The question of public safety deserves the focused efforts of everyone, including abolitionists. Fear of "dangerousness" is at the heart of public acceptance for holding hundreds of thousands of people captives: the accused in jails, the convicted in the grip of indeterminate sentences and under the thumb of parole boards, and the released under surveillance on the streets. In the interest of justice, it is imperative, therefore, that the question of "dangerousness" and its predictability be thoroly explored and clarified.

As abolitionists, our hope is to reduce significantly all violent behavior, including the act of caging. Our assumptions regarding definitions, causes and predictability of "dangerousness" are central to determining the solutions we advocate.

Long conditioned to the belief that problems of criminality lie mainly within the individual, society has fastened its attention on "dangerous individuals" largely ignoring the learned nature of behavior. Once educated to the notion that human behavior is significantly shaped by social interaction and subtle learning processes as well as the broader structure of society itself, we can begin to transform the institutions and values that are conducive to violent behavior.

In our view individuals cannot be accurately or reliably classified as "dangerous" or "not dangerous," tho the violent acts they commit can. While individuals and their acts obviously are related, the assumption that a status of dangerousness can reliably be attached to a given person has been greatly overemphasized. Because psychiatric prediction is unreliable, owing to the tendency to over-predict, and because the definition of "the dangerous offender" is unclear, we had best discard the classification and focus on the acts rather than the actors. Our next task is to define as specifically as possible the violent crimes that require physical restraint for a period of time.

Acts which cause bodily harm, whether committed individually or collectively, by private citizens, corporate entities or the state, can be clearly classified as crimes on the basis of harm done to the victim. However, only a very small percentage of all lawbreakers cause bodily harm.

Those who do exhibit persistent patterns of behavior defined as dangerous, require restraint or limited movement for specific periods of their lives. This restraint should be subject to carefully drawn procedures. The goal of such "last resort" procedures should be to work out the least restrictive and most humane option for the shortest stated period of time.

Individual rights, safeguards and due process must be guaranteed to those who threaten public safety. The judiciary should bear the burden of proving in evidentiary hearings that no acceptable alternative to physical restraint exists for the present.

Focus should be on improving the life of the lawbreaker with the help of peer groups and the community. No person should be excluded from participation in as many decisions about his/her life as possible. The opportunity for changing violent, physically harmful behavior should always be present.

Small community restraining and re-education centers are needed. These centers should be controlled by peers of those who will be served. Such centers do not now exist, tho projects such as Delancey Street, Synanon and House of Umoja provide some criteria of what they might be like.

Confinement in peer centers should be considered as imprisonment because-at least for some-confinement will not be voluntary. However, intentional family-type structures in the community should be vastly superior to the iron bars, isolation cells, controlling drugs and arbitrary decision-making that are the standard of imprisonment today.

The politics of dangerousness

The selective and arbitrary process of labeling dangerousness is inherently political. Such labels are the basis of "preventative detention" and other forms of "treatment" which result in the violent (non)solution of caging. It is crucial that abolitionists examine the political implications and reliability of "dangerous" labels and predictions:

We might expect the origins of the word "danger" to be related to... its current use in denoting physical objects and events that might damage property or injure people. Surprisingly ... the term seems to have shaped out of linguistic roots that signified relative position in a social structure, a relationship between roles on a power dimension. The root is found in Latin in a derivative of dominium, meaning lordship or sovereignty .... The implication ... leads us ... into the conception of danger as a symbol denoting relative power in social organization .... Those persons or groups that threaten the existing power structure are dangerous. In any historical period, to identify an individual whose status is that of member of the "dangerous classes," (i.e., the classes that threaten the dominium or power structure) the label "criminal "has been handy.

-Theodore R. Sarbin, The Myth of the Criminal Type, pp. 16-17

People do not come into the world labeled "chattel" and "not chattel," "schizophrenic" and "not schizophrenic," "dangerous" and "not dangerous." We-slave traders and plantation owners, psychiatrists and judges-so label them.

