpassed his teacher; he did go on to kill those whom he was punishing. Perhaps he was also influenced by the fact that he had seen his father killed in front of his eyes by two relatives … thus learning that that is how to treat people when you become an adult.)

Another inmate was a multiple murderer who had killed several people in his home town in a southwestern state, and then continued to kill people–his fellow inmates–after a race war erupted in the prison to which he had been sentenced. At that point he was transferred to the Massachusetts prison system, in exchange for a prisoner whom Massachusetts wanted out of its prison system, and entered the mental health program I directed there (following which he has remained completely non-violent). In exploring the roots of his violence, he described how as a child his mother had repeatedly assaulted him by throwing him out the window, attacking him in his sleep with an axe, setting him on fire, and on and on. After recounting this, he stated, more bemusedly than with anger, “I guess she wanted to kill me–but I just didn’t die.” But of course what he learned from these experiences was to kill other people.

In short, what I learned from decades of clinical experience with the most violent people our society produces is that many of those who murder others are survivors of their own attempted murder, or of the murders of their closest relatives; their fathers, mothers, sisters or brothers, whose murders they often witnessed. If punishment did inhibit or prevent violence, then these men would not have become violent in the first place, for they had already experienced the most severe punishments that it is possible to inflict on people without actually killing them.

But it was not only their past or childhood history that confirmed this relationship between punishment and violence. I saw the same relationship exemplified and acted out in the prisons on a daily basis: the more severely prisoners were punished by the prison authorities, the more violent they became, and the more violent they became, the more severely they were punished. One of the more extreme punishments the prison officers would inflict on a prisoner was to place him in a solitary confinement unit. They would turn off his light, remove his mattress, back up his toilet (which was a hole in the floor), so that he would sleep on concrete, surrounded by his own excrement and the vermin that are naturally attracted to such an environment. They would then, in effect, bury him alive by closing a solid steel door so that he was in complete darkness and silence. This punishment could go on for months, even years. But clearly it did not succeed in preventing or deterring the prisoners’ violence, or else the unit would not have been perpetually filled. What happened instead was that the prisoners I observed became so enraged and bitter that they did not care whether they themselves lived or died, if only they could get back at their tormentors, or at any other target on whom they could vent their rage.

I will give two examples. One inmate, who was in prison for having killed two people in the community, had been in solitary confinement for two years, mainly because every time the officers opened his door he lunged out to attack them. When he finally hurt one of the officers badly enough, they decided to have him charged with assault and battery in an outside court, at which time his lawyer asked me to conduct a psychiatric evaluation of him. Upon examining him it became clear that he was hallucinating, delusional, and showing other psychotic symptoms of which he had no prior history, presumably precipitated by the conditions of sensory deprivation and social isolation in which he had been confined. (This phenomenon has been observed so frequently and for so long in prison environments that there is a slang term for it: “stir-crazy.” It has been known since the 1950s that even mentally normal research subjects will sometimes start hallucinating after relatively brief periods of complete sensory deprivation.) After the court agreed with me that he was not guilty by reason of insanity, he was transferred to the prison mental hospital I directed, where he lived in an ordinary residential unit with other patients and engaged in individual and group psychotherapy. His psychotic symptoms resolved, he remained completely non-violent, and he even volunteered for useful work that made him a constructive addition to the hospital community. We learned that when he had the opportunity to express himself through words rather than violent actions, and was offered treatment rather than punishment, his violence disappeared.

The other example ended tragically rather than constructively. A young man who had been sentenced to a minimum-security prison for a non-violent crime (breaking and entering) and had no history of violence was, nevertheless, obnoxious and rebellious toward the prison authorities. The guards knew only one way to respond to his constant breaking of petty rules and regulations–to punish him more and more severely in the hope that when he was punished enough he would finally obey the rules. Exactly the opposite happened, however, and he was placed in progressively more punitive environments until finally he spent the last two years of his sentence in solitary confinement under the conditions I have described above. Since the prison authorities had placed him under maximum-security confinement, the parole board would not parole him. He eventually served out his entire sentence, with the result that he not only spent more years in prison than he otherwise would have, he also was given no support in the community upon his release (such as a parole officer, a halfway house, employment counseling, etc.). Instead, his solid steel door was opened and he walked, blinking into the sunlight, out the front door of the prison with nothing but the clothes on his back and nowhere to go. Within two days of his release he killed one of the first people he came upon while hitch-hiking, and attempted to kill another–two college students who had offered him a ride. Clearly, punishment had not rendered this man less violent, nor had it increased the safety of the public. In fact, it would seem to have had exactly the opposite effect. (As one of the older and wiser correction officers said to me afterwards, “You can lock a dog in a closet for a month, but I don’t want to be the one who’s standing there when you let him out.”)

What about the most extreme and irrevocable form of punishment: capital punishment? Does that prevent or deter violent crime? Of course it prevents the man who is executed from committing further murders. But that is not the relevant consideration from the perspective of public health and preventive medicine. The relevant question is: what effect does capital punishment have on the incidence of violence–especially lethal violence–in society as a whole? For if it stimulates the public in general to commit more violence, then the preventive effect of executing people who have already committed murder is outweighed by its effect in stimulating more murders by those who have not yet done so.

