George L. Kelling is a professor of criminal justice at Northeastern University and a fellow of Harvard University in criminal justice.

“The metaphor,” José Ortega y Gasset once wrote, “is perhaps one of man’s most fruitful potentialities.... Its efficacy verges on magic.” Sometimes it verges on black magic: Last year, in a depressed section of Buffalo, eight schoolgirls were raped. They were victims, though no one realized it at the time, not only of a rapist, but of a metaphor.

What happened was this: Over a 15-month period, 11 girls ranging in age between 11 and 16 years old were raped. After the third attack, police determined that a serial rapist was stalking young girls on their way to school. The police had a description and a good sense of the rapist’s method of operation. Yet they did not notify parents, or circulate a sketch of the rapist, until five months and eight rapes later.

This was not an ordinary case of dereliction of duty. The chief of detectives, who withheld the information, was doing his duty as he saw it. Openly defending the department’s decision to angry residents, he explained that “sensational” news coverage might have hindered the investigation of the crimes and apprehension of the suspect.

During that fearful five-month rape spree, Buffalo police believed they were doing their job. But what was their job? Was it to investigate crimes and apprehend criminals? Then perhaps they were doing it.

But what if their job was to prevent more schoolgirls from being raped?

If that had been their primary goal, surely the police would have behaved in an entirely different way. Not only would they have released a description of the rapist to the press, but they would have worked with the community, advising residents, school officials, and children. They would have mobilized parents and children to take preventive action. They would have acted as community leaders, helping the neighborhood overcome fear and suspicion so that it could resist the rapist. This may sound a bit utopian, but sophisticated private security companies do such things all the time.

The police failed to do these things primarily because police in Buffalo, as in most major U.S. cities, have come to believe that “doing their job” means playing their proper role in something that they, like the rest of us, have learned to call the “criminal justice system.” The particular job of the police is to serve as the “front end” of the system, for which they “acquire” criminals. The criminals are then fed further into the system, where they become the problem of somebody else: judges, juries, and jailers.

In this vision, it is not the particular job of the police, or any other single part of the system, to prevent crime. Preventing crime is one of several goals that the system as a whole is supposed to achieve.

This notion—that a police officer’s primary job is to acquire criminals for “the system”—governed not only Buffalo’s chief of detectives, but a whole generation of police, prosecutors, and prison officials.

So pervasive has the notion of the criminal justice system become in the thirty years since the term was coined that few have noticed the obvious: The system—on which we pin our hopes for such overall goals as preventing crime—does not exist. The system is not a fact but a phrase, a metaphor that attempts to capture the complex and problematic interactions between criminals and police, courts, corrections, and other crime-control agencies.

If this sounds surprising at first, it should. The most damaging thing about this metaphor is that it is not recognized as such, but regarded as a literal description of reality. Yet recognizing that the criminal justice system is a metaphor, and a misleading one at that, is crucial to understanding why we are losing the battle against crime.

Crime-control ideology is rife with metaphor: “Wars” are fought against crime; detectives solve cases using Sherlock Holmes-like ratiocination; Justice is a robed, blindfolded woman bearing a scale; the police are a “thin blue line” protecting innocent civilians from criminal marauders.

At their best, metaphors are simple and vivid, bringing fresh life to abstract concepts. They are signposts that can tell us better than any treatise what to expect or how to behave. The blindfolded woman heralds impartiality; “war” signals total commitment; the “thin blue line,” military courage and defense of the innocent.

Yet metaphors can distort as well as reveal. Lost, for example, in the image of Justice blindfolded is that the courts’ most substantial business today is ratifying plea bargains. Lost in the imagery of detectives as sleuths is the process by which most crimes are solved: preliminary investigation of crimes by patrol officers. And lost in the military image of the thin blue line, perhaps the second most important policing metaphor, is most of what good police do.

