Preventing Crime

Defining intent, not just acts, as criminal was one early move in the state's larger game of preventing crime. We have seen that deter­ring through public punishment was rudimentary at best.

The virtues of prevention could be argued at various levels. At the most general, social conditions caused crime. The immediate offenders could be deterred, incapacitated, or rehabilitated. But the ultimate culprit was society. Social conditions—whether poverty, inequality, exploitation, familial breakdown, or anomie—caused crime. Social reform therefore promised to diminish offending. The belief in the virtues of social engineering is ancient and has become evermore pervasive. Crime in this view is like disease, to be cured not punished. The prevention of crime aimed at by most active pol­icy makers has, in contrast, been narrower and more modest. It has sought not major social reform, much less revolution, but techniques to nab offending in the bud.

Offenders form their intent, prepare to carry it out, and finally commit the act. Laws sought to interrupt this causal chain at var­ious points: closer to the intent or to the act.

Prevention thus moved the philosophical basis of punishment away from the idea that offenders should make good the harm done and that retribution righted society's moral balance. With pre­vention, justice ceded pride of place to a victorious utilitarianism, concerned primarily with order and tranquility. The Enlightenment principles commonly thought to govern sanctions faded. Punish­ments no longer fit the crime. Nor were they announced clearly and determinately in advance for precisely specified behaviors so that, as sovereign citizens, potential offenders could calibrate their own conduct. Sentences grew increasingly indeterminate and discretion­ary, tailored to offenders, not to crimes, and fine-tuned for reha­bilitation more than for justice.¹⁶⁴ The clear relationship between offense and sanction was severed. Once the goal shifted from just retribution to effective prevention, the crime did not necessarily determine its punishment.¹⁶⁵ A crime might not be punished at all if that promised to spare society an offense to come. Or the state might impose punishment even on someone who had commit­ted no tangible act. Prevention overturned many of justice's usual assumptions. Criminals were defined not by what they had done but what they might perhaps do. Attending to intent broadened the state's remit.

Preventing crime meant identifying criminals before they acted. The easiest prediction was based on past behavior.

Modern methods of identifying people, undercutting their abil­ity to game the system by using aliases and the like, have allowed contemporary statistics to pinpoint just how much crime is commit­ted by how few.¹⁶⁸ Since the 1970s in the United States, 6 percent of offenders have committed half of all crime. In 1986, 3 percent of Minneapolis street addresses were the destination of half of all police dispatches. For robbery, criminal sexual conduct, and auto theft, fully 100 percent of dispatches went to 5 percent of all locations.¹⁶⁹ Conversely, 95 percent of urban space is altogether free of predatory crime.¹⁷⁰ By targeting crime hot spots, the authorities could antici­pate and prevent crime. By policing St. Giles, as a London police commissioner put it early in the nineteenth century, they were polic­ing St.

Prevention targeted not the offense but the kind of person who might commit it in the first place. It punished character more than crime. This approach had a venerable pedigree. Roman law allowed judges to accept as proven any information about offenders alleg­edly known by everyone to be true.¹⁷² Among the sixth-century Franks, someone identified as a criminal by upstanding community members could be convicted on that basis alone. Under Char­lemagne, the Rugeverfahren permitted judges to begin an inquest on the basis of an offender's reputation.¹⁷³ Early medieval law allowed only those of good reputation to swear oaths to clear themselves. For others, it demanded compurgators to stake their reputations alongside the suspect, many of them if the accused's character was not spotless. On occasion, more than a thousand were assembled. As the Roman law inquisitorial process was reintroduced in twelfth­century Europe, ecclesiastical courts accepted the doctrine of mala fama, “bad reputation.”¹⁷⁴ An ecclesiastical judge could now try a sus­pect without a specific accusation or accuser.¹⁷⁵ Freemen were sorted according to their reputations, leaving the bad eggs with less legal standing. Twelfth-century German courts could prosecute on repute (Leumund), with officials swearing to the blemished standing of the accused.¹⁷⁶ A notorious suspect could be accused without other evi­dence, required to swear a purgative oath, and punished on failure to do so. Priests suspected of living in sin and eventually heretics, too, were targeted by such techniques.¹⁷⁷

Prevention targeted the offender more than the offense. To get out in front of the act, the state had to grapple with the actor. But how? Blackstone lauded preventive over punitive justice as superior in all respects. The matter, he thought, was simple: preventive jus­tice merely meant obliging those persons whom there was “prob­able ground” to suspect of future crimes to assure the public that they would not offend.¹⁷⁸ Rarely had a seemingly innocuous phrase glossed over complications so glibly.