-Thomas Szasz, "On Involuntary Psychiatry," New York Times, August 4, 1975

Prisons have been used to limit the movement of persons labeled as "dangerous," "psychotic" or "disturbed," a labeling process which began in the community, in the bad schools and continued thru each stage of the criminal justice system. The result has been the destruction of thousands of lives. We have been so concerned with containment, with limiting movement, that we haven't looked for the real troubles in people, in communities, in our social and economic system.

-John Boone, Former Director of Corrections, State of Massachusetts, Fortune News, May 1975

It is no wonder that today preventive detention proposals are so intensely opposed by Black organizations. They recognize correctly that their movement for freedom and self-determination is seen as "dangerous" by established white America. We approach the concept of "dangerousness" with considerable skepticism, for it has little meaning apart from its social and political concept.

-Struggle for Justice, p. 78

Men in prison are dangerous because they are threatened with sophisticated forms of extinction in the hands of simple minded wage earners who claim they are "only doing their duty" or "just following orders" as five or six of them are wrestling you to the floor to stick a needle in your arm or ass.

-Howard A. Lund, prisoner, NEPA News, March 1974

The defenders of these treatment models refuse to acknowledge that society, thru its injustices which are magnified inside prison walls, remains the principle impetus to violent behavior. Almost inevitably, those prisoners who refuse to accept the authoritarian, dehumanizing conditions of prison and who organize disruptive political behavior, exhibit repeated, angry "acting out" behavior, and flood the courts with litigation are the prisoners deemed candidates for DSU (Departmental Segregation Unit) or other "special offender" programs.

-Donna Parker, NEPA News, June 1974

"Dangerousness" and predictability

"Dangerousness" is difficult to define. Definitions always hinge on the unstated assumption that it is possible to predict which persons will commit violent acts in the future. The ability to make such predictions has not been demonstrated. Judges, parole board members, psychiatrists and others who attempt to predict "dangerousness" err grossly on the side of overprediction.[1] This results in the needless imprisonment of the many out of fear of the few.

I charge that this so-called diagnostic study is a fraud, consisting of nothing more than the random and whimsical guesses and speculations of a team of men, most of whom know nothing at all about what they are doing. It would be just as valid to make judgements and assignments of prisoners on the basis of their astrological sign, their hat size or the last two digits of their social security number.

-Professor William Ryan, at hearings for the proposed Massachusetts Departmental Segregation Unit and Classification Rules and Regulations, NEPA News, September 1974

The thing we have to get thru our skulls is that we cannot predict with any degree of accuracy who is going to be dangerous in the future. That is the one hang-up that the system has to get over. Every time they attempt to do this, it over-predicts to such a degree that the injustice practice far outweighs the protection gains.

-John Irwin, "Rehabilitation vs. Justice," Stanley L. Brodsky, ed., Changing Correctional Systems. Center for Correctional Psychology, University of Alabama, p. 57

Psychiatrists are rather inaccurate predictors--inaccurate in an absolute sense-and even less accurate when compared with other professionals such as psychologists, social workers, and correctional officials, and when compared to actuarial devices such as prediction or experience tables. Even more significant for legal purposes, it seems that psychiatrists are particularly prone to one type of error-overprediction. They tend to predict antisocial conduct in many instances where it would not, in fact, occur. Indeed, our research suggests that for every correct psychiatric prediction of violence, there are numerous erroneous predictions. That is, among every group of inmates presently confined on the basis of psychiatric predictions of violence, there are only a few who would, and many more who would not actually engage in such conduct if released.

-Alan Dershowitz, "The Psychiatrist's Power in Civil Commitments: A Knife that Cuts Both Ways," Psychology Today, February 1969, p. 47

Political considerations may also enter into the decision to overpredict dangerousness .... If psychiatrists consistently erred in their judgement by predicting that patients would not become violent, when in fact some did, the psychiatrists would lose the power and right to exercise their expertise in court. By overpredicting they avert that tragedy, and no one pays any attention to the 20 or more harmless people locked up to prevent the 21st from committing violence.

