• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

Post-Modern Meditations on Punishment: On the Limits of Reason and the Virtues of Randomization (A Polemic and Manifesto for the Twenty- First Century)

Posted on: Tuesday, 11 September 2007, 06:00 CDT

By Harcourt, Bernard E

Enlightenment is man's emergence from his self-incurred immaturity. Immaturity is the inability to use one's own understanding without the guidance of another. -Immanuel Kant, "An Answer to the Question: 'What is Enlightenment?'" (1784)

IN AN ESSAY BEARING THE SAME TITLE, MICHEL FOUCAULT READ KANT'S text against the backdrop of his critique of reason, published just three years earlier. It is precisely at the moment that we assert ourselves as mature beings, Foucault observed, that it is most important to recognize the limits of reason. "It is precisely at this moment that the critique is necessary, since its role is that of defining the conditions under which the use of reason is legitimate in order to determine what can be known, what must be done, and what may be hoped" (Foucault, 1997: 308). Foucault warned us, with Kant, that reliance on reason beyond its proper bounds would merely set the clock back: "Illegitimate uses of reason are what give rise to dogmatism and heteronomy, along with illusion" (308). The careful and critical use of reason, in contrast, is what enlightens and leads forward, out of the shadows of illusion. But it could only do so by relying on itself. The modern period would thus embrace a strong conception of self-reliance-carefully bounded by critique.

Foucault also saw in Kant's essay a new philosophical attitude consisting of genuine reflection on the "present"- a turning of the more traditional, eternal gaze of the philosopher onto the contemporary moment, and, with that, an associated task of theorizing knowledge in relation to current times. Foucault dubbed this "the attitude of modernity" and located it later in the writings of nineteenth- and twentieth-century authors, starting foremost with Charles Baudelaire. "By 'attitude,'" Foucault wrote, "I mean a mode of relating to contemporary reality; a voluntary choice made by certain people; in the end, a way of thinking and feeling; a way, too, of acting and behaving that at one and the same time marks a relation of belonging and presents itself as a task" (Foucault, 1997: 309). This attitude brought together philosophical inquiry and critical thought focused on contemporary historical actuality. Philosophical training and reflection would now apply themselves to the contemporary moment-most notably, the French Revolution-and concentrate on the task of reasoning through the present. Foucault saw in Kant the origin of a modern attitude that would run through Hegel, Nietzsche, Marx, Durkheim, Rusche, and Kirchheimer.

In another essay bearing the same title, Jurgen Habermas adds: "Surprisingly, in the last sentence of his lecture Foucault includes himself in this tradition" (Habermas, 1994: 150).

Once again, the attitude of modernity triumphed over the critique of reason. In these pages, I argue that the two strands that Foucault identified in Kant's essay-the crucial moment of critical reason and the modern attitude-collided throughout the nineteenth and twenti- eth century, and that the modern attitude repeatedly prevailed. Even when the moderns were engaged in the most critical of enterprises, the attitude gained the upper hand and offered new ways of conceptualizing and making sense of the present, consistently beyond the limits of critical reason. Never daunted by those warnings about illusions, never chastened by the foolish excesses of earlier generations, modern thinkers continued to theorize contemporary historical actuality beyond reason's bounds.

I propose that we finally abandon the misguided attitude of modernity. It will mean, no doubt, leaving much to chance and randomization. This is all for the better. Let me explain.

1.

The moderns posed three questions of punishment. The first, born of the Enhghtenment itself, sought to identify and define a rational basis for punishing. As men freed themselves from the shackles of religious faith, this first question took shape: If theologians can no longer ground political and legal right, then on what foundation does the sovereign's right to punish rest? On what basis does the state have a right to punish its citizens?

Naturally, the question was not entirely innocent-no good questions ever are. It was animated by a desire to locate the righteous limits of the sovereign's punitive power at a time that was marked-at least in the eyes of many of the first modern men of reason-by excessive punishments. The right to punish, it turns out, would serve to limit punishment.

"Here, then, is the foundation of the sovereign's right to punish crimes," a young, 25-year-old Cesare Beccaria would declare in 1764: "the necessity of defending the repository of the public well-being from the usurpations of individuals" (Beccaria, 1995: 10). The origin of the right, Beccaria explained, derived from the sovereign's duty to promote "the greatest happiness shared among the greater number" (Beccaria, 1995: 7). Through his disciple, Jeremy Bentham, Beccaria's writings would translate into more conventional theories of utilitarianism and deterrence and, later, economic models of social welfare maximization. Other early moderns would derive the right to punish elsewhere-in the autonomy or dignity of the moral agent, in the interests of enlightened self-development, in the harm principle, and in those other traditional expressions of legal right.

This first line of inquiry endures well into the present. In the Anglo-Saxon tradition, the answers draw heavily on a functional analysis of the criminal sanction. In the classic debate over the legal enforcement of morality, for instance, all the major contributors-from H. L. A. Hart (1963: 14) and Patrick Devlin (1965: 2-3), to Ernest Nagel (1968: 138-39), Norval Morris (Morris and Hawkins, 1970: 4), Joel Feinberg (1987, 1984) and others-use as the starting point for their analysis the statement of function articulated in the Wolfenden Report of 1957: "the function of the criminal law . . . is to preserve public order and decency, to protect the citizen from what is offensive or injurious, and to provide sufficient safeguards against exploitation and corruption of others." The liberal theorists tend to narrow the function to preserving order, and draw more heavily on John Stuart Mill's earlier pronouncements in On Liberty, while legal moralists broaden the scope of the functions to include, centrally, the promotion of morality as a principal function of the criminal law. "The criminal law as we know it is based upon moral principle," Devlin argued. "In a number of crimes its function is simply to enforce a moral principle and nothing else" (Devlin, 1965: 7). On both sides of the debate, though, the sovereign's "right to punish" derived precisely from identifying the proper functions of the criminal sanction.

Although this first discourse continues today, it did not take long for men of knowledge-as Nietzsche described himself-to spot the error in this line of inquiry. The right to punish, after all, was precisely what defined sovereignty and, as such, could hardly serve to constrain sovereign power. The first question had gotten things backwards: the "right to punish" was what the sovereign achieved by persuading its members that it could best promote the legitimate goals of punishment. To seek the origin of the right to punish by analyzing the purposes or utilities of punishment would lead nowhere. "The 'purpose of law,'" Nietzsche declared, "is absolutely the last thing to employ in the history of the origin of law: . . . whatever exists, having somehow come into being, is again and again reinterpreted to new ends, taken over, transformed, and redirected by some power superior to it" (Nietzsche, 1967: 77).

