Federal Offense

The politics of crime control is, therefore, sometimes a simple matter of arithmetic. There are, of course, more crime victims than criminal defendants, particularly among the voting public, so there are likely to be more conservatives than liberals on the subject of crime--many more. Polling data indicate that nearly 80 percent of the American public supports the death penalty in the absence of an alternative such as life without parole, and this in turn suggests that nearly 80 percent of Americans fear being murdered more than they fear being convicted of murder.

Conservative approaches to crime thus have a natural advantage over liberal ones--an advantage that has been decisive during the recent years of intense social anxiety. Liberalism held sway briefly during the 1960s, at least until the Nixon law-and-order campaign of 1968, not because crime was down (it rose sharply) but, in part, because hope was up. Hope fueled the War on Poverty, the civil-rights movement, and feminism; even the angriest protests of the Vietnam War reflected hope for the possibility of peace.

Today hope seems as out-of-date as beehive hairdos, reaching improbably toward the sky. The hopeful notion that prisons might rehabilitate people has long been dismissed as naive, displaced by a belief in retributive justice and the demand that prisons serve as places of near permanent exile for the incorrigible among us. Some liberals still protest America's uniquely high incarceration rate, tirelessly pointing out that we imprison more people per capita than any other country in the world, but a majority of Americans favor building more prisons, despite their cost, and believe that sentencing practices are excessively lenient.

At the state and federal levels this year's answer to violent crime has been the three-time-loser statute, imposing mandatory life sentences without parole on repeat violent offenders, without much regard for the nature of their crimes or their characters. In Washington, which last November became by referendum the first state to enact a three-time loser law, qualifying "violent" felonies include drunk driving, promoting prostitution, and petty theft. One unintended consequence of this law may be an increase of violence on the streets: preliminary anecdotal reports from the police suggest that when cornered, offenders may shoot their way out rather than surrender and face life in prison. Three-time-loser laws have also been widely criticized as arbitrary and as burdensome to the nation's prisons and courts. (Such laws are bound to increase the number of trials, since people aren't likely to plead to life imprisonment, resulting in still more delays throughout the system.) Nevertheless, the drive to imprison more people for longer periods of time seems unstoppable, fueled not just by fury and fear but by a sense of resignation. It is as if all we can do is warehouse people until they die or are too old and decrepit to threaten anyone on the outside again.

AN UNCHANGEABLE FACT OF LIFE?

Public dismay about crime poses more than a rhetorical challenge to an Administration that came to power urging people to vote their hopes, not their fears. Adopting violence-prevention programs, as Attorney General Janet Reno has advocated for many months, requires hope that the federal government in partnership with localities can treat the social pathologies that give rise to crime. Indeed, the assumption that there are identifiable, treatable connections between these pathologies--for example, the neglect and abuse of children--and crime reflects the hopeful notion that violent people are made, not born, and can sometimes be unmade, if intervention comes early enough. Along with faith in the curative powers of good government, programs to treat and prevent violent behavior reflect faith in the malleability of human beings and the capacity of distressed people and communities for self-improvement.

Despite his sunniness about America, Ronald Reagan presented a much darker view of criminality, which still holds popular appeal. There are no social solutions to crime, Reagan asserted in 1981, because crime is not a social problem; "it's a problem of the human heart." Reagan cited what he viewed as a dual liberal fallacy about crime--the conviction that ameliorating poverty might reduce crime and the assumption that "there [is] nothing permanent or absolute about man's nature."

This bleak vision of crime as an unchangeable fact of life, implicitly equating crime with original sin, dismisses liberalism as utopian. Government can respond to the symptoms with arrest and incarceration, but only God can treat the disease, Reagan implied.

The public-policy expert James Q. Wilson offered a more thoughtful, less biblical variation on this theme in his landmark book Thinking About Crime, in which he chided liberals for seeking to treat the root causes of crime. In 1968 John Lindsay, then the mayor of New York City, gave a typical definition of these causes: "the poverty that grips over thirty million of our citizens." Lindsay added, "If we are to eliminate the crime and violence in this country, we must eliminate the hopelessness, futility, and alienation from which they spring." It's not that social reasons for crime don't exist, Wilson wrote, but they are practically impossible to address. "The more we understand the causes of crime, the more we are drawn into the complex and subtle world of attitudes, predispositions, and beliefs, a world in which planned intervention is exceptionally difficult." In Wilson's view, the liberal fallacy was the notion that "no problem is adequately addressed unless its causes are eliminated."

