The “social-science view” of crime is thought by many, especially its critics, to assert that crime is the result of poverty, racial discrimination, and other privations, and that the only morally defensible and substantively efficacious strategy for reducing crime is to attack its “root causes” with programs that end poverty, reduce discrimination, and meliorate privation. In fact, however, at the time when their views on crime were first sought by policymakers (roughly, the mid-1960’s), social scientists had not set forth in writing a systematic theory of this sort. I recently asked three distinguished criminologists to nominate the two or three scholarly books on crime which were in print by mid-1960 and which were then regarded as the most significant works on the subject. There was remarkable agreement as to the titles: Principles of Criminology, by Edwin H. Sutherland and Donald R. Cressey, and Delinquency and Opportunity, by Richard A. Cloward and Lloyd E. Ohlin. Agreement was not complete on the validity of the views expressed in these books. Quite the contrary; criminologists then and now debate hotly and at length over such issues as the cause of crime. But these two books, and others like them, are alike in the way questions are posed, answers are sought, and policies are derived—alike, in short, not in their specific theories of delinquency, but in the general perspective from which those theories flow. And this perspective, contrary to popular impression, has rather little to do with poverty, race, education, housing, or the other objective conditions that supposedly cause crime. If anything, it directs attention away from factors that government can control, even if only marginally, to move beyond the reach of social policy altogether. Thus when social scientists were asked for advice by national policymaking bodies on how to reduce crime, they could not respond with suggestions derived from and supported by their scholarly work. In consequence, such advice as they did supply tended to derive from their general political views rather than from the expert knowledge they were presumed to have.

_____________

I

In the 1960’s the prevalent social-science perspective on crime found its most authoritative development in the treatise by Sutherland and Cressey whose seventh edition appeared in 1966, just after President Johnson appointed his crime commission. In this work Sutherland and Cressey reviewed various “schools of criminology” and faulted all but the “sociological” approach, according to which criminal behavior is learned by a person in intimate interaction with others whose good opinion he values and who define crime as desirable. The “classical” theories of Bentham and Beccaria were rejected because their underlying psychological assumptions—that individuals calculate the pains and pleasures of crime and pursue it if the latter outweigh the former—“assume freedom of the will in a manner which gives little or no possibility of further investigation of the causes of crime or of efforts to prevent crime.” The hedonistic psychology of Bentham, in short, suffered from being “individualistic, intellectualistic, and voluntaristic.” Theories based on body type, mental abnormality, or mental illness were also rejected because the available data were inconsistent with them. Criminals were no more likely than law-abiding persons to have a certain stature, to be feeble-minded, or to suffer from a psychosis.

As for poverty—defined as having little money—Sutherland and Cressey’s references to its impact were few and skeptical. Sutherland was quoted from his earlier writings as observing that while crime was strongly correlated with geographic concentrations of poor persons, it was weakly correlated (if at all) with the economic cycle. That is, crime might be observed to increase as one entered a poor neighborhood, but it was not observed to decrease as neighborhoods generally experienced prosperity. Furthermore, Albert K. Cohen (to whom Sutherland and Cressey refer approvingly) had shown that much of the delinquency found among working-class boys was “non-utilitarian”—that is, consisted of expressive but financially unrewarding acts of vandalism and hell-raising—and that these acts were more common among this group than among middle-class boys. If economic want were the cause of crime, one would predict that delinquency for gain would be more common among those less well-off and delinquency for “fun” more common among the better-off. Yet the opposite seemed to be the case. “Poverty as such,” Sutherland concluded, “is not an important cause of crime.”

Nor could being a member of a minority group and experiencing the frustrations produced by discrimination explain crime for Sutherland and Cressey: while the experience of Negroes, whose crime rate was high, might support such a theory, that of the American Japanese, whose crime rate was low, refuted it. Poverty and racial segregation might serve to perpetuate crime, however, to the extent that these factors prevented persons from leaving areas where crime was already high and thus from escaping those personal contacts and peer groups from which criminal habits were learned.