How to anticipate the future offender? Two approaches prom­ised help: psychology and sociology. Offenders might be predicted psychologically if discernable traits revealed a propensity to crime. Lombroso's theory of the born criminal was influential in the late nineteenth century, harnessing biology and its degeneration to explain how the criminal psychology was formed. Countless other now discarded theories such as phrenology also claimed to discern transgressive proclivities through somatic or psychological indica­tors.¹⁷⁹ More recently, criminology has sought to predict dangerous­ness, identifying those who threaten to become offenders. In the 1960s and 1970s, measures targeting dangerousness, apart from any actual offense, were introduced in response to shocking crimes com­mitted by released inmates. Sociopaths were identified as those who inflicted suffering and injury for fun, lacked compassion and impulse control, saw themselves as victims, and resented authority. Nonethe­less, having failed in the past reliably to identify the criminal mind, now chastened psychiatrists also cautioned that testing for danger­ousness was almost impossible and that standard psychiatric diag­nosis was largely irrelevant. A major study of a large group of the supposedly criminally insane in 1966 in New York revealed them to be sad old men far more often than predatory psychopaths.¹⁸⁰

Besides magic and witchcraft, torture has been the most venerable attempt to probe the soul's secrets. The psychological sciences sought to follow suit more methodically, yet their promise disappointed. Fearful of false positives, clinicians tended to forecast conserva­tively who among their patients might offend, thereby hobbling their predictive acumen.¹⁸¹ Excepting the seriously mentally ill and perhaps drug abusers, few psy-forecasts managed accurately to pre­dict future offending.¹⁸² Even new technologies have done little to improve this track record.

Hopes of reading others' minds likely only shortly postdated Eve's encounter with the serpent. The correlation between lying and arousing the sympathetic nervous system was operationalized already in ancient China. Suspected liars had to take a mouthful of rice and spit it out.¹⁸⁴ Those with dry mouths found the task harder than normally salivating innocents. It was a trope of racist ideology that Blacks and Jews—both allegedly incapable of blushing—were inferior because their interior states were less visible to others.¹⁸⁵ Mechanized lie-detection technology began with Lombroso's hydro- sphygmograph to measure pulse rates, followed by John Larson's polygraph in the 1920s, correlating deception with increased blood pressure. The polygraph was sold to police and public as a means of getting at the truth that sidestepped the need for more primitive and violent techniques.¹⁸⁶ Other methods followed, all sharing the assumption that somatic reactions revealed inner states through a Pinocchio response, whether respiration rates, epidermal conduc­tivity, voice stress, heat around the eyeballs, fleeting facial micro expressions, or millisecond hesitations.¹⁸⁷

MRI scans and electroencephalography went further, measur­ing something seemingly closer to the inner workings of the brain, though still merely a heightened metabolic activity in certain cere­bral areas.¹⁸⁸ Such technologies assumed that lying was more ardu­ous (a greater cognitive load) than telling the truth, thus stimulating the responses detected. They all were unable to distinguish dissem­bling as such from other brain activities. Nor were they able to locate deception in specific and specialized areas of the brain. Moreover, liars could also game the system, all the more so when they knew what investigators were looking for. On the basis of no particular evidence, it is, for example, widely believed that liars avert their gaze, so those hoping to appear truthful now catch and hold their interlocuter's eye.¹⁸⁹ Perhaps we are entering a new era of predictive investigation, but attempts to discern individual thoughts by invol­untary physical responses have yet to prove very successful.

If the psy-sciences did not fire prediction's magic bullet, that left sociology. Potential offenders could also be identified by extrapolating from statistically characteristic behavior of groups they belonged to. Regularities were identified to induce forecasts of conduct, using the logic of social science developed by the nineteenth-century Belgian astronomer Adolphe Quetelet, who thought that society followed patterns as discoverable as nature's.¹⁹⁰ Offending was linked statisti­cally with various characteristics to develop a sociological criminal type. Membership of a group was associated with certain behav­iors, which in turn could be used to define the group.¹⁹¹ The group could thus be targeted, and individual members deemed potentially culpable—though merely for sharing the features for which the col­lective name was shorthand. Armed with—at best—probabilistic social science, prevention thus focused on the criminal more than on the crime. Did offenders belong to high-risk groups? Were they recid­ivists? Did they have a propensity for certain crimes? A penchant for repeated transgressions? A disposition for specific victims? Most gen­erally, were they dangerous, posing an ongoing threat? The answers determined the appropriate punishment more than did the nature of any (eventual) offense, thus weakening retributive justice's link between act and desert. Retributionists' focus on culpability for the act treated each offense alike, even those committed subsequently by the same person. Targeting some people as especially dangerous and even treating recidivists more harshly, in contrast, meant punishing the person as much as the act.

