Printer Friendly

Beyond rehabilitation: a new theory of indeterminate sentencing.


Indeterminate sentencing--that is, sentencing offenders to a range of potential imprisonment with the actual release date determined later, typically by a parole board--fell into disrepute among theorists and policymakers in the last three decades of the twentieth century. This sentencing practice had been closely associated with the rehabilitative paradigm in criminal law, which also fell from favor in the 1970s. In the years that followed, most states eliminated or pared back the various devices that had been used to implement indeterminate sentencing, especially parole release. Yet, sentencing remained indeterminate in most places to varying degrees, and now parole and similar mechanisms are staging an unexpected comeback. However, despite its perseverance and apparent resurgence, indeterminate sentencing has lacked any clear theoretical foundation since the demise of the rehabilitative paradigm. Indeed, indeterminate sentencing is commonly thought to conflict with retributivism, the dominant approach to punishment theory today. The lack of a clear theoretical foundation has likely contributed in recent decades to the ad hoc expansion and contraction of parole in response to short-term political and fiscal pressures.

In the hope of bringing greater stability and coherence to what seems once again an increasingly important aspect of our penal practices, this Article proposes a new normative model for indeterminate sentencing that is grounded in a retributive, communicative theory of punishment. In essence, the model conceives of delayed release within the indeterminate range as a retributive response to persistent, willful violations of prison rules. The Article explores the implications of this model for prison and parole administration and for punishment theory.

 A. Major Themes in Contemporary Retributivism
 B. Duff's Communicative Theory
 A. ISS and Limiting Retributivism
 B. Concerns with ISS
 A. Backup Sanctions and the Probation-Prison Analogy
 B. Responding to Objections
 1. Violations of Prison Rules Are Not Breaches of the
 Conditions of Community
 2. Parole Processes Do Not Fit Retributive Purposes
 3. Delayed Release Is Retributively Excessive


Parole is making a comeback. (1) Although it was a universal feature of the American criminal justice system as recently as forty years ago, (2) parole fell into precipitous decline over the final three decades of the twentieth century. By 2000, fifteen states and the federal government had abolished parole altogether, while twenty additional states had formally restricted its availability. (3) Since 2000, however, at least thirty-six states have enhanced release opportunities for prison inmates (4) (although some still resist the "parole" label for their new programs). (5) Highlighting the revival, the Supreme Court's May 2010 decision in Graham v. Florida prohibited the sentence of life without parole for juveniles who have not committed homicide offenses. (6) Indeed, the sentence of life without parole is under assault globally on human rights grounds, (7) and Graham may presage a broader recognition of a constitutional right to earned release from long prison terms in the United States. (8)

The resurgence of parole and similar release mechanisms has the effect of making sentencing more indeterminate: it is harder to know at the moment a prison sentence is imposed exactly how long the offender will be incarcerated. (9) Recent developments thus mark a retreat from the ideal of determinate sentencing, which had gained wide currency in the late twentieth century under the label "truth in sentencing." (10)

Although numerous factors contributed to the decline of parole and the rise of determinate sentencing beginning in the 1970s, (11) among the most notable were the discrediting of rehabilitative programs for prison inmates and the growth in retributive theorizing about punishment. (12) Parole and indeterminate sentencing had been closely associated with the rehabilitative paradigm that dominated American criminology through the middle decades of the twentieth century. (13) The opportunity for accelerated release was embraced as a way to encourage and recognize the rehabilitative efforts of prison inmates. (14) By the mid-1970s, however, researchers were proclaiming that existing rehabilitative programs produced little demonstrable reduction in recidivism rates, (15) which contributed to a broader rejection of the rehabilitative paradigm in criminal law. (16) At the same time, retributive theories of punishment, which assert that society ought to respond to crime with condemnation instead of treatment, attracted growing interest among scholars and policymakers. (17)

The recent revival of parole and indeterminate sentencing may owe less to such broad shifts in penal theory than to the fiscal pressures created by burgeoning prison populations. Although retribution maintains its preeminent position among theorists, (18) policymakers have been forced to confront the budgetary legacy of the sentencing severity revolution of the late twentieth century, (19) and they have had to do so recently against a backdrop of economic turmoil and stagnant revenue. (20) Not surprisingly, policymakers have found early release to be an attractive option, particularly to the extent that it can be implemented without obvious public-safety hazards. (21)

An apparent disconnect has thus emerged between our dominant theory of criminal punishment and an increasingly important aspect of our practice. To the retributivist, the legitimacy of punishment is based on its character as a morally fitting response to a criminal offense. Yet, indeterminate sentencing typically contemplates that an offender's actual release date from prison (and hence the actual severity of the sentence) will be based on considerations other than just the gravity of the offense, such as the offender's disciplinary record in prison.

Consider an example. Two brothers, Jesse and James, jointly plan and perpetrate a robbery. After police apprehend them, they are tried and convicted of the same offense. Neither has a prior record, and both receive a nine-year indeterminate sentence, with release from prison to occur sometime between three and six years. (22) Jesse does well in prison and is released at the earliest possible time (i.e., at the three-year mark). James, however, proves an incorrigible troublemaker and is held until the latest possible release date (at the six-year mark). Retributivists would be uncomfortable with such an outcome because it appears that two offenders who committed the same crime have received dramatically different punishments--one spends twice as much time in prison as the other. (23) Prima facie, it appears that Jesse has received an overly lenient punishment, or James has received an overly harsh punishment, or both.

Yet, the indeterminacy-retribution disconnect may not be as great as the illustration seems to suggest. My purpose in this Article is to describe one way that indeterminate sentencing may be reconciled with a retributive approach to punishment. Since the breakdown of the traditional rehabilitative paradigm in the 1970s, indeterminate sentencing has lacked a clear theoretical foundation. Such a foundation might help to bring greater coherence to parole policies and to diminish the ad hoc expansion and contraction of parole in response to fiscal pressures and anticrime political frenzies. (24) Such a foundation may also help to bring greater coherence to the emerging Eighth Amendment jurisprudence requiring release opportunities for at least some classes of offenders. My hope is that this Article will contribute to the construction of such a foundation by showing how parole and indeterminate sentencing may be connected to our dominant theoretical approach to punishment. (25) To be clear, though, I am not making the strong claim that retributivism must be implemented through indeterminate sentencing, or even the somewhat weaker claim that retributivism is best implemented through indeterminate sentencing; rather, I advance the more modest claim that, taking indeterminate sentencing as a given, variability in release dates can be conceptualized and structured in a retributively acceptable fashion.

Because the terms "indeterminate sentencing" and "parole" have acquired many diverse connotations over the years, I should make clear at the outset that my subject--the practice that I will try to reconcile with retributivism--is earned early release from prison to community-based supervision through a regularized bureaucratic decision-making process. (26) "Earned" means that release depends in large part on performance while in prison. "Early release" means that release occurs before the completion of some maximum period of possible confinement contemplated by the sentence. "Community-based supervision" means that release is conditional and monitored; the violation of release conditions may result in sanctions up to and including reincarceration. "Bureaucratic decision-making process" means that early release is not merely an ad hoc act of grace, like executive clemency, but instead occurs pursuant to reasonably transparent, consistently applied standards and procedures.

Understood as a mechanism to implement this sort of earned release, indeterminate sentencing's apparent conflict with retributivism rests largely on the mistaken premise that retributivism necessarily and inflexibly demands that the same degree of suffering be imposed on all criminals who are guilty of the same offense. As I will show, however, retributivism--at least in some of its more sophisticated contemporary formulations--focuses less on the suffering of offenders per se than in the communication of particular sorts of messages of censure. (27) Indeed, one influential strand of "communicative" retributivism, perhaps best articulated by the philosopher Antony Duff, envisions criminal punishment as a form of secular penance. (28) It is this formulation of retributivism that I will particularly draw on in order to show how indeterminate sentencing may be conceived and structured as a morally fitting response to crime, and not merely as a balancing of various utilitarian considerations.

The Article proceeds as follows. Part I briefly reviews contemporary "communicative" theories of retribution, focusing particularly on Duff's work. Part II considers a leading recent effort to reconcile retributivism and indeterminate sentencing, Steven Chanenson's "Indeterminate Structured Sentencing" proposal. Chanenson's model is a very helpful starting point, but it may not provide a wholly satisfactory account of how the full amount of time served in prison under an indeterminate sentence is retributively justified. Part III attempts to provide such an account, making use of Duff's concepts of secular penance and backup sanctions, and address a variety of objections to this account. Part IV concludes the Article with reflections on the implications of the analysis for sentencing and corrections reform and for punishment theory.


Although there is considerable variation in the views of retributive theorists, (29) several common themes unite many of the leading contemporary approaches to retribution. In this Part, I will first lay out some of these common themes and then discuss in a bit greater detail the one particular theory (Duff's) that will provide the foundation for my retributive account of indeterminate sentencing.

A. Major Themes in Contemporary Retributivism

In the popular imagination, "retribution" may be equated with vengeance or the talionic law of "an eye for an eye," but contemporary retributive theories do not aim for some sort of equalization of the suffering of victims and offenders as an end in itself. Rather, many leading contemporary theorists see punishment (if properly understood and structured along retributive lines) as a form of communication (30)--as a morally fitting way for society to express its condemnation of criminal wrongdoing and to vindicate the worth of victims. (31) This communicative aim is seen as capable of justifying punishment without regard to the sorts of crime-control benefits that utilitarians emphasize as the purpose of punishment. (32)

What many retributive thinkers aim to communicate through punishment involves core liberal values like individual autonomy and equality before the law--indeed, the instantiation of these values helps to define what makes a punishment morally fitting. (33) Consistent with these values, retributivists emphasize both uniformity (treating similar offenses similarly) and proportionality (avoiding excessive penal harshness and lenience relative to the severity of the offense) as important goals of sentencing. (34) But retributivists also recognize that uniformity and proportionality can be difficult to operationalize, and do not necessarily treat them as absolute requirements that must be elevated over all other considerations for sentencing purposes. (35)

Among retributivists, the conceptual difficulties in implementing proportionality have to some extent been addressed by a shift in focus from cardinal to ordinal proportionality. (36) Ordinal proportionality is oriented to system-wide practices and demands that relatively more serious offenses be punished with greater severity than less serious offenses. (37) Some indeterminacy remains, though, with respect to "anchoring points" (what severity of punishment should be associated with the most and least serious offenses?) (38) and what criteria should be used to determine offense severity. (39)

B. Duff's Communicative Theory

Antony Duff's communicative theory echoes many of the important themes in contemporary retributivism. Although I do not mean to suggest that Duff's views are shared by all retributivists in all respects, (40) Duff's work is undeniably among the most influential and sophisticated writing on punishment theory in the past generation, (41) and I will make it the theoretical foundation for most of what follows.

As Duff understands things, a political community that punishes communicatively aims to censure its criminal offenders in a way that treats the offenders both as responsible moral agents and as fellow members of the community. Communicative punishment has both a backward- and a forward-looking character. It looks backward because
 it takes the primary communicative purpose of punishment to be the
 communication to offenders of the condemnation they deserve for the
 wrongs they have committed, and explains that purpose in
 backward-looking terms of what we, as a polity, owe to victims, to
 offenders, and to ourselves as a political community ... as a
 response to such wrongs.... (42)

Yet, communicative punishment also looks forward because of the hope that censure will cause the offender "to understand, and so to repent, [his or her] wrong as a wrong both against the individual victim (when there was one) and against the wider political community to which they both belong." (43) This understanding is intended to accomplish a sort of "moral rehabilitation" of the offender and lead to the performance of reparative actions, even if only in the articulation of an apology, and to reconciliation with the victim and the community. (44)

Despite this forward-looking character, communicative punishment still differs from utilitarian and other traditional consequentialist approaches because it seeks to persuade the offender in a respectful fashion as an autonomous moral agent and not simply to control the offender through brute coercion. A hope that the offender will repent and reform is internal to the practice of censuring; such a hope does not imply that censuring is being thought of in purely instrumental terms. (45) Indeed, the communicative approach recognizes important ethical constraints on the pursuit of efficient crime control:
 If I try to persuade another person to do what (I think) she
 morally ought to do, my aim is of course that she should come to do
 right. But if I am to treat her as a moral agent, as a member of
 the moral community to which we both belong, my aim cannot be
 simply to find some efficient means of bringing it about that her
 conduct conforms to what morality requires.... Such an aim could in
 principle be achieved by deceiving her or bullying her or
 manipulating her emotions or beliefs--methods that fail to respect
 her as a moral agent. (46)

For present purposes, four (overlapping) aspects of Duff's account of communicative punishment bear particular emphasis. First, punishment ought to function as a form of secular penance: "It is a burden imposed on an offender for his crime, through which, it is hoped, he will come to repent his crime, to begin to reform himself, and thus reconcile himself with those he has wronged." (47) Punishment should focus the offender's attention on his wrong in a sustained fashion. (48) It should also help the offender to see what he must do to reform himself in order to avoid such wrongs in the future. (49) Finally, it should offer the offender an opportunity to seek reconciliation with those he has wronged through a process of apologetic reparation. For serious offenses, at least, the offender must normally undertake some burdensome work that "expresses to those concerned her repentant recognition of the wrong she has done." (50)

Second, there should be a substantive fit between the punishment and the crime. (51) Different modes of punishment (say, a fine or community service) carry different meanings. Sentencers ought to select a mode in each case based on the suitability of its meaning as a response to the specific offense; we must ask "not just how different modes of punishment can serve the general aim of communicating censure to offenders, but also how they can help to communicate a more substantive understanding of the wrongfulness and the implications of particular kinds of offense." (52) Among other implications, this concern for fit suggests a much diminished use of imprisonment, which must be reserved for only the most serious offenses. (53) "The message of imprisonment is that the offender has not just damaged or threatened, but has broken, the normative bonds of community. He has made it impossible for us to live with him in the ordinary community of fellow citizenship unless and until he has undergone this penitential punishment." (54)

Third, there must be proportionality between the severity of the offense and the severity of the punishment. (55) In light of fit considerations, however, "negative proportionality" is a more suitable objective than "positive." Positive proportionality demands greater rigor in ensuring that "the" proportionate quantum of punishment is imposed, while negative proportionality seeks a range of possible sentences, any of which would be "good enough." (56) The negative approach is preferable because it "gives sentencers more room to attend to the concrete particularities of the crime, without worrying about rendering it commensurable with all other crimes in terms of its seriousness, and to choose between a wider range of penalties, without worrying about rendering them all commensurable in terms of their severity." (57)

Finally, punishment should aim to persuade offenders in a way that leaves them the freedom to remain unpersuaded. (58) This follows from the requirement that "punishment must address the offender as a member of the liberal polity whose autonomy must, like that of any citizen, be respected." (59) Thus, while "offenders are forced to hear the message that punishment aims to communicate and to undergo a penal process intended to persuade them to accept it, they are not forced to listen to that message or to be persuaded by it." (60) Among other things, this consideration means that sentences should not be extended indefinitely for offenders who remain unrepentant. (61)


Having now identified some of the key themes in contemporary retributivism and described their elaboration in one particular retributive theory, I return in this Part to the central question of this Article: can indeterminate sentencing be conceived and structured in a way that is consistent with retributivism? Steven Chanenson has offered the most thoughtful recent attempt to reconcile indeterminate sentencing with retributivism. In this Part, I will first describe Chanenson's proposed penal scheme (which he calls "Indeterminate Structured Sentencing," or "ISS"), (62) and then lay out some criticisms of the model.