The National Academy of Sciences, in a review of this question, concluded that there is no conclusive evidence that capital punishment deters or prevents violent crime (Silberman, 1978). It is both empirically impossible to control all the variables that are relevant to the level of violence in a society, and morally unacceptable to engage in the kinds of controlled experiments that would be necessary to test the hypothesis that capital punishment actually increases the overall level of violence. However, there is a great deal of evidence that is at least consistent with the conclusion that capital punishment is more likely to stimulate violence than to prevent it. For example, the main risk suffered by the crowds who attended the public hangings of pickpockets in early modern England was having their pockets picked by the hordes of pick-pockets who were not inhibited in the least from plying their trade, or were even stimulated to do so, by the horrible punishment of it that was being carried out right in front of their eyes! Indeed, this paradoxical phenomenon became so notorious that it became one of the main arguments in favor of eliminating public executions in England, and eventually even private ones. The counter-productive nature of punishment was simply too obvious to be ignored forever–at least, in England. (Apparently we in America do not learn from experience quite as quickly as the English; or perhaps it is just that we are a younger country, and therefore have not had as long an experience of capital punishment as the British have.)

The United States is the only Western democracy that still practices capital punishment. If it were truly effective in deterring or preventing murder, one would expect that we would have a much lower murder rate than those other countries. But instead the opposite is true–our murder rate is five to ten times higher than those of every other such nation. Even within the United States, the murder rate is almost twice as high in the states with capital punishment as it is in those without it. And when capital punishment began to be used again in 1976, the states that used it most frequently experienced a large increase in their murder rates. Those states that used it less often experienced a smaller increase; and those that did not return to capital punishment at all experienced decreased murder rates. In fact, there are far more data consistent with the conclusion that capital punishment stimulates violence, and inconsistent with the conclusion that it prevents it, than there are for the opposite conclusions. How can we understand why capital punishment “backfires” in this way? I will begin by examining the psychology of homicide–both in its legal form (which is called capital punishment), and in its illegal form (which is called murder).

To begin with, it has been known for many years that many more murderers kill themselves than were ever killed by the state, even when death was the usual punishment for murder. In fact, the suicide rate of murderers, in every country in which these data have been collected, is from several hundred to several thousand times as high as in the general population (Wolfgang, 1958). These people feel so desperate and destroyed that they already feel dead inside; psychologically, spiritually, and even physically, so that they feel like the walking dead (as the most violent men in our prisons have told me they do). They do not care whether they themselves live or die, in the conventional (physical) sense of those words, since they already feel psychologically dead. They frequently decide they would rather be physically dead as well, since they feel psychologically dead. The notion that such men would be deterred from committing murder or other violent crimes by the death penalty is based on complete and utter ignorance of the psychology of violent people. If anything, they welcome death, and either provoke it or actively seek it. For example, the death penalty in America was reactivated in 1976, after a moratorium of several years, when one murderer, Gary Gilmore, actively sought to persuade the state to execute him and refused to cooperate with his lawyer’s attempts to save his life. Capital punishment for him, as for many murderers, was simply a vicarious form of suicide. Imagining that executing such people will prevent or deter violence is as naive a fantasy as believing that threatening the terrorists who act as suicide bombers, or the Kamikaze pilots of World War II, would deter them from committing their homicides.

Another example: the explosive six-fold increase in the homicide rate among American teen-agers over the forty years between 1955 and 1994 was paralleled by a five-fold increase in suicide among the same group during the thirty-five years from 1956 to 1990. In other words, whatever the social and historic forces are that cause an increase in homicide among our young men, they also seem to cause an almost identical increase in suicide among the same group. Perhaps the only lesson we can draw from this is that when people are desperate enough to kill someone else, they are just as likely to be desperate enough to kill themselves (as well, or instead, or by proxy–that is, by provoking the state to do it).

Why is capital punishment so counter-productive, if our goal is to prevent violence? One of the best explanations can be discovered by applying the insight of an opinion by the late Supreme Court Justice Brandeis, who wrote that

Our government is the potent, the omnipresent teacher. For good or ill, it

teaches the whole people by its example. Crime is contagious. If the

government becomes a lawbreaker, it breeds contempt for laws; it invites

every man to become a law unto himself; it invites anarchy….

To paraphrase the last two sentences: Violence is contagious. If the government commits cold-blooded murder, i.e., capital punishment, it breeds contempt for life; it invites every man to become a law unto himself (a murderer); it invites anarchy (unlimited violence). Violence is as contagious as many other deadly diseases because it can be taught so easily. It differs from the others only in that the vector of transmission is not microorganisms but teaching by example, whether by means of childrearing, the criminal justice system and the wider legal and political system of which it is a part, or any other means of acculturation and socialization. The most powerful teaching tool–compared with which words hardly matter–is example. Or, more pointedly, behavior.