_____________

There were in 1966 other theories of crime in addition to Sutherland and Cressey’s. Most of these were reviewed in their treatise and though criticisms were sometimes made, the governing assumptions of each were quite compatible with what the authors described as the sociological approach. Sheldon and Eleanor Glueck, for example, produced in the 1950’s a major effort to predict delinquency, and while the idea of predicting delinquency became controversial on grounds of both fairness and feasibility, their empirical data on factors that helped cause delinquency were not seriously challenged. They argued and supplied data to show that among the key variables distinguishing delinquents from non-delinquents were those related to family conditions—chiefly stability, parental affection, and the discipline of children. Walter B. Miller also argued that delinquency was in large part an expression of the focal concerns of lower-class youth. Toughness, masculinity, “smartness,” the love of excitement, and a desire for personal autonomy were valued by lower-class persons to a greater degree than by middle-class ones, and acting on the basis of these values, which were maintained by street-corner gangs, inevitably placed many lower-class boys (and some girls) in conflict with the laws of the middle class. Albert K. Cohen further suggested that delinquency was in part the result of lower-class youth striving, not simply to assert their focal values, but to repudiate those middle-class values which they secretly prized.

These and other sociological theories of crime, widely known and intensely discussed in the 1960’s, had certain features in common. All sought to explain the causes of delinquency, or at least its persistence. All made attitude formation a key variable. All stressed that these attitudes were shaped and supported by intimate groups—the family and close friends. All were serious, intelligent efforts at constructing social theories, and while no theory was proved empirically, all were consistent with at least some important observations about crime. But none could supply a plausible basis for the advocacy of public policy.

This was true for several reasons. By directing attention toward the subjective states that preceded or accompanied criminal behavior, the sociological (or more accurately, social-psychological) theories directed attention toward conditions that cannot be easily and deliberately altered. Society, of course, shapes attitudes and values by its examples, its institutions, and its practices, but only with great difficulty, slowly, and imprecisely. If families inculcate habits of virtue, law-abidingness, and decorum, it is rarely because the family is acting as the agent of society or its government, but rather because it is a good family. If schools teach children to value learning and to study well, it is not simply because the schools are well-designed or generously supplied, but because attitudes consistent with learning and study already exist in the pupils. One can imagine what government might do if it wished to make good families even better or successful pupils even more successful: more resources might be offered to reduce burdens imposed by want, but the gains, if any, would likely be at the margin.

If it is difficult by plan to make the good better, it may be impossible to make the bad tolerable so long as one seeks to influence attitudes and values directly. If a child is delinquent because his family made him so or his friends encourage him to be so, it is hard to conceive what society might do about this. No one knows how a government might restore affection, stability, and fair discipline to a family that rejects these characteristics; still less can one imagine how even a family once restored could affect a child who by now has left the formative years and in any event has developed an aversion to one or both of his parents. Government could supply the lower class with more money, of course, but if a class exists because of its values rather than its income, it is hard to see how, in terms of the prevailing theory, increasing the latter would improve the former.

If, similarly, the lower class has focal concerns that make crime attractive or even inevitable, it is not clear how government could supply it with a new set of values consistent with law-abidingness. Indeed, the very effort to inculcate new values would, if the sociological theory is true, lead the members of that class to resist such alien intrusions all the more vigorously and to cling to their own world-view all the more strongly. Peer groups exist, especially for young people, as a way of defending their members from an alien, hostile, or indifferent larger society and for supplying them with a mutually satisfactory basis for self-respect. A deviant peer group—one that encourages crime or hell-raising—would regard any effort by society to “reform” it as confirmation of the hostile intent of society and of the importance of the group.

_____________

II

The problem lies in confusing causal analysis with policy analysis. Causal analysis attempts to find the source of human activity in those factors which themselves are not caused—which are, in the language of sociologists, “independent variables.” Obviously nothing can be a cause if it is in turn caused by something else; it would then only be an “intervening variable.” But ultimate causes cannot be the object of policy efforts precisely because, being ultimate, they cannot be changed. For example, criminologists have shown beyond doubt that men commit more crimes than women and younger men more (of certain kinds) than older ones. It is a theoretically important and scientifically correct observation. Yet it means little for policymakers concerned with crime prevention since men cannot be changed into women nor made to skip over the adolescent years. Not every primary cause is itself unchangeable: the cause of air pollution is (in part) certain gases in automobile exhausts, and thus reducing those gases by redesigning the engine will reduce pollution. But social problems—that is to say, problems occasioned by human behavior rather than mechanical processes—are almost invariably caused by factors that cannot be changed easily or at all, because human behavior ultimately derives from human volition—tastes, attitudes, values, or whatever—and these aspects of volition are in turn formed either entirely by choice or are the product of biological or social processes that we cannot or will not change.