Wilson argued that policy is most effective when it focuses on objective matters like the costs and benefits of crime, not the realm of "the subjective and the familial." He posited criminals as essentially rational human beings who would be deterred from committing crimes when the associated costs became impractical.

It's impossible to know how many violent offenders weigh the consequences of their actions. (It's likely that many who do engage in cost-benefit analyses never get past assuming they won't be arrested.) Certainly thirteen-year-old boys with guns and a wide range of neurotics, psychotics, and apparently sane, smart killers like Gary Gilmore tend to act impulsively, without regard for harsh sentencing laws. Nor would the prospect of a prison sentence necessarily deter a young male for whom imprisonment has become a rite of passage or a haven from the street: "three hots and a cot." A sentence might not even distress his parents. One Boston defense attorney remembers the first time the mother of a client encouraged her not to free him from prison. "At least I'll know where he is," she said.

The poignancy of this appeal, and the threatened destruction of several young generations, may force liberals and conservatives alike to reconsider some of their most basic assumptions. "Those mothers in the ghetto love their children," Senator Orrin Hatch asserts, with all the force of revelation, and he acknowledges that poverty is related to crime (as is a welfare system that "does not stimulate the desire to make something of yourself"). "Some people become locked in poverty and become embittered in the process," Hatch says. "That's why drugs are so rampant in our society. Kids in the ghetto can make more money pushing drugs than working for the minimum wage." Meanwhile, Jesse Jackson is preaching self-help, an ethic of individual and communal responsibility, and telling African-Americans to "look inward to go forward." President Clinton is emboldened to exhort an African-American congregation to address the violence within their own community.

MANDATORY MINIMUMS

The new talk about crime control has not, however, been a prelude to new programs. It has long been clear that the President is willing to give people the "tough" law-enforcement measures they seem to want--death penalties and three-time-loser laws--and the Senate has been even more anxious to demonstrate toughness on crime. The Senate's $23 billion crime bill, promising more prisons and police officers, longer sentences, and some fifty new federal death penalties, was passed in a few frenzied weeks last fall, after state and local elections that seemed to crystallize voters' concern about crime.

"If someone came to the Senate floor and said we should barbwire the ankles of anyone who jaywalks, I think it would pass," Senator Joseph Biden said, shortly after voting with ninety of his colleagues to impose mandatory life sentences on people with the bad luck or bad judgment to commit their third "violent" felony on federal property. This bill would have particularly harsh effects on Native Americans who live on federally owned reservations, but it would not affect the majority of offenders nationwide. The small number to whom it could apply would be bound to include people who were more hapless than vicious, given the expansive definitions of "violent" incorporated in the bill. The only vote against this proposal, which promised to be both harsh and ineffective, was cast by Bob Packwood, conscience of the Senate.

Three-time-loser laws need not, of course, be written this stupidly. (The House has offered a narrower, smarter law.) But the more narrowly these laws are drafted, the more redundant they may be. Violent crimes already carry lengthy sentences, and many states have career-criminal laws that impose strict mandatory terms on repeat felony offenders. (In the states that have career-criminal laws, crime rates have generally stayed in line with those in the rest of the country.) The problem that is supposed to be addressed by three-time-loser laws--the early release of violent felons- might best be solved not by inflexible life sentences but by more rational allocations of prison space, decreases in the sentences of nonviolent offenders, and increases in strictly supervised alternatives to prison. More felons are put back on the street by prison overcrowding than by soft-hearted judges or legislators--creatures more mythic than real.

If, however, it is easy to imagine an arguably reasonable, narrowly drawn three-time-loser law, it is practically impossible to imagine state or federal legislators actually drafting one. Limiting the application of a three-time-loser law also limits its political appeal. A legislator who voted to allow for the release of geriatric prisoners after twenty-five or thirty years, or to exempt from the list of qualifying felonies burglaries or robberies in which no one is harmed, would probably be labeled soft on crime. And the recent history of congressional action on sentences demonstrates that lawmakers are less concerned with reducing crime than with signaling their concern about it.