What about imprisonment? Does that prevent or deter violence? Studying the effect of imprisonment on rates of violence is complicated, because prisons perform a variety of functions. Clearly, disarming someone and removing him from the community to a locked residential facility (whether it be a prison or a hospital) can restrain him from committing violence (whether toward himself or others). To the all-too-limited extent to which prisons simply restrain people without punishing them, treat them with respect rather than contempt, and make available to them the tools that can enable them to gain sufficient self-respect to outgrow their need to commit violent acts (such as education, psychotherapy, employment, treatment for alcoholism, and so on), prisons could (and there are some fortunate exceptions that do) actually prevent violence.

However, that is clearly not a remotely realistic description of the vast majority of American prisons. Traditionally, the purpose of imprisonment has included not only restraint (or incapacitation) and rehabilitation, as in the model I just summarized, but also punishment. Punishment in this case is the deliberate infliction of pain on the prisoner (beyond the pain that is unavoidable as a result of depriving him of his liberty). It is pain that is unnecessary to inflict on him, inasmuch as he has already been rendered harmless by being disarmed, confined, and removed from the public as a threat. That is what punishment is, and to the extent that prisons serve punitive purposes–as they manifestly do, with very few exceptions–they stimulate far more violence than they prevent.

For example, for a quarter of a century, between 1942 and 1966, the American imprisonment rate averaged about 100 per 100,000 population (which is where it had been since the beginning of the century), and our murder rate averaged 5 per 100,000 (per year). During the quarter of a century following 1972, we embarked on an unprecedented massive social experiment with higher and higher rates of longer and longer prison sentences under increasingly punitive conditions. During this time, both our imprisonment rates and our murder rates increased to the highest levels ever recorded in our history. Our average imprisonment rate grew six times during the first three-quarters of the twentieth century, reaching an all-time record of more than 600 per 100,000 (the highest in the world). Our murder rate doubled from where it had been during the previous quarter of a century, to an average of 10 per 100,000 (also an all-time U.S. record). To put it another way, for more than a quarter of a century, from 1970 to the late 1990s, throughout the time that we were constantly increasing our imprisonment rate year after year, the United States experienced an epidemic of criminal violence. During this time, the murder rate ranged between 8 and 11 per 100,000, a level that was twice as high as it had been during the previous quarter of a century. It never dipped below the level of 8 during that entire time. If the enormously expanded use of increasingly punitive prisons actually prevented violence, rather than increasing it, that should not have happened. That conclusion is also supported by comparisons with the other developed nations. Our rates of imprisonment and of murder are both, on average, five to ten times higher than those of any other developed nation (Western Europe, Canada, Australia, New Zealand, and Japan). If imprisonment prevented violence, one would think that our murder rates would be five to ten times lower than the other nations’, rather than five to ten times higher.

Is there any evidence that the increases in our imprisonment rates have decreased our murder rates? In 1970, when our national incarceration rate was exactly where it had been, on average, for the previous seventy years–about 100 people per 100,000 population–our national murder rate was 8.3 per 100,000. Fifteen years later, in 1985, our incarceration rate had doubled to just over 200 prisoners per 100,000 population. What was the murder rate that year? Totally unchanged: 8.4 per 100,000. Eleven years after that, our imprisonment rate had doubled again, to more than 400 per 100,000, by which point the murder rate was still exactly where it had started: 8.3 per 100,000! In other words, the doubling and even quadrupling of our imprisonment rate did not produce the slightest demonstrable decrease in our murder rate.

If increasing our imprisonment rate did not bring down the murder rate, what was it that finally enabled it to dip slightly below the epidemic range of 8 to 11 per 100,000, by the last two years of the twentieth century? If the increase in the imprisonment rate was responsible, the decrease in the murder rate should have started long before the end of the 1990s, since the imprisonment rate had been increasing without pause for the previous thirty years. So what had changed by the end of the 1990s that might be able to account for our beginning emergence from this epidemic of violence for the first time in thirty years? First, the unemployment rate dropped to its lowest level in thirty years. Second, both the median wage and the minimum wage increased for the first time in thirty years. Third, the rates of relative poverty began decreasing for the first time in thirty years, especially among the demographic groups that are most vulnerable to engaging in violence, such as the young and the most impoverished minority groups. Rates of violence correlate throughout the world and throughout the United States with those kinds of social and economic variables (Brenner, 1973 and 1977; Hsieh, Ching-Chi and M.D. Pugh, 1999), whereas they do not correlate in any consistent way with imprisonment rates anywhere (Christie, 1993). (Also, the violence associated with selling crack cocaine began to diminish, as dealers finally divided up the market and agreed on marketing boundaries. This made a truce in the drug wars possible–wars that would not have begun in the first place, of course, if the sale of those drugs had not been criminalized and threatened with punishment, just as Prohibition stimulated the bootlegging wars of the Roaring Twenties).

Using the average prison time that is served per violent crime as a measure, the National Academy of Sciences’ Panel on the Understanding and Control of Violent Behavior reached the same conclusion: that the tripling of this measure between 1975 and 1989 had had “Apparently, very little” effect on the rate of violent crime overall, which showed no evidence of any decline whatsoever. As they concluded, “if tripling the average length of incarceration per crime had a strong preventive effect, then violent crime rates should have declined in the absence of other relevant changes” (Reiss, 1993, p. 6).