It is the failure to understand this point that leads statesman and citizen alike to commit the causal fallacy—to assume that no problem is adequately addressed unless its causes are eliminated. The preamble to the UNESCO charter illustrates the causal fallacy: “Since wars begin in the minds of men it is in the minds of men that the defenses of peace must be constructed.” Yet the one thing we cannot easily do, if at all, is change, by plan and systematically, the minds of men. If peace can only be assured by doing what we cannot do, then we can never have peace. If we regard any crime-prevention or crime-reduction program as defective because it does not address the “root causes” of crime, then we shall commit ourselves to futile acts that frustrate the citizen while they ignore the criminal.

Sutherland and Cressey commit the fallacy; yet, being honest scholars, they provide evidence in their own book that it is a fallacy. “At present,” they write, “the greatest need in crime prevention is irrefutable facts about crime causation and sound means for transforming that knowledge into a program of action.” Suppose it could be shown that their own theory of crime causation is irrefutably correct (it may well be). That theory is that individuals commit crime when they are members of groups—families, peers, neighborhoods—which define criminal behavior as desirable. The policy implication of this, which the authors draw explicitly, is that the local community must use the school, the church, the police, and other agencies to “modify” the personal groups in which crime is made to appear desirable. No indication is given as to how these agencies might do this and, considering what the authors and other sociologists have said about the strength and persistence of family and friendship ties, it is hard to see what plan might be developed.

But we need not merely raise the theoretical difficulties. A series of delinquency-prevention programs have been mounted over the decades, many if not most of which were explicitly formed on the strategy of altering primary-group influences on delinquents. Almost none can be said on the basis of careful, external evaluation to have succeeded in reducing delinquency. Sutherland and Cressey describe one of the most ambitious of these, the Cambridge-Somerville Youth Study in the late 1930’s. The differences in crime between those youths who were given special services (counseling, special educational programs, guidance, health assistance, camping trips) and a matched control group were insignificant: “‘the treatment’ had little effect.” Perhaps a better program would have had better results, though it is striking that for some a “better” youth project is one that moves beyond merely providing concentrated social-welfare services to deliquents because these services do not address the “real” cause of crime. William and Joan McCord, in Origins of Crime, for example, draw the lesson from the Cambridge-Somerville study that the true causes of delinquency are found in the “absence of parental affection” coupled with family conflict, inconsistent discipline, and rebellious parents. They are quite possibly correct; indeed, if I may speak on the basis of my own wholly unscientific observation, I am quite confident they are correct. But what of it? What agency do we create, what budget do we allocate, that will supply the missing “parental affection” and will restore to the child consistent discipline supported by a stable and loving family? When it comes to the details of their own proposals, they speak of “milieu therapy” in which the child is removed from his family and placed in a secure and permissive therapeutic environment of the sort developed by Dr. Bruno Bettleheim for autistic children. Conceding that such a program is frightfully expensive, they urge that we attempt to reach fewer children than under conventional programs, and presumably keep each child for a relatively long period. That parents, children, taxpayers, or courts might object to all this is not considered.

Attempts to explain the causes of crime not only lead inevitably into the realm of the subjective and the familial, where both the efficacy and propriety of policy are most in doubt; they also lead one to a preference for the rehabilitative (or reformation) theory of corrections over the deterrence or incapacitation theories. Sutherland and Cressey recognize this: “On a formal level it may be observed that attempts to explain criminal behavior have greatly abetted at least the official use of the treatment reaction.” One may deter a criminal by increasing the costs or reducing the benefits of crime, but that strategy does not deal with the “causes” of criminality, and hence does not go to the “root” of the problem. Stated another way, if causal theories explain why a criminal acts as he does, they also explain why he must act as he does and therefore they make any reliance on deterrence seem futile or irrelevant. Yet when Sutherland and Cressey come to consider the consequences of treating criminals in order to reform them, as opposed to punishing in order to deter them, they forthrightly admit that “there is no available proof” that treatment increased or decreased crime, and that “the methods of reformation . . . have not been notably successful in reducing crime rates.” Careful reviews of the major efforts to rehabilitate criminals amply support this judgment.¹