A. ISS and Limiting Retributivism

The ISS model relies on the late Norval Morris's theory of "limiting retributivism." (63) Consistent with Duff's concept of negative proportionality, Morris discussed the proportionality required by retributivism in terms of a range. (64) Between the minimum and maximum levels of punishment that were retributively appropriate, Morris believed, a sentencing judge could permissibly take other considerations into account, including utilitarian crime-control objectives like rehabilitation and incapacitation. (65)

Although Morris himself was an outspoken critic of parole, (66) Chanenson observes that Morris's theory can provide a helpful framework for thinking about the respective roles of the court and the parole board in a system of indeterminate sentencing. (67) First, the judge imposes a sentence based on retributive proportionality considerations (68); consistent with Morris's view that retributive judgments cannot be precise, the sentence is expressed as a range, such as two-to-four years. (69) Then, the parole board determines when exactly an inmate is released within that range. (70) The board decides subject to parole guidelines that "encourage" release when the minimum amount of time has been served (e.g., two years in the example), but that permit release to be deferred, based, for instance, on an "inmate's violation of certain prison rules or a clear conclusion that the inmate would present an unusually high risk if released." (71) The board's authority is thus exercised so as to advance the sorts of utilitarian aims that Morris indicated could appropriately be pursued within the constraints of negative proportionality. Chanenson notes research suggesting that the availability of parole promotes rehabilitation by creating incentives for inmates to participate in prison programs, and also that our ability to measure recidivism risk accurately has improved since Morris first offered his critique of parole. (72)

Chanenson elaborates his proposal with much greater richness of thought and attention to detail than I can convey here; for present purposes I will note just one additional aspect of the proposal: a "super commission," which is charged with developing guidelines at the state level for both the sentencing and parole functions. (73) Although the guidelines include some flexibility for exceptional cases, they are intended to ensure at least a "reasonable" degree of uniformity in the treatment of similarly situated offenders. (74)

B. Concerns with ISS

Kevin Reitz has noted a number of objections to the idea of adapting limiting retributivism as the basis for a system of indeterminate sentencing. (75) Reitz's criticisms focus particularly on potential weaknesses of indeterminate sentencing from a utilitarian, crime-control perspective. He questions, for instance, whether information regarding in-prison performance can help parole boards to make reliable determinations of future dangerousness. (76) Although some research does support this proposition, (77) Reitz characterizes the work as "preliminary and only suggestive," (78) and notes that future study is warranted. (79)

Chanenson responds that the most effective way to find out whether parole boards really can serve rehabilitative and incapacitative ends through the exercise of guided discretion is to adopt a system in which parole boards actually attempt to do so and then study the results. (80) Still, Reitz's observations regarding the current state of the empirical research are a helpful reminder that we cannot take for granted that incapacitive and rehabilitative objectives will offer a stable, persuasive justification for indeterminate sentencing--it may be that further research will ultimately tend to vindicate Morris's critique of parole from forty years ago.

Indeed, even assuming incapacitation and rehabilitation benefits, it is possible that other aspects of indeterminate sentencing would lead to net costs from a crime-control perspective. In particular, the uncertainties surrounding real sentence length may diminish the deterrent value of punishment. If the bottom end of the proportionality-driven range imposed by the ISS sentencing judge normally achieves lesser deterrent effects than the top of the range, then judges might actually maximize crime-control benefits by simply imposing a determinate sentence at the top of the range. Although a risk-averse prospective offender might be deterred in the ISS scheme by the possibility of ending up at the top of the range, there is reason to think that criminals tend not to be especially risk-averse (81) and might instead decide to take their chances with the possibility that a parole board would release them closer to the bare minimum required by proportionality. (82)

The point here is not that crime-control utilitarianism is incapable of justifying ISS, but simply that it is not clear whether ISS would perform better than a determinate sentencing alternative (for instance, one that would permit the sentencing judge to select a determinate sentence somewhere within the retributively appropriate range or that would simply require the sentencing judge to impose the least--or most--severe sentence that is consistent with the proportionality requirement). As Chanenson argues, this uncertainty is not necessarily a basis for rejecting ISS, (83) but it is possible that experience with ISS will diminish, not strengthen, confidence in its crime-control capabilities.

At the same time, given the number of variables and complex social and psychological dynamics involved, the uncertainty surrounding the deterrence, incapacitation, and rehabilitation effects of ISS (or any other possible sentencing regime) will probably never be fully resolved. The uncertainties have likely contributed to the instability of existing indeterminate sentencing schemes in the face of fiscal and political pressures, and they might cause similar instability in an ISS scheme. Pro-parole advocates likely overestimate the ability of increasingly sophisticated instruments to provide a persuasive ground for increasing the generosity of parole. Consider, for instance, this assertion by the Justice Policy Institute: "Decisions based on evidence rather than anecdotes would help people in prison avoid perpetual denial of timely parole because of the single incidents--statistical outliers--that have dominated decision-making to date." (84) It is not clear, though, that public outrage over violent crimes committed by parolees would be made any less intense by reassurances that the incidents were only "statistical outliers." Moreover, no regression analysis can really supply a response to the argument that any crime committed by a parolee is one crime too many. Put differently, while risk-assessment instruments can give us a sense of the dangerousness of any given inmate, they cannot tell us what level of danger, if any, is tolerable.

All of this underscores the need (which Duff and many other retributivists would insist upon in any event) (85) to establish some firm ethical foundation for ISS beyond its crime-control potential--some reason to view ISS as an intrinsically (and not just contingently) appropriate response to crime. To be sure, ISS is designed to satisfy proportionality and uniformity requirements, but retributivism--properly understood--may demand more.

Imagine an offender sentenced to a term of two-to-four years under ISS guidelines. At the end of two years, the parole board denies release because it reaches (as Chanenson puts it) "a clear conclusion that the inmate would present an unusually high risk if released." (86) On that basis, the offender remains in prison until the four-year mark. What are we to make of the incremental two years from a retributive standpoint? By stipulation, it does not make the overall punishment excessively severe--anything in the range of two-to-four years would be consistent with the demands of proportionality. But the communicative character and appropriateness of a punishment is not merely a function of its severity.

From a retributive perspective, what seems most problematic about the incremental two years is that it may appear to be nothing more than preventive detention. (87) Although the two years were within the range of what could have been justified based on the severity of the offense, they were instead framed as a means to prevent the offender from committing additional crimes. This is troubling on at least two grounds. First, ISS does not seem to contemplate any difference in the conditions of confinement based on whether an inmate is serving the initial, retributively justified portion of the sentence or is being preventively detained by the parole board. Paul Robinson and John Darley have raised important ethical concerns about such conflations of penal incarceration and preventive detention:
 The conditions of confinement upon commitment under the criminal
 justice system are conditions of punishment. Yet, the justification
 for confinement under an incapacitation strategy is not punishment
 but prevention, akin to the system of preventive detention that we
 use for those with infectious diseases or mental illness that is
 likely to lead to violent behavior. Systems of preventive detention
 are morally ambiguous, but certainly we are most comfortable with
 them when they involve detention conditions that are not punitive
 in nature, involve "treatment" efforts that attempt to remove the
 elements in the individual that cause the presumed dangerousness,
 and continually reassess the dangerousness of the individual who is
 incarcerated. This is not a good description of the workings of the
 prison system in the United States. (88)

To be sure, this objection might be overcome if we offered inmates appropriate, improved conditions of confinement once they moved past the minimum retributively required portion of their sentences and into the preventive-detention portion. Yet, creating and maintaining separate facilities for such inmates might be costly, further diminishing the likelihood that ISS could be persuasively justified on utilitarian grounds.

Second, the parole board finding of "high risk" that forms the basis for incremental incarceration implies a measure of inevitability regarding the inmate's recidivism, which in turn raises doubts about the extent to which an inmate's conduct is determined by social or psychological circumstances beyond his or her control. There is thus at least some tension between a "high risk" finding and the underlying conviction and sentencing judgment that punishment is retributively justified; after all, retributive punishment is "suited only to beings capable of moral agency." (89) This is not to say that a finding of high risk is intrinsically logically inconsistent with retribution, but it may impose a burden of sorts on the parole board to explain its judgment in ways that do not undermine confidence in the inmate's moral responsibility for the underlying offense.

This task may be made more difficult by the use in risk determinations of so-called "static factors" (that is, factors an inmate cannot change, such as age, gender, and childhood experience). (90) Such factors are important to take into account if a parole board wishes to make the most accurate possible predictions of risk, (91) which may be required of a morally legitimate system of preventive detention. Yet, the parole board's reliance on static factors as a basis for denying parole may undercut the credibility of the criminal justice system in at least two respects. First, to the inmate denied parole based on factors present from the start of his or her incarceration, the system will seem dishonest for having held out the possibility of parole at the time of sentencing. Second, the implicit recognition that the inmate labors under circumstances that substantially increase his or her likelihood of falling into crime may cause underlying determinations of full moral responsibility at trial and sentencing to appear superficial or disingenuous. Ultimately, the legitimacy of the criminal justice system is fairly put into question when two of its central actors, the sentencing judge and the parole board, speak in quite different terms of the offender's capacity for self-control and prospects for release. (92)


If we are seeking a new theoretical model of indeterminate sentencing that takes retributivism seriously, ISS seems like a helpful starting place. Chanenson's proposal respects the key principles of proportionality and uniformity. Recognizing some flexibility in these concepts, however, Chanenson would permit a parole board to determine actual release dates based on rehabilitation and incapacitation considerations. Thus, other than establishing some (possibly broad) parameters within which the release date must be set, it is not clear that retributive thinking actually does much work in the ISS model. Indeed, as I suggested in Part II, ISS might in practice function as a pure preventive detention scheme within the retributive parameters. Such a scheme might be objectionable on grounds having to do with both efficacy and ethics, and would be challenging to implement in a way that did not appear incoherent or dishonest.

ISS may thus benefit from a bit of refining and reframing in order to provide a more stable and retributively satisfying theoretical foundation for indeterminate sentencing. The challenge, more specifically, is to provide a retributive account of what might justify a parole board's decision to defer release beyond the minimum period of incarceration required by negative proportionality.

In order to provide such an account, I will draw on Antony Duff's concept of backup sanctions. Although Duff developed this account to explain why incarceration might sometimes be a retributively appropriate response to violations of probation rules, I will argue that an analogy might be drawn to delayed release as a response to prison violations. To be clear, I do not discount the possibility that indeterminate sentencing might be retributively justified on other grounds, for instance, as a mechanism for exercising mercy on behalf of the genuinely repentant (93) or a way to achieve the right punitive intensity at the least cost. (94) Moreover, while the approach I describe in this Part seems to me the most promising, I acknowledge that it is subject to several important objections, several of which I identify and address in Section B below.

A. Backup Sanctions and the Probation-Prison Analogy

In order to understand Duff's concept of backup sanctions, we must first consider his account of probation as a model of punishment. Although probation is sometimes characterized as an alternative to punishment, Duff has sought to reclaim probation as a retributively appropriate sanction. (95) Consistent with his communicative aims, he identifies the "meaning" of the two basic elements of probation: supervision and conditions. (96) Supervision, which involves regular reporting to a probation officer, counts as punishment insofar as it constitutes a court-imposed burden on the offender for his offense. (97) But it is not just any burden; supervision communicates a particular type of censure:
 [T]his punishment tells [the offender] that his commission of the
 crime cast doubt on his commitment to the community's public values
 (the values embodied in the criminal law), threatening to undermine
 the mutual trust on which the community depends; he must therefore
 subject himself to a kind of supervision that other citizens need
 not accept. (98)

Conditions, the second basic element of probation, also convey meaning, although the meaning will vary depending on the specific conditions imposed. For instance, some conditions may aim to facilitate supervision, such as residency-related conditions, and the meaning they convey complements the basic meaning of supervision. (99) Other conditions may attempt to weaken crimogenic influences in the offender's life, or remove the offender from situations that seem likely to result in further criminality. (100) Examples would include mandatory drug treatment or rules against associating with certain people. (101) Conditions of this sort may convey a "message that the offender has shown by his crime that he cannot be trusted to deal with his problems in a way consistent with the criminal law." (102) The problem with conditions, of course, is that they may not be satisfied. What then? Duff would distinguish "willful defaulters from those whose default is due to incompetence or lack of resources, or to the fact that their original sentence was not realistic or feasible." (103) As to the latter, Duff suggests help instead of punishment. (104) But as to the former, the willful defaulters, some "backup sanction" is necessary, and, as Duff observes, "it is hard to see what alternative to imprisonment there would be in such cases." (105) Thus, in Duff's view, a relatively short prison term might be an acceptable and proportionate punishment, at least for an offender who persistently violates his conditions of probation. (106) "Such a persistent refusal to comply with what is penally required becomes a serious enough breach of the conditions of community to warrant imprisonment." (107)

Before developing the probation-prison analogy, it is important to clarify one point. Although parole is sometimes framed as "early release," this convention is artificial. We could just as easily say that the normal release date for a prison inmate is the time when he or she is first considered for parole, and an initial denial of parole amounts to "delayed release." This is how Chanenson discusses indeterminate sentencing in his ISS proposal--as providing for delayed, not early, release--and I will use the same framing here.

What might justify delayed release? In current practice, parole boards sometimes delay release based on their view of offense severity. (108) Yet, offense severity should be fully taken into account by the judge when the indeterminate sentence is imposed. Unless there is some reason to think that parole boards are better positioned to make retributive judgments regarding offense severity than sentencing judges, it would seem awkward and redundant for a parole board to reweigh the same offense characteristics that have already been evaluated by the sentencing judge. Because few, if any, parole-board advantages are apparent, (109) I will assume (as Chanenson does) that the parole board ought to focus its attention on matters other than offense severity.

Chanenson notes two different sorts of grounds that might justify delayed release in his ISS scheme: high risk and violations of prison rules. (110) I've already discussed objections to the first ground, but what about the second--can this justify delayed release? Prison performance has traditionally been a very important part of the parole calculus. (111) At least a couple of utilitarian explanations for this are possible: poor prison performance may indicate heightened recidivism risk upon release, and the threat of delayed release may deter rule violations. Duff's concept of backup sanctions, though, provides an alternative, retributive justification for delaying release based on disciplinary infractions.

More specifically, prison rules may be thought of as the functional equivalent of probation conditions. Many are designed to facilitate supervision of inmates, such as rules limiting access to certain areas or requiring presence in the cell at designated times. Other rules take aim at the factors that contribute to criminal misconduct by inmates, such as possession of drugs and weapons. Still other rules bar conduct that would be criminal even on the outside, such as violent assaults. Prison rules can thus be seen as conveying the message to inmates that they have put their trustworthiness into doubt, as Duff characterizes the basic message of probation, and thereby aiming to persuade "offenders to face up to and repent their crimes." (112) The inmate's efforts to comply with this regime can be seen as a secular penance--burdensome work that "expresses to those concerned [the offender's] repentant recognition of the wrong she has done." (113) On the other hand, persistent, willful violations of prison rules warrant backup sanctions for the same reason that violations of probation conditions warrant such sanctions. (Additional) imprisonment seems no less appropriate a sanction in the prison than the probation context.

This, then, is the key refinement I propose for the ISS model (the refinement that permits a more fully retributive justification for delayed release): violations of prison requirements must become the dominant consideration in parole board decision-making. (114) A parole board would delay an inmate's release if, and only if, the board found that the inmate had engaged in persistent, willful violations of prison rules. In such cases, delayed release--not indefinitely delayed, but within the parameters of the indeterminate sentence--can be conceived as a retributively appropriate response to the inmate's failure to engage in an even minimally acceptable fashion with the penitential regimen of prison life. Although ISS, thus refined, would work differently than existing indeterminate sentencing schemes typically do, its reliance on prison performance gives it at least a close family resemblance to current practices--close enough, we might hope, that this vision of indeterminate sentencing could provide a realistic and helpful normative model in ongoing policy debates regarding sentencing and imprisonment.

B. Responding to Objections

In the interests of setting forth the model and its basic logic succinctly, I have bypassed several important objections. These objections, although varied in form and discussed separately below, all ultimately point to a single overarching challenge: whatever superficial resemblances exist between my retributive model of indeterminate sentencing and current practices, closer scrutiny may reveal that my model rests on premises so wildly at odds with reality that the model really only offers, at best, an interesting thought experiment for a rainy afternoon. (115)

1. Violations of Prison Rules Are Not Breaches of the Conditions of Community

To some, it may seem odd to suggest that violators of prison rules are properly punished for "breach[ing] of the conditions of community." (116) Although probationers still reside in the community, prisoners seem excluded from it. As a result of the profound separation from the life of the community that we impose on prisoners, perhaps we lose the sort of moral claim on them that would permit us to punish them for the rules that we adopt for the management of our prisons. We might appropriately exercise control over prisoners, much as we might seek to exercise control over combatants from an enemy nation on the battlefield or in a prisoner-of-war camp, but--one might argue--the rhetoric of punishment is not morally fitting when a prisoner resists our efforts to control him. (117) Indeed, prison administrators themselves commonly employ a rhetoric of control, and avoid the language of retributive punishment, when they discuss what they do. (118) Consistent with that rhetoric, prison rules and disciplinary actions often have a petty or self-serving flavor. (119) Indeed, a profound status inequality between inmates and corrections officials seems manifest in the wide discretion afforded officials and the seeming arbitrariness with which they sometimes exercise their power. (120) Whatever else might be said for or against such arrangements, they may threaten the legitimacy of punishment (properly understood in its retributive sense) for violations of prison rules.