In this logic, those predicted to recommit were to receive differ­ent sentences from onetimers.¹⁹² Kleptomaniacs and drug addicts should be treated more harshly than one-off opportunistic offend­ers even though, acting under a psychological compulsion, they were—according to a retributive logic—less culpable. Sentences of treatment rather than punition also made sense if they promised less crime.¹⁹³ Inchoate crimes were by definition those that had not been consummated. Of interest, therefore, was perpetrators' intent and what it revealed about their state of mind, inclinations, and personality. Someone who would attempt, conspire about, solicit, or incite a crime posed a threat. Anyone willing to attempt an offense was likely to repeat it. Punishing attempts aimed foremost to neutralize threatening individuals and only secondarily to deter their offenses. That offenders were dangerous mattered more than whether they were guilty.¹⁹⁴

Sociological or actuarial prediction raised particular problems. It was static. The characteristics that defined someone as likely to offend— poverty, unemployment, residence in certain neighborhoods—could at best predict lifetime, not imminent, risk.¹⁹⁵ Conversely, predicting lifetime offending on the basis of recidivism, as three-strikes laws claimed, was hampered by the tendency for transgressing to drop off with age. Former serial offenders were locked up just as they lapsed into crimogenic senescence.¹⁹⁶ To avoid unacceptable false negatives—releasing some apparently harmless persons who then went on to offend—meant tolerating many false positives, which kept innocents behind bars.¹⁹⁷ The former set off political fireworks; the latter were rarely heard from again.

Not only did actuarial forecasting overpredict, but its logic was also circular: those who had stolen were likely to have been unem­ployed; the former unemployed were therefore potential thieves. Not-yet offenders thus had their future behavior forecast on the basis of demographic, economic, and social traits they shared with already offenders. They were held accountable for characteristics that had proven correlative for past culprits, though not yet shown to be causally determinative for them. In effect, they were punished for the crimes of others. Even sophisticated sociology could not sur­mount the dilemma famously identified by David Hume: how to break out of correlation into causality. However refined the actu­arial calculations, they still took a leap of faith from past behavior to future actions.

Since the popular mind confused correlation with causality, crime prevention reinforced the conceptual shorthands we call ste­reotypes. That the overwhelming majority of rapists are men does not mean that all men are rapists, except perhaps in the fevered imagination of Val in Marilyn French's pathbreaking feminist novel The Women's Room. But that African American men are charged with crimes disproportionately to their presence in the population has been used to justify the racial profiling that subjects them to more than their fair share of preventive encounters with the police. Sim­ilar discrimination confronts Middle Eastern travelers at airports. Such prejudices have been conceptually cemented throughout his­tory, leaving traces in our vocabulary. Vandals, barbarians, philis­tines, peons, banshees, troglodytes, plebians, Huns, sycophants, thugs, villains, beggars, buggers, coolies, bohemians, berserkers, boors, and the other n-words of yore—all were used at one time or another for eth­nic, national, or (quasi-) occupational groups now immortalized for their pejorative traits. In 1682, Louis XIV banished all “Bohemians” and “Egyptians” from France, by which he meant Roma.¹⁹⁸

The inherent unfairness of this sociological logic meant that laws based on it faced obstacles. That many women, not just prostitutes, were swept up in police dragnets and inspected for venereal disease in nineteenth-century England helped marshal opposition to the Con­tagious Disease Acts. Vagrancy laws are ancient, permitting police to harass or jail largely anyone found in public.¹⁹⁹ Antiloitering laws give police broad discretion to target otherwise legal activity, such as “wandering or strolling around from place to place without any law­ful purpose or object.”²⁰⁰ Broadly speaking, any car driver is fair game for police attention.²⁰¹ In recent years, courts have struck down some such statutes as too vague, all-inclusive, and broad, thereby justifying the targeting of racial, ethnic, and sexual groups only tangentially correlated to the relevant crimogenic characteristics. In response, the statutes have been refocused on more specific behaviors, such as loitering with various intents. Or they have been aimed at crime hot spots rather than at entire cities—all to avoid violating the rights of innocent bystanders who happen to share certain characteristics with the gang members or other groups being targeted.²⁰²

Psychological and sociological techniques of predicting and preventing crime alike raised problems. Modern law enforcement has therefore tended to retreat to the simplest means of forecast­ing of all: past behavior. Past actions were empirically and morally a more solid foundation for predicting future acts than actuarially based demographic or sociological correlations.²⁰³ How often and how seriously someone had already offended have become the most heavily weighted factors in predicting future transgression.²⁰⁴ Absent a belief in atonement or remorse (hard to test for its sin­cerity), past offenders were assumed likely to repeat their crimes, thereby condemned to a vicious circle.

However prevention formulated its predictions, it thus focused more on the criminal than on the crime. Unlike reactive policing, with culprits who could in theory be identified, prevention did not know who was going to offend. Yet it had to narrow its focus to only some citizens. Even before any crime had been committed, it too needed suspects. Bad characters, habitual offenders, dangerous classes, social parasites, objective enemies, recidivists: through the ages, such designations were shorthand for characteristics consid­ered indicative of likely criminality.²⁰⁵ Profiling is today's word for this use of extrapolation from demographic, social, ethnic, eco­nomic, behavioral, or other indicators to identify likely offenders.