Congress has traditionally committed itself to the promise of strict, sure sentencing, with little regard for consequences or facts. Consider federal sentencing laws. From 1984 to 1990 Congress every two years demonstrated its toughness on crime by enacting mandatory minimum sentences for various drug and firearms offenses, including drug offenses committed near schools, violent crimes (defined broadly) or drug crimes involving the use of a firearm (the firearm triggers a mandatory sentence in addition to the sentence for the underlying offense), the possession of more than five grams of crack cocaine, and any degree of involvement in a drug conspiracy.

These laws have had no demonstrable effect on drug- or gun-related violence but they have greatly increased the number of people taking up space in federal prisons for nonviolent, low-level drug offenses. According to one study, in the year ending September, 1992, more than 3,000 drug offenders with no record of violent crime in the previous fifteen years were sentenced to minimum terms of at least five years. From 1984 to 1990 the proportion of people sent to federal prison for drug offenses rose 12 percent, while the proportion of people sentenced for violent crimes and property offenses declined, according to a study by the Federal Judicial Center. Today more than 60 percent of all federal prisoners are serving time for drug offenses, including the simple possession of marijuana or cocaine.

These statistics are hardly news to legislators or policymakers, and stories about people spending five to ten years in federal prison for playing minimal roles in drug conspiracies have begun to appear in the mainstream press. Last November, The New York Times featured the case of a twenty-four-year-old man serving ten years in a federal prison because he agreed to help a federal undercover agent find someone selling LSD at a Grateful Dead concert. Relatives of people like this have formed a lobbying group, Families Against Mandatory Minimums, headed by Julie Stewart, whose brother is serving five years for a first offense--growing marijuana at home. Critics of mandatory minimums range from Janet Reno to Orrin Hatch. Hatch does not oppose mandatory minimums in principle, but he acknowledges that they have been "overused" for low-level, nonviolent crimes. Reno has been critical of mandatory minimums (though her criticisms do not appear to have had much effect on Administration policy).

There is also strong opposition to mandatory sentencing from federal judges, both Republican and Democratic appointees. In one survey a majority of judges and a majority of probation officers, in addition to most defense attorneys, viewed mandatory minimums unfavorably, as did about a third of federal prosecutors.

Widespread, bipartisan dissatisfaction with mandatory minimums, however, is not likely to effect any meaningful reform. Senator Biden, who does not generally support mandatory minimums, considers proposals to reform them utterly unrealistic: "I can count: eighty percent of the Congress and eighty-five percent of the public still believes the misinformed rhetoric about sentencing."

Much of what people don't know about mandatory minimums is included in a 1991 report on them by the U.S. Sentencing Commission, which gently recommended their repeal. The Sentencing Commission was established by Congress in 1984 and charged with promulgating guidelines for sentences in federal cases. This represented a revolution in sentencing which was initially sparked by liberal protests about undue sentence disparity resulting from the broad discretion exercised by judges and parole boards.

A bill to establish a commission had been introduced by Senator Ted Kennedy in 1975. It took nearly ten years for Kennedy's bill to pass the Senate, and by the time the commission was established, the political climate had changed considerably. As Kate Stith and Steve Koh point out in a 1993 article in the Wake Forest Law Review, sentence reform was "conceived by liberal reformers as an anti-imprisonment and antidiscrimination measure, but finally born as part of a more conservative law-and-order crime control measure."

Federal sentencing guidelines are stringent and are generally reviled by defense attorneys. They are also quite rigid and complex. Judges are required to rely on a sentencing table, established by the commission, which lists generic offense-severity levels in a vertical column and offenders' characteristics (mainly criminal histories) horizontally, creating a grid on which the judge locates the appropriate sentence range, which is fairly narrow. Judges are not supposed to consider mitigating factors, such as a defendant's family and community ties, employment history, or emotional and mental state. Sentencing becomes a technical task, not one requiring thinking or judgment. Judges have only limited discretion to depart from the guidelines, which have thus always been quite controversial. (The guidelines were challenged on constitutional grounds but upheld by the Supreme Court in 1989.) Still, sentencing guidelines don't eliminate all judicial discretion from sentencing, as the mandatory minimums do. If Congress had stopped at establishing the Sentencing Commission and not gone on to enact so many new mandatory minimums, it would have avoided much of the controversy about sentencing today.