What is clear is that the U.S. experience confirms what the Chairman of the European Council’s expert committee on crime and punishment concluded in 1982: “there is no direct relation between the level of crime and the number of imprisonments or [the rate of imprisonment] at any particular point in time” (Hans Henrik Brydensholt, quoted in Christie, 1993, p. 34). That is one reason why the nations of Western Europe have chosen not to rely on imprisonment as their main tool for controlling crime; and that in turn is one reason why their rates of murder and other serious violent crimes are so much lower than ours.

Of course it is true that if we imprisoned all young men between the ages of 14 and 39, we would experience a massive drop in the rates of the kinds of violence that the law defines as criminal, since that is the group that commits almost all such violence. There was very little crime in Mussolini’s Italy; indeed, he was the only Italian ruler who was able to control the Mafia. And there was very little crime in Hitler’s Germany, or Stalin’s Russia. If we created a police state, as those dictators did, and imprisoned enough people, we could conceivably diminish the amount of “criminal” violence in our society. One problem with that type of solution, however, is that the state can commit much more violence, and is much more dangerous, than all the so-called “criminals” put together. When the state has that much of a monopoly on both power and violence, the state itself becomes the criminal, and commits war crimes, crimes against humanity, and other forms of collective violence, rather than the sporadic individual crimes committed by individual murderers. And once a police state loses or relinquishes its power, all the social tensions that had been suppressed rather than resolved erupt in an explosion of violence. This can be seen in the Russian crime wave that followed the end of Stalinism, and the epidemic of genocide and ethnic cleansing in the Balkans after the end of Tito’s dictatorship.

There is another field of research that also sheds light on the relationship between punishment and violence. We now have at least ninety years worth of research on child-rearing, focusing on the effects of different parenting and disciplinary practices on the development of moral reasoning, conscience, loyalty, honesty, violence and aggression, other antisocial behavior, and capacities for empathy and altruism, and so on. Child-rearing is such an inherently complicated and ambiguous enterprise, and involves so many thousands of different variables not all of which are possible to control, that it is not surprising that there are few findings from research that are consistently replicated. But there is one finding that has been so consistent that there is a substantial consensus among researchers. The more severely children are punished, the more violent they become–both during childhood itself, and later, in adulthood. For example, as Roger Brown summarized this research in a standard textbook on the subject,

Very few associations have been reliably established between child-rearing

practices and child personality. The association between physical

punishment and an aggressive child has more evidence behind it than any

other…. the evidence relating punishment and aggression is … better

[than for any other association, such as that between early independence

training and a strong achievement motive in the child]…. Severe

punishment went with more aggression (Brown, 1965, p. 387).

In another review of this field, Berelson and Steiner summarized the scientific findings on human behavior for which there was such a solid empirical basis that they were met with broad consensus among behavioral scientists of all major schools of thought. They summarize the research on punishment by saying, “The specific technique of [discipline called] punishment does seem to carry a boomerang effect …” (1964, p. 72). They then quote a classic and comprehensive review of the research findings in this field:

The unhappy effects of punishment have run like a dismal thread through our

findings. Mothers who punished toilet accidents severely ended up with

bed-wetting children. Mothers who punished dependency to get rid of it had

more dependent children than mothers who did not punish. Mothers who

punished aggressive behavior severely had more aggressive children than

mothers who punished lightly…. Harsh punishment was associated with high

childhood aggressiveness (Sears, Maccoby and Levin, 1957, quoted in

Berelson and Steiner, 1964, p. 72)

Another study they quote concluded that “the more severely boys were punished for aggression by their mothers, the more aggressive they were in preschool” (Sears, Whiting, Nowlis and Sears, 1953, quoted in Berelson and Steiner, 1964, 73).

Finally, they conclude that:

The more the control of the child is love-oriented, rather than based on

physical punishment, the more effective is the parents’ control over

desired behavior and the stronger the development of the child’s guilt

feelings for improper behavior…. The less the parental warmth … or the

more the parental punishment, the slower the development of conscience [in

the child]…. The more severe the punishment for aggression in infancy and

childhood, the more … aggression later…. The less use of physical

punishment in childhood and the more use of reasoning, the less likely the

child or adolescent [is] to engage in delinquent behavior (Berelson and

Steiner, 1964, 77-82).

Roger Brown also points out that

… the severity of the child’s conscience … is not proportionate to the

severity of the parent’s punishment but actually tends to be inversely

related…. What were Freud’s words? “A child which has been very leniently

treated can acquire a very strict conscience.”

How well does physical punishment [as opposed to “simply restraining a

child or distracting him”] work as a check on aggression against parents?

Not well. The parents who punished had the more aggressive children.

Apparently then punishment fails to produce a generally strong conscience

and fails also to check the particular form of wrongdoing against which it

is most often directed–aggression against parents.

… The best-established proposition … is that rejection and physical

punishment by parents tend to produce aggressive children with

under-developed superegos (1965, 394).