Still, it may not be necessary for our corrections practices to be what they so often are--such exclusion and degradation may not be intrinsic to incarceration. (121) Duff argues, for instance, that prison may be administered in such a way as to constitute a form of penance (that is, as a "burden through which [the offender] can, if he does repent, express his repentant and apologetic understanding of crime, or by which, even if he does not repent, he is restored to normal community as if he had repented"). (122) Duff contends that we ought to recognize an important distinction between exile and imprisonment: the former is mere exclusion from the community, while the latter "subjects [the offender] to a particular regime. A temporary exile is left to make his own way outside the community from which he is exiled, whereas imprisoned offenders are required or forced to live within a prison, under the rules by which it is organized." (123)

Although Duff does not assert that existing prison rules are suitable to penitential purposes, he does believe that a prison regime might be penitential if it "treat[ed] those subject to it, as far as possible consistently with their detention, as members of the normative community--as citizens, not as outcasts." (124) For instance, he urges that programs and provisions imposed on inmates should be directed at their criminal conduct, much as probation conditions should be tailored so as to communicate appropriate, particularized penal messages to offenders who are on probation. (125) Moreover, to soften the message of exclusion and provide hope for reconciliation with the community, prisoners should be offered help to equip them for life after their terms are up. (126) A prison regime along these lines could perhaps provide a morally satisfactory foundation for the model of indeterminate sentencing I have proposed.

And perhaps we are not so profoundly far from such a regime. (127) Certainly, it is possible to overstate the extent to which contemporary imprisonment is exclusionary, degrading, or arbitrarily administered. Indeed, while considerable variation among institutions must be acknowledged, (128) imprisonment is probably on the whole much less these things today than it was a generation ago. Modern media and electronic communications permit inmates to retain a richer sense of contact with the outside community. (129) Recent trends in prison design facilitate association among inmates, diminish the visibility of restraints and symbols of official power, and generally aim for normalization of the physical environment. (130) Court intervention beginning in the 1970s has improved inmate access to courts and lawyers, led to more transparent policies and procedures, and generally made prisons more humane places to be. (131) Even after a period of retrenchment in the law, (132) it remains well accepted now that "prison walls do not form a barrier separating prison inmates from the protections of the Constitution." (133) Also under court pressure, inmate classification schemes have become far more refined, (134) which might be thought of as a move towards Duff's idea of tailoring prison conditions to fit the offense. Finally, even though the political system does not typically give them adequate resources for this purpose, (135) thoughtful prison administrators do recognize the importance of rehabilitative programming for inmates and have moved beyond the pessimism of the 1970s. (136)

The dueling opinions of Chief Justice Burger and Justice Stevens in the important prisoner rights case Hudson v. Palmer (137) nicely illustrate the competing views of what prison is and should be that are suggested by the discussion in this Section. In Hudson, the Court held that prison inmates do not have a reasonable expectation of privacy in their cells and thus may not invoke Fourth Amendment protections against unreasonable searches and seizures of property in their cells.138 Because the legal analysis turned on "whether a prisoner's expectation of privacy in his prison cell is the kind of expectation that society is prepared to recognize as reasonable," (139) the majority opinion of Burger and the dissenting opinion of Stevens focused on what the Justices perceived to be the essential realities of life in contemporary American prisons.

Burger emphasized the "otherness" of inmates: "they have shown an inability to regulate their conduct in a way that reflects either a respect for law or an appreciation of the rights of others." (140) After reciting statistics on prison violence, Burger observed, "[w]ithin this volatile 'community,' prison administrators are to take all necessary steps to ensure the safety of not only the prison staffs and administrative personnel, but also visitors." (141) The scare quotes around "community" are noteworthy: Burger's point is that the term "community" cannot be used to describe prisons other than in an ironic sense. Since inmates have an "inability" to respect law and the rights of others, they are seemingly incapable of living in a relationship with others that can properly be called "community." In Duff's terms, then, Burger is depicting the prison as a place of exile--it makes little sense, as Burger sees it, for us to think about inmates as fellow citizens who, although temporarily separated from the larger community in a physical sense, retain their membership in the community in a moral sense. Inmates cannot be counted on to respond to the sense of mutual respect that normally sustains peaceful social relations, so there is nothing that we can do but seek to control them in a direct, physical way. Burger thus concluded that it would be "literally impossible" to ensure safety in prisons if inmates retained a "right of privacy" in their cells. (142)

By contrast, Justice Stevens emphasized not the otherness of inmates, but their humanity and their hopes for full restoration to the outside community:
 From the standpoint of the prisoner, [the] residuum [of privacy
 that can be retained in the cell] may mark the difference between
 slavery and humanity.... Personal letters, snapshots of family
 members, a souvenir, a deck of cards, a hobby kit, perhaps a diary
 or a training manual for an apprentice in a new trade, or even a
 Bible--a variety of inexpensive items may enable a prisoner to
 maintain contact with some part of his past and an eye to the
 possibility of a better future. (143)

Stevens also noted the network of legal protections accorded prisoners by the wider community under state property law, the Due Process Clause, the Eighth Amendment, and the First Amendment. (144)

Stevens argued that Fourth Amendment protections should also be extended to prisoners so as to promote self-respect and reform. (145) He noted that prison administrators themselves recognized that arbitrary seizures of inmate property are contrary to institutional interests, and that this view was already embodied in the rules of many prison systems. (146) Moreover, in considering whether society is prepared to recognize the privacy expectations of prisoners as reasonable, he suggested that we ought to take into account the fact "that most inmates have family or friends who retain an interest in their well-being"; "one must acknowledge that millions of citizens may well believe that prisoners should retain some residuum of privacy and possessory rights." (147)

Stevens was not persuaded by the majority's depiction of prisons as powder kegs constantly on the verge of explosion, and countered the majority's statistics on prison violence with his own statistics showing that the rates of violent crime in prison are actually lower than in many large cities. (148) He noted that "literally thousands upon thousands of former prisoners are leading constructive law-abiding lives." (149) In ignoring the great diversity and rehabilitative potential of prisoners, the majority improperly accepted the view of prisons as "sterile warehouses." (150)

In sum, the vision of prison offered by Stevens is one in which inmates retain important, reciprocal, affective connections with the wider community; inmates are not seen as uniformly incorrigible, but as generally capable and desirous of resuming law-abiding lives outside prison; inmates share many of the legal rights possessed by citizens outside of prison; prison administrators are devoted to the reform of inmates; and, in their interactions with inmates, prison employees are bound by rules intended to achieve not merely physical security in the prison, but also to preserve the dignity of inmates. A prison system closely reflecting this vision would provide an experience for inmates that is profoundly different from mere exile, and that might well supply a morally satisfactory foundation for my retributive indeterminate sentencing model.

To be sure, Stevens lost in Hudson, which might be taken as a rejection by the legal system of the desirability or feasibility of Stevens' vision of the prison. On the other hand, three other Justices joined Stevens in a close 5-4 decision, his position was the dominant position in the lower courts pre-Hudson, (151) and, most importantly, Hudson has not been followed by similar holdings that categorically remove prisoners from the protection of other core constitutional rights. (152) In light of the subsequent quarter-century of jurisprudence, Burger's sweeping opinion is arguably no less anomalous than Stevens'. (153)

Interestingly, if Stevens' view (like Duff's) is that prison should not be thought of as a place of mere warehousing or exile, but as a place where inmates actively lay a foundation for reconciliation and reintegration into the community, this vision seems to be quite congruent with public preferences too. For instance, in his anthropological research in a diverse array of communities across the United States, Donald Braman found the same basic critique of imprisonment articulated again and again: inmates should not be idle, but should be required to engage in constructive, pro-social activities. (154) Public opinion surveys reflect the same view: large majorities support mandatory literacy, job-training, and correctional-labor programs for offenders. (155)

Braman's view may also find support in the surprising durability of "good-time" and "earned-time" sentence-reduction programs for good behavior and participation in pro-social programming, even as parole release went into steep decline. (156) Indeed, many of the recent reforms adopted by cash-strapped states have involved the creation or expansion of good-time and earned-time programs. (157) Although fiscal pressures doubtlessly helped to fuel these initiatives, the fact that so many states adopted these particular cost-saving measures (rather than, say, across-the-board sentence reductions for inmates convicted of nonviolent crimes) may indicate an underlying desire to strengthen the penitential character of prison.

In short, the public does not seem to want mere warehousing of offenders--the public cares about what inmates do while they are in prison, and what is desired of inmates is that they do engage in the sorts of pro-social, rehabilitative activities that are capable of expressing repentance and a desire to address the underlying deficiencies that contribute to offenders' criminality. (158) This provides hope that the sort of prisons envisioned by Stevens and Duff (different in some respects, (159) but far closer in spirit to one another than either is to Burger's) would find robust public support. Thus, while it is certainly arguable whether current prison practices provide a morally satisfactory foundation for retributive parole, it does not seem wholly unrealistic to think that prisons might be made increasingly penitential over time. (160)

2. Parole Processes Do Not Fit Retributive Purposes

I have suggested that indeterminate prison terms can be reconceptualized such that delayed release is seen as a retributively appropriate response to an inmate's failure to comply with prison rules. However, parole decisional processes do not seem well designed to advance such retributive ends. There are at least four difficulties here. First, prisons maintain separate disciplinary systems for rule infractions. To deny parole on the basis of misconduct that has already been sanctioned may have the appearance of double punishment for the same offense. Second, parole decisions are flamed as holistic assessments of the offender's prospects for successful reintegration, not as responses to discrete acts of misconduct. Third, parole decisional processes lack the formal attributes, such as confrontation and jury-trial rights, that we normally expect of processes leading to state-administered punishment. Finally, a decision to deny parole does not impose on the inmate some particular quantum of punishment, but instead simply delays release for an uncertain period of time. I will consider each of the four objections separately.

First, double-punishment concerns should be alleviated if the initial sanctioning is self-consciously restrained so as to permit further consideration of appropriate responses at the parole hearing, and if the parole board then takes into account what sanctions have already been imposed when determining whether an inmate's violations warrant delayed release. Nor does using such a bifurcated approach necessarily seem inefficient. An initial prison disciplinary process facilitates the creation of a permanent record of the circumstances of the infraction while memories are still fresh; ensures that the inmate is put on notice of prison requirements in a clear, timely fashion; and provides punishment for minor infractions that are not likely to warrant delayed release. Meanwhile, the later parole hearing provides a superior opportunity for determining whether the inmate has gone beyond isolated or inadvertent violations into the realm of persistent, willful violations. Such a determination requires a nuanced exploration of state of mind and patterns of misconduct over time that may be unnecessary and impractical in the context of the focused, expedited procedures typically used for prison discipline. (161) Moreover, the parole board may provide a needed independent perspective on the extent to which prison personnel and circumstances themselves contributed to a pattern of violations. (162) Thus, both phases of the bifurcated process are capable of making a distinctive contribution to penal objectives. Additionally, knowledge that a parole board will later revisit and address persistent violations might diminish over-reliance by prison officials on disciplinary segregation (which often involves conditions of extreme isolation and restriction that seem quite inconsistent with the penitential ideal of prison). (163)

Second, although we are accustomed to thinking about the parole heating as an open-ended inquiry into an inmate's record, there is no reason why the hearing could not be structured around the questions of whether the inmate has been a persistent, willful violator and, if so, what penal response is appropriate. Indeed, some of the considerations that are currently seen as relevant to rehabilitation and dangerousness might still be considered at a retributive parole hearing, albeit to somewhat different ends. For instance, an inmate's mental illness or mental capacity might be considered insofar as it shed light on the willfulness of violations, as might gang membership or other aspects of the inmate's social circumstances.

Third, nothing in my indeterminate sentencing model precludes greater procedural formality surrounding release decisions, although additional procedural safeguards would admittedly raise cost concerns. The difficult question is what sort of process would be minimally necessary for retributive parole decisions to meet constitutional requirements and, more generally, to achieve legitimacy as a mechanism for extending punishment.

As to constitutional requirements, the Supreme Court held in Greenholtz v. Inmates of Nebraska Penal and Correctional Complex (164) that parole decisions do not necessarily trigger procedural due process protections, although such protections may sometimes arise by virtue of the way that a particular state chooses to structure its parole system. (165) Even when the Due Process Clause has been activated, though, the Court has not required extensive protections. For instance, in Greenholtz, the Court held that no evidentiary hearing was required before parole was denied; it sufficed for the parole board merely to interview inmates who were eligible for release and to accept letters written on their behalf. (166) On the other hand, the Court's decision was expressly premised on the traditional rehabilitative model of parole. (167) It is possible that the Court would rule differently if presented with an avowedly retributive delayed-release scheme. Yet, Greenholtz also embraced the view that current inmates have a less important interest in being released than non-inmates have in remaining outside prison (168)--a distinction that might help to provide constitutional justification for relatively informal release procedures regardless of the purpose of the release mechanism. (169)

Nor are matters much clearer if the question is reframed from what is constitutionally required to what is necessary for legitimacy. The fact that plea bargaining has become by far the most common process for obtaining convictions in this country (170) suggests that the public does not demand anything like full-blown trial procedures as a condition of legitimate punishment. Yet, opinion surveys also indicate that the public is far from enthusiastic about plea bargaining. (171) Drawing on a large body of social scientific research on legitimacy and procedural justice, I have elsewhere proposed plea-bargaining reforms that would enhance the perceived legitimacy of the process. (172) More specifically, I urge that prosecutors use transparent, objective criteria in deciding what deals to offer; provide defendants with an opportunity to tell their side of the story; and explain their decisions. (173) Similar principles could also be observed in parole decision-making without imposing on the system anything approaching the costs of traditional trial-type procedures. (174)

As to the final procedural objection, that parole denials do not associate some particular term of incarceration with a particular offense, there might well be uniformity concerns--after all, for one inmate, a denial might lead to only a relatively short further stay in prison, while, for another, a denial might lead to many years of additional incarceration. The ultimate effect of a denial will be a function of the extent to which there is further misconduct in prison (and hence subsequent denials), as well as the maximum incarceration authorized by the original sentence. The concerns may be alleviated if those denied parole are given new hearings at relatively short intervals (say, annually) and if the new hearings focus especially on conduct since the last hearing--that way, a long period of additional incarceration is unlikely to occur unless it can fairly be conceptualized as a series of penal responses to a series of separate violations, as opposed to a single indeterminate response to conduct occurring before the first hearing. (175)

3. Delayed Release Is Retributively Excessive

Assuming that parole denials are reconsidered annually, the direct effect of a denial would be one more year in prison for the recalcitrant inmate. In many cases, though, a year in prison might be seen as an excessive punishment. Although some denials might be based on conduct that would be regarded as a substantial criminal offense even if perpetrated outside of prison, such as an assault or theft, other denials might be based on conduct that is only punishable because it violates prison rules, such as persistent failures to be present for cell counts, nonparticipation in mandatory therapeutic programs, or possession of contraband. If willful and persistent, such rule violations seem to warrant some punishment, but a full year in prison might appear excessive.

Duff himself emphasizes that, as a backup sanction, prison terms ought to be "relatively short" (although he has not associated any particular length with this principle). (176) If excessive, the incremental incarceration would raise two overlapping concerns. First, excessive incarceration might violate the requirement of proportionality and hence communicate something dishonest and unjust regarding the offender's deserved censure. (177) Second, excessive incarceration might violate the requirement that punishment leave offenders with the freedom to remain unpersuaded. Recall Duff's admonition that, while "offenders [may be] forced to hear the message that punishment aims to communicate and to undergo a penal process intended to persuade them to accept it, they [may not be] forced to listen to that message or to be persuaded by it." (178) We should thus be concerned about the possibility that a system of retributive parole might attach such harsh consequences to a parole denial that the system would cross that uncertain line into unacceptable coerciveness.

These concerns should be somewhat diminished by the existence of a maximum end date for release established by the underlying sentence: assuming he does not commit a fresh, prosecutable criminal offense, the truly defiant inmate will have the reassurance that he can simply wait out the original sentence without having to cooperate with any of the penal measures that are designed to give his imprisonment a penitential character.

The excessiveness concerns might also be mitigated by providing for some flexibility in the frequency of parole redeterminations. For instance, in a case involving relatively minor violations, a new hearing might be ordered in six months instead of a year.

But perhaps neither of these mitigating features is adequate. If the underlying sentence is a long one, then the imprisonment difference between the first and last possible release dates may be many years--no matter how persistent and willful, we might think that violations that are not serious crimes in themselves cannot possibly warrant such long incremental incarceration. Moreover, there are doubtlessly important practical constraints on how frequently parole denials can be reconsidered.