"PATHETIC CASES COME ALONG"

The establishment of a sentencing commission, however, isn't nearly as dramatic an anti-crime measure as the passage of mandatory-minimum statutes carrying sentences of many years. So, in the same year that Congress empowered the commission to create sentencing guidelines, it enacted new mandatory minimums for drug offenses. This wasn't quite a revolution in sentencing. Federal mandatory minimums date back to 1790. Today, the Sentencing Commission points out, federal penal law comprises a patchwork of about a hundred mandatory-minimum provisions contained in sixty statutes, which make up a little social history of crime. Like the federalization of local crimes, the enactment of mandatory minimums is a congressional show of concern for the crimes of the times. In 1790 piracy carried a mandatory life sentence; causing a vessel to run aground by use of false light carried ten years. In 1888 the bribery of a harbor inspector in Baltimore or New York was made subject to a mandatory minimum sentence of six months. In 1913 a mandatory prison sentence or fine was imposed on commodities-price fixing. In 1934 homicide or kidnapping committed during a bank robbery or larceny was made subject to a mandatory minimum sentence of ten years. In 1965 mandatory life imprisonment was applied to the murder of a President or a member of the executive staff. In the 1980s Congress turned its attention to drug offenses.

All these laws are still on the books; few are enforced. According to the Sentencing Commission, nearly 60,000 offenders were sentenced under mandatory minimums from 1984 to 1990; of the sixty mandatory minimum statutes only four accounted for 94 percent of these cases, which related to drug offenses or carrying a firearm during a drug-related or violent crime. Of the nearly 7,000 defendants sentenced in fiscal year 1990 under mandatory minimums, 91 percent were primarily drug offenders.

There are, then, essentially two federal sentencing systems in place. People who are not charged under statutes carrying mandatory sentences are sentenced under the commission guidelines. Since the guidelines are stringent, mandatory minimums seem at best redundant when applied to serious cases. At worst, and quite often, they're arbitrary and excessively harsh.

"Pathetic cases come along," the federal appeals court judge Stephen Breyer remarked in a speech to the American Bar Association last August. "No one will ever formulate a system of law for which you don't have to have exceptions." Breyer tells the story of a bank-robbery case involving "a man with the IQ of a seven-year-old who got a toy gun, went to a bank, got seventy dollars to get an operation for his dog, his best friend in the world, turned himself in to the FBI, and the dog died anyway. What should we do? Give him life?"

As Judge Breyer's story suggests, mandatory minimum sentences hold people strictly liable for their acts, regardless of any mitigating circumstances, which is another way of saying that mandatory minimums preclude individualized determinations of accountability--which is another way of saying they're un-American.

Virtually all of us act immorally if not illegally at least on occasion. We'd like to believe that if we ever get caught or feel compelled to confess, we'll be judged not simply by what we've done but by why we did it and who we are. In other words, we want our acts to be judged in the context of motivation and personal history, and a concept of character that involves more than our worst offenses. People do sometimes commit bad acts for good reasons, which means that guilt--and especially sentences -ought to be determined by considering the actor as well as the act. Two people may behave similarly under different circumstances, calling for different standards of accountability. A retarded man who holds up a bank with a toy gun in order to save his dog ought to be treated more leniently than, say, a successful college student who holds up a bank with a toy gun in order to buy a new car. Even in hard cases involving felony assaults or homicides, motive and character should play some role in sentencing. A woman who shoots her husband because he has been beating her and her children for several years seems less culpable than a woman who shoots her husband for money. Imposing the same mandatory sentence in both cases would be imposing punishment without regard to guilt.

LAW VERSUS JUSTICE

It's not surprising that people who bring a sense of justice to the system of mandatory minimums often find themselves at odds with the law. Judge Breyer observes that mandatory-minimum-sentencing schemes encourage disrespect and disregard for the law among the people charged with enforcing it. "You cannot tell human beings to do things they think are totally unfair. They won't do it. They'll figure a way out." The judge may be constrained from imposing a lesser sentence, but "the prosecutor won't prosecute, the juries won't convict." In fact, according to Breyer, many federal prosecutors are using their discretion not to prosecute under mandatory minimums. Mandatory-minimum statutes do not eliminate discretion or disparities in sentencing. They shift discretion from judges to prosecutors.