It was fruitless to look for the right to punish in its purposes, utilities, or functions- whether from a utilitarian or deontological perspective. "[PJurposes and utilities are only signs that a will to power has become master of something less powerful and imposed upon it the character of a function," Nietzsche emphasized (1967: 77). The proper question to ask of the "right to punish," then, was not "on what ground,""of what origin," or "from where" but rather: "How does the sovereign's act of punishing get perceived as legitimate?" Or better yet, "Under what conditions does the sovereign's exercise of that power of punishing not trigger sufficient resistance to undermine sovereignty itself?" That question, however, did not call for philosophical debate over rights, deontological argument about autonomy, or econometric analyses of deterrence. It called for historical, sociological, political, and genealogical research about acts of resistance, social movements, transformative moments, ideology, and social cohesion. It called for a genealogy of morals, law, and power- in sum, a genealogy of punishment.

With the birth of the social sciences in the late nineteenth century, this critical impulse gave rise to a second line of inquiry. More skeptical, more critical, the questions probed and excavated deeper processes and forces: If the rational discourse over the right to punish is mere pretext and serves only to hide power formations, then what is it exactly that punishment practices do for us? What is the true function of punishment? What is it that we do when we punish? From Emile Durkheim to Antonio Gramsci and the later Frankfurt School, Michel Foucault, and/in-de-siede trends in penology, twentieth-century moderns struggled over social organization, economic production, political legitimacy, governance, and the construction of the self-turning punishment practices upside down, dissecting not only their repressive functions but more important their role in constructing the contemporary subject and modern society. At the apex of this second line of inquiry, Michel Foucault would articulate and enumerate, in his magisterial book, Surveiller et punir: Naissance de la prison, the central tasks and rules of engagement. First and foremost, "Do not concentrate the study of the punitive mechanisms on their 'repressive' effects alone ___ [R]egard punishment as a complex social function" (1979: 23; 1975: 28). Foucault explicitly acknowledged that this second project built on the work of Emile Durkheim, citing him alone on that same page (1975: 28 n.l), and owed much to the Frankfurt School. In the immediate passage following his enumeration, Foucault emphasized that "the great book of Rusche et Kirchheimer, Punishment and Sodai Structures, offers a number of essential reference points" (1975: 29):

We must first rid ourselves of the illusion that penality is above all (if not exclusively) a means of reducing crime. . . . We must analyse rather the "concrete systems of punishment," study them as social phenomena that cannot be accounted for by the juridical structure of society alone. . . ; we must situate them in their field of operation, in which the punishment of crime is not the sole element; we must show that punitive measures are not simply "negative" mechanisms that make it possible to repress, to prevent, to exclude, to eliminate; but that they are linked to a whole series of positive and useful effects which it is their task to support (Foucault, 1979: 24; 1975: 29-30).

All of these reference points originated with the Frankfurt School, and the key task that emerged from the second line of inquiry was to unearth the deeper forces and relations of power that, through the means of punitive practices, shape us as contemporary subjects. To explore, in effect, "How a specific mode of subjugation could give birth to man as an object of knowledge for a discourse with scientific status" (Foucault, 1975: 28-29). To discover and trace the deeper forces that shape our punitive practices and, through them, our knowledge of ourselves.

However, a series of further critiques- critiques of metanarratives, functionalism, and scientific objectivity - would chasten this line of inquiry and nudge it around the cultural turn, helping shape a third discourse on punishment. This line of inquiry would focus not on what punishment is doing/or us, but on what punishment tells us about ourselves: What do our punishment practices tell us about our cultural values? What is the social meaning of our institutions of punishment? Less metatheoretical, less critical-theoretic, this final set of questions would build on, while simultaneously trying to avoid, the searing critique of the construction of knowledge. The questions were intended to be less normative. A description at most. A compelling interpretation. Something to make sense of our world and ourselves. Something to ground, perhaps later, an evaluation of those punishment practices.

The difference was subtle, but important. The second set of questions- especially as they evolved over the course of the twentieth century-had become increasingly focused on the constructed nature of knowledge, what has come to be known as the "power/ knowledge" critique: How, exactly, do we come to believe what we hold as true? How is it, for instance, that we come to believe a progress narrative of punishment? What institutions and practices shape us to believe in the idea of the "delinquent"-or, for that matter, in the idea that we could possibly "rehabilitate" or "correct" that "delinquent"? How have our own disciplinary practices contributed to shaping our beliefs? By the late twentieth century, this second set of questions had begun to revolve entirely around the formation of knowledge and to constitute an acid test for all knowledge claims regarding punishment.

In contrast, the third set of questions-the product, as I mentioned, of a critique of metanarratives-tried assiduously to avoid the power/ knowledge critique. It cut a more humble profile. It sought only to reflect on what our punishment practices tell us about ourselves, our values, our society- as a mere preliminary to a better understanding of punishment, to make possible, later, a better evaluation of our practices and institutions. David Garland's book, Punishment and Modern Sodety (1990), though ostensibly a pedagogic treatment of the four leading voices in the sociology of punishment, reflects well this third line of inquiry. "The social meaning of punishment is badly understood," Garland contends. What is needed is "a descriptive prolegomenon which sets out the social foundations of punishment, its characteristic modern forms, and its social significance" (Garland, 1990: 9). The social meaning of punishment "needs to be explored if we are to discover ways of punishing which better accord with our social ideals" (Garland, 1990: 1).

This third line of inquiry represents, in Garland's words, "a deliberate attempt to shift the sociology of punishment away from its recent tendency-engendered by Foucault and the Marxists-to view the penal system more or less exclusively as an apparatus of power and control" (1990: 1-2). The task is to develop "a pluralistic, multidimensional approach,""a rounded, completed image; a recomposition of the fragmentary views developed by more narrowly focused studies" (Garland, 1990: 280). To explore "multiple causality, multiple effects, and multiple meaning" (Garland, 1990: 280). Garland explains:

Values, conceptions, sensibilities, and social meanings-culture, in short - do not exist in the form of a natural atmosphere which envelopes social action and makes it meaningful. Rather, they are actively created and recreated by our social practices and institutions - and punishment plays its part in this generative and regenerative process (Garland, 1990: 251).