There are many other confirming studies. The Gluecks (1950) found severe physical punishment to be one of the major factors associated with delinquency in young boys. Bandura and Walters (1959) found physical punishments and paternal rejection associated with experiencing less guilt, being hyperaggressive, and being in trouble with the law…. Furthermore, they found that “the fathers had been rejecting long before the boys became exceptionally aggressive” (1965, 388-389).

But why would this apparently paradoxical or counter-intuitive relationship exist between punishment and behavior? To answer that question, Brown begins by observing that

This form of discipline has a peculiar and interesting property: it is

itself an instance of the behavior it is designed to eradicate. Punishment,

the response to aggression, is itself aggressive. What will the child

learn? … If … he learns by imitating what others do, he may learn to be

aggressive. He might even learn a subtler lesson that incorporates all the

information: Do not be aggressive to parents, since that is punished, but

do be aggressive to those smaller or subordinate to yourself as parents

successfully do.

… [Many studies, e.g.,] Bandura, Ross and Ross, 1961, have demonstrated

that children very readily imitate the aggressive actions of another person

so we cannot doubt that punishing parents could create aggressive children.

Whichever has priority in the child’s history, punishment or

aggressiveness, it seems likely that in a short time the two variables must

constitute a mutually reinforcing system [i.e., a literally “vicious”

cycle]. Punishment by angering the child and providing him with an

aggressive model must increase his own aggression and that increase would

stimulate the parents to further violence.

Thinking of punishment as itself a form of aggression has suggested to us

that punishment can engender aggression because it constitutes an imitable

model of aggression. In a parallel manner the “technique” called

“withdrawal of love” can be reconceived as a model of non-aggression under

provocation…. [E.g.,] A parent who will not allow himself to hit a child

or to scream at a child is nevertheless hurt, frustrated and angered….

Taking care not to “blow up” [is itself a message]…. the withdrawal of

love can … be described as an imitable model of non-aggression under

stress, and imitation of the model would produce a non-aggressive child.

What are the causes of psychopathy? … The most reliable antecedent of

adult psychopathy is … [that] the psychopath was severely rejected by his

parents and in many cases brutally beaten (McCord and McCord, 1956).

… Parents who beat their children for aggression intend to “stamp out”

the aggression. The fact that the treatment does not work as intended

suggests that the implicit learning theory is wrong. A beating may be

regarded as an instance of the behavior it is supposed to stamp out. If

children are more disposed to learn by imitation or example than by

“stamping out” they ought to learn from a beating to beat. That seems to be

roughly what happens (1956, 389).

… parents who beat their children for being aggressive nevertheless have

aggressive children, in fact children more aggressive than those of parents

who administer no beatings. This is not the way things should go if direct

reward and punishment were the only determinants of behavior. It is the way

things should go if children learn by example (1956, 395).

… aggressive children learn by the example of their parents and by the

example of aggression shown by the mass media (1956, 396).

But to understand more deeply why punishment stimulates violence rather than preventing it, it will be helpful to examine the psychological causes of violence. In a book in which I summarized observations and conclusions derived from clinical work with violent individuals and groups (Gilligan, 1996) I isolated the pathogen that is necessary (but not sufficient) for the development of violence, just as specifically as the tubercle bacillus is necessary (but not sufficient) for the development of tuberculosis. In the case of violence, however, the pathogen is an emotion, not a microbe–the emotion of shame and/or humiliation. This emotion is so powerful and pervasive, and so central to the experience of so many people, that there are forty synonyms for it, just as the Inuit were said to have forty words for snow because of its centrality in their experience. These are feelings of inferiority and inadequacy; feelings of being slighted, insulted, disrespected, dishonored, disgraced, disdained, slandered, treated with contempt, ridiculed, teased, taunted or mocked, rejected, defeated, subjected to indignity; feelings of being weak, ugly, incompetent or a failure, of “losing face,” of being treated as if one were insignificant, unimportant or worthless, or any of the numerous other forms of what psychoanalysts call “narcissistic injuries.” People become indignant (and may become violent) when they suffer an indignity; our language itself reveals the link between shame and rage.

This link first became apparent to me in my psychotherapeutic work with violent criminals when I discovered that I kept receiving the same answer when I would ask one man after another why he had assaulted or even killed someone: “Because he disrespected me.” In fact, they used that phrase so often that they abbreviated it into the slang phrase, “He dis’ed me.” Now, whenever people use a word so often that they abbreviate it, it is clearly central to their moral and emotional vocabulary. That experience and others like it convinced me that the basic psychological motive of violent behavior is the wish to ward off or eliminate the feeling of shame and humiliation–a painful, intolerable and overwhelming feeling–and replace it with its opposite, the feeling of pride (another master feeling for which there are also many synonyms–self-esteem, self-love, self-respect, feelings of self-worth, etc.).

Just as vulnerability to tuberculosis is influenced by the state of the body’s defense mechanisms (the immune system), so vulnerability to violence is influenced by the state of people’s psychological defense mechanisms. Especially important and relevant in this respect is the degree to which they have developed the capacity for an emotion that is antagonistic to shame, and inhibits the violence toward others that shame stimulates, namely, guilt and remorse. This is a capacity that the most violence-prone individuals and groups notably lack.