Pursuant to mandatory-minimum-sentencing laws, prosecutors essentially decide how defendants will be sentenced when they decide how defendants will be charged. This clear usurpation of judicial authority is justified as a way of providing prosecutors with an important bargaining tool: they use the threat of indictment under a mandatory minimum statute to persuade defendants to cooperate and inform.

But as a practical matter, the defendants most likely to have information for which prosecutors will bargain tend to be more rather than less culpable in the criminal enterprise: the lower you are in a drug deal, the less you're likely to know, and the less bargaining power you have to stave off an indictment. Federal law also allows judges to depart from the mandatory minimums after conviction and impose lesser sentences if prosecutors request them, claiming that defendants have provided "substantial assistance"--which, once again, tends to reward higher-level defendants and penalize the lowly.

It's not surprising that under this system sentence disparities abound. According to a 1992 report by the Federal Judicial Center, as late as the first half of 1990 nearly 50 percent of offenders who appeared eligible for a mandatory minimum term received a lesser sentence. In addition to disparate treatment of higher- and lower-level defendants, there are regional and racial differences in sentencing practices. The Federal Judicial Center report observed that although blacks have always tended to receive longer sentences than whites, sentence disparities increased following the introduction of new mandatory minimums in the 1980s. Examining data from 1989 and 1990, the Sentencing Commission found that 68 percent of blacks received sentences at or above the minimum, whereas only 57 percent of Hispanics and 54 percent of whites did. Significant racial differences in sentencing remained when the commission took into account the severity of the offense. (In contrast, the commission found, offense behavior did seem to account for differences in the sentencing of males and females.) Calling for further research, the commission nevertheless reported that its findings imply that race plays a role in sentence disparities, which have yet to be acknowledged, much less addressed, by Congress.

If Congress were genuinely interested in truth in sentencing, it would make clear to the public how erratically and arbitrarily mandatory minimums are enforced. It would point out that federal sentencing guidelines already call for very tough sentences for serious crimes. It would explain that defendants are not faceless or fungible: they and their circumstances vary, even when they are charged with the same crime, and judges must have at least limited discretion to vary their sentences.

But there are so few ways for Congress to demonstrate toughness on crime. The Senate passed new mandatory minimums last November, with no apparent opposition from the Administration. As Orrin Hatch concedes, "Mandatory minimums are a political response to violent crime. Let's be honest about it. It's awfully difficult for politicians to vote against them."

DURANCE VILE

It's equally difficult to address rationally the problem of prison overcrowding without rational sentencing policies. But the Senate voted to build new prisons anyway. It allocated $3 billion for the construction of high-security regional prisons, to be made available to the states on the condition that they adopt "binding sentencing guideline systems" and require violent felons to serve at least 85 percent of their sentences.

Informed opinion is divided over the merits of regional prisons (which house people from several states). Some say that they're needed to break up prison gangs that form when people are imprisoned locally. Some argue that regional prisons will separate people not just from fellow gang members but also from their families. Positive familial and community ties are considered primary deterrents to recidivism. The public is not involved in the debate about regional prisons, but it supports a general increase in prison space. Sixty percent of people responding to a Time/CNN poll last year said they favored building new prisons even at the cost of a significant tax increase. More striking is the degree of public dissatisfaction with the courts, based on the belief that judges are soft on crime. Eighty-one percent of people surveyed by the National Opinion Research Center agreed that "courts do not deal harshly enough with criminals." In fact, most public anger at the criminal-justice system seems to be anger at the courts, not at legislators, police officers, or corrections personnel. The perception that a great many violent felons are serving very little prison time is driving the crime debate.

It's a perception based primarily on anecdotes--which doesn't mean it's false. But periodic stories about people who commit horrific crimes while they're on probation or parole or awaiting trial mostly provide political pressure to build more prisons, not empirical evidence that more prisons will make us safer.