In this sense, the third line of inquiry calls for richly textured, thick descriptions of our punishment practices intended to expose their social meaning and their role in shaping the fabric of society-all this to serve as a preparatory to normative analysis, to provide "a proper descriptive basis for normative judgments about penal policy" (Garland, 1990: 10).

It is not entirely clear, though, how the third and final line of inquiry could escape the power/knowledge critique. If Foucault's disciplinary hypotheses were themselves susceptible, surely an interpretation of the "social meaning" of punishment practices and institutions would also be vulnerable. Any interpretation would tell us more about the interpreter and her belief systems than about the meaning of the practice itself. Surely the semiotic enterprise would reveal more about the modes of reasoning, beliefs, and ethical choices held by the individual interpreter than about the social meaning of the punishment practices themselves.

The closing paragraphs of Garland's book are revealing in this respect. Modern societies, Garland writes, should expect less from punishment and "might be encouraged to treat it instead as a form of social policy which should, where possible, be minimized" (Garland, 1990: 292). The goal should be to socialize and integrate young citizens, not punish them: "a work of social justice and moral education rather than penal policy. And to the extent that punishment is deemed unavoidable, it should be viewed as a morally expressive undertaking rather than a purely instrumental one" (Garland, 1990: 292). These, I take it, are significant normative commitments that, in all likelihood, bleed into and color the cultural critic's interpretation of the social meaning of punishment practices.

As dusk fell on the twentieth century, modern writings on punishment continued to reflect more on the authors than on the punishments. Somehow, despite the reformulation of the questions, the texts still told us more about the interpreter's beliefs, intuitions, and ethical choices than they did about the practices of punishment and their social meaning.

2.

What do we do now-now that we have seen what lies around the cultural bend and realize, painfully, that the same critiques apply with equal force to any interpretation of cultural meaning that we could possibly slap on our contemporary punishment practices? Should we continue to labor on this final set of questions, return to an earlier set, or, as all our predecessors did, craft a new line of inquiry? What question shall we-children of the twenty-first century- pose of our punishment practices and institutions?

The answer, paradoxically, is that it does not matter. The formulation of the questions themselves never really mattered, except perhaps to distinguish the analytic philosopher from the critical theorist, the positivist from the cultural critic-minor differences that reflected nothing more than taste, desire, personal aptitude, upbringing, and training. Yes, new questions were formulated and new discourses emerged, but the same problem always plagued those modern text.

In all the modern texts, there always came this moment when the empirical facts ran out or the deductions of principle reached their limit-or-both-and yet the reasoning continued. There was always this moment, ironically, when the moderns-those paragons of reason-took a leap of faith. It is no accident that it was always there, at that precise moment, that we learned the most-that we could read from the text and decipher a vision of just punishment that was never entirely rational, never purely empirical, and never fully determined by the theoretical premises of the author. In each and every case, the modern text let slip a leap of faith-a choice about how to resolve a gap, an ambiguity, an indeterminacy in an argument of principle or fact. The inevitable space between theoretical or empirical premises and the final judgment derives, in the end, from that imperceptible fissure in the human sciences between the not falsified, the not yet falsified, the apparently unfalsifiable, the verified but only under certain questionable assumptions, and truth. In the empirical domain-no less than in philosophical discourse, legal analysis, and public policy debates-proof never followed mathematical deduction, but rested instead on assertions-whether empirical or logical-that may well have been true, but for which other entirely reasonable hypotheses could have been substituted. The key issue was always which hypothesis to believe from among the many possible hypotheses, all of which were consistent with the data; which subprinciple to uphold from among all the possible subprinciples that were theoretically coherent with the guiding principle. What the moderns chose to believe, ultimately, told us more about them than it did about the world around them. It was always the answers that moderns gave to the questions-regardless of the question itself-that revealed the most about them and their intuitions about just punishment.

Ironically, this gap is precisely what made possible the moment of enlightenment at the very heart of critical theory-what Raymond Geuss refers to as that reflective opening that "gives agents a kind of knowledge inherently productive of enlightenment and emancipation" (Geuss, 1981: 2). Once we lifted the veil from our eyes and realized fully that our rational belief in certain theories or premises were no better than religious faith-that we had taken a leap of faith to arrive at our conclusion-it then became possible to trace the genealogy of how we took that leap. It became possible to explore how we came to believe what we did believe and at what price. That is precisely what the great critical thinkers of the nineteenth and twentieth century did along the three principal dimensions of radical thought-power (from Nietzsche through Foucault to Agamben), economic production (from Marx through the Frankfurt School to Althusser), and desire (from Freud through Lacan to Zizek). Not surprisingly, identifying the gap is what also gave birth to the American Legal Realist movement in the early twentieth century and the Critical Legal Studies movement at midcentury. It also made possible deconstruction at the end of the century-perhaps the fullest instantiation of the insight.

In this respect, Jacques Derrida-no hero of mine, I assure you, far too ambiguous and playful for my taste-was entirely right when he wrote that the foundation of law itself rests on a leap of faith- what he refers to as "a performative force, in other words always an interpretive force with an appeal to faith" (Derrida, 1994: 32). Legal authority traces to this act of auto-authorization, itself never subject to a legal evaluation of right or wrong-not simply, though certainly, because the legal framework itself postdates the founding moment, but also and more important, because the judgment that a punishment is just must always overcome the gap between theoretical premises and final judgment. The act of reaching the legal conclusion-the just punishment, the sentence, the execution- represents "a stroke of force, a violent performative act, and thus an interpretation that is in itself neither just nor unjust" (Derrida, 1994: 32-33). And it is precisely in this sense that Derrida concludes, paradoxically, that the structure of the law is what opens the door to the very possibility of deconstruction itself. Thus, his playful hypothesis that justice makes possible deconstruction (Derrida, 1994: 36). Though addressing law andjustice, Derrida's point applies equally well to the other disciplines that form the field of crime and punishment, such as sociology, politics, economics, and public policy.

I said "ironically" earlier because it is precisely the moment of critical perception and enlightenment that simultaneously undermines the claims of the radical critical theorists-though not necessarily those of the deconstructionists.

3.

What do these gaps, ambiguities, and indeterminacies look like? What does it mean, exactly, that the moderns inevitably took a leap of faith? "The empirical facts ran out, the deductions of principle reached their limit, and yet the reasoning continued." What does that reasoning sound like? Let me stop for a moment here and give some illustrations. Let me demonstrate some of the gaps and ambiguities.