The thin blue line metaphor was coined in the Fifties by Los Angles Police Chief William Parker, an important police reformer of his era. (Its origin is the thin red line, first used by London Times journalist Walter Russell to depict British infantrymen in the battle of Balaclava.) In Parker’s day, police reformers’ most important goal was to transform the nation’s police forces from sloppy, patronage-ridden, often corrupt municipal agencies with a distracting array of duties into highly disciplined, professional anticrime forces. To do this, they tried to make police forces more like military units. Parker’s image of a “’thin blue line’ arrayed against the forces of evil” captured the imagination of police and public alike and helped build that military ethos. Decades after Parker coined it, the image still wields considerable power, showing up not only in self-congratulatory speeches at union meetings, but in police textbooks as well.

This vivid image has many implications. It splits the world into friend and foe. On one side are citizens; on the other, forces of evil; between them, the thin blue line of heroes. The metaphor urges a passive citizenry: Police keep the beast at bay, while citizens are noncombatants, ceding streets to barbarians and police.

The great urbanist Jane Jacobs captured the meaning of the thin blue line perfectly: “The police precinct captain . . . admonished residents not only about hanging around outdoors after dark, but has urged them never to answer their doors without knowing the callers.”

This invitation to passivity is a disaster. In safe and orderly neighborhoods, citizens, not police, are the prime keepers of the peace. As Jacobs put it:

The first thing to understand is that the public peace—the sidewalk and street peace—of cities is not kept primarily by the police, necessary as police are. It is kept primarily by an intricate, almost unconscious, network of voluntary control and standards among the people themselves, and enforced by the people themselves.

In Buffalo the thin blue line model was well in place. To the police operating in that terrorized neighborhood, the residents were not citizens but civilians, not a resource but a potential inconvenience. A strict division was maintained between combatants and noncombatants. Though the war was being fought for the civilians’ ultimate benefit, police automatically assumed that their short-term interest—preventing the very next rape—might have to be sacrificed, exactly the trade-off friendly military forces make while turning the villages of their own countrymen into battlefields.

But even if one believes, as I do, that the Buffalo police acted wrongly in following the thin blue line paradigm, this much can be said for this particular metaphor: It is easily recognized and therefore easily critiqued when it leads us astray.

The power of metaphor to shape public policy reaches its peak when a metaphor becomes so ingrained that it drops out of sight. In other words, metaphors gain ideological power as their literary power fades. When metaphors lose their capacity to attract attention—when they become a linguistic habit—they become dangerous: A trick of language becomes an intellectual trap.

By these standards, the “criminal justice system” is a dangerous metaphor indeed. For over the last thirty years this metaphor, largely unrecognized as such, has radically transformed the way police define their jobs, revolutionizing both police missions and tactics, and powerfully distorting the way we think about crime control. The results have been disastrous, not only for those eight girls in Buffalo, but for millions of other Americans living in large cities.

The idea of a criminal justice system was invented during the 1960s. Until then, phrases such as “penal systems,” “police systems,” and “correctional systems” were bandied about, but the idea that these agencies together comprised a “criminal justice system” originated in a decade-long research project of the American Bar Foundation (ABF), research that yielded a report and five very influential books on the operation of criminal justice agencies, including police, in the United States.

Most of the ABF’s work was remarkably useful. In fact, it was the first serious work of social science to demonstrate that good police work does not consist primarily in making a lot of arrests, but in using a whole array of less formal techniques to keep neighborhoods orderly and safe. Nevertheless, with one fell metaphor the ABF report undid years of its own meticulous research. For the “big” idea from the ABF research that proceeded to capture popular and professional imaginations was the “criminal justice system.”

ABF researchers noted that arrestees went from agency to agency: police to prosecutor, to court, to prison, and so forth. To the accused being processed, these agencies must have appeared somewhat like a system. Moreover, policies and events in one agency affected others. If police increased their number of arrests, prosecutors’ offices had a greater workload. That seemed like a system too.