This same conclusion concerning the psychological cause of violence has been reached by scholars from a wide range of behavioral sciences, including clinical psychoanalysis, experimental psychology, sociology, criminology and law-enforcement. The psychoanalyst Heinz Kohut, for example, wrote that “The deepest level to which psychoanalysis can penetrate when it traces destructiveness [is to] the presence of a serious narcissistic injury, an injury that threatened the cohesion of the self” (1977). Another analyst, Gregory Rochlin made the same point when he emphasized “the relation of injured narcissism to aggression [and of] humiliation to violence,” and concluded that “The question … is … what makes people … so prone to feeling vulnerable and humiliated, and therefore ultimately what causes violence” (1973, viii).

Experimental psychologists have reached the same conclusion. Many individual studies and several reviews of the published research literature have been devoted to the study of aggressive behavior and simulated violence elicited under experimental conditions in psychological laboratories. These consist of, for example, experiments in which attempts are made to convince a subject to press a button that he is told will administer painful, potentially injurious or even lethal electrical shocks to another person. The consensus that has emerged from this work is that the most potent stimulus of aggression and violence, and the one that is most reliable in eliciting this response, is not frustration per se (as the “frustration-aggression” hypothesis has claimed), but rather, insult and humiliation. In other words, the most effective way, and often the only way, to provoke someone to become violent is to insult him. Feshbach, for example, after reviewing the literature on this subject, concluded that “violations to self-esteem through insult, humiliation or coercion are … probably the most important source of anger and aggressive drive in humans” (1971, 285). (It should be stressed that coercion, as a violation of autonomy, also produces feelings of shame, as Erik Erikson (1963) stressed; that is, pride is dependent on being independent, and coercion is the direct negation of autonomy.) Geen concluded that personal insult was more powerful in provoking aggressive behavior than frustration per se (1968). Sabini, in another review of the literature, generalized that

frustration per se does not lead to anger. If frustration is not the cause

of anger, what is? According to Aristotle, the perception that one has been

insulted leads to anger. … Curiously, when psychologists have tried to

produce anger in the laboratory, even when they have written about their

results in terms of the consequences of frustration, they have not relied

very much on frustrating people but have much more commonly insulted

people–possibly because it is very difficult to make adults angry just by

frustrating them (1978, 347).

The only situation in which frustration without deliberate insult was found to elicit anger was when the frustration was unjustified (e.g., a bus driver deliberately by-passing a bus stop). This does not constitute an exception to the principle that anger and violence are caused by feeling shamed, however, for the perception that one has been a victim of injustice elicits feelings of shame. Shame is felt from being valued so little by the other person, and for being too weak to change their behavior. In fact, the Latin word for injustice, iniuria, also means “insult” (as well as “injury”). One does not need to add insult to injury, or to injustice; it is already contained within both of those experiences, as it is in the words used to refer to them. (Karl Marx, another investigator of people’s responses to perceived injustice and exploitation, said, “shame is the emotion of revolution.”) The perception that one has been a perpetrator of injustice, by contrast, elicits feelings of guilt.

A number of sociologists have arrived at the same explanation of the psychological roots of human violence. Scheff and Retzinger, for example, wrote that “a particular sequence of emotions underlies all destructive aggression: shame is first evoked, which leads to rage and then violence” (1991, 3). The criminologist David Luckenbill analyzed the step-by-step escalation of the confrontations between victim and perpetrator that led to all seventy murders that occurred in one California county over a ten-year period, 1963-72. He found that in all cases, the murderer had interpreted his violence as the only means by which to save or maintain “face” and reputation and demonstrate that his character was strong rather than weak, in a situation that he interpreted as casting doubt on that assessment of himself. What started this process was some behavior by the victim that the perpetrator interpreted as insulting or disparaging to him. It would cause him to “lose face” if he “backed down” rather than responding with violence–even when the victim was only a child who refused to stop crying when ordered to (1977).

Nor is it only behavioral scientists and academicians who have reached these conclusions. The same findings have been reported by law-enforcement officers who have investigated the motives of murderers and other violent criminals. John Douglas, for example, was a “profiler” with the FBI whose career was devoted to studying the personalities and discerning the motives of the most violent and dangerous criminals in the United States. What he concluded was that any ultimate violent act “is the result of a deep-seated feeling of inadequacy,” and that these men attempt to diminish their low self-esteem by blaming others for their own real or imagined shortcomings, which were often caused, he discovered, by the way they were treated by overly authoritarian fathers (1999).

It is understandable, in terms of the etiological principles just discussed, why punishment would stimulate violence rather than inhibiting it. Punishment increases feelings of shame (it is humiliating to be punished, as it is intended to be) and decreases feelings of guilt (as it is intended to do). The whole purpose of legal punishment, after all, is to be the means by which the criminal expiates his guilt and thus “pays his debt to society,” at the completion of which process he is no longer guilty in the eyes of the law, nor presumably in his own. That punishment relieves guilt and leads to humiliation or shame is also apparent from the fact that those are precisely the purposes for which the religious sacrament of penance, or self-punishment, was institutionalized by the Church. Penance, or self-humbling, is the ritual that alleviates the feeling of sinfulness, or guilt (as it is intended to do). As we have just seen, violence is most likely when shame is maximized and guilt is minimized–exactly the conditions created by punishment.