-Henry J. Steadman and Joseph J. Cocozza, "We Can't Predict Who Is Dangerous," Psychology Today, January 1975, p. 35

The conclusion to emerge most strikingly from these studies [predicting violence] is the great degree to which violence is overpredicted.. . Of those predicted to be dangerous, between 65 percent and 95 percent are false positives-that is, people who will not, in fact, commit a dangerous act. Indeed, the literature has been consistent on this point ever since Pinel took the chains off the supposedly dangerous mental patients at La Bicetre in 1792, and the resulting lack of violence gave lie to the psychiatric predictions that had justified their restraint.

-John Monahan, "The Prediction of Violence," Duncan Chappell and John Monahan, eds., Violence and Criminal Justice, p. 20

... Our data while not conclusive, indicated that the "deviant offender" existed more in the minds of those responsible for labeling him as such than he does in the real world ... if anything [the data] indicated those labeled deviant by the prison staff were not significantly different than "normal" inmates in any respect except slightly more depressed. And one need not wonder why that should be the case .... In a statistical sense as far as the data showed the "deviant" shared no other characteristics with other "deviants" except the name and treatment afforded him by the prison staff ... it was my conclusion in looking at the data that far from being a group in any respect the "deviants" are as different from others in that group as they are the same. Obviously, a proposal to "treat" this group-since it is not a group in any meaningful respect-is nonsense, at least with regard to the inmates we saw.

-Joan Smith, Dartmouth professor, letter to NEPA News, April/May 1974

Counteracting belief in predictability

It is clear after examining the data, that "experts" cannot predict dangerousness, either among prisoners or among the accused. In both cases, over-prediction means that untold numbers of innocent persons remain imprisoned, needlessly punished.

Despite our awareness of problems with predictability, the public's real emotional problem still remains. How can they be told that the "experts" they look to for protection cannot define who needs to be kept off the streets? The burden falls on judges, mental health workers and prison changers to take the lead in dispelling the myths of predictability of dangerousness.[2]

In the short range, we suggest three ways to begin to dispel the myths and deal with the concrete realities of defining, categorizing and responding to violent behavior:

(1) Encourage research to reveal statistics on overprediction of dangerousness. Utilize the findings for public education.

(2) Limit discretion by shifting the emphasis from "dangerous people" to violent behavior. Raise consciousness about cultural and institutionalized violence and support statutes that categorize violent crimes on the basis of harm done, rather than the individual lawbreaker's personal characteristics.

(3) Actively challenge the concept of "special prisons" and classification procedures in general, which label certain prisoners as "special offenders." Such labels focus on the individual as a predictable, unchanging, "sick" and dangerous object requiring treatment rather than as a human being exhibiting behavior generated in part by the violent society of prison.

Research challenging overprediction

Research can be cited which points to the myth of dangerousness. Many studies have established the lack of proof of predictive skills on the part of psychiatrists and others. In advocating decarceration and excarceration strategies, the following studies are useful:

The American Psychiatric Association Task Force on Clinical Aspects of the Violent Individual. [3] This report, released in November 1974, concluded that predictions of dangerousness are fundamentally of very low reliability. With few exceptions they are predictions of rare or infrequent events.

The likelihood of the expected behavior, such as violation of parole by a released prisoner whose previous crime was violent or the possibility of serious assault being committed by a released mental patient, would be very slight. Even if an index of violence proneness could be developed to correctly identify prior to release 50 percent of those individuals who will violate parole by committing violent offenses, the actual employment of such an index would identify eight times as many false positives as true positives. This means that eight of the nine persons retained in prison as a result of application of the index would not have committed such offenses if released.