In weighing these difficulties, it is important to bear in mind that a parole denial resulting in an incremental year of prison likely differs quite a bit experientially from the upfront imposition of a year in prison as a sentence. Much research points to a common pattern of adaptation to prison life. "While most new prisoners (especially first-timers) are disoriented and fearful on arrival, there is now substantial research evidence that over time they gradually work out ways of coping with this strange social world." (179) Surveys of prison inmates, for instance, show that after a few difficult weeks at the outset, positive feelings tend to increase and negative feelings tend to decrease steadily over a period of years and may eventually reach the normal range. (180) Other measures of inmate quality of life are also in accord with these findings. (181)

In light of adaptation effects, it may be inappropriate to judge the severity of a parole denial from the point of view of someone outside prison--to an inmate already acclimated to prison life, the incremental year may be far less daunting than it would appear to others, (182) Moreover, the social scientific research also indicates that parole release may not be nearly so positive an experience as one might expect relative to prison life. Even after a short stay in prison, the research suggests that ex-inmates typically face extraordinary, long-term challenges to reintegration and a return to the level of well-being they enjoyed before prison. (183) The difficulties facing former inmates include high rates of joblessness, high rates of mental and physical illness, legal disabilities resulting from their convictions, and onerous conditions of release. (184) Exacerbating all of these challenges is the social stigma experienced by former prisoners and their families, which seems a surprisingly powerful force even in communities with high rates of crime and incarceration. (185) There should be little wonder, then, that recidivism and reincarceration rates for parolees are high. (186) Nor should we be surprised to hear evidence that a growing proportion of parole-eligible inmates are waiving their hearings and opting to finish their sentences in prison. (187) These considerations might also cause us to see a parole denial as less severe a sanction than if we are thinking more generically about the difference between living in prison and living in the outside community. (188)

To be sure, the retributive suitability of a punishment must be assessed by reference to social understandings of the punishment, and studies of the relative happiness or quality of life of prisoners and parolees should not necessarily control the analysis. (189) Yet, there is good reason to think that the social understanding of prison and parole will, over the long run, tend to reflect, at least obliquely, the way that they are experienced. As Justice Stevens pointed out in Hudson, the reality of mass incarceration in the United States is such that a large proportion of the general population has direct personal connections with the incarcerated population. (190) Moreover, the general phenomenon of "hedonic adaptation" (that is, the return to normal experiences of emotion after unpleasant events, such as being sent to prison) has been documented in a wide variety of settings, (191) and is likely familiar to most people from their life experiences. Perhaps an instinctual appreciation of the phenomenon has contributed to the American "popular punitiveness" that has recently led both to much longer prison terms and to a resurgence of interest in shaming sanctions. (192) In any event, based on these trends, one might plausibly conclude that the current social understanding is that an incremental year of prison is not an especially stiff punishment. (193) Moreover, although parole is doubtlessly seen as more lenient still, there are even signs that the public is developing a better appreciation of the extraordinary hardships faced by returning prisoners. (194) As these hardships are better recognized, and the constraints of prison life are seen in many respects as differing more in degree than in kind from the constraints of post-prison life, parole denials are less likely to appear excessive in severity.


As a practice, indeterminate sentencing emerged from an era when the classical rehabilitative paradigm dominated American penological thinking, but the practice has long outlived the ideology on which it once rested. The practice likely owes its durability to some combination of institutional inertia, its usefulness as a prison-management tool, (195) continued (albeit more modest) beliefs that at least some inmates may be rehabilitated in prison, (196) fears that eliminating indeterminacy may cause corrections budgets to escalate wildly, (197) and the flexibility of indeterminate systems to increase and decrease release rates in response to political and fiscal pressures. All of this leaves indeterminate sentencing with a troublingly incoherent character--both over time (as release is made easier or harder on an ad hoc basis) and in the disconnect between sentencing decisions (which now commonly emphasize retributive considerations) and parole decisions (which either disregard retributive considerations or take them into account only by reconsidering and possibly rejecting the retributive judgments made at sentencing). In order to address these concerns, I have proposed a new normative model of indeterminate sentencing that rests not on a rehabilitative, but on a retributive, theory of punishment. The basic components of this model include: an indeterminate prison term whose parameters are established at the time the sentence is imposed by reference to the principle of negative proportionality, a presumptive release from prison at the earliest possible time within those parameters, and a possibility of delayed release as a type of backup sanction for violations of prison rules.

I conclude now with a few thoughts on the implications of my analysis for practice and theory. First and most immediately, in the face of the current fiscal pressures, the retributive model provides a basis for thinking critically about the many cost-saving reforms currently being considered. Reforms that bring practice closer to the model seem at least prima facie more attractive than reforms that move in the opposite direction. By this way of thinking, among the measures currently drawing interest, the most attractive might include the following:

* Reducing sentences or otherwise expanding release opportunities for non-violent offenders (198): inmates serving time for nonviolent offenses seem especially likely to have sentences that exceed what is minimally necessary (and perhaps even what is maximally acceptable) from the standpoint of retributive proportionality, particularly if they have been sentenced under "three strikes" laws or other similarly harsh and inflexible mandatory minimum statutes. (199)

* Establishing or expanding geriatric release for long-serving inmates (200): as with reforms targeting nonviolent offenders, reforms targeting long-term inmates seem especially likely to reach offenders whose underlying sentences were longer than necessary.

* Enhancing "good-time" credits for compliance with prison rules (201): such programs make release dates more dependent on performance in prison, consistent with my retributive indeterminate sentencing model.

* Enhancing "earned-time" credits for participation in programs designed to address problems contributing to inmates' underlying offenses (202): functionally, these programs might be thought of as imposing requirements for inmates to address their issues while in prison and imposing a backup sanction (that is, not providing earned-time credits) for violations of the requirement.

To be sure, with all of these reforms, it is important that they do not drop punishment severity in individual cases below what is minimally required by negative proportionality. There will surely be much disagreement about what that minimal amount is in many cases, but the dramatic increase in American penal severity over the past four decades (203) offers a good reason for confidence that a great many sentences can be reduced to a considerable extent without fear of unduly deprecating the seriousness of the underlying offense. (204) In any event, if the support for these reforms is to outlast the current fiscal crisis, it is important that they be increasingly framed in ethical and retributive terms, as I think they can be.

Other recent cost-cutting reforms that seem far more troubling include those that make prison life harsher and that diminish access to rehabilitative programming. (205) Such reforms give imprisonment more of the character of warehousing or exile, rather than secular penance, and make incremental imprisonment seem less morally fitting as a backup sanction for inmates.

In addition to providing a new method for evaluating and discussing immediate responses to fiscal pressures, the retributive model of indeterminate sentencing also underscores the importance of pursuing a variety of other longer-term reforms. Such reforms might include: the adoption of sentencing guidelines built around proportionality principles, abandonment of harsh mandatory minimum statutes (such as three-strikes laws) that preclude sentencing judges from taking into account the full range of retributively appropriate considerations, depoliticization of parole boards, (206) and adoption of parole guidelines that focus attention on prison performance.

The analysis in this Article also has implications for the work of retributive theorists. Many leading retributive thinkers have been disappointed that the retributive revival of the 1970s was followed by an extraordinary increase in sentence lengths (207)--recall that retributive principles are intended not merely to justify punishment, but also to limit its severity. Retributive theorists may have been successful in helping to undermine the rehabilitative paradigm and reinforcing the view that criminal punishment should be based on ideals of individual accountability and moral responsibility, but they seem to have been much less successful in bringing reason and restraint to penal practices. This failure may have something to do with their orientation toward "front end" aspects of the criminal justice system (208): establishing a justification for punishment, (209) identifying principles of criminal liability, (210) and working out how to translate proportionality and uniformity ideals into sentencing rules. (211) Duff's brief commentary on incarceration notwithstanding, retributive theorists have devoted far less energy to the back end, particularly to the implementation of sentences of imprisonment. (212) What would a retributive prison look like, and how ought a prison term be managed based on retributive principles? I've suggested a few thoughts in this Article, but I've only barely scratched the surface. (213)

Donald Braman has argued that what appears as a public desire for ever-tougher treatment of criminals--the phenomenon of "popular punitiveness"--is more accurately understood as a deep dissatisfaction with the perceived failure of prisons to promote true accountability, as opposed to mere detention. (214) For instance, as noted in the previous Part, public-opinion research shows very strong support for mandatory literacy, job-training, and work programs for offenders. (215) But, Braman contends, policy-makers have failed to read the nature of public dissatisfaction accurately, and so have simply delivered more and more prison time. (216) If Braman is correct, then retributivists who are concerned about popular punitiveness might do well to focus their attention on the penal back end and develop a reform agenda that would translate their ideals of individual accountability and moral responsibility into the prison setting. Getting imprisonment right, whether through a new retributive model of indeterminate sentencing or otherwise, might help to break the cycle of pathological politics and bring about a more humane, cost-effective, and genuinely retributive criminal justice system.
States That Enhanced Early Release Opportunities for Prison Inmates,

Alabama 2008: established new program permitting early
 release of certain older inmates with serious medical
 conditions (217)

 2003: authorized early release of nonviolent
 offenders to reduce prison overcrowding 218

Arkansas 2005: authorized transfer of inmates to
 community-based transitional housing up to one year
 before parole eligibility (219)

 2009: expanded earned-time opportunities; permitted
 parole to be granted without a hearing in some

California 2010: authorized medical parole for inmates who are
 permanently incapacitated (221)

Colorado 2009: expanded number of good-time credits that can
 be earned per month; authorized Department of
 Corrections to deduct additional sixty days from
 sentence of nonviolent inmates for good behavior

 2010: expanded eligibility for earned time (223)

Connecticut 2005: expanded parole eligibility; authorized
 transfer of inmates to community-based residences
 within final eighteen months before release date;
 established compassionate release program for inmates
 who are elderly or i11 (224)

Delaware 2008: authorized earned-time credit for participation
 in prison programs (225)

 2010: expanded eligibility for good-time credits

Georgia 2009: expanded eligibility for work release and
 transitional centers for violent offenders during
 final year of incarceration 227

Illinois 2007: authorized earned-time credits for inmates who
 earn GED229

Indiana 2010: authorized rehabilitation-based discharge for
 certain long-serving inmates (229)

Kansas 2007: authorized earned-time credits for completion
 of certain programs by low-level offenders (230)

 2010: authorized medical parole for terminally ill
 inmates (231)

Kentucky 2008: permitted shift to home confinement for
 nonviolent offenders within 180 days of release date;
 authorized earned-release credits for completion of
 drug treatment and education programs (232)

 2009: expanded parole eligibility for inmates
 convicted of certain low-level felonies (233)

 2010: authorized certain inmates to be released to
 home confinement when they have 180 days left in
 their terms (234)

Louisiana 2001: reestablished parole eligibility for certain
 nonviolent offenders 235

 2008: increased earned-time availability for inmates
 serving consecutive sentences (236)

 2009: authorized parole for inmates serving life
 sentences for heroin offenses (237); authorized
 earned-time credits for participation in approved
 treatment programs (238)

 2010: made good-time credits retroactive to 1992 for
 some inmates; eliminated requirement that parole
 board be unanimous in support of release (239)

Maine 2007: permitted inmate work hours to be applied
 toward good time (240)

 2009: expanded eligibility for release for terminally
 ill inmates (241)

Maryland 2004: authorized early release of certain offenders
 for drug treatment in the community (242)

 2007: established parole eligibility for certain
 offenders sentenced to a mandatory minimum (243)

 2009: expanded conditional release eligibility for
 nonviolent offenders (244)

Massachusetts 2010: authorized early release for some inmates
 serving mandatory terms for nonviolent drug

Michigan 2002: established parole eligibility for certain drug

Minnesota 2005: authorized early release of certain nonviolent
 drug offenders 247

Mississippi 2001: reduced amount of time that first-time,
 nonviolent offenders must serve before becoming
 eligible for parole; created programs permitting
 certain inmates to obtain earned-time credits (248)

 2004: authorized transfer of terminally ill inmates
 to community supervision; increased sentence
 reduction based on participation in trusty programs

 2005: expanded parole eligibility for certain drug
 offenders (250)

 2008: restored parole eligibility for nonviolent
 offenders; authorized release of terminally ill
 inmates regardless of time served (251)

 2009: expanded early-release credits for
 participation in education programs (252)

Nebraska 2010: increased amount of good-time credit that could
 be earned by jail inmates (253)

Nevada 2007: increased maximum number of possible good-time
 credits (254)

 2009: permitted parole violators to earn good-time
 credits; authorized parole without a hearing for some
 offenders; expanded eligibility for residential
 confinement (255)

New 2010: established presumptive parole for certain
Hampshire inmates (256)

New Jersey 2010: authorized courts to waive or reduce parole
 ineligibility for certain drug offenses (257)

New York 2004: authorized certain drug offenders to petition
 for resentencing; increased merit-time credits (258)

 2005: expanded resentencing possibilities to new
 class of drug offenders (259)

 2009: expanded eligibility for early release through
 Shock Incarceration Program; authorized medical
 parole for inmates suffering permanent disabilities;
 increased merit-time credits (260)

 2010: expanded earned-time program (261)

North 2001: expanded programs through which earned-time
Carolina credit might be obtained (262)

 2008: established procedures for medical parole (263)

Ohio 2008: streamlined procedures for medical and other
 early-release programs (264)

Oregon 2009: expanded earned-time credit opportunities (265)

 2010: expanded eligibility for education-based credit
 toward early release (266)

Pennsylvania 2008: adopted risk-reduction sentences to give
 inmates early-release opportunity based on
 participation in programs designed to reduce
 recidivism (267)

 2010: authorized parole release for some inmates who
 have not completed prescribed programming (268)

South 2010: required that certain inmates be released to
Carolina mandatory supervision 180 days before their prison
 release dates; expanded parole eligibility (269)

Tennessee 2007: authorized earned-time credits for inmates who
 complete certain programs (270)

Texas 2009: eliminated life without parole for juveniles
 (271); authorized Department of Criminal Justice to
 restore lost good-time credits (272)

Vermont 2010: expanded furlough program (273)

Virginia 2001: expanded eligibility for geriatric release (274)

Washington 2003: increased amount of potential good-time credit
 from 33 to 50% of total sentence (275)

 2007: authorized earned-time credits for
 participation in programming (276)

 2009: expanded eligibility for early release for
 medical reasons (277)

 2010: authorized release of certain inmates to
 partial confinement during final twelve months of
 sentence in order to complete parenting program (278)

West Virginia 2010: instituted accelerated parole for inmates who
 complete rehabilitative program (279)

Wisconsin 2009: adopted risk-reduction sentences to give
 inmates early-release opportunity based on
 participation in programs designed to reduce
 recidivism; established "positive adjustment" time;
 expanded eligibility for medical parole (280)

Wyoming 2008: authorized earned time and medical parole (281)

(1.) Parole is sometimes confused with probation. Both involve supervision of an offender in the community for a specified period of time, but probation is a sentence given in lieu of imprisonment, while parole is a form of discretionary release that occurs after an offender has already served some portion of a sentence of imprisonment. JOAN PETERSILIA, REFORMING PROBATION AND PAROLE IN THE 21ST CENTURY 3, 111 (2002). Parole is implemented through a system of indeterminate sentencing. In its classic form, indeterminate sentencing in the United States required the sentencing judge to impose a sentence in the form of a range, with upper and lower limits expressed in terms of years (e.g., a sentence of incarceration for two to four years). JEREMY TRAVIS, BUT THEY ALL COME BACK: FACING THE CHALLENGES OF PRISONER REENTRY 13 (2005). A parole board was then authorized to release the offender at a point within the range depending on the offender's progress towards rehabilitation. Id. "The former prisoner then serve[d] the remainder of his sentence in the community, but [could] be returned to prison if he fail[ed] to observe the conditions that the parole board placed on his liberty." Id.

(2.) Joan Petersilia, Parole and Prisoner Reentry in the United States, in PRISONS 479, 492 (Michael Tonry & Joan Petersilia eds., 1999).

(3.) Id. at 495-96. To be sure, even in the abolishing jurisdictions, parole boards have remained active to handle cases from before the effective date of the abolition. See, e.g., U.S. PAROLE COMM'N, U.S. DEP'T OF JUSTICE, HISTORY OF THE FEDERAL PAROLE SYSTEM 2-3 (2003) (discussing repeated extensions of life of the United States Parole Commission beyond the effective date of abolition of federal parole, November 1, 1987).