_____________

Policy analysis, as opposed to causal analysis, begins with a very different perspective. It asks, not what is the cause of a problem, but what is the condition one wants to bring into being, what measure do we have that will tell us when that condition exists, and finally what policy tools does a government (in our case, a democratic and libertarian government) possess that might, when applied, produce at reasonable cost a desired alteration in the present condition or progress toward the desired condition? In this instance, the desired condition is a reduction in specified forms of crime. The government has at its disposal certain policy instruments—rather few, in fact—that it can use: it can redistribute money, create (or stimulate the creation of) jobs, hire persons who offer advice, hire persons who practice surveillance and detection, build detention facilities, illuminate public streets, alter (within a range) the price of drugs and alcohol, require citizens to install alarm systems, and so forth. It can, in short, manage to a degree money, prices, and technology, and it can hire people who can provide within limits either simple (e.g., custodial) or complex (e.g., counseling) services. These tools, if employed, can affect the risks of crime, the benefits of non-criminal occupations, the accessibility of things worth stealing, and the mental state of criminals or would-be criminals. A policy analyst would ask what feasible changes in which of these areas would, at what cost (monetary and non-monetary), produce how much of a change in the rate of a given crime. He would suspect, from his experience in education and social services, that changing the mental state of citizens is very difficult, quite costly, hard to manage organizationally, and may produce many unanticipated side-effects. He would then entertain as a working hypothesis that, given what he has to work with, he may gain more by altering risks, benefits, alternatives, and accessibility. He would not be sure of this, however, and would want to analyze carefully how these factors are related to existing differences in crime by state or city, and then would want to try some experimental alterations in these factors before committing himself to them wholesale.

In sum, the criminologist, concerned with causal explanations and part of a discipline—sociology—which assumes that social processes determine behavior, has operated by and large within an intellectual framework that makes it difficult or impossible to develop reasonable policy alternatives and has cast doubt, by assumption more than by argument or evidence, on the efficacy of those policy tools, necessarily dealing with objective rather than subjective conditions, which society might use to alter crime rates. A serious policy-oriented analysis of crime, by contrast, would place heavy emphasis on the manipulation of objective conditions, not necessarily because of a belief that the causes of crime are thereby being eradicated, but because behavior is easier to change than attitudes, and because the only instruments society has for altering behavior in the short run require it to assume that people act in response to the costs and benefits of alternative courses of action. The criminologist assumes, probably rightly, that the causes of crime are determined by attitudes which in turn are socially derived, if not determined; the policy analyst is led to assume that the criminal acts as if crime were the product of a free choice among competing opportunities and constraints. The radical individualism of Bentham and Beccaria may be scientifically questionable, but it is prudentially necessary.

_____________

The other most important work of the 1960’s in the field of criminology, Delinquency and Opportunity by Richard A. Cloward and Lloyd E. Ohlin, would appear to be an exception to the general criminological perspective of the day. Writing in 1960, Cloward and Ohlin developed an influential theory of delinquency in big cities. A delinquent gang (or “subculture”—the terms are used, for reasons not made clear, interchangeably) arises in response to the conflict that exists between socially-approved goals (primarily monetary success) and socially-approved means to realize those goals. Certain youths, notably of the lower class, desire conventional ends but discover that there are no legitimate means to attain them; being unable (unwilling?) to revise these expectations downward, they experience frustration and this may lead them to explore illegitimate (“nonconforming”) alternatives. Some lower-class youth may aspire to middle-class values (“money and morality,” as the authors put it) while others may aspire only to success in lower-class terms (money alone). The barriers to realizing those aspirations are found in part in cultural constraints derived from the immigrant experience (Southern Italians and Sicilians, for example, allegedly do not value schooling highly), but in larger part in structural difficulties, chiefly the fact that education is costly in money outlays and foregone earnings.

In its brief form, the theory of Cloward and Ohlin would seem to be in sharp contrast to the general sociological perspective. Delinquency may in their view be learned from peers, but it is learned because of the gap between aspirations and opportunities, and opportunities in turn are objective conditions determined by government and the social system. Education, they claim, is the chief source of opportunity. One therefore expects them to end their book with a call for cheaper, more readily available educational programs. But they do not. Indeed, less than one page is devoted to policy proposals, amounting essentially to one suggestion: “The major effort of those who wish to eliminate delinquency should be directed to the reorganization of slum communities.” No explanation is offered of what “slum reorganization” might be, except for several pages that decry “slum disorganization.” Their analysis leads the reader toward the material desires of life as the key factor (indeed, that is all the lower classes are supposed to value), but stops short of telling us how those material desires are to be realized. Their theory states that “each individual occupies a position in both legitimate and illegitimate opportunity structures” (they rightly note that this is a “new way” of viewing the problem), but they do not speak of the costs and benefits of illegitimate as opposed to legitimate opportunities. Instead, the individual who is confronted with a choice among kinds of opportunities does not choose, he “learns deviant values” from the “social structure of the slum.”