Deterrence of Juvenile Offenders

First, let's examine a claim of deterrence. The trouble with most research on deterrence is that it is extremely difficult to divorce the effects of deterrence from those of incapacitation-from the fact that increased law enforcement will also result in more imprisonment and thus greater incapacitation of criminal offenders. The National Academy of Sciences appointed a blue-ribbon panel of experts to examine the problem of measuring deterrence in 1978-led by Alfred Blumstein, Jacqueline Cohen, and Daniel Nagin-but the results were disappointing: "Because the potential sources of error in the estimates of the deterrent effect of these sanctions are so basic and the results sufficiently divergent, no sound, empirically based conclusions can be drawn about the existence of the effect, and certainly not about its magnitude" (Blumstein, Cohen, and Nagin, 1978: 42; see also Nagin, 1978: 95, 135; Spelman, 2000: 97). Little progress has been made since then. As Steven Levitt suggested in 1998, "few of the empirical studies [regarding deterrence of adults] have any power to distinguish deterrence from incapacitation and therefore provide only an indirect test of the economic model of crime" (Levitt, 1998: 1158 n.2).

Levitt nevertheless contends that juveniles and young adults are responsive to increases in punishment. In order to demonstrate this, Levitt takes a state-level dataset of criminal offending rates and classifies states into three categories: first, states that have a more severe adult than juvenile criminal justice system; second, states that have similar levels of severity for their adult and juvenile criminal justice systems; and third, states that have a more lenient adult than juvenile criminal justice system. Levitt then compares the relative offending rates of young adults as they turn from juveniles to adults-as they reach majority age and become subject to the adult criminal justice system.

Levitt finds that juveniles who have turned adult in the first category of states-those with relatively more severe adult systems- offend less in their first year of majority than they did in the previous year, whereas those juveniles in states with relatively more lenient adult systems offend more than they did the previous year (Levitt, 1998: 1175). Levitt concludes from this that deterrence, rather than simply incapacitation, is at work: "Sharp drops in crime at the age of majority suggest that deterrence (and not merely incapacitation) plays an important role" (Levitt, 1998: 1156).

Why, exactly, do the data confirm the deterrence hypothesis? The answer, Levitt suggests, is that "if deterrence is at work, then one would expect an abrupt change in behavior associated with passage to adult status. If, on the other hand, incapacitation is the primary channel, then one would expect longer delays in the transition from the juve- nile equilibrium to the adult equilibrium due to lags in the timing of arrest and sentencing. ... It seems likely that large immediate changes in behavior associated with the age of majority are likely to primarily reflect deterrence" (Levitt, 1998: 1172). The logic of the argument, then, rests on the assumption that deterrence works more speedily than incapacitation at the transitional period around majority.

The trouble with this logic, though, is that there is no metric to test the speed of either mechanism alone, nor is there any metric to compare the speed of the two competing theories (Harcourt, 2005: 219-226). There is no way, a priori, to determine how fast either effect would take-whether it is a month, two months, three months, six months, nine months, twelve months, eighteen months, two years, or more. Levitt's model uses an annual measure of crime. Yet the incapacitation time lag-if there is one-may very well be shorter than that. In fact, if true incapacitation theory is correct-the idea that about 6 percent of young adults are responsible for about 50 percent of their cohort's criminal activity-one would expect that strict enforcement would have an immediate and sharp incapacitative effect precisely at the moment of the release of delinquent youths turning to majority.

Here, then, is the gap: there is no measure, no metric, no standard against which we could declare that an effect on crime- deterrence or incapacitation-is abrupt or delayed. Nor is there any way to determine how the two effects would compare. We do not have a measure for the incapacitation effect, and a separate one for the deterrence effect. We just have one number, and have to guess whether it seems relatively immediate or relatively delayed. Since we do not know how long the incapacitation effect takes, there is no way of knowing from annual crime data whether the effect looks more immediate or more delayed-whether it is incapacitation or deterrence.

Why is it that Steven Levitt is prepared to skip over this gap and confirm the deterrence hypothesis? It does not really matter. I would tend to emphasize taste, desire, training, and professional advancement; but there may be other explanations. What does matter is that there is a gap and a leap of faith-of faith in rationality- that we can identify. Here it is a gap of the not yet falsified type. A theory that is consistent with the data, but does not exclude other competing hypotheses. It would be wrong to base public policy on these empirical findings. Racial Profiling

A number of economists contend that the use of racial profiling can improve the efficiency of policing by increasing the number of successful searches (Knowles, Persico, and Todd, 2001; Hernandez- Murillo and Knowles, 2003). Assuming that people respond rationally to the increased cost of offending-assuming rational action theory- targeting more police resources at a higher-offending population will reduce their rate of offending (given the greater likelihood of being detected and punished). If we assume, in addition, that minorities have a higher offending rate than whites, then the optimal level of profiling occurs when the offending rate of minorities declines to the same level as the offending rate of whites. At that point, the police will maximize the number of successful police interventions and have no legitimate interest in profiling minorities to any greater extent. The economists verify these conclusions with accurate mathematical equations and economic models.

Even under these assumptions, however, racial profiling may increase the overall societal rate of offending (Harcourt, 2007a: 111-144). It all depends on the relative responsiveness of the two groups-the profiled minorities and the nonprofiled whites-to policing. If minorities are less responsive to policing then whites, then their decrease in offending will be outweighed, in absolute numbers, by the more elastic responsiveness of whites-i.e. by the increased offending of whites in response to the fact that they are being policed less. This is true despite the fact that the overall number of successful police interventions increases - despite the fact that the police are detecting and punishing more crime. I demonstrate this with accurate mathematical equations and economic models in Against Prediction (2007a: 132-136).

The economists had essentially assumed in their model of racial profiling that minorities are as responsive to policing as whites, if not more. If they had not made that crucial assumption, then their own models would demonstrate that racial profiling may increase the amount of crime in society-which is definitely not an efficient outcome. Their claims are nonfalsifiable but only under dubious assumptions. They are mathematically verified, but only if we assume something about the relative elasticity of the two groups that we have no ground to assume. (In fact, if minorities have a higher offending rate than whites, it is far more likely that the cause of that difference, perhaps lower employment opportunities, would lower their responsiveness to policing in comparison to whites).