(4.) These are listed in an appendix to this Article. To be sure, fear of political backlash seems to be causing corrections officials in many states to be more conservative than anticipated with their new early release authority. See, e.g., Ben Poston, Sentencing Reform Results Fall Short, MILWAUKEE J. SENTINEL, July 10, 2010, at A1 (discussing Wisconsin); see also Cara Buckley, Law Has Little Effect on Early Release for Inmates, N.Y. TIMES, Jan. 30, 2010, at A17, available at (discussing California, Alabama, and New York).

(5.) For instance, when Wisconsin enhanced its early-release opportunities in 2009, it replaced its parole board with a new agency called the "Earned Release Review Board." Gregory J. O'Meara, Compassion and the Public Interest: Wisconsin's New Compassionate Release Legislation, 23 FED. SENT'G REP. 33, 34 (2010); see also PETERSILIA, supra note 1, at 115-16 (noting rejection of "parole" term in Ohio, Minnesota, and Washington).

(6.) Graham v. Florida, 130 S. Ct. 2011, 2034 (2010).

(7.) Dirk van Zyl Smit, Outlawing Irreducible Life Sentences: Europe on the Brink?, 23 FED. SENT'G REP. 39, 46 (2010).

(8.) Although the holding of Graham was limited to juveniles, the Court's reasoning gave considerable weight to the value of providing long-term inmates with hope that they may eventually be able to secure release based on demonstrated rehabilitation. See, e.g., 130 S. Ct. at 2030 ("What the State must do, however, is give defendants like Graham some meaningful opportunity to obtain release based on demonstrated maturity and rehabilitation."). Over time, one might imagine this principle being used to pare back the categories of offenses and offenders eligible for life without parole, much as the Court has done with respect to the death penalty. See, e.g., Atkins v. Virginia, 536 U.S. 304, 321 (2002) (banning execution of the mentally retarded); Enmund v. Florida, 458 U.S. 782, 798 (1982) (holding death penalty impermissible for felony-murder defendant who did not kill and did not intend to kill). For a more thorough discussion of the path the Court might follow in expanding Graham, see Michael M. O'Hear, The Beginning of the End of Life Without Parole?, 23 FEB. SENT'G REP. 1, 2 (2010).

(9.) The term "indeterminate sentencing" is often confused with discretionary sentencing. Discretionary sentencing refers to a system in which the sentencing judge has discretion to select a sentence within a wide statutory range associated with the offense of conviction, while indeterminate sentencing refers to a system in which other actors (classically, the members of a parole board) have substantial authority post-sentencing to determine the actual release date from prison. Steven L. Chanenson, The Next Era of Sentencing Reform, 54 EMORY L.J. 377,382-84 (2005). Thus, an indeterminate sentencing system might or might not be a discretionary one, and vice versa.

(10.) See, e.g., Thomas J. Hammer, The Long and Arduous Journey to Truth-in-Sentencing in Wisconsin, 15 FED. SENT'G REP. 15, 15 (describing truth-in-sentencing legislation adopted in Wisconsin in 1998).

(11.) See Petersilia, supra note 2, at 492-97 (noting that major criticisms of the parole system included, inter alia, claims that uncertainty relating to release dates was inhumane to inmates and parole boards exercised their discretion in a racially biased fashion).

(12.) Although retributivism has deep historical roots, see, e.g., Robert Blecker, Less Than We Might: Meditations on Life in Prison Without Parole, 23 FED. SENT'G REP. 10, 11 (2010) (noting expression of retributive idea of proportionality in Deuteronomy and the Magna Carta), much of the work of contemporary retributive theorists differs markedly from the "eye for an eye" way of thinking that we associate with the Old Testament, see, e.g., id. (noting importance of culpable mental state to retributivists operating in the Kantian tradition); see infra Part I.A. (describing communicative theories of retributivism).

(13.) PETERSILIA, supra note 1, at 135 (2002).

(14.) Id. at 133.


(16.) Michael M. O'Hear, The Original Intent of Uniformity in Federal Sentencing, 74 U. CIN. L. REV. 749, 757-58 (2006).

(17.) See Andrew von Hirsch, Penal Theories, in THE HANDBOOK OF CRIME AND PUNISHMENT 659, 666 (Michael Tonry ed. 1998) ("Criminologists' interest in desert dates from the mid-1970s.... Once broached, the idea of desert quickly became influential.").

(18.) John Bronsteen et al., Happiness and Punishment, 76 U. CHI. L. REV. 1037, 1070 (2009); Russell L. Christopher, Deterring Retributivism, 96 NW. U. L. REV. 843, 845-47 (2002); cf. AMERICAN LAW INSTITUTE, MODEL PENAL CODE: SENTENCING, TENTATIVE DRAFT NO. 1, at 30 (2007) (noting that proposed revisions to sentencing provisions of the Model Penal Code are based on Norval Morris's theory of limiting retributivism, although the revisions do not use the term "retributivism").

(19.) See CHRISTINE S. SCOTT-HAYWARD, VERA INSTITUTE OF JUSTICE, THE FISCAL CRISIS IN CORRECTIONS: RETHINKING POLICIES AND PRACTICES 3 (2009) ("Second only to Medicaid, corrections has become the fastest growing general fund expenditure in the United States.... In FY2008 ... states spent an estimated $47 billion of general funds on corrections, an increase of 303 percent since 1988.") (citations omitted).

(20.) See id. ("Forty-three states were facing an aggregate budget gap in FY2009 of more than $100 billion...."). The fact that the federal government has displayed less interest than many states in strengthening early-release opportunities doubtlessly owes something to the facts that federal corrections spending is a comparatively small portion of the federal budget, see Kevin Johnson, 2011 Budget Gives Federal Prisons $528M, USA TODAY, Feb. 4, 2010, available at (noting that the Obama Administration sought $528 million for the Bureau of Prisons, out of total budget of $3.8 trillion), and that budget deficits create less pressure at the federal level due to an absence of balanced-budget requirement and federal control over the monetary supply, see National Conference of State Legislatures, BCSL Fiscal Brief: State Balanced Budget Provisions 2 (2010), available at documents/fiscal/StateBalancedBudgetProvisions2010.pdf ("Most states have formal balanced budget requirements with some degree of stringency....").

(21.) SCOTT HAYWARD, supra note 19, at 9-11. Policymakers' comfort with early release has doubtlessly been enhanced by a new generation of risk assessment and management techniques, see, e.g., JUDITH GREENE & MARC MAUER, THE SENTENCING PROJECT, DOWNSCALING PRISONS: LESSONS FROM FOUR STATES 34-36 (2010) (describing the Michigan Prisoner Reentry Initiative, in which the Michigan Department of Corrections recently achieved increased parole approval rates following institution of measures designed to increase confidence of the parole board, including data-driven risk assessment and improved case management of parolees), as well as the development of new forms of rehabilitative programming that have moved criminologists beyond the despair of the 1970s, see Michael Tonry & Joan Petersilia, American Prisons at the Beginning of the Twenty-First Century, in PRISONS 1, 8 (Michael Tonry & Joan Petersilia eds., 1999).

(22.) Different jurisdictions have different rules with respect to parole eligibility. For purposes of this example, I follow the old federal parole rules, which generally provided for release after between one-third and two-thirds of the sentence had been served. KATE STITH & JOSE A. CABRANES, FEAR OF JUDGING: SENTENCING GUIDELINES IN THE FEDERAL COURTS 20 (1998). Other jurisdictions have adopted rules that are structurally similar, but different in the release-date specifics. See, e.g., JUSTICE POLICY INSTITUTE, THE RELEASE VALVE: PAROLE IN MARYLAND 7 (2009) (discussing the Maryland system and noting that "[p]eople convicted of a nonviolent offense are eligible for parole after serving a quarter of their sentence, while individuals convicted of a violent offense are eligible after serving half of their sentence for the violent offense, or one quarter of the total aggregate sentence if there is more than one conviction").

(23.) See R.A. Duff, Dangerousness and Citizenship, in FUNDAMENTALS OF SENTENCING THEORY: ESSAYS IN HONOR OF ANDREW VON HIRSCH 141, 148-49 (Andrew Ashworth & Martin Wasik eds., 1998) (noting this retributive objection to parole).

(24.) Cf. Kevin R. Reitz, Questioning the Conventional Wisdom of Parole Release Authority, in THE FUTURE OF IMPRISONMENT 199, 206 (Michael Tonry ed. 2004) ("Through the 1980s and 1990s, parole boards across the country became visibly more risk averse in their release decisions, often jolted a single but terrifying episode of criminality....").

(25.) I do not here attempt any systematic defense of retributivism, and I acknowledge that there are many important objections to retributivism, perhaps most importantly those founded on the link between the retributive urge and what Nietzsche referred to as the emotions of ressentiment (resentment, fear, anger, cowardice, and so forth). See MICHAEL MOORE, PLACING BLAME: A GENERAL THEORY OF THE CRIMINAL LAW 119-20 (1997) ("I think the most serious objection to retribution as a theory of punishment lies in the emotional base of retributive judgments."). For an eloquent account of how the ugly emotions associated with retributivism have negatively affected American penal policy for the past generation, see generally James Q. Whitman, A Plea Against Retributivism, 7 BUFF. CRIM. L. REV. 85 (2003). Michael Moore has offered the most effective response to the Nietzschean critique of retributivism. MOORE, supra, at 140-52. Yet, Moore also emphasizes the need for self-consciously humane institutional design in order to minimize the possibility that punishment will be used as an excuse for the expression of ressentiment. Id. at 152. My project here may be thought of as preliminary attempt to meet that challenge as to one of the most important of our criminal-justice institutions, the prison.

(26.) I thus focus on the availability of parole release, and not on the way that parolees are supervised once they are returned to the community. An important subject of debate regarding parole has been the frequency of revocation and reincarceration, see TRAVIS, supra note 1, at 47-51, but that debate lies outside the scope of this Article.

(27.) See infra Part I.A.

(28.) See infra Part I.B.

(29.) See Matt Matravers, Introduction, in PUNISHMENT AND POLITICAL THEORY 1, 1 n.2 (Matt Matravers ed., 1999) ("[T]he labels 'consequentialist' and 'retributive' are of increasingly little use as the theories that they are meant to group together have become so diverse.").

(30.) Communicative approaches are sometimes subdivided into expressive theories, which tend to emphasize the role of punishment in "giving voice to the abhorrence of the community," and more "genuine" communicative theories that focus on the idea that punishment offers "some form of argument ... to the offender to explain that she has done wrong and to challenge her to answer that claim." Matt Matravers, "What to Say?" The Communicative Element in Punishment and Moral Theory, in PUNISHMENT AND POLITICAL THEORY 108, 109 (Matt Matravers ed., 1999). It is the latter approaches, of which Duff's theory is a leading example, that are of greatest interest to me in this Article.

In recent decades, among retributive theorists, the main competitor to the communicative approaches has been the "unfair advantage" theory, which holds that the wrongfulness of crime lies in the offender's decision to take extra liberty for himself beyond that which has been permitted to him through our system of mutual restraint; punishment is then seen as a taking back of liberty that restores the balance. Steven P. Garvey, Punishment as Atonement, 46 UCLA L. REV. 1801, 1836 n.149 (1999). Herbert Morris provided a seminal statement of this theory in Persons and Punishment, 52 MONIST 475 (1968). However, this approach has been widely criticized for failing to account for what really strikes us as wrong about many crimes; "unfairness," for instance, does not seem really to capture what is wrong about rape or murder. Garvey, supra, at 1836 n. 149.

To be sure, critics have also noted difficulties with the communicative approaches. See, e.g., Matravers, supra, at 109 (noting that diversity of moral understandings in society "raises a prima facie problem for the communicative account"), 111 ("[I]t would seem that punishment must condemn certain actions as wrong and give reasons for that judgment if it is to have a communicative element. Yet, our confidence in our beliefs about right and wrong ... has been undermined.").

(31.) See, e.g., R.A. Duff, Punishment, Communication, and Community, in PUNISHMENT AND POLITICAL THEORY 48, 48 (Matt Matravers ed., 1999) ("One theme in the 'retributive revival' of the last two decades has been that of punishment as a communicative practice.... [W]hat crime deserves or makes appropriate is a response which punishment communicates to the criminal."): Jean Hampton, Correcting Harms Versus Righting Wrongs: The Goal of Retribution, 39 UCLA L. REV. 1659, 1686 (1992) ("In short, retribution is a response to a wrong that is intended to vindicate the value of the victim denied by the wrongdoer's action through the construction of an event that not only repudiates the action's message of superiority over the victim but does so in a way that confirms them as equal by virtue of their humanity."); John Kleinig, The Hardness of Hard Treatment, in FUNDAMENTALS OF SENTENCING THEORY: ESSAYS IN HONOR OF ANDREW VON HIRSCH 273, 275 (Andrew Ashworth & Martin Wasik eds., 1998) ("[P]unishment has a distinctively expressive character. It involves a stigmatizing condemnation of the punished. It does so, because the person punished has been judged to be guilty inter alia of some moral wrongdoing, that is, of violating basic conditions of our human engagement."); Dan Markel, Executing Retributivism: Panetti and the Future of the Eighth Amendment, 103 NW. U. L. REV. 1163, 1184 (2009) ("Retributive punishment for legal wrongdoing is justified in part because it communicates to the offender a respect for his dignity by treating him as a responsible moral agent."); von Hirsch, supra note 17, at 666-67 ("Punishment, [contemporary theorists] point out, is a blaming institution.").

(32.) See, e.g., Anthony Bottoms, Five Puzzles in von Hirsch's Theory. of Punishment, in FUNDAMENTALS OF SENTENCING THEORY: ESSAYS IN HONOR OF ANDREW VON HIRSCH 53, 80 (Andrew Ashworth & Martin Wasik eds., 1998) ("[Retributivist theories] may justify punishment even when it has no preventive utility."); Christopher, supra note 18, at 847-48 ("Simply put, retributivism justifies punishment, or the suffering by the punished, not on any actual good consequences that might be attained, but solely because the punished deserve it.").

(33.) See, e.g., Duff, surpa note 31, at 56-60 (defending his theory of punishment against charges of illiberalism and noting that the theory is "fully consistent ... with a proper regard for individual autonomy, freedom, and privacy"); Markel, supra note 31, at 1186 ("Even against a quickly repentant offender, retributive punishment is warranted to effectuate our commitment to the principle of equal liberty under law. In a liberal democracy, punishment serves to fulfill part of equality's promise because, as citizens, we are each burdened by an obligation to obey those laws reasonably crafted, enforced, and applied."); Matravers, supra note 29, at 1 ("The retributive revival developed not merely from a dissatisfaction with [] consequentialist goals, but also from the reemergence of rights based and contractualist political theory more generally."). On the other hand, as some commentators have observed, there may be some tension between liberalism and the moralism that seems central to much retributive theorizing. See, e.g., Leo Zaibert, Punishment, Liberalism, and Communitarianism, 6 BUFF. CRIM. L. REV. 673, 685 (2002).

(34.) See, e.g., Markel, supra note 31, at 1192 ("[A] desire to avoid randomness, arbitrariness, and invidious bias in the death penalty are core commitments of the institutional and communicative account of retributivism I have described."): von Hirsch, supra note 17, at 666 ("The central organizing principle for deciding the quanta of punishments is the principle of proportionality....").

(35.) See, e.g., R.A. DUFF, PUNISHMENT, COMMUNICATION, AND COMMUNITY 139 (2001) (defending the flexible "negative" version of proportionality because it "gives sentencers more room to attend to the concrete particularities of the crime, without worrying about rendering it commensurable with all other crimes in terms of its seriousness, and to choose between a wider range of penalties, without worrying about rendering them all commensurable in terms of their severity."); NORVAL MORRIS, THE FUTURE OF IMPRISONMENT 75 (1974) ("The concept of desert is a necessary but not sufficient condition of the punishment of the crime. Desert is, of course, not precisely quantifiable."); von Hirsch, supra note 17, at 668 ("[T]here do not seem to be definite quanta of severity associated with our desert-based judgments."). Not all proponents of desert-based sentencing, however, share the view that desert is so difficult to operationalize. See Paul H. Robinson, The A.L.I's Proposed Distributive Principle of "Limiting Retributivism": Does It Mean Anything in Practice Other Than Pure Desert?, 7 BUFF. CRIM. L. REV. 3, 10 (2003) ("Contrary to the assumption of the original advocates of 'limiting retributivism'--that desert provides only vague outer limits on punishment--desert has quite specific demands...."). For reasons that I have explored elsewhere, I suspect that Robinson's claim is overstated. Michael M. O'Hear, Plea Bargaining and Procedural Justice, 42 GA. L. REV. 407, 441 (2008); see also Alice Ristroph, Desert, Democracy, and Sentencing Reform, 96 J. CRIM. L. & CRIMINOLOGY 1293, 1308 (2006) ("[T]he remarkable consistency with which people speak of punishments as deserved, even as those punishments expand in scope and severity, suggests that the concept of desert is quite elastic.").