These superficial appearances were completely misleading. As a metaphor, “system” is originally drawn from the natural sciences, particularly biology and astronomy, where it refers to a coherent natural order, such as the solar system, or the ecosystem, or even the nervous or digestive systems. The parts of a system are not autonomous but interdependent and subordinated to the ends of the system as a whole. Their interactions are governed by built-in feedback systems and a few natural laws.

The notion of systems migrated from the natural sciences to other sciences, especially the social sciences, and to technology, to planned or man-made order. We speak, for example, of computer systems, data systems, and school systems. The system metaphor is especially common in the analysis of business organizations. Peters and Waterman’s In Search of Excellence is heavily influenced by this and other biological metaphors: Organizations inhabit an environment, an ecological niche; organizations adapt in response to other organizations; and internal regulatory processes maintain the system’s distinctive form and character.

None of this is a very good description of our collection of criminal justice agencies and the people passing through them. They obviously are not part of a natural order, but more relevantly, they are not designed or governed by a single or even a collective intelligence. The courts, for instance, govern the police not in pursuit of crime control, but of justice, or even more abstractly, of “constitutionality.” And, of course, they are not the only force governing the police.

These various agencies do not share coherent goals. Judges, juries, and defense attorneys are not “crime fighters.” Nor is it the goal of the prisons to prevent crime; their goals are to rehabilitate, to punish, and to keep order among a dangerous population. Even if one could claim that collectively these agencies aim to prevent crime, that certainly is not their only collective goal.

The several parts of the system do affect each other: More arrests mean more court procedures, which mean more prisoners. More criminals let out on parole or probation mean more marauders on the streets and more overworked cops. But these crude results have little to do with the elegant, self-regulating feedback mechanisms of an ecosystem, or even a good computer system. If the criminal justice system were a computer program, it would be crashing constantly.

Because this collection of agencies and people does not have the characteristics of a system, it cannot do the job of a system: to regulate itself and its parts so that all those parts work together toward a coherent goal.

Nevertheless, the inherent power of the system metaphor was so great that it overwhelmed reality and was soon the cornerstone of the dominant crime-control paradigm. President Johnson’s Commission on Law Enforcement and the Administration of Justice adopted the system metaphor and widely disseminated it in its prestigious report, The Challenge of Crime in a Free Society. The report presented a flow chart entitled “A General View of the Criminal Justice System,” which has appeared in virtually every criminal justice text since then.

Staff of the ABF and the commission went on to prominent careers in powerful institutions that affected criminological theory and criminal justice policy, spreading the system notion to think tanks, foundations, and universities. Between 1967 and 1981, federal agencies that bought into the metaphor pumped $8 billion into various state and local outposts of the “criminal justice system.”

The police loved the idea. Transformed into the front end of a system, they were liberated from acting as a city agency, responsible for a broad array of services or for actually solving problems on their own. Instead, their main job was to introduce offenders into a system by arrest, after which they were somebody else’s problem. The quantity and quality of arrests became the dominant means of evaluating departments and officers. The score could always be counted at the end of the day.

Concentrating on processing criminals through the system, police sloughed off many traditional responsibilities. Traffic enforcement was either largely ignored (as in Boston) or handed off to other agencies (as in New York). Emergency medical care went to other newly created agencies. The job of maintaining order was largely ignored. To the extent that police continued to provide these services, they did so unofficially—their actions neither recorded, recognized, nor rewarded.

The criminal justice system metaphor not only misleads police about their job, it encourages them to ignore the way most citizens experience crime: how it affects their lives, and what they fear. Citizens certainly are distressed about the serious crimes that warrant not only arrest but processing through the rest of system, such as murder, assault, rape, robbery, and burglary. But even in our high-crime era, these remain relatively uncommon occurrences in the daily lives of most citizens.