Thus, when the authors come to speak of policy, they have little to say about what determines the choice of illegitimate opportunities (nobody has chosen anything, he has only “learned” or “assimilated”), and thus they have no theoretical grounds for suggesting that the value of legitimate “opportunities” should be increased (e.g., better-paying jobs for slum youth), or that the benefits of illegitimate ones should be decreased (e.g., more certain penalties for crime), or that “opportunities” for goal gratification be replaced by direct goal gratification (e.g., redistributing income).

_____________

III

Explaining human behavior is a worthwhile endeavor; indeed, for intellectuals it is among the most worthwhile. Those who search for such explanations need not justify their activity by its social utility or its policy implications. Unfortunately, neither intellectuals nor policymakers always understand this. If the government becomes alarmed about crime, it assumes that those who have studied crime most deeply can contribute most fully to its solution. Criminologists have rarely sought to show statesmen the error of this assumption. Much of their writing is “practical,” much of their time is “applied.” To a degree, of course, criminological knowledge may assist criminologists’ actions: careful study and conscientious learning can help one avoid obvious errors, attack popular myths, and devise inventive proposals. But it is also likely that the most profound understanding may impede or even distort, rather than facilitate, choice, because much of this knowledge is of what is immutable and necessary, not of what is variable or contingent.

In the mid-1960’s, when the federal government turned toward social scientists for help in understanding and dealing with crime, there was not then in being a body of tested or even well-accepted theories as to how crime might be prevented or criminals reformed, nor was there much agreement on the causes of crime except that they were social, not psychological, biological, or individualistic. In fact, there was not even much agreement that crime was a major and growing problem—scholars noted the apparent increase in crime rates, but (properly) criticized the statistical and empirical weaknesses in these published rates. While these weaknesses did not always lead the critics to conclude that crime was in fact not increasing, some scholars did draw that conclusion tentatively and their criticisms encouraged others to draw it conclusively.

Nor were scholars very farsighted. Having established beyond doubt that crime rates were strongly related to age differences, few scholars (none that I can recall) noted the ominous consequences for crime of the coming-of-age in the 1960’s of the products of the postwar “baby boom.” Similarly, while some scholars had shown by cross-sectional studies that the proportion of a city’s population that was nonwhite was powerfully correlated with assaultive crimes, few to my knowledge drew the obvious implication that, unless this correlation was spurious, the continued in-migration of blacks to large cities would inflate crime rates. Once the various national commissions were underway, however, scholars associated with them (notably the group associated with the Task Force on the Assessment of Crime, under the direction of Lloyd Ohlin) began to work vigorously on these issues and produced a number of reports that showed vividly the impact of demographic changes on crime rates.

The major intellectual difficulty governing the relationship of social scientists to policymakers with respect to crime was not the presence or absence of foresight, however, but rather the problem of how to arrive at policy proposals in the absence of scientific knowledge that would support them. The crime commission did not develop new knowledge as to crime prevention or control; as Professor Ohlin later described it, existing “social-science concepts, theories, and general perspectives were probably of greater utility to the staff and the commission in forming the final recommendations than the inputs from new knowledge development efforts.” What were these “concepts, theories, and general perspectives”? One, cited by Ohlin, consisted of “grave doubts” about the effectiveness of the criminal-justice system and of rehabilitation and treatment programs. From this, Ohlin and his colleagues drew the conclusion that “the criminal-justice system should be used only as a last resort in the control of undesirable conduct.” From that inference, in turn, the commission adopted the view that offenders should be “diverted” from the system and recommended a broad policy of “deprisonization.”

There are no doubt ample grounds in humane sentiment for finding fault with prisons, but at the time of the commission’s work there were scarcely any well-established scientific grounds. That “treatment” had failed seemed clear, but “non-treatment” had failed just as clearly: persons on probation might be no more likely to “recidivate” than those in prison, but neither were they much less likely. As for deterrence, there was, when the commission deliberated and Professor Ohlin advised, virtually no scientific material on whether prison did or did not deter. It was not until 1966, fifty years after criminology began as a discipline in this country and after seven editions of the leading text on crime had appeared, that there began to be a serious and sustained inquiry into the consequences for crime rates of differences in the certainty and severity of penalties. In any case, the commission scarcely dealt with the deterrence or incapacitation functions of prison.