It's true that prisons are overcrowded. Representative Barney Frank observes that in some circumstances liberals should support new prison construction, in order to make prisons more humane. The trouble is that prisons are a bit like highways: build them, and a great many people will come. It's also true that many prisoners do not serve out their full terms. In Texas, for example, people in noncapital cases generally serve only about 20 percent of their sentences, according to an American Bar Association report. And it's true that many felons (roughly half, according to the Bureau of Justice Statistics) convicted by the states are not imprisoned at all.

These are dramatic statistics, but by themselves they're not terribly informative. It means little to say that convicted felons serve only a fraction of their sentences unless the length of their sentences is known. Judges may take early-release systems into account when they impose sentence, so that, for example, if they want someone to serve six years, they may sentence him to ten. And the fact that many people convicted are not imprisoned may be an argument for improving probationary supervision, not for increasing the prison population, which already stands at nearly one million (more than 700,000 in state prisons and the remainder in federal custody).

"What's most scarce today aren't prison cells but nonprison supervisory programs," according to Mark Kleiman, a professor of public policy at the John F. Kennedy School of Government, at Harvard University. With three times as many people on probation or parole as are in prison, Kleiman says, "we're spending too much on the inpatients and not enough on the outpatients." He believes that with a few exceptions, such as the people who regularly commit violent crimes and the relatively small proportion of people who commit single crimes that shock the conscience of the community, convicted felons should be subject to heavily supervised probation and alternative sentences that stress mandatory, enforceable drug abstinence. "The trouble with most nonprison programs today," Kleiman says, "is that they're not backed up with the threat of imprisonment. I'd be happy to spend three billion dollars for more prison cells to use as backup to sanction people who don't comply with the terms of their probation or parole. They don't need to be imprisoned for a long time, but they need to be imprisoned every time they don't comply."

Common sense suggests that prescribing long prison sentences for a relatively small proportion of offenders (which is all we can reasonably expect to do) will not deter potential offenders who have no expectation of ever being imprisoned at all. A recent report by the National Research Council (an offshoot of the National Academy of Sciences) estimated that "a 50% increase in the probability of incarceration would prevent twice as much violent crime as a 50% increase in the average term of incarceration." Of course, increasing the probability of incarceration would require increasing the probability of arrest and prosecution, not merely a reallocation of prison space. But Mark Kleiman's proposal to use new prisons as backup sanctions for people on probation and parole would at least increase the probability of imprisonment for the great number of "outpatients" who are already caught up in the criminal-justice system. The Senate proposal, which makes the allocation of prison funding conditional on the states' adoption of binding sentencing guidelines, focused instead on further increases in the length of incarceration. Federal lawmakers can, after all, do little to increase the efficiency of local police departments, prosecutors, and courts, even if they can write lots of strict sentencing laws.

THE PRISON DEBATE

Lengthening prison sentences is not a new idea, or one that has proved effective. From 1975 to 1989 the average prison time served per violent crime tripled, according to the NRC. And the prison population nearly tripled during that same period--because of an increase in sentence length--reflecting, in part, the increased use of mandatory minimums, not an increase in the chance of incarceration, which remained fairly constant. Violent crime declined in the early 1980s but rose in the second half of the decade. The NRC tentatively concluded that the increased imprisonment rate of the 1980s had little effect on violent crime and that further increases would have little effect as well. Liberals agree, sometimes pointing out that violent felons, like athletes, peak early and become less violent as they age, obviating the need for long sentences. Conservatives, however, argue that violent-crime rates would have soared during the eighties had imprisonment rates remained stable. "America would have been a much more dangerous place today if we hadn't built the prisons we built in the 1980s," says Paul McNulty, now the executive director of the First Freedom Coalition, a conservative lobbying group, though he concedes that his claim is impossible to prove--as is the liberal claim that increased incarceration has had no effect or a negative effect on crime.

The prison debate is driven more by ideology than by science. Cost-benefit analyses of increasing prison terms are suspect, because the secondary costs and benefits of both victimization and imprisonment are quite difficult to quantify. As the NRC observed, we have only imprecise estimates of prison's effect on crime. We also have no rational system for identifying and diverting people who aren't dangerous.