This, I take it, is a gap within their own model: even assuming deterrence (which itself is, for many, a leap), there is a gap over which these economists took another leap of faith. Why? Again, it does not matter. I would speculate that it is because they desire a clean, parsimonious, mathematical model that affirms rationality. Maybe that is why they became economists. But again, why they took a leap of faith does not really matter. What matters is that they took it and that we can identify it.

Order Maintenance

For a third illustration, let me turn to a modern policing practice. In the early 1990s, several major US cities began implementing order-maintenance strategies-most notably New York City, in 1994, where Mayor Rudolph Giuliani and his first police commissioner, William Bratton, put in force the "quality-of-life initiative." The order-maintenance strategies rested on the "broken windows" theory: the idea that minor neighborhood disorder like graffiti and loitering, if left unattended, will cause serious criminal activity (Wilson and Kelling, 1982: 31; Harcourt, 2001: 23- 27; Harcourt and Ludwig, 2006: 278-287).

During the 1990s, several proponents of order-maintenance declared that the broken windows theory had been empirically verified (Kelling and Coles, 1996: 24). They rested this assertion on the findings of a 1990 study titled Disorder and Decline. Subsequent research discovered several gaping flaws in the study that undermine confidence in the findings (Harcourt, 2001: 59-78). Even putting those gaps aside, the 1990 study used a static dataset to test a dynamic hypothesis: the data consisted of disorder and robbery victimization at one point in time, whereas the broken windows theory posited a developmental sequence over time. The statistical analysis could not-and as a result, did not-falsify the broken windows hypothesis. As Ralph Taylor succinctly observes, the 1990 study was simply off the mark "because these data are cross- sectional, and the thesis is longitudinal" (Taylor, 2006: 1626). The gap was between the not-yet-falsified-because-not-really-tested and truth. Again, it was inappropriate at the time to form public policy on its basis.

Today, the social scientific support for the broken windows theory rests principally on a 2001 study coauthored by George Kelling and William Sousa. In their study, Kelling and Sousa focus on the 75 police precincts in New York City over the period 1989 to 1998. They statistically compare the relationship between violent crime and broken-windows policing, as well as three other independent variables-unemployment, demographics, and crack cocaine consumption-simulating comparison groups by treating the city as 75 separate and comparable entities. Looking at the change over time, they find that the measure of broken-windows policing is significantly related to the drops in precinct violent crime over the 10-year period - in contrast to demographics, unemployment, and drug use patterns, which are not.

The trouble with the Kelling and Sousa study is that they do not control for what statisticians call "mean reversion." An examination of their data reveals just that: those precincts that experienced the largest drop in crime in the 1990s were the ones that experienced the largest increases in crime during the city's crack epidemic of the mid- to late 1980s. In other words, it may well be true that the precincts that received the greatest dose of broken windows policing in the 1990s experienced the largest declines in crime. But those precincts were precisely the ones that were hit hardest by the crack epidemic that fueled homicide rates in New York City from the mid-1980s through the early 1990s. Everywhere that crime skyrocketed as a result of the crack epidemic, crime declined sharply once the epidemic ebbed-which, it turns out, was also true across the country.

In a recent study with Jens Ludwig, we demonstrate that the declines in crime observed in New York City in the 1990s are exactly what would have been predicted from the rise and fall of the crack epidemic, even if New York had not embarked on its broken windows policing strategy (Harcourt and Ludwig, 2006: 315). Jens Ludwig and I call this Newton's Law of Crime: what goes up, must come down, and what goes up the most, tends to come down the most. What it represents, in effect, is a competing hypothesis that more fully explains the relationship between crime and policing.

There's a third gap-or fourth or fifth, I've lost track frankly. Kelling and Sousa infer the truth of the broken windows theory-and advocate a public policy of broken windows policing-on the basis of a not-falsified hypothesis that coincidentally fails to test for a competing explanation. But even if they had tested for mean reversion, we would still be left with a nonfalsified hypothesis: the broken windows theory has not been disproved by Jens Ludwig and my study of the New York City data, we have merely offered a different (though in our mind better) explanation of the crime trends.1 But who is to say which is right? It is, as Ludwig and I suggest, a Scotch verdict: not proven. That's a gap.

Why are Kelling and Sousa willing to take a leap of faith and advocate policies based on the broken windows theory-a theory that is at best not falsified? It does not matter. I think I know why, but of course I may be wrong: George Kelling, the coauthor of the original Broken Windows article, has a lot invested in its truth, especially now that he is running a consulting business, the Hanover Justice Group, that markets broken-windows policing methods to city mayors and councils. But again, it really does not matter. What matters here is that we have identified another gap and a corresponding leap of faith.

The Death Penalty

Modern writings on capital punishment also illustrate radical indeterminacy - those gaps and ambiguities that allow moderns to express, more than anything, their moral intuitions about just punishment. Take, for example, modern rational choice literature on the death penalty. Beccaria, the first true rational choice theorist, did not believe that the sovereign had the right to sentence to death. Capital punishment, according to Beccaria, fell within the domain of war and thus was governed by rules of necessity and utility. Beccaria did not believe, however, that the death penalty served either interest. It was not necessary because longdrawn-out punishments, such as penal servitude or slavery for life, were more effective and fear-inducing than the fleeting shock of death. It was also not useful because capital punishment had a brutalizing effect on society (Beccaria, 1995: 67). Jeremy Bentham- the very spokesman for the theory of marginal deterrence in the modern era-agreed entirely: "the more attention one gives to the punishment of death the more he will be inclined to adopt the opinion of Beccaria-that-it ought to be disused. This subject is so ably discussed in his book that to treat it after him is a work that may well be dispensed with" (quoted in Hart, 1982: 41). Fast forward to the present, Christmas Day 2005 (and note the irony of the timing). Here's Gary Becker, one of the world's leading rational choice theorists and Nobel prize economist, writing on his blog: "My belief in [the death penalty's] deterrent effect is partly based on these limited quantitative studies, but also because I believe that most people have a powerful fear of death. David Hume said in discussing suicide that 'no man ever threw away life, while it was worth living. For such is our natural horror of death. . .'. Schopenhauer added also in discussing suicide '. . . as soon as the terrors of life reach a point at which they outweigh the terrors of death, a man will put an end to his life. But the terrors of death offer considerable resistance. . . ."' (Becker-Posner blog, December 25, 2005). Richard Posner adds, also on Christmas Day: "I do not consider revenge an impermissible ground for capital punishment. Revenge has very deep roots in the human psyche" (Becker-Posner blog, December 25, 2005).