Every day, however, citizens confront, and are demoralized and frightened by, the incivilities of our streets and public spaces: drunkenness, prostitution, aggressive panhandling, roving youths, drug dealing, farebeating in subways. These experiences are not only morally repugnant, but menacing: Disorder, left untended, increases citizen fear and leads to more severe disorder, resulting, as a good deal of research has shown, in serious crime.

The system notion encourages police to focus on serious crimes and the formal criminal-justice proceedings to which they lead. But citizens want civil and orderly streets on which to conduct business or take their ease without fear or loathing. The relationship between these goals and the formal proceedings of the system is tenuous. If the police are primarily part of the system, they will be of little help in pursuing these goals. Can the system process disorder or incivility?

To the extent that police have recognized a broader definition of crime prevention, it has been through the creation of small, specialized “crime prevention” units that, for the most part, advise residents and small businesses about security hardware. These units are often referred to derisively as being staffed by the “empty holster crowd,” because they do not do “real policing.” “Real police” feed the system.

The criminal justice system concept assumes a theory of crime control that we now know to be untrue. The system consists of the front end (police who try to capture criminals), a middle (prosecutors who convict and judges who sentence), and a back end (prisons, which deter, punish, and rehabilitate). In this model of the criminal justice system, crime is supposedly prevented by primary deterrence (the criminal’s fear that he will be caught and punished), secondary deterrence (potential criminals’ learning from the experience of others who are caught and punished), incarceration (street crimes cannot be committed during imprisonment), or rehabilitation.

The system metaphor inevitably focuses attention on the back end of the process-the penalties that deter, or the programs that rehabilitate, criminals. The more convicted criminals are punished, deterred, or rehabilitated, the less crime there will be, according to the system theory. So for thirty years we have been pouring resources into the enormously expensive back end of the system: more courts, more prisons, more parole officers. The effects on citizens’ lives, and especially the disorder that disrupts urban life the most, have been negligible at best.

I do not mean to claim that deterrence, the primary business of the back end of the system, does not work. To the argument that enormous numbers of arrests and convictions have not slowed crime, the advocates of deterrence have answers: “We do not do deterrence right. We incarcerate for the tenth offense, a murder, rather than the first, a shoplifting, etc.” There is merit in these arguments, but they come down to saying that the system is not working. That being the case, do we really want the police to think of themselves primarily as a part of it?

Freed from the constraints of the system metaphor, we can begin to see that we may have been getting the crime-control equation backwards. Perhaps the answer is not to feed the back end of the “system,” but to exploit the capacity of police to control disorder, solve problems, and prevent crimes from happening. If the police are far more than the front end of a system, if they can, by their own actions, prevent crime, then the obvious strategy is to pour more resources into controlling crime before it happens, rather than controlling criminals afterwards.

What, over the last thirty years, has the “system” produced? An endless temptation to spend money. The image of a system induces us to try to create a fiscal balance between the parts. More police mean more criminals arrested, more arrestees mean more prosecutors and judges to convict, more convicts mean more prisons and more parole and probation offices. But perhaps that idea is wrong. Perhaps instead of spreading resources evenly over a system to process criminals, we need to concentrate them on the agencies that prevent crime. Perhaps, to put it bluntly, we need fewer prisons and far more cops—not cops who will feed the system, but cops who will starve it by helping communities protect themselves.

The system’s “back end,” courts and prisons, have become a source and a repository of frustration and cynicism. Police forces keep feeding that system, but the crime problem remains unsolved. How much police morale is wasted on complaints about the failures of the back end: the criminal released on technical violations, the prisoner freed for lack of space, the long waits in overcrowded courts, plea bargains, and blown prosecutions? Even worse, as anyone who spends a lot of time with cops can tell you, the system mentality provides police with an enervating plethora of excuses for why the streets don’t get any safer. After all, the police are doing their job; it is the rest of the system that is constantly screwing up. And since it is the system as a whole that is supposed to reduce crime, say the cops, critics had best look to the parts of the system that aren’t working.