(36.) Von Hirsch, supra note 17, at 668.

(37.) Id. Ordinal proportionality is not necessarily operationally incompatible with cardinal proportionality. One might adopt ordinal proportionality as the basic organizing principle for a penal scheme, yet retain cardinal proportionality as a negative constraint on punishment (punishment may not be so severe in any given case as to be unfair relative to the severity of the offense). Bottoms, supra note 32, at 61.

(38.) Von Hirsch, supra note 17, at 668.

(39.) One point of contention, for instance, has been the relevance of the offender's record of prior convictions. Id. at 670. In light of such difficulties, Anthony Bottoms argues--persuasively, I think--that ordinal, no less than cardinal, proportionality should be conceived in negative terms: avoiding injustice, instead of seeking to achieve "exact justice." Bottoms, supra note 32, at 62-63.

(40.) See, e.g., Andrew von Hirsch, Punishment, Penance, and the State, in PUNISHMENT AND POLITICAL THEORY 69, 69-70 (Matt Matravers ed., 1999) (noting agreement with Duff that punishment should be conceptualized in communicative terms, but disagreeing with Duff's claim that "that the hard treatment component of the penal sanction can itself be explained in reprobative terms"). For Duff's reply to von Hirsch's critique, see generally R.A. Duff, Response to von Hirsch, in PUNISHMENT AND POLITICAL THEORY 83 (Matt Matravers ed., 1999). To be sure, in light of the forward-looking aspect of his approach to punishment, see infra text accompanying notes 4344, not all theorists are convinced that Duff is best characterized as a retributivist. See, e.g., Zaibert, supra note 33, at 683. Duff, however, distinguishes his approach from consequentialist theories inasmuch as he "looks to a future aim to which [punishment] is related, not merely contingently as an instrumental technique, but internally as an intrinsically appropriate means." DUFF, supra note 35, at 89.

(41.) See, e.g., Kyron Huigens, On Aristotelian Criminal Law: A Reply to Duff, 18 NOTRE DAME J.L. ETHICS & PUB. POL'Y 465, 468 (2004) ("[N]o one working on punishment theory can ignore Duff's work."); Peter Western, Offences and Defences Again, 28 OXFORD J. LEGAL STUD. 563, 563 (2008) ("Antony Duff is widely regarded as the leading criminal law theorist in the Anglo-American world today...."); Zaibert, supra note 33, at 673 (referring to Duff as "without a doubt one of the preeminent criminal law theorists today").

(42.) R.A. Duff, Guidance and Guidelines, 105 COLUM. L. REV. 1162, 1182 (2005).

(43.) Id.

(44.) Id. at 1183.

(45.) DUFF, supra note 35, at 80.

(46.) Id. at 81.

(47.) Id. at 106. Duff is not alone among retributive theorists in thinking of punishment in penitential terms. See, e.g., MOORE, supra note 25, at 152 ("There is no contradiction in attempting to make a retributive punishment system humane; doing so allows penitentiaries to be faithful to their names--places for penance, not excuses for sadism, prejudice, hatred, and the like.").

(48.) DUFF, supra note 35, at 108.

(49.) Id.

(50.) Id. at 109.

(51.) Id. at 143.

(52.) Id. at 145.

(53.) Id. at 150.

(54.) Id.

(55.) Id. at 142.

(56.) Id. at 138.

(57.) Id. at 139.

(58.) Id. at 135.

(59.) Id. at 122.

(60.) Id. at 126.

(61.) Id. at 122.

(62.) Chanenson, supra note 9, at 380.

(63.) See id. at 451 (discussing limiting retributivism).

(64.) MORRIS, supra note 35, at 75.

(65.) Id.

(66.) Id. at 34-35. Morris's chief criticism of parole centered on the inability of parole boards to make reliable predictions about post-release performance. Id. at 35 ("Protracted empirical analysis has demonstrated, however, that predictions of avoidance of conviction after release are no more likely to be accurate on the date of release than early in the prison term.").

(67.) Chanenson, supra note 9, at 451.

(68.) Id. at 447.

(69.) Id. at 433.

(70.) Id. at 434.

(71.) Id. at 447.

(72.) Id. at 454-55.

(73.) Id. at 433. The super commission would differ from traditional parole boards in at least two important respects: it would develop guidelines for parole release to limit discretion and bring greater transparency to the early-release process, and it would better integrate front-end sentencing and back-end release decisions by ensuring consistency between sentencing and parole guidelines. Id. at 450-51.

(74.) Id. at 459.

(75.) Reitz, supra note 24, at 216. On the other hand, Reitz endorsed "good-time" allowances for inmates "in the neighborhood of 20 or 25 percent," so it is not clear that his position would be much different from Chanenson's in practice. Id. at 200.

(76.) Id. at 208.

(77.) See, e.g., JOAN PETERSILIA, WHEN PRISONERS COME HOME: PAROLE AND PRISONER REENTRY 70-71 (2003) (discussing study showing better success by inmates released through discretionary parole than as a result of mandatory release requirements, even controlling for offense type and other variables, and arguing that the study "buttresses" the view that discretionary parole "ultimately leads to greater public safety, since it encourages both inmates and prison officials to focus more heavily on reintegration programs"); Mandeep K. Dhami et al., Adaptation to Imprisonment: Indigenous or Imported?, 34 CRIM. JUST. & BEHAVIOR 1085, 1085 (2007) ("[D]isciplinary infracations in prison have been found to be predictive of recidivism, whereas participation in prison education programs and maintenance of family ties are associated with reductions in recidivism." (citations omitted)).

(78.) Reitz, supra note 24, at 208.

(79.) Id. at 215.

(80.) Chanenson, supra note 9, at 455.

(81.) Chad M. Oldfather, Heuristics, Biases, and Criminal Defendants, 91 MARQ. L. REV. 249, 255-56 (2007).

(82.) On the other hand, there are good reasons to doubt that offenders are very responsive to differences in potential periods of imprisonment; the more immediate risks of getting caught and convicted likely overwhelm the significance of the more remote and speculative risks of being incarcerated for, say, four years instead of two years. See Paul H. Robinson & John M. Darley, The Utility of Desert, 91 NW. U. L. REV. 453, 458-64 (1997) (reviewing social science research explaining why "lengthy prison terms are not a particularly useful method of increasing deterrence effects" id. at 462); Tonry & Petersilia, supra note 21, at 8 ("There has, however, been relatively little research in recent years on deterrence and incapacitation effects, and most authoritative reviews of both subjects conclude that, while such effects exist, they are probably modest.").

(83.) Chanenson, supra note 9, at 455.

(84.) JUSTICE POLICY INSTITUTE, supra note 22, at 16.

(85.) DUFF, supra note 35, at 216-17 n.12 (noting disagreement with Morris's theory of limiting retributivism).

(86.) Chanenson, supra note 9, at 447.

(87.) See Duff, supra note 23, at 149 ("[Respect for citizens' autonomy] is inconsistent with a system of purely preventive detention which details those deemed to be dangerous: for such a system does not leave citizens free to decide for themselves whether to obey the law."). Duff, however, has offered a moral defense of a limited form of "special selective detention" for persistent, serious, violent offenders. Id. at 159-62. He does not endorse the aggregative, actuarial approach that is commonly used to justify selective incapacitation. Id. at 144.

(88.) Robinson & Darley, supra note 82, at 467.

(89.) Von Hirsch, supra note 17, at 667. There is a rich and contentious philosophical literature on the question of whether it is compatible to say both that an act is determined (say, by social pathologies) and that the actor is morally responsible for the act. For a helpful review and insightful critique of the literature, see generally Anders Kaye, Powerful Particulars: The Real Reason the Behavioral Sciences Threaten Criminal Responsibility, 37 FLA. ST. U. L. REV. 539 (2010). "Compatibilists" would likely discount the concern I articulate here. As Kaye argues, however, compatibilists may assume too quickly that their position reflects common moral intuitions. Id. at 557-61. Empirical research suggests that our moral judgments in the face of evidence of determinism are highly dependent on the specific facts of the case and how the facts are presented. Id. at 562-66.

(90.) Reitz, supra note 24, at 208-09.

(91.) Robinson & Darley, supra note 82, at 466-67. I assume here that a parole board would use "actuarial," instead of "clinical," risk assessment; actuarial approaches have proven far more reliable. John Monahan, The Future of Violence Risk Management, in THE FUTURE OF IMPRISONMENT 237, 253 (Michael Tonry ed., 2004). Many states currently use actuarial parole guidelines that are based on the United States Parole Commission's "salient factor score," which has demonstrated reliability in identifying high-risk offenders. PETERSILIA, supra note 77, at 71. The seven variables considered in this system are largely static, including criminal history, age, and whether the underlying offense or conviction constituted a violation of probation, parole, or confinement. Id. Newer risk assessment instruments, such as the Level of Service Inventory-Revised, make more use of dynamic factors, but still rely on many static factors. Id. at 72-73.

(92.) Moreover, much social psychology research suggests an important connection between the perceived legitimacy of legal actors and actual compliance with the law. O'Hear, supra note 35, at 420-24. This research, in turn, may suggest that at some level, ethically sound practices are complementary, not in opposition, to utilitarian crime-control objectives. Robinson and Darley have made a similar point, arguing that desert-based punishment better serves crime-control interests than more conventionally utilitarian approaches. Robinson & Darley, supra note 82, at 456.

(93.) Mercy is a controversial proposition among retributivists, see, e.g., Dan Markel, Against Mercy, 88 MINN. L. REV. 1421 (2004), but may be compatible with retributive principles, Bottoms, supra note 32, at 68-69. Duff has argued that mercy may appropriately reduce punishment severity, but only in unusual circumstances in which the offender has suffered a serious harm unrelated to the crime. DUFF, supra note 35, at 214-15 n.46. In general, though, Duff's account "does not imply that repentance should entitle the offender to a lighter sentence." Id. at 120.

(94.) It is sometimes said that proportionality is concerned with achieving the right severity, of punishment, and not per se with a particular mode or duration of punishment. Thus, for instance, Robinson and Darley imagine a research-based equivalency table that would permit trade-offs between, say, X amount of time in prison and Y amount of time in house arrest. A sentencing judge might freely choose between the two options based on a given offender's need for the closer supervision provided in prison. Robinson & Darley, supra note 82, at 491. Parole might be conceived in the same way, but with the trade-off made by a parole board instead of a sentencing judge. Imagine that some table of equivalence is worked out between prison and parole supervision: say, one year of prison equals four years of parole. A parole board might then treat its task as deciding whether or at what time to convert an inmate's remaining prison term to a parole term of four times that length. If such a conversion could be effected without compromising public safety, there might be fiscal benefits for the state and/or improvements in the offender's prospects for successful reintegration--benefits that, on Robinson and Darley's view, could be obtained without giving the offender any less severe a punishment than he or she deserves.

On the other hand, developing the prison/parole equivalence formula might be quite difficult. Relevant research is still "in its nascent stages." Christopher Slobogin & Mark R. Fondacaro, Juvenile Justice: The Fourth Option, 95 IOWA L. REV. 1, 56 (2009). Moreover, Dan Kaban argues that there is a crucial expressive incommensurability between imprisonment and other punishments: "Imprisonment is an extraordinarily potent gesture of moral disapproval; because of the symbolic importance of individual liberty in American culture, there is never a doubt that society means to condemn someone when it takes that person's freedom away." Dan M. Kahan, Social Influence, Social Meaning, and Deterrence, 83 VA. L. REV. 349, 384 (1997); see also id. at 362-63. For that reason, Kahan is skeptical of the claim that "all forms of punishment are interchangeable along the dimension of severity or 'bite.'" Dan M. Kahan, What Do Alternative Sanctions Mean?, 63 U. CHI. L. REV. 591, 593 (1996). If Kahan is correct, then efforts to find a prison/parole equivalence formula may be futile, at least with respect to the offenses that society most wishes to condemn.

(95.) DUFF, supra note 35, at 100-02.

(96.) Id. at 101.

(97.) Id.

(98.) Id.

(99.) Id. at 102.

(100.) Id.

(101.) Id.

(102.) Id.

(103.) Id. at 152.

(104.) Id.

(105.) Id.

(106.) Id.

(107.) Id.

(108.) See, e.g., GREENE & MAUER, supra note 21, at 33 (discussing practice in Michigan); W. David Ball, Heinous, Atrocious, and Cruel: Apprendi, Indeterminate Sentencing, and the Meaning of Punishment, 109 COLUM. L. REV. 893, 902-03 (2009) (discussing practice in California).

(109.) Reitz, supra note 24, at 204. To be sure, it is possible to imagine some potential advantages. For instance, I have observed elsewhere that group decision-making (as on a parole board) may help to diminish the effects of certain cognitive biases that may commonly afflict the individual sentencing judge. Michael M. O'Hear, Appellate Review of Sentences: Reconsidering Deference, 51 WM. & MARY L. REV. 2123, 2158 (2010). But I offered this observation in support of a more rigorous review of sentencing decisions by collegial appellate courts, and it would seem preferable for appellate courts to perform any necessary reweighing in the course of the normal appellate process rather than to have a parole board do the reweighing after the offender has already served much time in prison, which may create a longer period of uncertainty and which misses the opportunity to provide meaningful relief for excessively harsh sentences. Additionally, as I have argued at length elsewhere, the individual sentencing judge's knowledge of the local community in which the offense occurred may provide the judge with an important advantage in assessing offense severity that often outweighs any disadvantages relative to a more centralized body. Id. at 2152-55.

Reitz likewise considers, but finds unpersuasive, the view that "judges are too close to their cases and sometimes cannot free themselves from the raw emotions and public visibility that may surround a sentencing proceeding," noting "'grave doubt" that parole boards are not affected by the "tugs of publicity and victim sentiment." Reitz, supra note 24, at 205-06.

(110.) Chanenson, supra note 9, at 447.

(111.) See Greenholtz v. Inmates of Neb. Penal and Corr. Complex, 442 U.S. 1, 15 (1979) (noting that "the behavior record of an inmate during confinement is critical" to parole decisions).

(112.) DUFF, supra note 35, at 145.

(113.) Id. at 109.

(114.) Disciplinary infractions are not an uncommon fact of life for many inmates. For instance, one survey of federal inmates, who had served on average less than four-and-a-half years, found that the inmates had been charged on average with 2.16 infractions while in prison. Dhami et al., supra note 77, at 1096.

(115.) To be clear, I focus only on objections to the delayed release mechanism. Objections at a more general level to the "Duffian" framework I use might also be made and responded to, see, e.g., supra notes 30 and 40, but to do so at length would take us beyond the scope of this Article.

(116.) DUFF, supra note 35, at 152.

(117.) See Alice Ristroph, Respect and Resistance in Punishment Theory, 97 CALIF. L. REV. 601, 604-05 (2009) ("[W]e respect the criminal by acknowledging that punishment, though perhaps justified by societal interests, is hardly in the condemned man's interest or legitimate from his perspective. The right to resist [punishment] grounds an account of punishment that is arguably more honest, more egalitarian, and more uniformly respectful than the familiar retributive and utilitarian accounts.") (citation omitted).

(118.) Blecker, supra note 12, at 13-15.

(119.) See, e.g., Cutter v. Wilkinson, 544 U.S. 709, 716 (2005) ("Congress documented, in hearings spanning three years, that 'frivolous or arbitrary' barriers impeded institutionalized persons' religious exercise.") (citation omitted); Turner v. Safley, 482 U.S 78, 98 (1987) (holding unconstitutional the prison rule prohibiting inmates from getting married without a superintendent's permission and noting the existence of "obvious, easy alternatives ... that accommodate the right to marry while imposing a de minimis burden on the pursuit of security objectives") (citation omitted); Franklin E. Zimring & Gordon Hawkins, Democracy and the Limits of Punishment: A Preface to Prisoners' Rights, in THE FUTURE OF IMPRISONMENT 157, 170 (Michael Tonry ed., 2004) ("[W]hat [prison administrators] tend to favor is systems where small comforts are available to prisoners but only at the discretion of those who run the prisons.").