The Research Center of the National Probation and Parole Institute [4] of the National Council on Crime and Delinquency. A study released in October 1972 follows the success and failure of more than 50,000 men thruout the country who were paroled in 1969. The rate of return for major crimes is not nearly as high as commonly believed -somewhere between five and eight percent in the first year, and presumably less after that, since the recidivism rate declines the longer parolees are on the street. Offenses involving violence--homicide, manslaughter, forcible rape and aggravated assault -accounted for less than one percent (.79 percent) of the men returned because of new commitment or allegation of violent offense. Another 1.1 percent were returned for potentially violent offenses (armed or unarmed robbery). The bulk of returns are for various forms of theft and violation of alcohol and narcotics laws.

The Baxstrom Studies [5]. In 1966, the U.S. Supreme Court ruled in Baxstrom v. Herold that Johnnie K. Baxstrom, who was sent to an institution for second-degree assault, could not be held in a maximum-security hospital for the criminally insane without proper judicial review for a longer period than he would have served in prison for the same offense. The decision resulted in the transfer of 967 patients from New York's two hospitals for the criminally insane to regular civil hospitals. These patients had committed or allegedly committed crimes, were considered dangerous, and were widely feared by the hospital staffs who had to house them in regular security facilities. Mental health officials were convinced that most of them would be so dangerous they would have to be returned to maximum-security hospitals, and if released, would be a threat to the community.

These patients were followed for more than four years after their transfer. Only 26 of them became troublesome enough to be returned to hospitals for the criminally insane. In a sample of 98 patients who were released, 20 were arrested, 11 convicted, but only two of the offenses could be considered dangerous: a robbery and an assault. If it hadn't been for the Baxstrom decision, almost 1,000 persons would have spent another five, ten or more years in hospitals for the criminally insane while only a tiny minority of them would have exhibited dangerous behavior after release.

Other studies on similar populations have also found low rates of dangerous behavior. In 1968 P.G. McGrath reported the results of his study of 293 murderers who were released from Broadmoor Hospital in England. Not one killed again. Four years later only one had killed again. Moreover, said McGrath, in the past 50 years about 140 patients were released each year from Broadmoor, and only two had been convicted of murder since release. [6]

Uniform Parole Report Studies of Murderers.[7] Nationwide statistics on parole performance, compiled by National Council on Crime and Delinquency, consistently show that paroled murderers are the best parole risks. Because, from the viewpoint of the general public, the murderer is perceived as the most dangerous type of offender, it might be supposed that murderers as a group present grave risks on parole. In fact, this is simply not the case. Parole Risks of Convicted Murderers, a special UPR study of 6,908 paroled murderers released during 1965-1969 across the nation, showed that 21 (0.3 percent) committed murder again during the first year of parole. A total of 122 (1.77 percent) were found guilty of new major offenses. Compare this failure rate with that of 9.03 percent for all other type of lawbreakers.

The Center for the Care and Treatment of Dangerous Persons.[8] A team of five mental health professionals, including two psychiatrists, made clinical examinations of individuals who had been convicted of serious assaultive crimes, often sexual in nature. These lawbreakers were assigned to special treatment programs after conviction and, at the time of the study, were eligible for release. Based upon the examinations, extensive case histories and the results of psychological tests, the team attempted to predict which individuals would commit assaultive crimes if released. These predictions of dangerousness were made prior to the court hearings at which the ultimate release decisions were made. Of 49 patients considered by the evaluating team to be dangerous and therefore not recommended for release, but who nevertheless were released after a court hearing, 65 percent had not committed a violent crime within five years of returning to the community. In other words, two-thirds of those predicted to be dangerous by a team of professionals did not, in fact, turn out to be dangerous.