In short, criminology could not form the basis for much policy advice to the commission. Yet that did not prevent criminologists from advising. Professor Ohlin is entirely honest about this: “The relevant social-science literature was descriptive and analytical. There were relatively few experimental or controlled studies of the effectiveness of particular programs or policies. . . . Sociologists serving as consultants to the commission proved reluctant to draw out . . . action recommendations. . . . When they did try to do this, the recommendations were often more influenced by personal ideological convictions than by appropriately organized facts and theories . . .” [emphasis added].

Social scientists did not carry the day on the commission (they could not, for example, get their view on marijuana accepted), but the effect of their advice, based on personal belief rather than scholarly knowledge, was clear. Working with sympathetic commission members in small task forces, the advisers stimulated and participated in a process that—as Professor Ohlin later put it—“led to far more liberal recommendations by the commission than one would have thought possible at the outset given the conservative cast of its membership.”

There is nothing whatsoever wrong with social scientists trying to persuade others of their policy beliefs, just as there is nothing wrong with lay commission members trying to persuade sociologists of their beliefs. There is something wrong with a process of persuasion colored by the mistaken notion that one party is an “expert” whose views are entitled to special consideration because of their evidentiary quality. There is no way of knowing to what extent commission members believed what the sociologists were saying was true, as opposed to merely plausible or interesting. But based on my own experience in advising national commissions, including the crime commission, I am confident that few social scientists made careful distinctions, when the chips were down, between what they knew as scholars and what they believed as citizens, or even spent much time discussing the complex relationships between knowledge and belief. I certainly did not, and I do not recall others doing so.

_____________

IV

Having alluded to my own role as a policy adviser, let me amplify on that experience to reinforce, by self-criticism, the point I am making. I was not in 1966 a criminologist, nor am I now. I came to crime, if I may put it that way, as a consequence of my study of police administration and its political context, and found myself labeled an “expert” on crime because of that interest and perhaps also because of the desire of governmental consumers of “expertise” to inflate, by wishful thinking, the supply of such persons to equal the demand for their services.

Once I found myself, willy-nilly, in the crime business, I found that my ideas on the subject—apart from those formed by my own empirical research on policing—were inevitably influenced by the currents of academic opinion about me. The effect of these currents is not to persuade one of what is true, but to persuade one of what is important. In my case, I did not absorb from criminological writings a set of policy conclusions about whether criminals can be deterred or rehabilitated, but I did absorb a set of interesting facts about crime: for example, that crimes are age-specific, that victims contribute to their victimization in most assaultive crimes, and that published crime rates are unreliable. All of these things were (and are) true, but of course they are not directly related to the policy question of what is to be done about crime.

In short, I did not, any more than Professor Ohlin, have in 1966-68 empirically supported policy advice to offer statesmen dealing with crime. What I then realized, as did Professor Ohlin, was that many of those seated about me, urging in the strongest tones various “solutions” to crime, were speaking out of ideology, not scholarship. Nor was this only true of my colleagues on the crime commission. Walter Reckless, for example, in the 1967 edition of his text, The Crime Problem, states flatly that punishment “does not . . . prevent crime,” though he adduces no systematic evidence to warrant such a conclusion. Charles R. Tittle and Charles H. Logan provide other examples of this unsupported assertion in their review of the more recent literature on deterrence, a review that nevertheless concludes by observing that “almost all research since 1960 supports the view that negative sanctions are significant variables in the explanation of conformity and deviance. . . . Sanctions apparently have some deterrent effect under some circumstances.”

What I only later realized was that criminologists, and perhaps sociologists in general, are part of an intellectual tradition whose focal concerns are with those aspects of society that are, to a great extent, beyond the reach of policy and even beyond the reach of science. Those matters that are within the reach of policy have been, at least for many criminologists, defined away as uninteresting because they were superficial, “symptomatic,” or not of “causal” significance. Sociology, for all its claims to understand structure, is at heart a profoundly subjectivist discipline. When those who practice it are brought forward and asked for advice, they will say either (if conservative) that nothing is possible, or (if liberal) that everything is possible. That most sociologists are liberals explains why the latter reaction is more common even though the presuppositions of their own discipline would more naturally lead to the former.

¹ See, for example, Leslie T. Wilkins, Evaluations of Penal Measures (Random House, 1969), and Robert Martinson, “What Works? Questions and Answers About the Rehabilitation of Prisoners” (Public Interest, Spring 1974).