What then can we do about crime and punishment? To ask that question is to ask what we can do about violence–whether it takes the form of crime or punishment. The first and most important principle might be to start doing nothing; in other words, to stop doing the things we already do that only stimulate violence, such as responding to the kind of violence that we call crime with the kind that we call punishment. I am not suggesting that it would make sense to let those who are actively raping and murdering others walk the streets. Physical restraint of those who are currently and actively physically violent–including confining or “quarantining” them in a locked facility–is at times the only way we have, in our ignorance, so far discovered to temporarily prevent further violence. But to punish people–that is, to deliberately cause them pain–above and beyond the degree that is unavoidable in the act of restraining them, only constitutes further violence (on our part), and only causes further violence (on the part of the “criminals” we punish).

Because restraint itself unavoidably involves coercion and physical force, and will inevitably be experienced by some (though not all) as a form of punishment, it would make sense to utilize it only for those who are physically violent themselves. In other words, it is time that we stopped overcrowding our prisons, bankrupting our economy, and subjecting the non-violent to violence and teaching them to become violent, by placing people in prisons for non-violent crimes. To use our prisons for those who have committed crimes against property, drug offenses, or have offended against someone’s sense of morality, as with prostitution or gambling, is self-defeating. Our goal ostensibly is to decrease the amount of violence in our society; yet the most effective way to turn a non-violent person into a violent one is to send him to prison. This does not always work, of course, as is shown by the examples of Thoreau, Gandhi, Martin Luther King, Nelson Mandela, and others who have survived the violence of imprisonment without becoming violent themselves. However, the mass of men who are sent to prison for non-violent offenses are not sent there because they had engaged in principled political protests. They are sent there for minor (often victimless) crimes born of weakness, poverty and despair, such as drug addiction or (in the case of women) prostitution. Even though they may not have harmed anyone, most of them do not have the personal strength or the moral resources of a King or a Mandela. I have seen far too many of them who have concluded that the best defense (against being brutalized by their guards and their fellow inmates) was a violent offense–a strategy they carried with them into the community after they were released from prison.

Prisons themselves could actually start preventing violence, rather than stimulating it (as they now do). We could take everyone out of them, demolish the buildings, and replace them with a new and different kind of institution: namely, a locked, secure residential school and college, whose purpose and function would be primarily educational and therapeutic. It would make sense to organize such a facility as a therapeutic community, with a full range of treatments for substance abuse and whatever other medical and mental health services are needed to help the individual heal the damage that deformed his character and stunted his humanity.

If it seems utopian to replace prisons with schools, please let me remind you that prisons already are schools, and always have been–schools in crime and violence, in deliberate and systematic humiliation, degradation, brutalization, exploitation and dehumanization, not in peace and love and dignity. I am merely suggesting that we replace an already existing type of school with another. Such a program would enable those who have been violent to develop non-violent means for developing feelings of self-esteem and self-respect. This could enable them also to be respected by others, and be able to take legitimate and realistic pride in their skills and knowledge and achievements, which all human beings need (and which most of us already have, at least to whatever is the necessary minimum degree). Such an atmosphere is necessary if they are to be able to find alternatives to violent behavior when their self-esteem is threatened (as it is for everyone, at one time or another). It would also enable them to become employable and self-sufficient, and to make a productive contribution to society when they return to the community. But before that can happen, we will have to renounce our own urge to engage in violence–that is, punishment–and decide that we want to engage in educational and therapeutic endeavors instead, so as to facilitate maturation, development, and healing.

But I do not wish to leave the impression that either the cause or the cure of violence and crime is to be found primarily in the criminal justice or penal system. I do not believe that prisons are the main causes of violence (though they are among the main ones); nor that they could be the main cures, even if they were abolished and replaced with something better. Rather, all violence–both the kind that is called crime and the kind that is called punishment–is a symptom of an epidemic social pathology, which has been called structural violence since it is a product of our social and economic structure. That is, both crime and punishment are symptoms of the social pathology that results from the division of society into “higher” and “lower” social classes, castes and age groups, with huge differences of power, wealth, prestige, honor and privilege (the rich versus the poor, whites versus people of color, and the old versus the young). The relative poverty that is suffered by those at the bottom of our social structure causes incomparably more loss of life in this country than all the violent behaviors put together, far more than cancer and heart disease combined, and more than AIDS or any other contagious disease. But it is also the main cause of the violent behavior that does occur (Gilligan, 1999). An additional cause is the division of our society into polarized, asymmetrical gender groupings. One consequence of this has been that men have traditionally been honored for being violent and shamed for refusing to be, while women have been shamed for being aggressive, competitive or independent even in non-violent ways, and honored only to the degree that they remained dependent on and submissive and subordinate to “their man” (first their father, then their husband).

As Lawrence M. Friedman put it,

… [T]he criminal justice system cannot … –in our society–even hope to

crush crime. Crime is far too complicated; its roots are too deep … in

the wellsprings of culture … [The sources of crime] do not lie within the

criminal justice system itself.