It would almost be funny, if it were not so sad, to watch these moderns twist and turn and contort themselves to justify their own ethical intuitions about killing other people. The only issue for a rational choice theorist is whether the death penalty actually deters homicides, net of any other effect. Becarria chose to believe that the brutalizing effect outweighed the deterrent effect. Becker chose to believe that people fear death. The empirical literature is all over the lot (see Donohue and Wolfers, 2005). Yet Becker and Posner decide to believe those economists who find a deterrent effect. It is remarkable to watch-though disheartening for those who once believed in the critical project of reason.

This is not to suggest that the rational choice theorists alone exhibit raw choice. Listen to Hegel: "Beccaria's endeavor to have capital punishment abolished has had beneficial effects. Even if neither Joseph II nor the French ever succeeded in entirely abolishing it, still we have begun to see which crimes deserve the death penalty and which do not. Capital punishment has in consequence become rarer, as in fact should be the case with this most extreme punishment" (Hegel, 1981: 247 [emphasis added]). These are telling words. What they tell us, though, is not the right way to formulate the inquiry, nor the correct answer to the proper question, but something more fundamental about the personal convictions of the author-and how it is, exactly, that authors bridge the inherent indeterminacy of their own principles.

4.

These gaps and ambiguities will bury the modern period-or at least, they should. Even the sharpest of critics, the most radical thinkers have never been able to escape the overpowering urge to build some new construct, a new edifice, some bridge to get to the other side of knowledge. Neither the followers of Nietzsche, Durkheim, or Marx, nor the cultural critics were able to resist the lure of reconstruction, always cobbling together the "best evidence" to soften their landing. Not even Michel Foucault-that wisest of moderns-could resist displacing our faith in rehabilitation with a genealogical story-one that required just as great a leap of faith. Tragically, this is as true of the cultural critics as it was of the two earlier inquiries.

Many have argued over the ages-and still do-that we should simply continue to live with our structure of knowledge and adjust our expectations of truth: that the not-yet-falsified simply is the best model-which is, obviously, hard to dispute-and that we should continue to deploy reason to select the most robust empirical inferences and the most coherent deductions of principle. But the idea that we could distinguish between different hypotheses consistent with the data or principles based on what "makes the most sense,""sounds the most reasonable," or "seems the most coherent," is simply fantastic. Those types of judgment are so culturally determined and so highly influenced by our particular time and place, it is inconceivable that any rational being today could possibly continue to make those statements at this late stage of modernity-at least, with a straight face.

No more. It is too embarrassing to watch as one generation after another of moderns, under the banner of reason, hop, jump, and skip over the gaps of knowledge. One would have thought that phrenology would have been sufficient to stop us in our tracks, but, no, instead we got biological determinist theories of social behavior applied to male rape, moral poverty theories of delinquency applied to super-predator black males, rational action theories applied to suicide bombers-and the list goes on and on of theories that require so many caveats and exceptions that even a child would question our modern claim to rationality.

We can no longer leap over the not yet falsified. It is no better than turning the clock back and resurrecting faith in divine providence. Foucauldian geneaology does not solve the problem: tracing the formation of belief represents nothing more than seducing us to believe another explanation for which there is hardly stronger evidence. It too requires a leap of faith. The cultural turn also solves nothing. The idea that we could interpret the social meaning of a contemporary punishment practice or institution- let alone a practice that occurred in a completely different historical and cultural context-is complete fantasy. The fact that the cultural interpretation is persuasive to us tells us a lot more about what we find convincing-how we categorize, what kind of evidence we find persuasive, what disciplines we defer to-than it does about the "social meaning" of the practice itself. And even when we do come to a rich description that makes sense of the world around us, even when we achieve that formidable task of symbolic interpretation, we are no closer to drawing normative conclusions. We are located precisely at the gap, forced to take a leap. The symbolic interpretation tells us nothing about how the practice came about, how to transform or change it, or how to modify its social meaning. Social meaning offers no purchase on action.

5.

Where does this leave us? The answer must be randomization. Where our social scientific theories run out, where our principles run dry, we should leave the decision making to chance. We should no longer take that leap of faith, but turn instead to the coin toss, the roll of the dice, the lottery draw-in sum, to randomization and chance. And we should do so, I almost hesitate to say, throughout the field of crime and punishment.

In the realm of searches, surveillance, and detection, law enforcement agencies should turn either to completeness or to random sampling. The Internal Revenue Service could audit tax returns at random using a social security number lottery system. The Transportation Security Administration could search every passenger at the airport, or randomly, select a certain percent based on a computer-generated algorithm using last names. The Occupational Safety and Health Administration could investigate compliance by employers randomly selecting on employer tax identification number. In these and other prophylactic law enforcement investigations, the agency could very easily replace profiling-which rests on uncertain assumptions about responsiveness and rational action-by randomization.

In choosing law enforcement priorities, government agencies should begin allocating resources by chance. The local district attor- ney's office, as well as the federal prosecutor's office, could select annual enforcement targets (as between, for instance, public corruption, insider trading, drug enforcement, or violent crimes) by lottery. State highway policing authorities could distribute patrol cars through a randomized mapping system using heavily trafficked roads and interstate highways. The Bureau of Alcohol, Tobacco, and Firearms could choose between equal-impact initiatives on the basis of an annual lottery draw.

And yes, even in the area of sentencing and corrections, courts and prison administrators should start tninMng about relying more heavily on chance. Judges could impose sentences, following conviction, based on a draw from within a legislatively prescribed sentencing range; the range could easily be determined, for instance, by felony classifications. The department of corrections could assign prisoners to facilities on a random basis within designated escape-risk or security-level categories. Prisoners in need of drug, alcohol, or mental health treatment could be assigned to comparable programs based on a lottery draw.

This is not as far-fetched as it may seem. It does, naturally, assume a sentencing scheme with specified ranges for different degrees of felony. The same kind of randomization, though, could be introduced at the legislative process to decide on actual ranges or to set mandatory sentences (if fixed sentences are preferred to ranges). So, for instance, legislators, having no scientific or principled way to distinguish between 6 or 12 months of imprisonment for an aggravated assault, could turn to chance. Randomization would allow those legislators to pick a mandatory sentence from within those bounds.