How much energy and initiative might be released if, instead, police morale were based on the ability of the police themselves to restore order and prevent citizens from becoming victims in the first place, on pursuing goals not utterly dependent on the back end of the system doing its job?

It is worth considering, especially if one examines the most responsive, and neglected, police agents: private cops.

In 1990 the ratio of private police to public police in the United States was estimated at 2.4 to 1. By the year 2000, according to some estimates, it will be close to 3 to 1. Yet for criminal justice experts, private security hardly exists. It has attracted very little important research. Its implications for crime control have largely been ignored by politicians, analysts, and the police themselves. Perhaps this is not too surprising: The President’s Commission, enchanted with the idea of the criminal justice system, virtually ignored private security.

Private security’s mission is not to feed a system, but to prevent losses from crime. It does this first by preventing crimes; second, by limiting the damage when a crime occurs; and third, by restoring the protected organization’s (or community’s) capacity to conduct its business effectively.

These priorities make sense. If Buffalo police had not been trapped in their role in “the system,” they might have acted differently. They might have remembered that the first responsibility of police in such a situation is to prevent children from being raped in the first place. If a rape occurs, their primary responsibility remains to prevent another child from being raped. Nor is preventing future rapes the only way of limiting the damage: A more comprehensive approach would include helping children and the neighborhood cope with fear. Children, aware of the danger of a sexual predator, are traumatized. Many are grieving for their friends. They need reassurance, information, and understanding from authority figures, including the police.

Moreover, in the problem-solving or private security paradigm, police would help in the healing of the community, not only because that is part of the job of limiting the damage caused by the criminal, or because healing is a good thing. They would do it also because a healthy community will better resist further assaults, because even a horrible crime need not be utterly meaningless if it can help the community to prevent future horrors. All these efforts require a much broader approach than the criminal-catching activities that currently dominate the police departments of Buffalo and other major cities.

The dominant metaphor for private cops is not the system but the (private) eye. Unlike the system, which processes the past into an endless future, the eye sees and guards and prevents in the present tense. If the prime mode of the criminal justice system is deterrence, that of the private eye is vigilance.

In romantic fiction, of course, the very allure of the private eye comes from what makes him seem dangerous as well as effective: He is truly “private”; he works alone, often bending and sometimes breaking the rules. He does not buck problems further up the system; he solves them here and now, his way, with the resources at his disposal. Sometimes he has to get rough.

We do not want our police to be private in that sense—lone romantic figures doing justice as they will. They would then breed more problems than they solve. But within the rules we do want them to take responsibility. We want them to solve problems, not alone, but by helping citizens stop crime.

Jane Jacobs, in describing how citizens help stop crime, wrote of “eyes on the street.” Those eyes were private too—they were citizens, not police—but not private in the sense of being alone. We want not a lone romantic figure, violating the rules of the community to save it, but vigilant aides, helping the community guard itself, with all eyes open.

It is not hard to understand the attractiveness of the “thin blue line” and the “criminal justice system” metaphors. It is far easier to count how many arrests officers make than how effective they are educating citizens or organizing a community. The idea of a system with inputs and outcomes seems more dynamic than informal peacekeeping. “Fighting crime” is much more dashing than “keeping peace.”

These metaphors, however, not only shape but constrain professional thinking about crime control. Not only Buffalo’s citizens but Buffalo’s cops are victimized by them. The metaphors set the police up for failure, for the police cannot deliver what the metaphors promise. They cannot place every citizen behind their thin blue line. They cannot feed criminals into the system fast enough to win a war of attrition: Upping the body count is no more workable a strategy in our streets than it was in Vietnam. We must pull our police forces out of the system and put them back into the community.


City Journal is a publication of the Manhattan Institute for Policy Research (MI), a leading free-market think tank. Are you interested in supporting the magazine? As a 501(c)(3) nonprofit, donations in support of MI and City Journal are fully tax-deductible as provided by law (EIN #13-2912529).

Further Reading

Up Next