(120.) Under the standard, deferential test used by the Supreme Court to determine the constitutionality of prison regulations, regulations are permissible if they are "reasonably related to legitimate penological interests" and are not an "exaggerated response" to such objectives." Beard v. Banks, 548 U.S. 521, 528 (2006) (plurality opinion) (quoting Turner, 482 U.S. at 87). We may also be appropriately concerned about exclusionary and degrading failures to exercise power, e.g., failures to protect inmates from sexual victimization at the hands of other inmates, which appears to be a substantial problem in some prisons. See Tess M.S. Neal & Carl B. Clements, Prison Rape and Psychological Sequelae: A Call for Research, 16 PSYCHOL. PUB. POL'Y & L. 284, 285 (2010) (discussing data on incidence of prison rape). Sexual victimization further stigmatizes and isolates some inmates within the prison culture. Id. at 288-89.

(121.) See JAMES O. WHITMAN, HARSH JUSTICE: CRIMINAL PUNISHMENT AND THE WIDENING DIVIDE BETWEEN AMERICA AND EUROPE 8 (2003) ("[In Europe], offenders who do wind up in prison are subjected to a regime markedly less degrading than that that prevails in the United States.... Prison uniforms have generally been abolished. Rules have been promulgated attempting to guarantee that inmates be addressed respectfully.... Rules have also been promulgated protecting inmate privacy.... [T]hese measures include ... the principle that life in prison should approximate life in the outside world as closely as possible.").

(122.) DUFF, supra note 35, at 149.

(123.) Id. (emphasis omitted).

(124.) Id. at 150.

(125.) Id.

(126.) Id. at 149.

(127.) Interestingly, prisons were intended from their start as penal institutions in the late eighteenth and early nineteenth centuries to be places of penance and moral transformation, although the reality of prison life even then often veered from the penitential ideal into simple warehousing. Kleinig, supra note 31, at 278.

(128.) For instance, the "supermax" prisons established by many states in the 1990s, which are designed to isolate problem inmates to the maximum extent, see Zimring & Hawkins, supra note 119, at 166-69, would seem pretty clearly not to be places where retributive punishment would be appropriate for disciplinary infractions.

(129.) Anthony E. Bottoms, Interpersonal Violence and Social Order in Prisons, in PRISONS 205, 208 (Michael Tonry & Joan Petersilia eds., 1999) (noting, for instance, that "in-cell televisions are no longer a rarity in prisons"). This is not to downplay the importance of the (probably unnecessarily high) barriers to contact between inmates and their family members, such as poor treatment of visitors by prison officials and exorbitant fees for telephone use. PETERSILIA, supra note 77, at 44-46. Moreover, it should be noted that some jurisdictions acted to restrict inmates' access to television in the 1990s. Zimring & Hawkins, supra note 119, at 165.

(130.) ROGER MATTHEWS, DOING TIME: AN INTRODUCTION TO THE SOCIOLOGY OF IMPRISONMENT 33 (2d ed. 2009) (describing "podular design"). Ironically, the new design approach was intended to further rehabilitative ends, but rehabilitation fell out of favor just as the new approach became established. Id. at 35.

(131.) Chase Riveland, Prison Management Trends, 1975-2025, in PRISONS 163, 172-74, 184 (Michael Tonry & Joan Petersilia eds., 1999).

(132.) See, e.g., id. at 174 (noting effect of Prison Litigation Reform Act in limiting inmate legal challenges to prison conditions).

(133.) Thornburgh v. Abbott, 490 U.S. 401, 407 (1989) (quoting Turner v. Safley, 482 U.S. 78, 84 (1987)). The Supreme Court embraced this principle and rejected the so-called "hands off" approach in the 1970s. See Hudson v. Palmer, 468 U.S. 517, 555 (1984) (Stevens, J., dissenting).

(134.) Riveland, supra note 131, at 174.

(135.) See PETERSILIA, supra note 77, at 93 ("[V]irtually all prison programs have long waiting lists.").

(136.) Riveland, supra note 131, at 178. For the contrasting glass-is-half-empty view, see, e.g., DAVID GARLAND, THE CULTURE OF CONTROL: CRIME AND SOCIAL ORDER IN CONTEMPORARY SOCIETY 178 (2001) ("The prison is used today as a kind of reservation, a quarantine zone in which purportedly dangerous individuals are segregated in the name of public safety."). I do not mean here to reject the pessimistic characterizations of Garland and other critics of our penal practices, but I would suggest that there are some hopeful signs, too, and that public preferences do not seem consistent with the vision of imprisonment as a form of exile or simple warehousing. See infra text accompanying notes 154-55.

(137.) 468 U.S. 517 (1984).

(138.) Id. at 530.

(139.) Id. at 525 (citation and internal quotation marks omitted).

(140.) Id. at 526.

(141.) Id.

(142.) Id. at 527.

(143.) Id. at 542 (Stevens, J., dissenting).

(144.) Id. at 545-47.

(145.) See id. at 552 ("'Without the privacy and dignity provided by fourth amendment coverage, an inmate's opportunity to reform, small as it may be, will further be diminished. It is anomalous to provide a prisoner with rehabilitative programs and services in an effort to build self-respect while simultaneously subjecting him to unjustified and degrading searches and seizures.'" (quoting Paul C. Gianelli & Francis A. Gilligan, Prison Searches and Seizures: "Locking" the Fourth Amendment Out of Correctional Facilities, 62 VA. L. REV. 1045, 1069 (1976))).

(146.) Id. at 551.

(147.) Id. at 553 n.28.

(148.) Id. at 553. Stevens' argument would in some respects be even stronger today: the homicide rate in prisons has actually fallen sharply since Hudson. Bottoms, supra note 129, at 225-26. Inmates themselves typically place a high value on order and safety in prisons, id. at 250, so it may be misleading to characterize prison security as purely a matter of guards exercising control over recalcitrant inmates, as Burger seems to do. To be sure, the accuracy of Stevens' more positive picture of prison life varies considerably from institution to institution and inmate to inmate. For instance, conditions in some prisons cause inmates to look for order not from prison officials, but from gang membership: unfortunately, "[t]hey soon learn that, eventually, having been identified as a gang member virtually always results in a worsened rather than improved quality of prison life." CRAIG HANEY, REFORMING PUNISHMENT: PSYCHOLOGICAL LIMITS TO THE PAINS OF IMPRISONMENT 219 (2006). Additionally, researchers have identified some groups of inmates with special vulnerabilities, such as those with mental illness or developmental disabilities, who may be more likely to experience prison "as an especially painful, intolerant, and even intolerable place." Id. at 241. Those who experience sexual victimization in prison may also be in a particularly precarious position. Neal & Clements, supra note 120, at 288-89.

(149.) Hudson, 468 U.S. at 553 (Stevens, J., dissenting).

(150.) Id. at 554.

(151.) Id. at 549 (citing cases). I am grateful to Michael Tonry for drawing to my attention the fascinating work of one federal district judge in the 1970s, James Doyle, who articulated with particular clarity and forcefulness the belief that courts ought to recognize the claims of rights violations by prisoners on no different footing than similar claims made by other citizens. See, e.g., Mabra v. Schmidt, 356 E Supp. 620, 630 (W.D. Wis. 1973) ("I believe that the time has come in this corner of human experience, as it came much earlier in others, for the courts to make with careful and deliberate firmness, the limits upon permissible governmental interference in the lives of the people. To the degree that the present correctional system cannot withstand this searching inquiry, it should be terminated, better soon than late.").

(152.) See, e.g., Johnson v. California, 543 U.S. 499, 509 (2005) (holding that inmate challenging racial segregation in prison could get benefit of strict scrutiny test under Equal Protection Clause).

(153.) The more common approach is for the Court to use the Turner v. Safley test, which recognizes inmates' rights, but permits restrictions using a deferential balancing test. See supra note 120.

(154.) Donald Braman, Punishment and Accountability: Understanding and Reforming Criminal Sanctions in America, 53 UCLAL. REV. 1143, 1145-46 (2006).

(155.) Id. at 1195. In addition to offering such views in public opinion surveys, the voters in at least one state have also expressed this position in the form of a ballot initiative. See TRAWS, supra note 1, at 177-79 (describing Oregon's Measure 17, which required that prisoners work forty hours a week in a job or in other activities that prepare them for a job).

(156.) For a discussion of the persistence of good time and earned time, see TRAVIS, supra note 1, at 53-54. Interestingly, Travis argues that inmates should be given good- and earned-time credits regardless of predictions of their future behavior, which means that his vision for expanded good- and earned-time opportunities would be quite closely compatible with what I envision for retributive indeterminate sentencing.

(157.) See infra notes 199-200.

(158.) A well-known example of the sort of program that seems to involve inmates in publicly approved, pro-social activities is the training of dogs. Colorado's program, for instance, is described here: http:// (last visited June 1, 2011).

(159.) Perhaps most importantly, Stevens sometimes invoked traditional rehabilitative considerations that are in tension with Duff's more retributive framework. See, e.g., Hudson v. Palmer, 468 U.S. 517, 552 (1984) (Stevens, J., dissenting) (discussing adverse consequences of denying inmates any residuum of privacy). But cf. id. at 558 (Stevens, J., dissenting) (closing opinion by invoking not consequences, but "the ethical tradition that I had thought was enshrined forever in our jurisprudence").

(160.) Despite the perceptions of idleness that Braman encountered, approximately 53% of inmates had work assignments in 2000, while an estimated 10% of inmates participate in prison educational or vocational programs. TRAVIS, supra note 1, at 160-61. On the other hand, it is also true that prisons are investing less in educational and vocational programs than in the past, which diminishes participation opportunities for many inmates. Id. at 161.

One potentially important symbolic measure that states could undertake to highlight the continued membership of inmates in the outside community would be to overturn disenfranchisement laws that take away the right to vote from inmates and even from convicts post-release. Only two states (Maine and Vermont) permit inmates to vote in prison, while thirty-one states prohibit parolees from voting, and eleven impose a lifetime ban on voting for anyone with a felony conviction. Id. at 254-56.

(161.) See, e.g., 28 C.F.R. [section] 541.15(b) (2010) ("Each [federal] inmate so charged [with violating a prison rule] is entitled to an initial hearing before the [Unit Discipline Committee], ordinarily held within three work days from the time staff became aware of the inmate's involvement in the incident.") (emphasis added); [section] 541.15(0 ("The UDC shall give the inmate a written copy of the decision and disposition by the close of business the next work day [after the hearing].") (emphasis added).

(162.) Although criminologists have long theorized that the prison environment contributes to inmate misconduct, little rigorous empirical work has been done to support the proposition. Scott D. Camp et al., Fed. Bureau of Prisons, THE INFLUENCE OF PRISONS ON INMATE MISCONDUCT: A MULTILEVEL INVESTIGATION 2--4 (2003), available at mgmt/oreprcamp_mis.pdf. However, one recent study of misconduct in federal prisons found statistically significant correlations between one or more forms of misconduct and several prison-level variables, including the average security level of inmates, crowding, and racial integration. Id. at 22-23. On the other hand, an even more recent meta-analysis concludes that "inmate crowding emerged as a weak predictor of violence and misconduct in prison." Travis W. Franklin et al., Examining the Empirical Relationship Between Prison Crowding and Inmate Misconduct: A Meta-Analysis of Conflicting Research Results, 34 J. CRIM. JUST. 401, 408 (2006). The authors instead suggest that "inmate violence and general misconduct stem from poor managerial skills, where guards are not properly trained and prison security is compromised." Id. Whatever the relative importance of crowding and managerial incompetence in contributing to broad patterns of misconduct across an institution, such considerations might well be relevant to determining willfulness in many individual cases.

(163.) For a description of disciplinary segregation and its effect on inmates, see COMM'N ON SAFETY AND ABUSE IN AMERICA'S PRISONS, CONFRONTING CONFINEMENT 57-59 (2006). The number of inmates in disciplinary segregation has been growing rapidly, and anecdotal evidence suggests that inmates are frequently segregated for minor rules violations and that stays in segregation can last for months or even years. See id. at 53.

(164.) 442 U.S. 1 (1979).

(165.) Id. at 12. This aspect of Greenholtz was cast into doubt, but not clearly overturned, by Sandin v. Conner, 515 U.S. 472 (1995). Sandin was a prison discipline, not a parole, case. However, in Sandin, the Court expressed considerable dissatisfaction with a line of procedural due process cases, including Greenholtz, in which it had parsed state statutes for discretion-cabining language in order to determine whether the plaintiffs had a constitutionally protected liberty interest. Id. at 479-84. On the other hand, the Court made clear that it was not overruling any of its prior holdings. Id. at 484 n.5. Moreover, the new test the Court formulated (whether a challenged disciplinary sanction "imposes atypical and significant hardship on the inmate in relation to the ordinary incidents of prison life," id. at 484) does not seem to contemplate parole decisions. In any event, even if the Greenholtz statute-parsing approach were no longer the law, the effect would not be to enhance, but only further to diminish, due process requirements in connection with parole decisions.

(166.) Greenholtz, 442 U.S. at 14-15.

(167.) Id. at 13-14.

(168.) Id. at 9.

(169.) Outside the parole context, the Court has not been entirely clear or consistent as to what triggers each of the core constitutional criminal procedure protections. For instance, a well-known recent line of cases holds that defendants are entitled to have "any fact that increases the penalty for a crime beyond the prescribed statutory maximum" found by a jury. Apprendi v. New Jersey, 530 U.S. 466, 490 (2000). However, the Court has declined to require jury fact-finding for mandatory minimum sentences, Harris v. United States, 536 U.S. 545, 551-52 (2002), or with respect to a decision to impose consecutive sentences, Oregon v. Ice, 555 U.S. 160, 160 (2009). Perhaps a decision to delay release could be similarly conceptualized as incremental punishment within an otherwise permissible maximum. Moreover, the Court has not as yet extended the Apprendi principle to other constitutional rights, such as the fight to confront accusers.

In a provocative recent article, David Ball argues that Apprendi ought to be understood as establishing a distinction between "retributive facts" and facts relating to public safety; in Ball's view, a jury must find retributive facts, that is, facts that form the basis for a judgment of condemnation. Ball, supra note 108, at 923-26. In his view, Apprendi is best read as precluding parole boards from extending incarceration based on their own views of the moral gravity of the crime of conviction. Id. at 938. Although Ball does not address a parole-board role such as I outline here, his reasoning might suggest that a parole board could not respond to in-prison misconduct in an expressly retributive fashion, at least in the absence of jury fact-finding. The argument, which is based on Bali's view that "the jury, the conscience of the community, is uniquely suited to make moral judgments," id. at 924, is not without appeal. But, as Ball concedes, the Court has not adhered to his interpretation of Apprendi in its more recent decisions. See id. at 910-11 (discussing Cunningham v. California, 549 U.S. 270 (2007)). As Ball also notes, any effort to extend Apprendi to parole decisions would run into the objection that the Sixth Amendment (the constitutional source of the right to jury fact-finding) only applies to "criminal prosecutions," and it is not clear that a parole decision ought to be regarded as part of the "prosecution." Id. at 932. Finally, it has long been established that the right to a jury trial does not extend to cases in which the defendant faces relatively less serious punishments. See, e.g., Blanton v. City of N. Las Vegas, 489 U.S. 538, 541 (1989) (finding no right to jury trial for DUI offense where maximum penalty is prison term of six months). Although the normal cut-off is exposure to no more than six months of punishment, id. at 543, the Court (following the distinction made in Greenholtz) might find a different cut-off to be appropriate in connection with a parole denial that merely keeps a person who is already in prison behind bars for awhile longer.

(170.) O'Hear, supra note 35, at 409 (noting that approximately 95% of convictions in the United States are obtained by way of a guilty plea).

(171.) Sergio Herzog, Plea Bargaining Practices: Less Covert. More Public Support?, 50 CRIME & DELINQ. 590, 590-91 (2004).

(172.) O'Hear, supra note 35, at 426-31.

(173.) Id. at 431.