No more than 20 to 25 percent of the people in prison are "dangerous career criminals," according to the National Council on Crime and Delinquency. Does this mean we're imprisoning too many people, as the NCCD suggests? Not necessarily. Dangerousness is notoriously difficult to predict, and according to the NRC, most violent crimes are not committed by repeat violent offenders ("dangerous career criminals"); they tend, rather, to be occasional events in long histories of crimes against property.

It is this problem of recidivism that most bedevils debates about prison policy. Conservatives and liberals agree in principle that some people in prison would have been more appropriately sentenced to an alternative correction program. They can even agree that a number of first-time, nonviolent offenders, particularly low-level drug offenders, should not be serving long mandatory minimum sentences. But they disagree about the need to imprison nonviolent repeat offenders. The consensus breaks down over recidivism, Paul McNulty confirms. He believes that with some exceptions, people should have only one opportunity to enjoy an alternative sentence. "If they waste that opportunity to obey the law," they should not, in most cases, receive another, he suggests. "We wouldn't raise children this way"--with empty threats.

If you believe in imprisoning recidivists, violent or not, "you win the argument about building more prisons every time," McNulty adds. "Surveys of state prisoners show that ninety-three percent of them are either violent criminals or recidivists. The number of people who ought to be released will never be larger than the number of people who are out on the streets and ought to be incarcerated."

But your view of who ought to be incarcerated depends on your ideology and your instincts about the dangers posed by nonviolent recidivists. In fact no one, liberal or conservative, can accurately say how many people in prison today would pose a danger to the community if they were released to strictly supervised parole, or how many would have posed less of a danger had they been sentenced to strictly supervised probation, including mandatory drug treatment and job training. The difficulty of identifying violent offenders is one reason that three-time-loser laws will probably have little overall effect on violent crime. Advocates argue that three-time-loser laws will decrease violence even if they apply to only a small percentage of offenders, because only a small percentage are responsible for the nation's violent crime. The trouble is, no one can say precisely which offenders make up that percentage.

The unpredictability of dangerousness can, however, be used as an argument for imprisoning practically everyone or practically no one, depending on how the due-process rights of offenders are balanced with those safety interests of the community which are served by incapacitating offenders (putting aside questions about retribution and deterrence). In recent years majorities have erred on what looks like the side of safety. But in addition to the nearly one million people in prison, some three million are on probation or parole. At some point ideology and speculation about recidivism become academic: we're not going to imprison four million people forever.

Occasionally Congress lapses into recognition of this--that there are limits to how many people we can incarcerate and for how long, just as there are limits to what imprisonment and other traditional law enforcement measures can accomplish. Crime bills passed last fall in both the Senate and the House provided funding for drug treatment for prisoners and gang members, diversion and counseling programs for juveniles, and boot camps--perhaps the trendiest of alternative sentencing proposals.

Boot camps combine the toughness on crime traditionally demanded by conservatives with traditional liberal expressions of compassion and an all-American belief in self-improvement through discipline and work. Historically, work has been viewed as character-building and has played an important role in the development of penal systems; it can symbolize both repentance and restitution to the community. The Senate allocated $3 billion for boot camps in what was presented as one of its more progressive anti-crime votes. Progressives, however, along with researchers, are a bit skeptical about the prospects for boot camps.

Ronald Hampton, a police officer in Washington, D.C., and the executive director of the National Black Police Association, observes that boot camps work for the military because people graduate from them into the military, with jobs to perform, rules to obey, and a disciplinary system to fear. Inmates of civilian boot camps will graduate into anarchy. Nor do boot camps work for everyone in the military, Hampton adds. Some people don't prosper in the service. "That doesn't mean they're not good citizens; it just means they're not good soldiers," he says. In fact, outcome studies of civilian boot camps have been discouraging: their rates of recidivism are no lower than the prisons' rates. Boot camps may pose particular problems for the juveniles they're intended to help if they allow for the mingling of juvenile and adult offenders. There is also concern that the quasi-abusive environment of boot camps will be counterproductive for young offenders with histories of familial abuse. As Mark Kleiman puts it, "If pushing these people around and screaming at them were therapeutic for them, they wouldn't be in boot camp."