Shifting the emphasis

The media constantly reinforces the belief that crime is a symptom of underlying psychic disturbance. This view has bolstered the assumption that criminality lies mainly within the individual. One of the difficulties with this conception of crime is that it is almost impossible to prove or disprove, at least in a systematic way.[9]

A primary theme in the sociology of crime emphasizes the learned nature of criminal behavior. Learning includes not only direct instruction, but also the long term influences of the socialization process. These are often quite subtle. All human behavior significantly reflects such influences, and criminal behavior is no exception.[10]

As an example, learned behavior is particularly evident with the violent crime of rape (considered at length in the next chapter). A sexist culture which devalues and objectifies women is certainly instructing consumers of that culture in violent sexual behavior. The problem of violent behavior will not be decreased or controlled merely by locking up rapists individually labeled "dangerous" while such practices in one form or another continue to be glorified by the culture. We can challenge many other obvious examples of societal instruction in criminal violence, not least among them the daily instruction in murder and assault on t.v. Our energies must focus on changing the violent message emanating from the culture. Cultural values and behavioral patterns can be changed through broad, systematic public re-education and resocialization.

Prison: More dangerous than prisoners

There is little disagreement that for those very few people who exhibit continual violent and aggressive behavior in society, temporary restraint is not only indicated but demanded. Review and monitoring procedures can be designed with adequate due process safeguards.

We believe the public can be educated to recognize that dangerousness cannot he clearly predicted, but that violent acts, both individual and collective, can he enumerated. We believe also that most citizens will support the constitutional guarantees that people are innocent until proved guilty, and that no one can be deprived of freedom for what they might do in the future only because of what they have done in the past.

The danger of needlessly denying an individual his/her liberty is far greater than the risk of freeing certain individuals who may again commit violent acts. The dangerousness of prison exceeds that of the combined dangerousness of each and all of its prisoners.

We are clear that no one should ever be excluded from humane conditions or the opportunity for changing violent, physically harmful behavior. Prisoners speak clearly to this point:

The guiding principles of the phaseout of the old] and introduction of the new but ever-adapting system are: No single individual must be excluded as an incorrigible problem. States must not ship out their "problem "prisoners to other places. That is not a solution; it is a cover-up for a fundamentally unworkable program.

It is a social atomisni; it is a rat psychology; it is the first stages of '84 and Clockwork Orange; it is fascism, the expendability or final solution of human beings. The so-called incorrigible prisoner, or "completely" insane person is precisely the measure of the depths of the challenge and must be faced and touched and transformed, no matter what the cost, for she or he is who we are in the furthest reaches of our humanity.

-The Action Committee, Walpole Prison, Massachusetts, NEPA NEWS, March! April 1975


1. Struggle for Justice, pp. 77-82.

2. Paul Warhaftig, "Prediction of Dangerousness-Does the Doctor Know Best? Or at All?" Pretrial Justice Quarterly, November 1975, p. 7.

3. John R. Lion and Donald P. Kenefick, et al., "Clinical Aspects of the Violent Individual," American Psychiatric Association News, November 20, 1974.

4. David F. Greenberg, "How Dangerous is the Ex-offender?" The Freeworld Times, January 1973, p. 11.

5. Henry J. Steadman and Joseph J. Cocozza, "We Can't Predict Who's Dangerous," Psychology Today, January 1975, p. 33. Also Henry J. Steadman and Gary Keveles, "Community Adjustment and Criminal Activity of the Baxstrom Patients: 1966-1970," American Journal of Psychiatry, Vol. 129, September 1972, pp. 304-310.

6. Ibid.

7. "Questions and Answers," Crime and Delinquency Literature, National Council on Crime and Delinquency, June 1974, p. 232.

8. Bruce Ennis and Thomas Litwack, "Psychiatry and the Presumption of Expertise: Flipping Coins in the Court Room," California Law Review, 62 (1974) p. 693. Also Harry L. Kozol, Richard J. Boucher and Ralph F. Garofalo, "The Diagnosis and Treatment of Dangerousness," Crime and Delinquency, October 1972, pp. 37192.

9. Edwin M. Schur, Our Criminal Society, pp. 6667.

10. Ibid. , pp. 96-97.