The common gesture running through all this is to question and, ultimately, to reject social engineering through criminal punishment. The desire to stop and refuse to take leaps of faith represents nothing more, in practice, than stopping to engineer persons and social relations through the criminal sanction. The central impulse is precisely to resist shaping people by means of punishment-and thereby to wipe the crime and punishment field clean of speculative social science and mdeterminate principle.

Critical reason reveals the limit of our reasoning abilities. It brings us to the gap where our predecessors always took their leap of faith. It sheds light on those theoretical constructs that the moderns used to bridge the gaps of knowledge. It should now also allow us to clear the field of these fabrications. It should free us to use the only unbiased device to decide our fate: randomization, lotteries, dice, and chance. And the point is not to roll the dice as between different theories all of which require a leap of faith, but instead to use critical reason to take those theories off the table. To eliminate them-and thereby to stop social engineering. 6.

How far exactly shall we go with this? Once we have begun to roll the dice, how will we know when to stop? If we use a lottery within sentencing ranges, why not then draw all sentences from the same urn? Why not determine guilt by the toss of a coin? Why not even decide whom to accuse by drawing lots? This is, after all, the whole point of Jorge Luis Borges' brilliant short story, The Lottery of Babylon (1998). Once you go down the path of chance, the road may well lead to hell. Once we have tasted the sweat nectar of chance, where will we find the right place to stop?

The Degree of Skepticism

The central claim is not that we can know nothing. No. We have some basic intuitive knowledge that no one can dispute. As an empirical matter, we know that if we execute someone, we are not going to be able to rehabilitate them. As a matter of principle, we know that murdering an innocent person is worse than stealing her wallet. We know that raping someone is worse than spraying graffiti. We know that punishing an entirely innocent person is wrong. And we can use these minimal ingredients of certainty to set limits to the use of chance. So, for instance, we do not draw punishments for murder and pickpocketing-or for rape and vandalism-from the same urn. We do not decide who to accuse by drawing lots. These elementary forms of knowledge allow us to rest our punishment practices minimally on very basic notions of proportionality. For instance, the convicted murderer and the person exceeding a speed limit are not to be treated the same. We impose proportionality constraints on the use of chance. Perhaps we create a category for homicide, another for serious bodily or psychological injury, and another for property damage. There are some natural limits to the use of randomization.

We only turn to chance when our social science and principles run out. The easy cases are where our social science findings rest on bad evidence, weak data, or faulty models, where there is no scientific evidence at all, or where there are competing and equally plausible hypotheses that are all similarly nonfalsified - in other words, when we do not have reliable social science findings to rely on. This, I take it, can hardly be contested. No one wants to affirmatively and intentionally punish another human being on the basis of bad science or no science at all.

What about the harder case where a reliable social scientific study exists that falsifies a null hypothesis? Here, I would argue that we need to draw distinctions between types of social scientific theory and save only those that involve social physics. By social physics I mean those claims that are necessarily true as a result of the physical nature of our mortal existence. Theories that depend on the intermediation of human consciousness and decision making should be set aside, left to deal with later when we have more leisure time or, perhaps, when we have made breakthroughs in those new consciousness studies. For the time being, though, we should focus on social physics only.

By way of illustration, consider four theories dear to the field of crime and punishment: rational choice theory, the broken-windows theory, legitimacy theory, and incapacitation theory. The first three operate through the intermediary of human consciousness. In each case, the theories depend on actors believing certain things and conforming their behavior accordingly. The first assumes that individuals pursue their self-interest or maximize their utility, and that, accordingly, when the cost of offending goes up, they will offend less. It is a theory that requires us to accept the idea that individuals - whether knowingly or unconsciously - conform their behavior to calculated expectations of success or failure. The second and third theories - broken-windows and legitimacy theories - also depend on people taking cues from their social or physical environment - a disorderly neighborhood in one case, a discourteous or insolent police officer in another - and adapting their behavior accordingly. All three of these theories require a defined process of the human intellect and a decision about behavior. They require the intermediation of human consciousness. They are neither true, nor false, just not yet falsified properly, nor clearly falsifiable in the near future.

In contrast, the fourth theory involves social physics. If we physically detain an individual and isolate her from the free world, she will not commit statutory offenses on the outside. This is a matter of social physics, not modern social science. Similarly, transportation made it physically impossible for a convict to offend in the original jurisdiction. These types of theories alone are respectable hypotheses for the twenty-first century. To be sure, it narrows the range of acceptable empirical and principled claims. But that is all for the better.

My intention is not to prove a social physics hypotheses, but to illustrate a point: critical reason may not necessarily eliminate all social science theories related to crime and punishment. There may still be room for some empirical findings that avoid the intermediation of human consciousness. It is important to note, though, that even those rare claims of social physics will not resolve the policy choices. The fact that incapacitation or transportation makes it physically impossible for the convict to offend (at least, in the original jurisdiction) does not tell us whether that effect will be outweighed by other consciousness- intermediating theories, such as a brutalizing effect on the convict or detrimental consequences on his family and community. It also does not tell us how much incapacitation we should have. It takes us to another empirical and theoretical gap that simply cannot be bridged. The triage and elimination of claims that rest on the intermediation of consciousness will leave us most often without any guidance, without any theory at all. There will be no "best evidence" to fill the void. How then shall we organize our political and social environment? Yes, you know the answer by now: through randomization.

Ultimately, the degree of uncertainty in the punishment field is not complete, but radical. The model of nonfalsification alone is not adequate to sort social science theories apart. We need, in addition, a mechanism that distinguishes between social science hypotheses that intermediate though consciousness and those that do not. The first are far too susceptible to ideology. Only the second can be a valid basis for decision making. I say "ideological" in the sense that social scientific theories that intermediate through consciousness are shaped by historical, social, and familial contexts that change over time and that are affected by the very punitive practices that we implement. In this area, there is a feedback mechanism - what Ian Hacking refers to as a "looping effect" (Hacking, 2006: 2). The practices and categories we deploy shape us as subjects and change the way we respond to those very practices. As a result, these theories will necessarily be filled with ideological commitments, biases, and prejudices. They need to be eliminated from the field of crime and punishment.