… [T]he furious building of prisons, the stiff laws, the cries for more,

more, more in the way of punishment–what has the upshot been? The effect

on crime–imperceptible. … [I]f the new toughness has had any effect on

crime rates, it is certainly hard to prove. Clear causal lines run the

other way…. But the higher the crime rate, the more people lean toward

“law and order.”

The sad fact is that no amount of tinkering, no amount of jail building or

amendments to penal codes will do the trick, at least not in this

society…. [T]he “crime problem” flows largely from changes in the culture

itself; it is part of us, our evil twin, our shadow; our own society

produced it. (1993, 446-7, 460-464)

That is why the solution to crime and the prevention of violence cannot be achieved through the penal system. The problems of crime and violence can only be solved by reforming our social and economic system, and reformulating the cultural and moral values that have produced that system and are in turn reinforced by it. A good first step in that direction would be to renounce the quixotic illusion that revenge (punishment) can inhibit or prevent violence.


Berelson, Bernard and Gary A. Steiner. Human Behavior: An Inventory of Scientific Findings. New York: Harcourt, Brace & World, 1964.

Brenner, M. H. Mental Illness and the Economy, Cambridge, MA: Harvard University Press, 1973.

Brenner, M. H. “Personal Stability and Economic Security.” Social Policy, 8 (1977): 2-4.

Brown, Roger. Social Psychology. New York: Free Press, 1965.

Christie, Nils. Crime Control as Industry: Towards GULAGS, Western Style? Routledge, 1993.

Douglas, John, and Mark Olshaker. The Anatomy of Motive: The FBI’s Legendary Mindhunter Explores the Key to Understanding and Catching Violent Criminals. New York: Lisa Drew/Scribner, 1999.

Erikson, Erik H. Childhood and Society. New York: W. W. Norton, 1963.

Feshbach, S. “The Dynamics and Morality of Violence and Aggression.” American Psychologist. 26 (1971): 281-292.

Friedman, Lawrence M. Crime and Punishment in American History. New York: Basic Books, 1993.

Geen, R.G. “Effects of Frustration, Attack, and Prior Training in Aggressiveness upon Aggressive Behavior.” Journal of Personality and Social Psychology. 9 (1968): 316-321.

Gilligan, James. Violence: Our Deadly Epidemic and Its Causes. New York: Grosset/Putnam, 1996; also published in paperback as Violence: Reflections on a National Epidemic. New York: Vintage Books, 1997.

–. “Structural Violence.” Violence in America: An Encyclopedia. Ed., Ronald Gottesman. New York: Charles Scribners Sons, 1999.

Glueck, Sheldon and Eleanor. Unraveling Juvenile Delinquency. Cambridge, MA: Harvard University Press, 1950.

Hsieh, Ching-Chi and M. D. Pugh. “Poverty, Income Inequality, and Violent Crime: A Meta-Analysis of Recent Aggregate Data Studies.” Criminal Justice Review, 18 (1993): 182-202, 1993.

Kohut, Heinz. The Restoration of the Self. New York: International Universities Press, 1977.

Piaget, Jean. The Moral Judgment of the Child. New York: Free Press, 1965.

Reiss, Albert J. and Roth, Jeffrey A., Eds. Understanding and Preventing Violence. Washington, D.C.: National Academy Press, 1993.

Rochlin, Gregory. Man’s Aggression: The Defense of the Self. Boston: Gambit, 1973.

Sabini, John. “Aggression in the Laboratory.” Violence: Perspectives on Murder and Aggression. Eds., Irwin L. Kutash, Samuel B. Kutash, and Louis B. Schlesinger. San Francisco: Jossey-Bass, 1978, pp. 343-371.

Scheff, Thomas and Suzanne Retzinger. Emotions and Violence: Shame and Rage in Destructive Conflicts. Lexington, MA: Lexington Books, 1991.

Sears, Robert R., John W. M. Whiting, Vincent Nowlis, and Pauline S. Sears. “Some Child-Rearing Antecedents of Aggression and Dependency in Young Children.” Genet. Psychol. Monogr. 47 (1953): 135-263.

Sears, Robert R., Eleanor E. Maccoby, and Harry Levin. Patterns of Child Rearing. Harper and Row, 1957

Silberman, Charles E., Criminal Violence, Criminal Justice, New York: Random House, 1978, p. 47.

Wolfgang, Marvin E., Patterns in Criminal Homicide (1958). New York: John Wiley & Sons (Science Editions), 1966.

James Gilligan is a Clinical Instructor in Psychiatry, Harvard Medical School. As a psychiatrist on the faculty of the Harvard Medical School, where he directed the Institute of Law and Psychiatry, he provided mental health services for 25 years to the Massachusetts prison system. He is currently conducting research on violence prevention in California jails and prisons, as Director of the Center for the Study of Violence. He is the author of Violence: Reflections on a National Epidemic (1997) and of the forthcoming Preventing Violence: An Agenda for the Coming Century (forthcoming 2001).

COPYRIGHT 2000 New School for Social Research

COPYRIGHT 2000 Gale Group