Misgivings about boot camps expressed by people who study them and others who work with young offenders, however, are dismissed by politicians who establish them. In the view of John Laub, a professor of criminal justice at Northeastern University, political enthusiasm for boot camps exemplifies the futility of research on criminal-justice issues. Knowledge isn't power in criminal-justice debates; knowledge is irrelevant. Politicians who support boot camps rely on "gut instincts" or "intuition" or fond memories of their own military experiences.

"This is why what we do doesn't matter," Laub says, pointing to Georgia Governor Zell Miller's proud proclamation of his intention to ignore research on boot camps. "Nobody can tell me from some ivory tower that you take a kid, you kick him in the rear end, and it doesn't do any good," Miller told The New York Times. "I don't give a damn what they say."

OUR DOUBLE STANDARD OF ACCOUNTABILITY

Protesting the influence of politics on policy, you feel a little like Claude Rains protesting gambling at Rick's. Still, it's hard not to be shocked- shocked!--by the utter politicization of criminal-justice debates. These are case studies in demagoguery. Informed debates among elites about public-safety issues seem more than usually irrelevant to public opinion. Politicians profess to believe what they must know is not true.

How are most citizens apt to learn about the criminal-justice system? Amy Fisher docudramas, TV cop shows and crime stories, and talk-show palaver about sensational cases seem to be primary sources of "information" for people who are spared firsthand experience with corrections and courts. The most attention is paid to the least typical cases. The acquittal in the Lorena Bobbitt case and the hung juries in the Menendez cases were followed by weeks of public hand-wringing about the "abuse excuse." We've become too tolerant of criminal behavior, people said, which must have come as a surprise to the roughly 2,800 people on death row and the petty offenders with drug habits, long untreated, who may be subject to life imprisonment as three-time losers. Jurors, and a large segment of the public which ordinarily objectifies less sensational defendants, seemed to empathize with Lorena Bobbitt and the Menendez brothers, treating them with great leniency. That is perhaps what was most remarkable about these cases. In general, as the intense demand for harsh, punitive sentencing shows, we're more likely to hold defendants to standards of accountability higher than those we'd apply to ourselves.

This is rarely recognized as hypocrisy. Many people who haven't been criminally prosecuted or endured the prosecution of a friend or family member tend to regard criminals as "the other." This unwillingness or inability to admit their humanity is facilitated by the anonymity of the vast majority of defendants, whose cases aren't televised; by the increasing senselessness of random violence (it's hard to empathize with a sociopath); and by the awful demographics of crime: the disproportionate involvement of black males in violent crime makes it easy for criminal defendants to be objectified by whites who fear being victimized and don't consider the fact that crime tends to be intraracial (that is, blacks are disproportionately represented as victims, too).

Our double standard of accountability becomes particularly clear when popular notions of criminal justice are compared with popular notions of personal development, which dominated the Bobbitt and Menendez cases. If you argue for leniency in a typical, uncelebrated criminal case, pointing out that the defendant suffered abuse in childhood or that the crime resulted from some post-traumatic stress disorder, you're likely to be dismissed as a bleeding-heart liberal. "We're tired of Twinkie defenses. No one takes responsibility for himself anymore," people say. Meanwhile, millions of Americans are complaining about their own histories of addiction and abuse, refusing to take responsibility for their own bad behavior. Many of us want most people charged with serious crimes to be considered actors, absolutely, so that we can punish them absolutely without qualms, but we often consider ourselves to be acted upon.

If only guilt and innocence were so clear and simple and absolute, instead of merely relative. Of course, all of us are the victims or beneficiaries of forces we never controlled, accidents of upbringing or birth; ultimately, none of us are absolutely responsible for our characters or our actions. But at some point we must assume and assign responsibility. At some point we must treat ourselves and one another as if our crimes were indeed our fault.

Sometimes we have to pass judgment on others, with modesty and reluctance, knowing that our judgments are flawed. That is one of the fearful necessities of a civilized society. Today, in general, we alternate between judging too harshly--condemning people to life sentences or death with an arrogant belief in our own righteousness--and not judging at all. Watch the talk shows and the cases on Court TV and then spend a few days in the criminal courts: they make clear that we lack a consistent, nuanced notion of moral agency, an ethic of relative accountability. How can we claim a sense of justice?