The Domain of Uncertainty

Does this extend to the natural sciences - or at least to the hard sciences in contrast to the human sciences? Does it apply to fields where the manipulations do not transform the objects of study or the underlying processes? In biology, for instance, there may be placebo and mental effects on health that intermediate through consciousness. But is it different in physics and chemistry? Even within the human sciences, there seem to be pockets of more reliable social scientific knowledge. Is the setting of fiscal policy or interest rates, for instance, more "certain" than setting the price of crime? Could it be that there is somehow a continuum in terms of the intermediation of conscience? Might it have something to do with the rational element in the decision-making process? In other words, is consumer spending related more to economic rationality than committing crime?

These questions have been a source of formidable debates for centuries. There are, naturally, important differences between the human and hard sciences, not the least of which is the very possibility and use of experimental designs in the natural sciences. Experiments and control groups afford different insights into correlation. But these issues are far too significant to address in a short essay - and they are far too distracting. Distracting, because they paralyze us and prevent us from doing what is right in the field of crime and punishment. Let us leave those questions aside - to be answered some other day - and limit ourselves here to the intentional infliction of pain as a form of punishment. The same reasoning may well apply elsewhere, but that is not our concern. Time is of the essence. Our brothers and sisters are being punished. Right now, we can simply limit these claims of critical reason to the carcerai zone.

Why Turn to Randomization?

Yes, of course, there are other alternatives. At the point of making that punishment decision, there are other options. We could simply stick with what we have done in the past. Do what we did before, use the same punishments as our mothers and fathers. We could heed the status quo. The problem is, their judgments were precisely the product of years and years of uncritical leaps of faith. We will have learned nothing from the exercise of critical reason. Alternatively, we could turn to the democratic process and allow the legislature to decide. But in the end, their vote will reflect nothing more than prejudice, ideology, bias, and, again, leaps of faith. We could decide simply to impose our tastes and aesthetic preferences; but that seems obnoxious and irrational. No, we must turn instead to randomization because we have no other choice. We must turn to the lottery because it is the only way to act within the bounds of critical reason. We must turn to the dice by default. Of course, randomization may have some positive values. It may remind us that our knowledge claims are limited. It may point out the frightening role of ideology in our punishment practices. It may help gather information. By using a form of random sampling, we may in fact learn a lot about the world of deviance that surrounds us. Randomization may offer more transparency in our policy making. And in fact, it may be more efficient than the alternative. But none of these are the reason we turn to randomization. We turn instead because there is no alternative that satisfies critical reason.

7.

Randomization is by no means foreign to the law - and not just in Francois Rabelais' vivid imagination (Rabelais, Tiers livre III: 39, 1999). A number of states today statutorily prescribe a coin toss to resolve election ties (Choper, 2001: 340 n.22). Louisiana law expressly states that "In case of a tie, the secretary of state shall invite the candidates to his office and shall determine the winner by the flip of a coin" (La. Rev. Stat. 46: 1410(C)(3) (2005)). In New Mexico, it is a poker hand that resolves a tie (Reuters, 2000). A number of courts also partition disputed land by lot (Zitter, 1999). Randomness also surfaces across a number of policing strategies, including sobriety checkpoints and the random selection of airline passengers for further screening at airports. It plays a large role in the detection of crime. Our society embraces a "detection lottery" (Duff, 1990: 26-27). Even fixed sentencing schemes have a significant element of chance. A lot turns on the luck of the draw regarding which judge - lenient or stern - presides over the sentencing (Harel and Segal, 1999: 292).

Efficiency and Deterrence

Nevertheless, a call for more randomization will undoubtedly meet with great resistance. Many will instinctively protest that the use of chance is far less efficient than profiling or targeting higher offenders - that it is wasteful to expend law enforcement resources on low-risk offenders. There is no point conducting extra airport security checks on elderly grandmothers in wheelchairs and families with infants - or "Girl Scouts and grannies," as one recent commentator writes (Sperry, 2005). As I demonstrate elsewhere with equations and graphs, however, profiling on the basis of group- offending rates may in fact be counterproduc- tive and may actually cause more crime even under very conservative assumptions regarding the comparative elasticities of the different populations (Harcourt, Against Prediction, 2007a, 129-132; 2007b). We have no good theoretical reason to believe that targeted enforcement would be efficient in decreasing crime or would increase, rather than decrease, overall social welfare.

More sophisticated economists may respond that targeting enforcement on groups that are more responsive, at the margin, would maximize the return of any law enforcement investment (Margoliath, 2007). But here we face an empirical void. What we would need is reliable empirical evidence concerning both the comparative offending rates and the comparative elasticities of the targeted and non-targeted populations. I derive the exact equation for this in Against Prediction (Harcourt, 2007a: 133). That evidence, however, does not exist. The problem is not the reliability of the evidence, it is that it simply does not exist.2 If there ever was a place to avoid taking leaps of faith, surely it would be here, where there is no empirical data whatsoever.

The conventional wisdom among law-and-economists is that increasing the probability of detection serves as a greater deterrent to crime than increasing the amount of the sanction because of the high discount rate imputed to criminals (Polinsky and Shavell, 1999: 12). Assuming this is true, the decision to embrace randomization in sentencing should have no effect on deterrence. Using a sentencing lottery to determine the length of incarceration from within a sentencing guideline range, rather than using a grid that profiles on prior criminal history, gun use, or other factors, would not change the certainty of the expected sentence and need not change the amount of the expected sentence.

Some behavioral law-and-economists had suggested that the certainty of a criminal sentence - the fact that the size of a criminal sanction is fixed and known ahead of time - may deter criminals more effectively than uncertain sentences and, on those grounds, had argued against sentencing lotteries (Harel and Segal, 1999: 280). However, more recent studies involving actual experimental research suggest that uncertainty regarding a sanction may be more effective at deterring deviant behavior. Experiments by Alon Harel, Tom Baker and Tamar Kugler reveal that a sentencing lottery may in fact be more effective at deterring deviant behavior than fixed sentences (Baker, Harel, and Kugler, 2003). Other psychological experiments have similarly shown individuals to be averse to ambiguity (Harel and Segal, 1999: 291).

Just Punishment and Moral Luck

Randomization in sentencing will likely meet much greater resistance, not just because of efficiency concerns but also because of considerations of just punishment and desert. In a 1987 Tanner Lecture, Jon Elster reviewed the arguments for randomization and discussed a number of legal areas where lotteries might make sense. Yet he refused to see

Source: Social Research

More News in this Category