(174.) As noted above, David Ball argues that one particularly costly aspect of traditional trial procedures (jury fact-finding) ought to be employed whenever there is punishment to be imposed based on "retributive facts." See supra note 169 (discussing "retributive facts"). His argument is premised on the greater suitability or competence of juries to function as the "conscience of the community" in comparison to judges and parole boards. Ball, supra note 108, at 924. Although the claim is plausible, the (somewhat stronger) assertion that punishment requires jury fact-finding in order to have legitimacy in our polity would have to account for the reality that punishment is routinely imposed in the United States in the absence of jury fact-finding (that is, pursuant to a plea deal). Moreover, it is not clear that juries (selected from pools of varying representativeness and winnowed by lawyers bent on tactical advantage, see Shari Seidman Diamond, Beyond Fantasy and Nightmare: A Portrait of the Jury, 54 BUFF. L. REV. 717, 733-36 (2006) (summarizing research on jury representativeness)) are substantially better qualified to act as the conscience of the community than parole boards appointed by politically accountable officials.

(175.) Even with this structure, however, the severity of the cumulative responses to continuing misconduct will be somewhat arbitrarily constrained by the severity of the underlying sentence. For instance, imagine two equally incorrigible inmates, one of whom has been sentenced to a short term for a relatively minor offense and the other of whom faces a much longer sentence for a much more serious offense. Although their behavior in prison may merit equal condemnation, the parole board will run out of opportunities to deny parole much more quickly as to the first of the two inmates. On the other hand, if the first inmate's prison violations are sufficiently serious, then it might be possible to prosecute them as fresh crimes in their own right (e.g., assaults of guards or fellow inmates). And if the violations are less serious, then satisfactory punishment may be available through standard prison disciplinary processes. This is not say that the perfect parity in the treatment of equally incorrigible inmates can consistently be achieved through or around a retributive delayed-release mechanism, but the magnitude of disparities does not necessarily seem large relative to current norms. See U.S. SENT'G COMM'N, FIFTEEN YEARS OF GUIDELINES SENTENCING: AN ASSESSMENT OF HOW WELL THE FEDERAL CRIMINAL JUSTICE SYSTEM IS ACHIEVING THE GOALS or SENTENCING REFORM xi-xvi (2004) (discussing ongoing disparity problems in the federal system).

(176.) DUFF, supra note 35, at 152.

(177.) Id. at 132.

(178.) Id. at 126.

(179.) Bottoms, supra note 129, at 270.

(180.) Bronsteen et al., supra note 18, at 1048.

(181.) Id. at 1048 n.67; see also Dhami et al., supra note 77, at 1096 (discussing survey results indicating that, controlling for a number of variables, longer-serving inmates participate in more programs than newer inmates and have fewer disciplinary infractions). This is not to say that successful adaptation is inevitable or that patterns of adaptation are uniform among all inmates. A large body of penological research has explored the adaptation phenomenon and identified several variables that seem important. For a summary of leading studies, see Bottoms, supra note 129, at 245-49. Much of the research highlights the significant role played by institutional variables and the "fit" between the inmate and the institution. Id. at 247-48.

(182.) Cf. Francis T. Cullen et al., Public Opinion About Punishment and Corrections, 27 CRIME & JUST. 1, 42 (2000) (discussing studies showing that offenders "rate certain intermediate penalties as more severe than limited stays in prison").

(183.) Bronsteen et al., supra note 18, at 1054.

(184.) Christy A. Visher, Returning Home: Emerging Findings and Policy Lessons About Prisoner Reentry, 20 FED. SENT'G REP. 93, 96 (2007); see also PETERSILIA, supra note 77, at 9 ("Since 1980, the United States has passed dozens of laws restricting the kinds of jobs for which ex-prisoners can be hired, easing the requirements for their parental rights to be terminated, restricting their access to public welfare and housing subsidies, and limiting their right to vote.").

(185.) DONALD BRAMAN, DOING TIME ON THE OUTSIDE: INCARCERATION AND FAMILY LIFE IN URBAN AMERICA 165--67 (2004); see also PETERSILIA, supra note 77, at 11 ("The stigma associated with a criminal past significantly affects one's chances of finding and keeping a job, personal relationships, and housing....); id. at 10 (noting increasing public availability of criminal records through the Internet).

(186.) See PETERSILIA, supra note 1, at 171 ("A majority of those being released to parole will not successfully complete their terms...."); see also PATRICK A. LANGAN & DAVID J. LEVIN, U.S. DEP'T OF JUSTICE, RECIDIVISM OF PRISONERS RELEASED IN 1994, at 1 (2002) (based on a study of prisoners released from fifteen states in 1994, concluding that 67.5% were rearrested for a new offense within three years of their release).

(187.) PETERSILIA, supra note 77, at 73.

(188.) I suggest here that the difficulties of post-prison life should be taken into account in assessing the retributive significance of an incremental year in prison. On the other hand, by my reasoning, to the extent that at least some of these difficulties are limited to a transitional period in the former inmate's life, the incremental year's delay in completing the transition and attaining a more satisfactory situation in the community also ought to be taken into account, which would concededly diminish the force of my argument on this point. I'm grateful to Antony Duff for drawing my attention to this difficulty.

(189.) Cf. Dan Markel & Chad Flanders, Bentham on Stilts: The Bare Relevance of Subjectivity to Retributive Justice, 98 CALIF. L. REV. 907, 969-70 (2010) ("Given our focus on the idea of communicating the wrongness to the offender, and, in the case of prison, using the custodial period to signal in part the degree of societal condemnation, the offender's post-supervision experience is less important. Because the polity did not intend, authorize, or proximately cause these contingent and speculative (even if foreseeable) post-punishment experiences or effects, they cannot plausibly carry any communicative message on behalf of the polity.").

(190.) Hudson v. Palmer, 468 U.S. 517, 553 n.28 (1984) (Stevens, J., dissenting).

(191.) See Jeremy A. Blumenthal, Law and Emotions: The Problems of Affective Forecasting, 80 IND. L.J. 155, 167-69 (2005) (summarizing various studies regarding hedonic adaptation).

(192.) For a more detailed description of popular punitiveness, see Braman, supra note 154, at 1181-83. Shaming sanctions include, for instance, "televising the names of the busted clients of prostitutes, or making drunk drivers parade around wearing placards describing their offenses, or leading Wall Street brokers out of their offices in shackles." WHITMAN, supra note 121, at 24.

(193.) Cf. Cullen et al., supra note 182, at 35 ("[T]hose surveyed [in public opinion surveys in the 1990s] generally agreed that 'the vast majority of prison inmates sit around all day, playing cards and watching TV....'").

(194.) Exemplified, for instance, by passage of the federal Second Chance Act, there has been considerable bipartisan interest in recent years in providing enhanced social services for returning prisoners. Beth A. Colgan, The Presidential Politics of Prisoner Reentry, Reform, 20 FED. SENT'G REP. 110, 111-12 (2007).

(195.) See Mary Zahn & Gina Barton, Tougher Sentencing Law Carries Hefty. Price; Estimated $1.8 Billion through 2025, MILWAUKEE J. SENTINEL, Nov. 21, 2004, at 1A, 20A (noting that Wisconsin inmates admitted under new truth-in-sentencing law were responsible for higher rate of major conduct violations than inmates who were eligible for parole).

(196.) See Mary Zahn, Inmates Less Motivated, Wardens Find, MILWAUKEE J. SENTINEL, Nov. 22, 2004, at 14A (noting view of prison wardens that adoption of truth-in-sentencing law in Wisconsin reduced the motivation of inmates to participate in drug and alcohol treatment programs and other rehabilitative opportunities).

(197.) See Zahn & Barton, supra note 195, at 1A ("A state law that gives criminals virtually no chance for early release will cost Wisconsin taxpayers an estimated $1.8 billion for inmates admitted through 2025 if current trends continue....").

(198.) As indicated in the Appendix, reforms along these lines recently

have been adopted in Arkansas, Colorado, Kentucky, Louisiana, Maryland, Michigan, Minnesota, Mississippi, and New York.

(199.) See Duff, supra note 23, at 161 (criticizing three strikes laws); Douglas N. Husak, Desert, Proportionality, and the Seriousness of Drug Offences, in FUNDAMENTALS OF SENTENCING THEORY: ESSAYS IN HONOR OF ANDREW YON HIRSCH 187, 218 (Andrew Ashworth & Martin Wasik eds., 1998) (discussing difficulty of justifying criminal punishment for drug possession offenses under principle of proportionality).

(200.) As indicated in the Appendix, reforms along these lines have recently been adopted in Alabama, Connecticut, Virginia, and Wisconsin.

(201.) As indicated in the Appendix, reforms along these lines have recently been adopted in California, Colorado, Louisiana, Maine, Nevada, New York, and Washington.

(202.) As indicated in the Appendix, reforms along these lines have recently been adopted in California, Delaware, Illinois, Kansas, Kentucky, Louisiana, Mississippi, New York, North Carolina, Oregon, Pennsylvania, Tennessee, Washington, Wisconsin, and Wyoming. For a consequentialist argument in favor of expanding good- and earned-time, as well as some helpful reform proposals for such programs, see TRAVIS, supra note 1, at 53-54.

(203.) See THOMAS P. BONCZAR, U.S. DEP'T OF JUSTICE, PREVALENCE OF IMPRISONMENT IN THE U.S. POPULATION, 1974--2001, at 1 (2003) (showing doubling of percentage of U.S. adults who had served time in state or federal prison between 1974 and 2001).

(204.) Consistent with this view, Duff argues that we can and should embark on a "'decremental strategy' of gradually reducing the general levels of penal severity." DUFF, supra note 35, at 134.
 The point of such a strategy will be to see how far we can lower
 our conceptions of what modes and levels of punishment will be
 adequate to communicate an appropriate kind and degree of censure
 to offenders.... I see no reason to believe that [the reductions]
 could not in the end be very significant.

Id. at 134-35.

(205.) For fiscal year 2010 alone, at least nine states have cut health services for prisoners, seven have cut food services, and twenty have eliminated or reduced other programs, or discontinued or renegotiated contracts for programming. SCOTT-HAYWARD, supra note 19, at 5.

(206.) See Reitz, supra note 24, at 206 (discussing recent history of parole-board politicization).

(207.) See, e.g., von Hirsch, supra note 17, at 667 (noting support of substantial reductions in penalty levels by von Hirsch, Singer, Duff, and Ashworth).

(208.) See Bottoms, supra note 32, at 75 ("[P]hilosophical discussions of criminal punishment usually focus either on the fundamental justification for punishment ... or on how much we should punish....").

(209.) See, e.g., Jean Hampton, The Moral Education Theory of Punishment, 13 Prim. & PUB. Am. 208, 209-21 (1984) (presenting moral-education theory as justification for punishment); Markel & Flanders, supra note 189, at 968. ("[Retributive justice] is a theory about the proper justification of punishment.... [R]etributive theories stand at their tallest when they explain why we punish....").


(211.) See, e.g., von Hirsch, supra note 17, at 668-70.

(212.) Because so much of the attention of theorists has focused on the implementation of proportionality and uniformity ideals, their principal interest in imprisonment typically seems to be in the quantitative aspects of the prison term, with remarkably little regard for variation along quantitative dimensions. See, e.g., Bottoms, supra note 32, at 57 ("[T]he severity of sentences of imprisonment can be fairly straightforwardly measured by their length...."). Then again, this orientation of retributive thinking may not be anything especially distinctive; as social critics have long recognized, the preoccupation with time in our penal practices--that is, the quantitative dimension of imprisonment--likely owes much to the historical path of twentieth-century capitalism and to the commodification of time in labor markets. MATTHEWS, supra note 130, at 10, 36-38.

(213.) For another preliminary treatment of the topic, see Kleinig, supra note 31, at 288-97. Robert Blecker has also offered a provocative reform proposal for prisons that is informed by retributive considerations. Robert Blecker, Haven or Hell? Inside Lorton Central Prison: Experiences of Punishment Justified, 42 STAN. L. REV. 1149, 1244-48 (1990). However, his understanding of retribution seems to focus on proportionality alone and without a broader communicative agenda. See, e.g., id. at 1164.

(214.) Braman, supra note 154, at 1192-98.

(215.) Id. at 1195.

(216.) Id. at 1199.




(220.) National Conference of State Legislatures, Significant State Sentencing and Corrections Legislation in 2009 (2010), available at [hereinafter National Conference of State Legislatures, 2009].



(223.) National Conference of State Legislatures, State Sentencing and Corrections Legislation in 2010, available at [hereinafter National Conference of State Legislatures, 2010].

(224.) KING, CHANGING DIRECTION, supra note 219, at 12.

(225.) National Conference of State Legislatures, Significant State Sentencing and Corrections Legislation in 2007 and 2008 (2010), available at = 12682 [hereinafter National Conference of State Legislatures, 2007-2008].

(226.) National Conference of State Legislatures, 2010, supra note 223.

(227.) National Conference of State Legislatures, 2009, supra note 220.

(228.) National Conference of State Legislatures, 2007-2008, supra note 225.

(229.) PORTER, 2010, supra note 221, at 9.

(230.) LAWRENCE, supra note 222, at 6.

(231.) PORTER, 2010, supra note 221, at 3.





(236.) National Conference of State Legislatures, 2007-2008, supra note 225.

(237.) PORTER, 2009, supra note 233, at 7.

(238.) Id. at 13.

(239.) VERA INSTITUTE OF JUSTICE, supra note 234, at 18, 19.

(240.) National Conference of State Legislatures, 2007-2008, supra note 225.

(241.) National Conference of State Legislatures, 2009, supra note 220.

(242.) KING, CHANGING DIRECTION, supra note 219, at 6.

(243.) KING, 2007, supra note 218, at 10.

(244.) National Conference of State Legislatures, 2009, supra note 220.

(245.) PORTER, 2010, supra note 221, at 9.

(246.) GREENE & MAUER, supra note 21, at 28.

(247.) KING, CHANGING DIRECTION, supra note 219, at 14.

(248.) KING & MAUER, supra note 235, at 5, 9.

(249.) KING, CHANGING DIRECTION, supra note 219, at 14.

(250.) Id.

(251.) KING, 2008, supra note 232, at 4-5.

(252.) PORTER, 2009, supra note 233, at 3.

(253.) National Conference of State Legislatures, 2010, supra note 223.

(254.) KING, 2007, supra note 218, at 14.

(255.) National Conference of State Legislatures, 2009, supra note 220.

(256.) PORTER, 2010, supra note 221, at 9-10.

(257.) Id. at 3.

(258.) KING, CHANGING DIRECTION, supra note 219, at 15-16.

(259.) Id. at 16.

(260.) GREENE & MAUER, supra note 21, at 25.

(261.) National Conference of State Legislatures, 2010, supra note 223.

(262.) KING & MAUER, supra note 235, at 10.

(263.) National Conference of State Legislatures, 2007-2008, supra note 225.

(264.) Id.

(265.) Id.

(266.) VERA INSTITUTE OF JUSTICE, supra note 234, at 17-18.

(267.) SCOTT-HAYWARD, supra note 19, at 11.

(268.) National Conference of State Legislatures, 2010, supra note 223.

(269.) VERA INSTITUTE OF JUSTICE, supra note 234, at 18-19.

(270.) National Conference of State Legislatures, 2007-2008, supra note 225.

(271.) PORTER, 2009, supra note 233, at 3.

(272.) Id. at 14.

(273.) VERA INSTITUTE OF JUSTICE, supra note 234, at 19.

(274.) CHIU, supra note 217, at 9.

(275.) LAWRENCE, supra note 222, at 3.

(276.) National Conference of State Legislatures, 2007-2008, supra note 225.

(277.) CHIU, supra note 217, at 9.

(278.) National Conference of State Legislatures, 2010, supra note 223.

(279.) VERA INSTITUTE OF JUSTICE, supra note 234, at 19.

(280.) National Conference of State Legislatures, 2009, supra note 220.

(281.) National Conference of State Legislatures, 2007-2008, supra note 225.

Michael M. O'Hear, Associate Dean for Research and Professor, Marquette Law School. Editor, Federal Sentencing Reporter. Author, Life Sentences Blog. B.A., J.D., Yale University. I am grateful to Antony Duff, Janie Kim, Michael Tonry, and participants at a Marquette Law School faculty workshop for comments on an earlier draft. [C] Michael M. O'Hear, 2011.
COPYRIGHT 2011 Georgetown University Law Center
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2011 Gale, Cengage Learning. All rights reserved.

 Reader Opinion




Article Details
Printer friendly Cite/link Email Feedback
Author:O'Hear, Michael M.
Publication:American Criminal Law Review
Date:Jun 22, 2011
Previous Article:Tax violations.
Next Article:Deal or no deal: why courts should allow defendants to present evidence that they rejected favorable plea bargains.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters