Printer Friendly

Vulnerability and just desert: a theory of sentencing and mental illness.

III. HOW VULNERABILITY MAY AFFECT THE DISTRIBUTION OF DESERVED PUNISHMENT

While philosophers, scholars, and policymakers have argued for centuries over the proper justification for state-imposed punishment, less attention has been paid to theories governing the allocation of punishment (177) Critically, the principles that justify the power of the state to impose punishment may differ from those that control the distribution of punishment--the type and quantity of punishment the state should order a particular offender to suffer for a particular crime relative to the penalties other offenders should receive for their offenses. (178) Given the great variety and sheer number of offenses and offender characteristics, and the competing norms at stake, elucidating a defensible distributive theory is a difficult question of tremendous practical significance. (179)

There are numerous strands of retributivism, (180) but one dominant viewpoint holds that state-imposed punishment serves to express blame or censure, (181) Professor Joel Feinberg first observed that punishment has a denunciatory aspect: "[P]unishment is a conventional device for the expression of attitudes of resentment and indignation, and of judgments of disapproval and reprobation, on the part either of the punishing authority himself or of those 'in whose name' the punishment is inflicted." (182) While few would disagree with this descriptive account, more controversial has been some scholars' assertion that the distribution of punishment can be justified as a means of expressing a certain quantum of censure. (183) Modern desert theorists have distinguished between justifying the censure of criminals and justifying the imposition of the hard treatment typically inherent in state punishment. (184) According to some expressive perspectives, the hard treatment inherent in a criminal sanction should reflect the degree of censure appropriate for an offender's blameworthiness. (185) This viewpoint inspired the development of just desert theory, which rose to prominence in the late 1970s as a means of curtailing sentencing discretion and bounding the state's coercive power over offenders. (186)

Just desert theory, which has been propounded most thoroughly and effectively by Professor Andrew von Hirsch, (187) purports to allocate punishment according to a proportionality equation involving culpability, seriousness of harm, and severity of penalty. (188) A chief aim of desert theory is to translate retributivism's general call for just punishment into a workable scheme to guide sentencing policy and individual sentencing decisions. (189) Desert theory gives "conceptions of justice a central role in sentencing policy" through the mechanism of proportionality, which is intended to ensure that penal sanctions fairly reflect the culpability of an offender and the harmfulness of his offense. (190) One especially relevant concern of desert theorists is whether, and how, to take into account the differential impact of certain penalties on vulnerable populations of offenders. (191) Desert theory, in light of its prominence and explicit aim to determine the proper amount of punishment in individual cases, should inform the current debate over retributivism's accommodation of foreseeable risk of harm.

As will be explained in Part III.B, just desert theory measures a penalty's severity by how the sanction typically affects an offender's interests and quality of life, from the perspective of the typical offender. (192) Given desert theory's focus on deprivations as typically experienced by offenders (as opposed to considering only those hardships intended by a sentencing authority), the theory appears to rely upon an understanding of punishment broad enough to encompass at least foreseeable, substantial risks of serious harm, proximately caused by the state in the context of incarceration. (193) This appears especially true for one strand of desert theory--that developing and applying the "equal-impact principle"--which is premised upon a recognition that particular sanctions typically pose significant and foreseeable (but presumably unintended) hardships for offenders with certain handicaps. (194) For that reason this Article will stipulate to and partially defend a definition of punishment ample enough to include foreseeable, substantial risks of serious harm that is proximately caused by the state. (195) Recently scholars have debated whether punishment should include foreseeable risks of harm and from whose perspective the severity of a penalty should be measured. (196) Offering a full defense of a particular definition or theory of punishment is beyond the scope of this Article, but the following section will highlight some of the main arguments in favor of extending punishment beyond intentional deprivations as well as offer additional support, derived from an expressive view of punishment, for why evaluation of punishment severity for purposes of sentencing should include foreseeable, substantial risks of serious harm that is proximately caused by the state in the context of incarceration.

A. "PUNISHMENT," RISK OF SERIOUS HARM, AND PENAL SEVERITY

Criminal punishment is, broadly speaking, the state's imposition of a typically unpleasant condition on an individual in response to that individual's violation of a legal rule. (197) Traditionally, philosophers and legal scholars have defined punishment as consisting only of hardships or deprivations intended and authorized by a legitimate sentencing authority. (198) Recently, Professors Dan Markel, Chad Flanders, and David Gray have defended a similar conception of punishment. (199) Under this formulation, the kinds of experience described in Part II--physical and sexual assault by inmates and prison guards, and mental deterioration from extended stints of solitary confinement--would not constitute punishment. Physical and sexual assault are both unlawful and are unintended by a sentencing authority. If these harms fall outside the realm of punishment, then a sentencing judge may be under no obligation to consider the likelihood of their occurrence when meting out proportionate punishment. (200)

Challenging that viewpoint, a number of scholars have recently stressed that the state should be morally responsible for the foreseeable results of its actions, and some have argued that assessments of a sanction's severity, for purposes of sentencing, should therefore include foreseeable risks of harm. (201) As Professor Alice Ristroph has pointed out, punishment is not designed and meted out by a single actor with a single intent, but rather consists of a set of practices, with one practice triggering the next. (202) Given the numerous and potentially conflicting intentions involved in imposing and administering punishment (including the intentions of members of a legislature and possibly a sentencing commission, a sentencing judge, and prison officials), (203) differentiating between intentional and foreseen punishment can be difficult, (204) and it is unclear whether any difference would hold moral salience. (205) Professor David Enoch has argued that states' reliance upon the distinction between intention and foresight is an attempt to evade moral responsibility and that state actors should instead feel heightened responsibility to take into account the foreseen effects of their actions. (206) Professor Jesper Ryberg has argued through the use of hypotheticals that attempts to exclude foreseeable side effects and other unintended aspects of punishment from evaluations of punishment severity do not accord with common intuitions of just deserts and are "implausible." (207) Professor Adam Kolber has made a compelling case for a justification-symmetry principle: namely, "any state actor who harms an offender in the name of just punishment must have a justification for doing so if you or I would need a justification for causing the same kind of harm to nonoffenders" under the criminal law. (208) Thus, state actors should be required to justify harms that they recklessly or negligently inflict upon offenders in the punishment process. (209) Finally, to the degree that an aim of punishment is to convey censure or blame, the state "cannot define the content of its messages by authorial fiat," in the words of Professors Bronsteen, Buccafusco, and Masur, but must consider reasonable interpretations of its message, which would include foreseeable harms associated with a given penalty. (210)

1. Additional Support for Conceptualizing Certain Foreseeable Risks of Harm as Punishment

An expressive understanding of punishment suggests additional reasons why evaluation of punishment severity should extend beyond intentional deprivations to consider foreseeable, substantial risks of serious harm, proximately caused by the state during confinement. As stated previously, this Article adopts the position that the penalty imposed in response to an offender's crime should communicate society's disapproval and censure. (211) Thus, it is critical that society view the medium of the message (the penalty) as censorious and the penal response as roughly proportionate to the seriousness of the offense. (212) For this reason, some scholars have concluded that a penalty should be measured by how society perceives the typical offender will experience it. (213)

It is reasonable, however, to expect that the widespread experience of a class of people subjected to a penalty will shape, at least to some extent, the public's perception of the severity and constitution of that penalty, so long as the experience of that class is brought to the public's attention. (214) This should be especially true when published, empirical studies document a group's experience and the class of persons affected can be readily identified. (215) If the public recognizes that offenders with major mental illnesses are more vulnerable to serious physical or mental harm in prison than non-ill offenders, then it should be inclined to evaluate the severity and constitution of carceral penalties for these offenders differently from those for non-disordered individuals. Indeed, evidence suggests that, at least to some degree, society does differentiate between the prison experiences of offenders with and without mental illnesses (216) and does believe that this experience merits mitigation of the sentences of seriously disordered offenders. (217) It is important to keep in mind, though, that the public's perception of desert is not static but highly malleable and susceptible to change through public education. (218) Therefore, as a normative matter, it is useful to analyze whether judges, as conduits for expressing society's disapproval, should pay attention to--and adjust sentences in response to--the ways in which offenders with major mental disorders, as a class, tend to experience serious harm in prison.

Retributivism's commitment to justice and respect for the individual offender suggest the proper resolution of this normative question. Justice is the chief concern of retributive theory, and, as a matter of justice and fairness, a punishment system should strive to avoid imposing punishments of differential impact on equally deserving offenders. (219) Ideally, the actual punishment experienced by an offender should equal in severity the penal response deemed deserved by his criminal act. Of course some degree of variance between predicted and actual experience is unavoidable. To the extent, however, that empirical evidence establishes that a penalty poses a foreseeable, substantial risk of serious harm, proximately caused by the state, (220) to a morally significant class of offenders, (221) this vulnerability should factor into sentencing. (222) Indeed, this approach is essential if the severity of a penalty experienced by an offender is to be fairly calibrated to the degree of censure intended by the state. (223)

The second argument in favor of reflecting risk of serious harm in sentencing derives from the role that offenders play as punishment's primary audience. (224) Again, society, as the sender of the communicatory message of punishment must, at base, view the penalty as censorious and roughly proportionate to the crime. (225) But society should tailor a punishment by how it anticipates that an offender in a given class will experience and understand a particular penalty. Professor R.A. Duff, who has articulated a powerful and distinctive communicative view of punishment, (226) has explained the importance of tailoring a punishment to an offender in these terms:

   If I am trying to communicate with someone, I must try to make the
   form and content of my communication appropriate to its context,
   its subject matter, and my interlocutor: I must do my best to
   ensure both that my communication does justice to its subject
   matter and also that it is so phrased and expressed that my
   interlocutor (given what I know about her) will have the best
   chance of understanding it. If we apply this idea to the context of
   punishment, it suggests that sentencers should look for (or try to
   create) that particular sentence that will express most appositely
   the censure merited by this offender's crime and which will be
   appropriate to this particular offender. Now this communicative
   ideal includes a requirement for proportional severity: the
   stringency of the censure we communicate must not be
   disproportionately severe (or lenient) in comparison to the crime
   we are censuring. It also, however, involves a requirement of
   substantive appositeness of "match" or "fit" between the particular
   substantive punishment and this particular crime and this
   particular criminal: is this the best kind of punishment by which
   to communicate to this offender an appropriate judgment on her
   particular crime? (227)


Treating the offender as the primary audience for society's censure reflects retributivism's preoccupation with, and commitment to honoring, the offender as a moral agent. (228) When empirical evidence demonstrates that a penalty poses a substantial risk of serious harm to an identifiable class of offenders, using class members' foreseeable experiences in part to measure the content and severity of punishment would be one way to express respect for the moral autonomy and dignity of the offender. (229) In addition, using the offender's anticipated experience as the benchmark for punishment severity presents the most rational way to achieve a key expressive or communicative aim of punishment: the hard treatment conveys to the offender that what you did was wrong, and this is how wrong it was. (230) The offender need not respond in any particular way to society's expression of disapprobation--epiphany and reform are hoped for but unnecessary results of punishment--but society assumes that competent offenders retain the ability to understand the message of disapproval. (231) Thus, from the moment a punishment is imposed to the moment it is completed, the way that an offender is likely to experience and understand a penalty should be of critical, though not controlling, (232) importance to the assessment of a penalty's content and severity.

The challenge of sentencing, then, is to accommodate both society's and offenders' views of the nature and severity of punishment. Most fundamentally, society must understand a proposed penalty as censorious and its severity as roughly proportionate to the seriousness of a given offense. Because offenders are the ultimate recipients of (and audience for) punishment's message of condemnation, society's proportionality analysis should consider how offenders are likely to experience a given penalty. Penalties need not be calibrated to every offender's idiosyncratic tolerance for punishment. But when an offender is an identifiable member of a vulnerable class, society's evaluation of a penalty's severity should reflect available evidence regarding how the penalty is likely to affect members of that class. At a sentencing hearing, (233) a judge should consider the empirically documented, substantial risks of serious harm that incarceration poses to offenders with major mental disorders and factor these risks into his sentencing calculus. (234) In particular, when evidence demonstrates that a particular sanction poses a substantial risk of serious harm to a seriously disordered individual, a judge should consider selecting an alternative penalty or taking other measures to avoid imposing a disproportionate or inhumane punishment. (235)

By considering foreseeable, substantial risks of serious harm, proximately caused by the state and posed by available criminal sanctions, the sentencing judge can take steps to ensure that the chosen penalty, as experienced, will equal the degree of condemnation actually warranted by an offender's criminal act. Of the myriad actors in the criminal justice system, the sentencing judge is the institutional player charged with selecting and conveying, within constraints established by the legislature and possibly a sentencing commission, the type and length of sentence that constitutes an offender's just deserts. (236) Because a judge typically cannot select the facility where an offender will serve a term of incarceration (237) (an aspect of an offender's sanction that will greatly affect its severity (238)), he possesses only a limited ability to ensure that an offender's ultimate punishment, as executed, is not harsher than intended. Often all a judge can do is attempt to predict, given data brought to his attention during the sentencing process, how an individual will fare if incarcerated (239) and then assess sentencing options accordingly. When an individual with a serious mental illness faces a possible term of incarceration, a sentencing judge should consider the defendant's foreseeable vulnerability to serious harm when weighing sentencing options to best ensure that the ultimate sentence ordered, as likely to be experienced, will convey the degree of censure warranted by his offense.

As this discussion suggests, expressive theory could potentially support a definition of punishment that encompasses a broad swath of foreseeable risks of harm, (240) and some scholars have suggested that punishment should be understood in this way. (241) However, primarily for prudential reasons, this Article takes the less radical position that foreseeable, substantial risks of serious harm, proximately caused by the state and occurring in the context of confinement, should factor into the distribution of punishment under a theory of proportionality. (242) For purposes of this Article, the term "serious harm" is limited to serious impairment of functioning and includes, but is not necessarily confined to, serious physical assault, serious sexual assault, substantial exacerbation of serious mental illness, and precipitation of a new serious mental disorder. This definition does not cover many of the harms and negative experiences suffered by offenders with serious mental disorders, but it does cover the most substantial harms experienced by at least a segment of this population.

2. Consideration of Prison Violence

The extent to which sentencing should reflect the foreseeable risk of prison violence merits separate discussion. Rape and assault in prison are certainly "not part of the penalty that criminal offenders pay for their offenses against society," and prison violence can never be condoned. (243) On this ground, Professors Gray, Markel, Flanders, and Mary Sigler have objected to the consideration of risk of prison violence in sentencing under a theory of proportionate punishment. (244) Their primary argument seems to be that, because a just punishment can never include rape or assault, (245) it would be immoral for sentencing authorities to consider the risk that these acts may occur and to provide a sentence reduction on the basis of that risk. (246) Reducing an offender's sentence in response to the fear that he might be raped in prison would, according to this viewpoint, convey official approval of or acquiescence to that violence. (247) Indeed, Gray has intimated that modifying a sentence based on the risk or realization of prison violence would render the victim ineligible for a remedy and the perpetrators of violence "immune" from prosecution. (248) And Markel has warned that taking the risk of prison violence into account in sentencing might mean that, if the predicted violence does not occur, an offender might deserve increased punishment through resentencing. (249)

Though some exaggeration is surely at work here--it is hard to understand how an exercise of sentencing discretion for the benefit of one individual could deprive a prosecutor of the authority to prosecute another individual for perpetrating a crime (250)--the arguments made by these scholars raise an important theoretical point, and it is worth examining it in some detail. For purposes of this discussion, it is necessary to distinguish between two types of sentencing accommodations that may be available as a means to respond to an offender's predicted hardship in prison: substituting a prison term with one or more noncarceral sanctions (251) and ordering a reduced prison term. (252) This Article will refer to the latter option as a sentencing "discount." To examine whether an evaluation of penalty severity should ever include foreseeable risk of serious harm, let us assume that, for the commission of a certain crime, offender A of a particular degree of blameworthiness deserves a term of imprisonment of ten years. Now assume that offender B commits the same crime with the same degree of blameworthiness. He is identical to offender A except that, for him, imprisonment predictably carries a substantial likelihood of serious sexual assault. Does responding to that risk through an adjustment in sentencing necessarily entail approval of violence? The answer is no.

First, it is possible that the risk of serious harm to offender B could be so great that incarcerating him for any length of time would be inhumane. Confining a person under conditions posing a high likelihood of serious sexual assault would be degrading, dehumanizing, and akin to torture, (253) and thus should be prohibited within a system of punishment premised on respect for the moral autonomy and dignity of the offender. (254) Therefore, consideration of the offender's vulnerability may compel the determination that imprisonment--for any period (assuming that risk of sexual assault does not markedly increase over time)--is not an available option, and a judge should order one or a combination of noncarceral penalties as a roughly equivalent alternative. (255) Part III.B.2 explores this process in more detail.

A second example of when a sentencing accommodation may be appropriate is when a judge believes that prison officials are likely to take measures to reduce an individual's vulnerability to predation, but those protective measures are likely to create harmful collateral effects. (256) For instance, in many jurisdictions, prisons protect vulnerable inmates by isolating them in protective custody. Conditions in protective custody often resemble those in disciplinary isolation, (257) with isolation for twenty-one to twenty-four hours a day. (258) Confinement under those conditions is certainly more onerous than confinement in the society of others. (259) A judge could order a reduced sentence for offender B to reflect the increased harshness of those conditions and the substantial risk of serious psychological harm those conditions pose. (260) Ordering a reduced prison term on the basis of the spartan conditions of protective custody would not constitute approval or anticipation of sexual assault, so the discount would avoid the charge of having incorporated risk of an inherently unjust penalty (the crime of sexual assault) into an offender's sentence. (261)

Other collateral consequences of vulnerability to victimization may also supply grounds for a sentencing discount, so long as the collateral consequences pose a foreseeable, substantial threat of serious harm and are susceptible to justification in light of the state's legitimate retributive goals. (262) For instance, perhaps a judge may find that offender B is likely to remain within the general prison population, where he will likely face the threat of assault but will not actually be harmed due to the successful protective efforts of prison guards. In this case, the discount could reflect the psychological harm that offender B is likely to incur by being confined in a dangerous environment. (263) Factoring this kind of anxiety into calculations of penalty severity bears some similarity to the government's policy of compensating employees for living in dangerous environments, and this example is useful for the limited purpose of demonstrating how receiving compensation for exposure to a dangerous environment (in our case, through a sentencing discount) does not deprive an inmate of any remedy if harm ultimately materializes. (264)

Currently, the U.S. Department of State increases an employee's compensation when that employee is stationed in a particularly dangerous environment--such as Iraq, Israel, or Lebanon (265)--where "civil insurrection, terrorism, or war conditions threaten physical harm or imminent danger to all U.S. Government civilian employees." (266) This "danger pay" varies by location according to the government's assessment of the physical threat posed by that environment. (267) In essence, living under a constant threat of physical danger is itself a form of harm that warrants compensation, even if the threat never materializes. The government's provision of danger pay does not serve to authorize, justify, or condone violence that may occur in these locations. If an employee is injured while on duty, he may be eligible for medical and wage loss benefits, schedule awards for permanent impairment, and vocational rehabilitation. (268) In addition, nothing would prevent injured individuals from seeking redress (or demanding that the U.S. government seek redress) against the perpetrators of violence, as permitted by the law of the relevant jurisdiction. (269) Similarly, a judge's acknowledgement of and responsiveness to harsher conditions likely to be experienced by a vulnerable offender--even if the offender's original vulnerability was to prison violence--would not serve to condone the actual occurrence of assault, so long as the discount is not based explicitly on the likelihood of the occurrence of that crime.

While discounts on the basis of harm that is capable of justification--including harm stemming from correctional efforts to reduce an individual's likelihood of predation--are morally permissible, (270) this Article agrees with Professors Sigler, Gray, Markel, and Flanders that it would be immoral for a judge to base a sentencing discount on the express likelihood of an offender's victimization. (271) In that case, the proportionality of the punishment would depend on the foreseeable occurrence of a criminal harm that can never be a moral part of a just punishment. (272) To be clear, the grant of a discount would not deprive an offender of any remedies he would otherwise have under the law (273)--indeed, the sentencing discount should alert prison officials to the offender's vulnerability to abuse, which should trigger officials' Eighth Amendment obligation to take reasonable steps to abate any substantial risk of serious harm posed to that prisoner. (274) However, that practical caveat does not take away from the important moral point that anticipated actual abuse should not factor into a sentencing discount. But, while the state should not include anticipated acts of abuse in its proportionality analysis, it can--and should--take into account the repercussions that inure to vulnerable prisoners from their foreseeable vulnerability when those collateral consequences pose a substantial risk of serious harm, so long as that foreseeable harm is justified. (275)

Even if judges were to exclude illegal acts and their collateral consequences from the definition of punishment and omit them from the sentencing calculus, judges should still consider foreseeable, substantial risks of other forms of serious harm that occur in prison. As discussed in Part II, many offenders with serious mental illnesses are unable to cope within prison or comply with the requirements of prison life and, as a result, are likely to violate prison rules and be isolated in solitary confinement, where they experience mental deterioration and exacerbation of illness. (276) At the very least, this predictable consequence of incarceration for offenders with serious mental illnesses must be considered in sentencing. (277) Arguably, most prisons are currently structured so that prisoners with serious mental illnesses are likely to fail, and, when they fail, they suffer serious and sometimes irreparable harm. Indeed, concern about how vulnerable individuals will fare within the typical and reasonably foreseeable structure of penal institutions has led some scholars to suggest that sentence mitigation may be appropriate for juveniles, (278) the elderly, (279) and persons with physical disabilities. (280) The same analysis should apply to offenders with serious mental illnesses. (281) The fact that penal institutions have a constitutional duty to provide mental health treatment to inmates with serious mental illnesses may make proactive attempts to safeguard mental health through sentencing even more compelling. (282)

B. JUST DESERT THEORY'S ACCOMMODATION OF VULNERABILITY

Assuming that punishment includes certain foreseeable risks of harm, just desert theory provides a framework for assessing how the vulnerability of offenders with serious mental illnesses should affect sentencing. (283) For a penalty to reflect properly the degree of censure warranted by an offense, three requirements must be satisfied. (284) First, criminal sanctions must take a punitive form so that deprivations are imposed in a manner that expresses censure or blame. (285) Second, the severity of a sanction should convey the degree of the censure. (286) Finally, to effectuate the second requirement, punitive sanctions should reflect "ordinal proportionality," (287) that is, they "should be arrayed according to the degree of blameworthiness (i.e., seriousness) of the conduct." (288) Blameworthiness is an amalgamation of the harm caused by an offense and the offender's culpability in effecting that harm. (289)

According to von Hirsch's analysis, there are three key aspects of ordinal proportionality. First, ordinal proportionality demands parity: individuals convicted of crimes of comparable seriousness should receive penalties of comparable severity. (290) Such offenders need not receive the same punishment, but they should receive penalties of substantially the same degree of onerousness. (291) Second, penalties should be ordered so that their relative severity reflects the degree of seriousness of their corresponding crimes. (292) Finally, penalties should be spaced so that the difference between two penalties' onerousness mirrors and calls attention to the difference between two crimes' seriousness. (293) Just desert theorists consider constraints of ordinal proportionality to be essential to justice. (294)

Ordinal proportionality, however, supplies at most relative proportionality, as it is concerned with the internal structure of a punishment scale. (295) To identify the particular penalty that should attach to a specific crime, it is necessary to determine the overall magnitude and anchoring points of the penalty system. "Cardinal magnitudes" refer to absolute severity levels that must be chosen for certain crimes in order to anchor a penalty scale. (296) For instance, one could argue--as a moral matter and without reference to sentences prescribed for other crimes--that the commission of an armed robbery of an inhabited dwelling at night warrants a term of imprisonment of five years. The cardinal magnitude of the penalty for this crime would thus be established, and appropriate punishments for other crimes could be derived based on their ordinal ranking relative to this one. Theorists are in general agreement, however, that it is impossible as a deontological matter to divine the precise quantum of punishment deserved by a specific crime. (297) This imprecision reflects the fact that the amount of censure conveyed by penal sanctions is a convention, and that conventions may vary among communities. (298)

It is, however, easier to identify unreasonable punishments, both those at the ends of a punishment scale and, to a lesser degree, for particular crimes within a scale. (299) The maximum punishment must be humane, (300) for instance, while the minimum penalty must constitute a sanction capable of conveying disapproval and blame. (301) When assessing whether a punishment scale is too severe and thus violates principles of cardinal proportionality, von Hirsch suggests that one must compare the interests affected by crimes with those disturbed by the corresponding punishments to determine whether "punished persons' vital interests are being trivialized." (302) When "drastic deprivations are used to convey merely a mild degree of censure," for instance, principles of cardinal proportionality are violated. (303) Thus, desert provides some, though far from definitive, guidance for how severe or lenient a punishment scale should be. (304)

This scheme of ordinal and cardinal proportionality supplies the necessary structure for assessing how the empirically demonstrated risk of serious harm for offenders with major mental illnesses may factor into the allocation and distribution of punishment, given the particular definition of punishment adopted in this Article for purposes of sentencing. Just desert theory suggests three primary distributional consequences for the punishment of offenders with serious mental illnesses, and the remainder of this Part will explore each in turn. First, Part III.B.1 assesses sentencing options flowing from the disqualification of inhumane penalties. Second, Part III.B.2 explores the use of a living standard analysis to identify noncarceral penalties for offenders with serious mental illnesses of equivalent punitive bite as carceral terms for standard offenders. Finally, Part III.B.3 evaluates the practice of granting discounted carceral terms for vulnerable offenders as a means of eliminating the differential effect of incarceration under the principle of equal impact. Future work will explore the theoretical and practical nuances of the sentencing options outlined below.

1. Inhumane Treatment and Cardinal Proportionality

Retributivism and desert theory are premised upon the moral dignity and personhood of the offender. (305) In the words of Professors Andrew von Hirsch and Andrew Ashworth, "The entire structure of modern desert theory is one that views offenders and potential offenders as persons whose capacity for moral judgment is to be respected, and whose rights are to be taken seriously." (306) Thus, retributivism cannot justify punishment that violates human dignity. (307) Retributive theory provides at least two reasons to prohibit degrading or dehumanizing punishments. First, in failing to recognize the personality and humanity of offenders, such punishments treat offenders as less than persons. (308) Second, brutal punishments that "approximate a system of sheer terror in which human beings are treated as animals to be intimidated and prodded" (309) are liable to transform offenders into purely reactive and survival-oriented beings incapable of exercising autonomy or understanding the disapprobation communicated through state-inflicted punishment. (310) As Professors Dan Markel and Chad Flanders have observed, to the extent that the state "breaks" an offender, or renders him incapable of understanding his punishment as punishment, the state forfeits its right to punish that individual. (311) In both respects, degrading punishments are inherently excessive and therefore disproportionate under just deserts principles. (312)

Assume, for the moment, that the elevated risk of physical and psychological harm from existing prison conditions in a relevant jurisdiction renders any term of incarceration within the general prison population (313) intolerably degrading or inhumane for any offender with a serious mental illness. (314) Also assume that the relevant prison system's means of protecting vulnerable inmates from abuse consists of confining inmates in protective custody, in isolation for twenty-three hours a day. (315) Further assume that evidence establishes that housing individuals with major mental disorders in these conditions, as a means to eliminate (or at least minimize) (316) the risk of sexual and physical violence, would pose an unjustifiably high probability of exacerbating offenders' mental illnesses and would also be inhumane. Assume the prison has no other practical means to care for these inmates. (317) As a result, incarceration of individuals with major mental disorders in this environment, for any crime, should be prohibited on moral grounds.

If prison were deemed an inhumane environment for offenders with serious mental illnesses, a legislature could respond in a number of ways. It may direct the prompt reform of the prison system or authorize a system of alternative, noncarceral penalties for these individuals. (318) One alternative, which would involve the straightforward application of principles of ordinal and cardinal proportionality, could consist of authorizing a nontraditional "guilty but mentally ill" verdict, whereby a convicted offender would be confined in a segregated facility for offenders with serious mental illnesses. (319) If a legislature created such a segregating mechanism, and it succeeded in eliminating the inhumane portion of a carceral term, then the penalty scale for offenders with serious mental illnesses would resemble that for nonvulnerable offenders, under principles of ordinal and cardinal proportionality. In particular, both sets of offenders could be confined for the same amounts of time for the same offenses, just in different facilities.

For a more complicated example, assume that confining offenders with major mental disorders in the segregated facility just described for more than forty years poses an impermissibly high likelihood of substantial exacerbation of mental illness. In this situation, carceral sentences of longer than forty years should be prohibited, and the entire penalty scale for these offenders should shift downward relative to the scale for the general population to preserve punishments' ordinal rankings and principles of ordinal proportionality. (321) To illustrate, assume that the jurisdiction in question has identified life in prison as the most severe penalty that is morally permissible (321) for any individual and sets a small fine as the lowest penalty on its punishment scale. At this point, the penalty scales for offenders with and without major mental disorder will be "anchored" by their most severe penalties (which will differ for each population) and least severe penalties (which will be the same). (322) Following the methodology outlined by von Hirsch, penalties for offenses along the criminal spectrum will be assigned based on considerations of ordinal proportionality, meaning that crimes and penalties should be ranked and ordered according to seriousness, and the spacing of penalties should reflect the differential gravity of offenses. (323) Because the most severe penalty (the upper anchoring point) is milder for offenders with serious mental illnesses, (324) the entire spectrum of penalties for this population will shift downward to preserve ordinal proportionality. As a result, the cardinal or absolute magnitudes of penalties along the entire penalty spectrum will be lower for offenders with serious mental illnesses than for other offenders. In other words, for any crime except the most trivial one, the corresponding penalty imposed on an offender with major mental disorder will be less severe than the penalty imposed on a less ill or non-ill offender. (325)

2. Alternative Penalties and the Living Standard Analysis

Beyond helping to discern and adjust penalty scales in light of inhumane punishments and violations of cardinal proportionality, just desert theory supplies guidance for how to identify penalties of equivalent punitive bite or onerousness. If, in our example above, the jurisdiction failed to provide a safer housing arrangement in which to confine inmates with major mental disorders, then it would need to identify alternative penalties to substitute for terms of imprisonment for various crimes. (326) This would involve the construction of a penalty scale with combinations of alternative penalties of roughly equivalent severity to the carceral terms available for non-ill offenders. Noncarceral penalties could include intermittent confinement at a state-designated facility, home detention with electronic supervision, a community service order, a treatment or residential order, a fine, or probation. (327)

To compare the onerousness of various penalties, Professors Andrew von Hirsch and Andrew Ashworth have proposed utilizing a "living standard analysis," which von Hirsch and Professor Nils Jareborg developed in the context of evaluating the severity of criminal offenses. (328) A living standard analysis focuses on "the means or capabilities for achieving a certain quality of life" (329) and compares the severity of various penalties by their degree of intrusion into offenders' interests. (330) In essence, penalties that typically impair interests more critical to quality of life (such as physical integrity) will be considered more severe than those that affect less significant interests (such as privacy). (331) Similarly, a penalty that typically threatens individuals' very subsistence will be understood as more severe than one that infringes upon their well-being to a lesser extent. (332) Thus, penalties' severities can be compared objectively by the extent to which the sanctions interfere with offenders' interests or with resources to which offenders have legitimate claims. (333) Penalties may affect multiple dimensions of human welfare, including physical integrity, freedom of movement, choice of activity and associates, material support and amenities, freedom from degrading treatment, future earning power, privacy, and autonomy. (334) In using this scale to compare penalties' severity, it is necessary both to identify the set of offender interests typically affected by a penalty and to classify the extent to which a penalty's impairment of those interests affects an offender's quality of life. (335)

Critically, a living standard analysis judges the severity of penalties from the standpoint of their anticipated, likely effects on offenders, and does not restrict its gaze to effects intended by a sentencing judge or legislature. Von Hirsch argues that penalties should be ranked "according to the degree to which they typically affect the punished person's freedom of movement, earning ability, and so forth." (336) "The importance of those interests," he continues, should "be gauged according to how they typically impinge on a person's 'living standard'...." (337) This focus on the actual "effects of the penalty on the quality of persons' lives" reflects the origin of the living standard analysis, (338) which was designed to measure the harms effectuated by criminal offenses. (339) The living standard analysis, in that context, measured harm by the way an offense typically set back a victim's interests and impacted his means of achieving a certain quality of life, (340) as assessed through the perspective of the typical victim. (341) It was normally of no consequence in an evaluation of criminal harm whether a perpetrator intended to effectuate a particular harm when carrying out an offense--harm and culpability are distinct components of the seriousness of crime. (342) In transplanting the living standard analysis into the context of gauging penalty severity, it should likewise be of no consequence whether a sentencing authority intended that an offender suffer a particular harm, or whether the harm was merely a foreseeable side effect of a penalty, proximately caused by the state. As von Hirsch notes, one benefit of evaluating penalties in this way is that the analysis accords with community sentiment about punishments' severity. (343)

Von Hirsch and others have argued that a living standard analysis should reflect "the means and capabilities that ordinarily assist persons in achieving a good life" and should not vary according to individuals' preferences or subjective perceptions of pain. (344) They articulate two main reasons for confining the living standard analysis to the "typical" offender: first, individualizing the standard would create unmanageable diversity; (345) and, second, "when one is talking about atypical harms, foreseeability diminishes." (346) Although scholars to date have focused largely on the average offender, when empirical evidence establishes that a given penalty will affect the interests of an identifiable, (347) vulnerable offender class more substantially than average offenders, strong arguments exist for conducting a separate analysis for individuals in that class. (348) Indeed, just desert theorists have indicated a willingness to perform living standard analyses in "nonstandard cases" by assessing "typical impacts for the members of that group." (349)

If certain penalties are prohibited for offenders with serious mental illnesses because those penalties are too likely to cause serious harm, a living standard analysis would provide a means of identifying sets of alternative penalties of roughly equivalent punitive bite, which could serve as plausible substitutes for this population. (350) Take, for example, the felony of possessing marijuana for other than personal use. Assume, under state law, the minimum penalty for this crime is confinement for one year and one day in prison. (351) Assuming that imprisonment were an inhumane penalty for individuals with serious mental illnesses, it would be necessary to authorize an alternative penalty or set of penalties for disordered offenders so that they could receive a sanction of roughly equivalent severity to that imposed on their non-disordered peers. (352) To identify penalties of roughly equivalent punitive bite for the two offender populations, one would first assess the degree to which the sanction of a year and a day in prison impairs non-disordered offenders' interests and affects their quality of life. Then one would work to identify alternative penalties that would impair seriously disordered offenders' interests and affect their quality of life roughly to the same extent. Penalties with roughly equivalent punitive bite for offenders with serious mental illnesses might include, for instance, two years of home detention with electronic monitoring, 4,000 hours of community service, or six years of probation. (353)

A judge could order mental health treatment to accompany some or all of these sanctions.

3. The Equal-Impact Principle and Carceral Discounts

While a living standard analysis provides a way to identify a slate of noncarceral sanctions for most crimes, some scholars have opined that, to convey the degree of censure warranted for the most serious crimes, it is necessary to impose the sanction of incarceration. (354) If that is the case--and in recognition of the reality that most jurisdictions authorize incarceration for many, if not most, criminal offenses--it is necessary to address whether proportionate punishment calls for reduced terms of incarceration for offenders with serious mental illnesses in light of the foreseeable risks of serious harms proximately caused by the state in the context of confinement. The following discussion assumes that conditions of incarceration in a given jurisdiction, though predictably harsher for an offender with a serious mental illness, do not reach the threshold of inhumanity. (355)

When incarceration would be significantly more onerous for a member of a vulnerable population, Professors von Hirsch and Andrew Ashworth have argued that the individual should receive a discounted term of confinement as a means to avoid disproportionate penal severity. The principle of equal impact, which resides in the "borderlands" of desert theory, (356) holds that, "when an offender suffers from certain handicaps that would make his punishment significantly more onerous, the sanction should be adjusted in order to avoid its having an undue differential impact on him." (357) The principle of equal impact derives from ideals of fairness and equality and is closely related to a principle of nondiscrimination. (358) At its essence, the equal-impact principle calls for members of certain vulnerable classes to receive penalties of roughly equivalent severity as nonvulnerable individuals, again identified through use of a living standard analysis. (359) It is important to emphasize that the equal-impact principle does not call for a reduction in punishment, but rather for equalizing the severity of penalties imposed on equally blameworthy offenders. (360) While scholars originally conceptualized the equal-impact principle as standing outside the bounds of proportionality, (361) desert theorists have argued that, at least in some cases, recognition of the equal-impact principle is necessary for the achievement of proportionality. (362)

As argued in the previous section, a living standard analysis is best understood as calling for a consideration of harms beyond those intended by a sentencing judge or legislature, (363) but it is even more clear that an equal-impact inquiry necessarily considers foreseeable harm. (364) Presumably, a sentencing judge never intends that an offender with a serious mental illness or a physical disability will suffer more than non-ill or non-disabled offenders when confined. Yet the very aim of the equalimpact principle is to acknowledge the foreseeable, typical, and serious side effects that certain penalties hold for these and other vulnerable populations and to adjust these penalties so their overall effect will not be disproportionately severe. (365) Restricting contemplated harms to those intended by a sentencing authority would obviate the very problem the equal-impact principle was created to remedy.

While the equal-impact principle may also apply to onerous sanctions such as community service or intensive probation, scholars typically raise the principle within the context of incarceration. (366) The below example illustrates its application in this context:

   Suppose the standard sentence for a given species of offence is
   three years' imprisonment. If this sanction is applied to a
   defendant in a wheelchair, he actually--under an interest
   analysis--has his interests set back to a greater degree. Reducing
   the sentence under an equal-impact theory would be designed merely
   to eliminate this increment in severity. (367)


Concerned about the foreseeable impairment of interests critical to offenders' quality of life, scholars have suggested--though provided no fulsome analysis to explicate the suggestion--that discounted terms of incarceration would be appropriate to effectuate proportionate punishment for mentally ill individuals, (368) the physically disabled, (369) and the elderly. (370) Scholars have also argued for categorical youth discounts to ensure proportionate punishment for juveniles. These discounts would reflect the fact that incarceration infringes upon important developmental interests of juveniles and is more likely to impair juveniles' self-esteem. (371)

Application of the equal-impact principle in the context of mental disorder is appropriate because incarceration affects the interests of offenders with serious mental illnesses and non-ill offenders differently. (372) For example, incarceration typically deprives a standard prisoner of privacy and autonomy by restricting the individual's ability to choose his activities and associates. When the offender has a major mental illness, however, he may suffer a much more extreme loss of autonomy due to the deterioration of his mental integrity. (373) The ultimate result of a carceral term (particularly of a long term) may be highly degrading, corroding his rationality and perhaps threatening his capacity for autonomous thought. As compared to non-ill offenders, the likely psychological toll of incarceration on offenders with major mental illnesses impedes a broader, more substantial set of interests and affects these interests to a more significant degree. In addition, offenders with serious mental illnesses have an interest in receiving mental health treatment to retain or recover their mental functioning and autonomy. (374) As juveniles have certain developmental interests that make imprisonment uniquely hard for them, (375) offenders with serious mental illnesses have health-related interests that are also negatively impacted by incarceration. These interests include receiving regular and adequate mental health treatment in a nurturing environment in which they can develop a relationship of trust with a mental health provider in order to function at an adequate level. (376) Mental health care in prison is often inadequate, and the environment is far from therapeutic. (377) Under the equal-impact principle, if incarceration is more onerous when undergone by offenders with serious mental illnesses, principles of cardinal proportionality require that judges reject standard terms of incarceration for this population. (378) To avoid disproportionality, judges should select a noncarceral sanction of roughly equivalent punitive bite (379) or potentially order a discounted term of incarceration. A living standard analysis would be useful in identifying the proportionate penalty in both instances.

Consistent with the conception of punishment adopted in Part III.A of this Article, if a term of imprisonment poses a foreseeable, substantial risk of serious harm to an offender with serious mental illness (380)--and noncarceral sanctions of equivalent punitive bite are not available or appropriate (381)--a sentencing judge should consider adjusting the term of imprisonment through a sentencing discount to avoid its having an undue differential impact on the individual. (382) Two caveats to the application of this principle are necessary. First, to the extent that a sentencing discount derives from concerns about an offender's vulnerability to physical or sexual abuse, (383) the discount should not rest on an anticipation of victimization, for reasons explained earlier. (384) Instead, the discount should reflect the increased harshness or severity of the penalty in light of the anticipated measures that prison officials will take to protect the offender (such as housing the offender in isolation or in conditions with fewer programming opportunities) or perhaps the anxiety the offender will likely experience while living in a dangerous environment, (385) if these collateral effects rise to the level of serious harm and can be justified. (386)

Second, judges should not reduce terms of imprisonment on the basis of foreseeable, nonserious harms. While it is clear that a living standard analysis is only concerned with diminution of objective interests (and not subjective distress), how serious the harms must be to warrant adjustment under the equal-impact principle is unclear. (387) The principle, as expressed, intends to provide adjustment for "handicaps that would make [an offender's] punishment significantly more onerous." (388) If a penalty causes serious harm when applied to a vulnerable population (and not when applied to nonvulnerable individuals), it is certainly "significantly more onerous" for the vulnerable population. Moreover, just desert scholars have emphasized a number of interests that, if impaired to a significant degree, would constitute serious harm consistent with the position taken in this Article. (389) However, desert scholars' discussion of the relevant experience of persons with mental disorders and how to tailor sentencing in light of that experience is sparse and vague, (390) and some discussion suggests that sentencing discounts could be granted on the basis of foreseeable risks of nonserious harms. (391) For instance, when illustrating the hardship faced by a physically disabled offender, Professors von Hirsch and Ashworth highlight the offender's "physical handicap that impedes his movement sufficiently to make the routines of being imprisoned more onerous." (392) They conclude:

   In such eases the "equal impact" principle ... might still apply.
   This would involve adjusting the penalty so as to take into account
   its more onerous character in the particular circumstances. The
   duration of a prison sentence might thus be reduced to some degree.
   The aim would [be to make]... adjustments in sentence[s] to deal
   with certain foreseeable differential impacts.... (393)


To the extent that equal-impact proponents intend judges to grant sentencing discounts on the basis of foreseen risks of nonserious harms, such an application would conflict with the position taken in this Article, which advocates for sentencing accommodation only for substantial risks of serious harm.

Other theories, beyond those of proportionality, may support responding to the foreseeable, nonserious suffering of vulnerable offenders through a sentencing discount. One such theoretical candidate would be mercy. The proper understanding of mercy and its relationship to justice are subjects of lively debate, (394) but some equal-impact scholars may be alluding to mercy when observing that the principle of equal impact "calls for reference to factors beyond strict proportionality to the gravity of the offence" and rests upon "a more flexible conception of justice which does not presume uniformity among offenders." (395) Another theoretical candidate may include equity mitigation. (396) Exploring the potential applicability of these theories falls beyond the scope of this Article.

IV. CONCLUDING THOUGHTS

Just desert theory suggests that judges should consider certain offenders' vulnerability when meting out punishment. This is an issue of tremendous practical significance: empirical evidence and data collected by human rights groups and others demonstrate that offenders with serious mental illnesses are disproportionately likely to suffer a range of harms in prison. These harms include physical assault, rape and sexual assault, isolation in solitary confinement, mental deterioration and severe psychological distress, and exacerbation and onset of new mental disorder. When imprisonment poses foreseeable, substantial risks of serious harm to an offender with a serious mental illness, a judge's sentencing calculus should include these risks to ensure that the individual receives a proportionate and humane sentence.

By design, the arguments offered in this Article may raise more questions than they answer. Future work will be devoted to probing the theoretical implications of including foreseeable harm within the definition of punishment for purposes of sentencing and assessing the practical implications of the sentencing options outlined here. For the moment, it is appropriate to recognize that factoring vulnerability into sentencing could have important and far-reaching effects. Broadening the definition of punishment to include foreseeable harm should help breach the divide between punishment in theory and practice and allow the philosophy of punishment to expand its relevance and practical effect. (397) Reforming sentencing to reflect offenders' anticipated experiences should yield more humane sentences, increase the use of alternative sanctions, and prompt the reform of prison conditions. Moreover, by taking susceptibility to harm into account, we will end the morally indefensible practice of overpunishing offenders with major mental disorders and will give these offenders, deemed competent and culpable by the criminal justice system, their just deserts.

(1) The "typical" offender does not have a serious mental illness, though individuals with mental illnesses are disproportionately represented in the criminal justice system. A 2009 study of more than 20,000 adults entering five local jails found that 14.5% of male and 31.0% of female inmates had a serious mental illness. Henry J. Steadman et al., Prevalence of Serious Mental Illness Among Jail Inmates, 60 PSYCHIATRIC SERVS. 761, 764 (2009). These rates are three to six times higher than those found in the general population. The Numbers Count: Mental Disorders in America, NAT'L INST. OF MENTAL HEALTH, http://www.nimh.nih.gov/health/publications/the-numbers-count-mental-disorders-in-america/index.shtml (last visited Nov. 11, 2012) (citing Ronald C. Kessler et al., The Epidemiology of Co-Occurring Addictive and Mental Disorders." Implications for Prevention and Service Utilization, 66 AM. J. ORTHOPSYCHIATRY 17 (1996), for the proposition that approximately 6% of Americans suffer from a serious mental illness). Since male prisoners constitute 93% of the prisoner population in the United States, this Article will focus on male prisoners. See HEATHER C. WEST ET AL., BUREAU OF JUSTICE STATISTICS, U.S. DEP'T OF JUSTICE, PRISONERS IN 2009 (2011), available at http://bjs.ojp.usdoj.gov/content/pub/ pdf/p09.pdf.

(2) See, e.g., LA. REV. STAT. ANN. [section] 14:62 (2007) (providing a maximum penalty for simple robbery of twelve years). The mean maximum sentence imposed on an offender without a mental illness in a state prison is 141 months, or three months shy of twelve years. DORIS J. JAMES & LAUREN E. GLAZE, BUREAU OF JUSTICE STATISTICS, U.S. DEP'T OF JUSTICE, MENTAL HEALTH PROBLEMS OF PRISON AND JAIL INMATES 8 (2006), available at http://bjs.ojp.usdoj.gov/content/pub/pdf/mhppji.pdf.

(3) See RICHARD S. JONES & THOMAS J. SCHMID, DOING TIME: PRISON EXPERIENCE AND IDENTITY AMONG FIRST-TIME INMATES 135 (2000) (describing a newcomer's challenge of "understanding the prison hierarchy and recognizing [his] place in it, learning whom to trust and whom to avoid, [and] determining how to evade trouble in a trouble-filled environment").

(4) See HANS TOCH, MEN IN CRISIS 146 (2007) (explaining that non-ill offenders adapt to prison by behaving in a stoic manner and observing behavior of other inmates).

(5) See John Bronsteen, Christopher Buccafusco & Jonathan Masur, Happiness and Punishment, 76 U. CHI. L. REV. 1037, 1046-49 (2009) (summarizing literature demonstrating prisoners' adaption to incarceration).

(6) This paragraph and the next offer fictitious descriptions of the experiences of two hypothetical prisoners. They are intended simply to highlight and dramatize how severely limited cognitive abilities may make one vulnerable to a predictable set of hardships in prison, thus giving prison a greater punitive bite. For in-depth portrayals of the experience of prisoners without serious mental illness, see WILBERT RIDEAU, IN THE PLACE OF JUSTICE: A STORY OF PUNISHMENT AND DELIVERANCE (2010), ANDREAS SCHROEDER, SHAKING IT ROUGH: A PRISON MEMOIR (1976), and GRESHAM M. SYKES, THE SOCIETY OF CAPTIVES: A STUDY OF A MAXIMUM SECURITY PRISON (1958). For a depiction and discussion of prisoners' adaptation to prison and coping mechanisms, see ANN CORDILIA, THE MAKING OF AN INMATE: PRISON AS A WAY OF LIFE (1983) and JONES & SCHMID, supra note 3, at 57-62 (describing inmates' survival strategies of territorial caution, impression management, selective interaction with other inmates, and partnership).

(7) Axis I disorders, as defined by the American Psychiatric Association, include clinical syndromes such as schizophrenia, bipolar disorder, and depression, as well as chronic brain diseases that cause extreme distress and interfere with social and emotional adjustment. See AM. PSYCHIATRIC ASS'N, DIAGNOSTIC AND STATISTICAL MANUAL OF MENTAL DISORDERS 13-24, 28 (4th ed. rev. 2000) [hereinafter DSM-IV-TR]. In this paper, "serious mental illness," "major mental illness," "major mental disorder," and "Axis I disorders" are used interchangeably.

(8) This Article takes as its subject prisoners who are found or assumed to be competent to stand trial and be sentenced. Cf. Dan Markel & Chad Flanders, Bentham on Stilts. The Bare Relevance of Subjectivity to Retributive Justice, 98 CALIF. L. REV. 907, 952-54 (2010) (arguing that if a person is not a fit interlocutor for the state's retributive message, then he does not warrant retributive blame and should not be punished).

(9) See TERRY A. KUPERS, PRISON MADNESS: THE MENTAL HEALTH CRISIS BEHIND BARS AND WHAT WE MUST DO ABOUT IT 20 (1999).

(10) Studies show that individuals with major mental disorders are disproportionately vulnerable to physical and sexual abuse in prison. See infra notes 64-74 (physical assault); notes 91-96 (sexual assault).

(11) The above narrative draws upon personal accounts published in Professor Terry Kupers's Prison Madness, information from the American Psychiatric Association, and other works in order to portray the experiences of inmates with schizophrenia and other serious mental illnesses in correctional facilities. See Jones 'El v. Berge, 164 F. Supp. 2d 1096, 1098-103 (W.D. Wis. 2001); DSM-IV-TR [section][section] 295.1-295.3, 295.90; John J. Gibbs, Disruption and Distress: Going from the Street to Jail, in COPING WITH IMPRISONMENT 29 (Nicolette Parisi ed., 1982); KUPERS, supra note 9, at 9-65; TOCH, supra note 4, at 144-55, 213-14.

(12) See, e.g., Lisa A. Goodman et al., Recent Victimization in Women and Men with Severe Mental Illness: Prevalence and Correlates, 14 J. TRAUMATIC STRESS 615, 627 (2001) (finding that women with severe mental illness were sixteen times more likely than those in the community sample to report violent victimization in the past year, and men with severe mental illness were ten times more likely than those in the community sample to report an assault); Linda A. Teplin et al., Crime Victimization in Adults with Severe Mental Illness: Comparison with the National Crime Victimization Survey, 62 ARCHIVES GEN. PSYCHIATRY 911, 917 (2005) (finding that more than one quarter of persons with severe mental illness had been victims of a violent crime in the past year, a rate more than eleven times higher than the general population rates).

(13) See infra notes 64-74 (physical assault), 91-96 (sexual assault).

(14) See infra notes 114-134.

(15) See infra notes 146-149.

(16) See infra notes 151-156.

(17) This Article takes the position that vulnerability should factor into sentencing and is not simply an issue appropriate for penal administration. See infra notes 236-239 and accompanying text.

(18) Prior to consideration of vulnerability at the sentencing phase, a defendant's mental illness may reduce his culpability at the guilt phase of his proceeding if he successfully mounts an insanity defense or presents diminished capacity evidence demonstrating that he lacked the necessary intent for a crime. Others have argued for expanding or reducing these defensive strategies. Compare Christopher Slobogin, An End to Insanity: Recasting the Role of Mental Disability in Criminal Cases, 86 VA. L. REV. 1199, 1246 (2000), with Laura Reider, Toward a New Test for the Insanity Defense: Incorporating the Discoveries of Neuroscience into Moral and Legal Theories, 46 UCLA L. REV. 289 (1998). This Article takes no position on these issues. Instead, this Article avers that sentence mitigation may be necessary for offenders with serious mental illnesses to receive proportionate punishments or punishments of appropriate punitive bite, even if mental illness has already factored into a culpability determination.

(19) Because an individual's capacity to understand reality may implicate his culpability for past acts and his ability to participate in the adversarial process, a number of steps are built into the criminal justice process to allow for consideration of a defendant's mental illness. These include opportunities to challenge an accused's competence to stand trial and his decisional competence to make the few choices allocated to him, such as whether to waive his right to counsel or to plead guilty. If the accused pleads not guilty, he may use evidence of mental illness to advance an insanity defense or demonstrate a lack of intent. See supra note 18. Typically, pretrial services will chronicle an accused's mental health history in the report it prepares for bail determination, and a probation officer will include a defendant's mental health history in the presentencing report created for the court. See infra notes 49-51 and accompanying text.

(20) Judges' familiarity with data regarding the vulnerability of prisoners with serious mental illness may vary with their sentencing options. Presumably, if a trial judge has the discretion to modify an individual's sentence on the basis of his perceived vulnerability, then the judge will, over time, learn about the sources and extent of vulnerability associated with various sanctions. Defense counsel will play a role in the judge's education. See infra notes 49-51 and accompanying text.

(21) See D.C. SENTENCING & CRIMINAL CODE REVISION COMM'N, VOLUNTARY SENTENCING GUIDELINES MANUAL [section] 5.2.3(8) (2011) (allowing a judge to sentence outside the guidelines where "the court determines that the defendant, by reason of obvious and substantial mental or physical impairment or infirmity, cannot be adequately protected or treated in any available prison facility"); see also 9 MINN. PRAC., CRIMINAL LAW & PROCEDURE [section] 36.30(k) (3d ed. 2001) (recognizing vulnerability as a mitigating factor though case law). In addition, several other jurisdictions include "excessive hardship" to the offender as a factor bearing on the appropriateness of imprisonment as a sanction. See, e.g., HAW. REV. STAT. ANN. [section] 706-621(2)(i) (LexisNexis 2007); IND. CODE ANN. [section] 35-38-1-7.1(b)(10) (West 2012); N.D. CENT. CODE [section] 12.1-32-04(11) (2012); N.J. STAT. ANN. [section] 2C:44-1(b)(1 l) (West 2005); UTAH SENTENCING COMM'N, 2011 ADULT SENTENCING AND RELEASE GUIDELINES 13, available at http://www.sentencing.utah.gov/sentencing_ archives.html (under "Adult Sentencing Guidelines" menu, select "Adult Guidelines 2011"). Under the U.S. Sentencing Guidelines, mental and emotional conditions are ordinarily irrelevant in determining whether a sentence should fall outside the sentencing range established by the Guidelines for a criminal offense. See U.S. SENTENCING GUIDELINES MANUAL [section] 5H1.3 (2010). However, under Guideline [section] 5H1.3, an offender's vulnerability due to mental or emotional conditions may justify a downward departure so long as such conditions "are present to an unusual degree and distinguish the case from the typical cases covered by the guidelines." Id. Some courts have also relied on Guidelines [section][section] 5K2.0, 5H1.4, and 5K2.13 to grant downward departures on the basis of suspected or demonstrated hardship in prison. See infra notes 172-175.

(22) See, e.g., KAN. STAT. ANN. [section] 21-6815(c)(1) (2007 & Supp. 2011) (providing a nonexclusive list of mitigating factors that may be considered in determining whether substantial and compelling reasons exist for a departure from a presumptive sentence); N.C. GEN. STAT. [section] 15A-1340.16(e)(21) (2011) (providing, as a factor supporting a mitigated sentence, "[a]ny other mitigating factor reasonably related to the purposes of sentences"). Indeed, some state judges have altered offenders' punishments in response to the perception that the defendants' serious mental illnesses could lead to intolerable suffering in prison. See People v. Jackson, Nos. 282708, 284430, 2009 WL 1361956, at *3 (Mich. Ct. App. May 14, 2009) (approving a trial court's grant of a downward departure due to "substantial and compelling reasons," one of which was defendant's history of suicide ideation); People v. Zung, 531 N.Y.S.2d 615, 615-16 (App. Div. 1988) (substituting, "as a matter of discretion in the interest of justice," a period of community service for incarceration because "uncontroverted medical documents indicate that a period of incarceration would be severely detrimental to this defendant's mental health, and could possibly exacerbate his suicidal tendencies").

(23) See, e.g., ALASKA STAT. [section] 12.55.155(d)(18) (2012) (allowing for imposition of a sentence below the presumptive range when "the defendant committed the offense while suffering from a mental disease or defect ... that was insufficient to constitute a complete defense but that significantly affected the defendant's conduct"); CAL. RULES OF COURT [section] 4.423(b)(2) (2012) (characterizing as a factor in mitigation that "[t]he defendant was suffering from a mental or physical condition that significantly reduced culpability for the crime"). The U.S. Sentencing Guidelines contain a similar provision. See U.S. SENTENCING GUIDELINES MANUAL [section] 5 K2.13 (2011) ("A downward departure may be warranted if (1) the defendant committed the offense while suffering from a significantly reduced mental capacity; and (2) the significantly reduced mental capacity contributed substantially to the commission of the offense.").

(24) See Richard C. Boldt, Rehabilitative Punishment and the Drug Treatment Court Movement, 76 WASH. U. L. Q. 1205, 1219-23 (1998).

(25) See, e.g., Russell L. Christopher, Deterring Retributivism: The Injustice of 'Just' Punishment, 96 NW. U. L. REV. 843, 844-45 (2002) ("Retributivism is all the rage. Whether it is a 'revival,' a 'resurgence,' or a 'renaissance,' retributivism's rapid 'rise' since the early 1970s has been remarkable."); R.A. Duff, In Defence of One Type of Retributivism. A Reply to Bagaric and Amarasekara, 24 MELB. U. L. REV. 411, 411 (2000) ("A striking feature of penal philosophising during the last thirty years has been the revival of retributivism."); Jean Hampton, Correcting Harms Versus Righting Wrongs: The Goal of Retribution, 39 UCLA L. REV. 1659, 1659 (1992) ("There has been a steady rise in the popularity of retributivism over the last decade, which is surprising given its near death in the 1950's and 1960's.").

(26) See Richard S. Frase, Punishment Purposes, 58 STAN. L. REV. 67, 76 n.22 (2005) (claiming that nearly every jurisdiction in the United States has promulgated sentencing codes consistent with Norval Morris's limiting retributivism); Paul H. Robinson, Competing Conceptions of Modern Desert. Vengeful, Deontological, and Empirical, 67 CAMBRIDGE L.J. 145, 145-46 (2008), available at http://ssrn.com/abstract=924917 ("In the US, a number of sentencing guidelines have adopted desert as their distributive principle, and it is increasingly given deference in the 'purposes' section of state criminal codes, where it can be the guiding principle in the interpretation and application of the code's provisions."). Contra Michael Tonry, Looking Back to See the Future of Punishment in America, 74 SOC. RES. 353, 363 (2007) ("In this first decade of the twenty-first century, there is neither a prevailing punishment paradigm in practice nor a prevailing normative framework for assessing or talking about punishment in principle.").

(27) See ANDREW VON HIRSCH & ANDREW ASHWORTH, PROPORTIONATE SENTENCING: EXPLORING THE PRINCIPLES 4 (2005) (explaining that the desert rationale of proportional sentencing underlying retributivism "rests on the idea that the penal sanction should fairly reflect the degree of reprehensibleness (that is, the harmfulness and culpability) of the actor's conduct"); Paul Butler, Retribution, For Liberals, 46 UCLA L. REV. 1873, 1884 (1999) ("Retribution measures just punishment by considering whether there is proportionality between crime and punishment."); see also infra Part III.B. While many retributivists believe that punishment is just so long as it is proportionate to the moral culpability of the offender and the wrong he committed, there are many variations of retributivism. See infra note 180.

(28) See infra note 199.

(29) See infra note 198.

(30) Exactly who is the punisher--the communicator of society's censure--for purposes of retributive punishment is a difficult and complicated question. See, e.g., Alice Ristroph, State Intentions and the Law of Punishment, 98 J. CRIM. L. & CRIMINOLOGY 1353 (2008) (arguing that punishment involves a wide array of actors and institutions with varying intent and that judgments about intent are always contestable, and therefore concluding that state actors' intent should be relevant to, but not dispositive of, whether a system is imposing punishment). Clearly, the sentencing judge plays an important role in communicating to an offender, through sentencing, society's stern disapproval of a criminal act. See R.A. DUFF, TRIALS AND PUNISHMENTS 235-36 (1986). Even in sentencing, however, the judge's communication is not unfettered: the content of his speech is usually restricted by statutory constraints set by a legislature and perhaps guidelines established by a sentencing commission. See Markel & Flanders, supra note 8, at 954-57 (arguing that legislatures must authorize sentencing options to ensure adequate condemnatory treatment). Additionally, the degree to which society's condemnatory message is communicated through the execution of a sentence, by prison guards and other correctional personnel, may play an important communicative role. Cf. DAVID GARLAND, PUNISHMENT AND MODERN SOCIETY: A STUDY IN SOCIAL THEORY 180-89 (1990) (exploring the effect of the rationalization and bureaucratization of the penal process on the social meaning of punishment). When multiple actors participate in the communication of a message, multiple intentions are often present, and the ultimate meaning of the message becomes muddled. See, e.g., William DeFord, Comment, The Dilemma of Expressive Punishment, 76 U. COLO. L. REV. 843, 857-60 (2005) (identifying the problem, for expressive punishment, of multiple institutional speakers and exploring the implications of this involvement for confusing the intent behind, and meaning of, the message of punishment). Exploring the ramifications for expressive punishment of the myriad institutional actors involved in the criminal justice system is beyond the scope of this Article.

(31) See infra notes 199-200 and accompanying text.

(32) David Gray, Punishment as Suffering, 63 VAND. L. REV. 1619, 1670 (2010); see also infra note 200.

(33) See infra note 200 and accompanying text.

(34) See Bronsteen et al., supra note 5, at 1068-80 (exploring the implications of hedonic adaption to retributive and mixed theories of punishment); John Bronsteen, Christopher Buccafusco & Jonathan S. Masur, Retribution and the Experience of Punishment, 98 CALIF. L. REV. 1463, 1480-81, 1482-95 (2011) (expanding upon their original argument and defending the assertion that post-prison outcomes are a component of retributive punishment). Although Professors Bronsteen, Buccafusco, and Masur support Kolber's position on the importance of subjective variance in punishment experience, see, e.g., Bronsteen et al., supra note 5, at 1039, they assert that their views on the importance of hedonic adaption to retributivism do not depend upon agreement with Kolber, see Bronsteen et al., supra, at 1464, and their focus is on the hedonic adaptation of the "typical" offender, id. at 1469.

(35) See Adam J. Kolber, The Subjective Experience of Punishment, 109 COLUM. L. REV. 182, 185-86 (2009) [hereinafter Subjective Experience] (arguing that retributivism, by virtue of its commitment to proportionality, must consider the variance in offenders' subjective experiences of punishment and therefore that sentencing decisions should reflect these variances); see also Adam J. Kolber, The Comparative Nature of Punishment, 89 B.U.L. REV. 1565, 1566-67 (2009) [hereinafter Comparative Nature] (arguing that the severity of punishment should be measured by deviance from subjects' baseline states).

(36) See Adam J. Kolber, Unintentional Punishment, 18 LEGAL THEORY 1, 15-16 (2012) (asserting that the "justification-symmetry principle" requires that state actors provide justification for any harm or risk of harm associated with punishment that would require justification if posed by an individual). Indeed, prior sociological accounts of judges' actual sentencing practices reveal their inclination to take into account the effect of the sanction on the offender. See STANTON WHEELER, KENNETH MANN & AUSTIN SARAT, SITTING IN JUDGMENT: THE SENTENCING OF WHITE-COLLAR CRIMINALS 22, 144-52 (1988) (exploring, through interviews, the white-collar criminal sentencing practices of federal judges in the mid-1980s, before the adoption of the U.S. Sentencing Guidelines, and finding concern for the effect of the sanction on the offender as a common consideration in sentencing).

(37) See, e.g., Kolber, Subjective Experience, supra note 35, at 184-87; Bronsteen et al., supra note 5, at 1050-55 (identifying a broad range of negative experiences as relevant to punishment severity); cf Markel & Flanders, supra note 8, at 908, 973-84 (arguing that finely calibrating offenders' punishment by their "ex post idiosyncratic tastes, capacities, and experiences" would threaten the core aims of retributivism).

(38) See Kolber, Subjective Experience, supra note 35, at 212-13.

(39) See id. at 187 n.5, 200.

(40) See Gray, supra note 32, at 1623 (observing that "these scholars often treat all

suffering as fungible"); Kenneth W. Simons, Retributivists Need Not and Should Not Endorse the Subjectivist Account of Punishment, 109 COLUM. L. REV. SIDEBAR 1, 7 (2009) (pointing out ambiguities as to which mental states should be treated as disvaluable). But see infra note 41.

(41) Professor Kolber has recently written an innovative piece discussing how technological advances in neuroscience are improving our ability to measure states of pain and distress. See Adam J. Kolber, The ExperientialFuture of Law, 60 EMORY L.J. 585, 635-40 (2011). He argues that these technologies should be used to calibrate penalty severity or monitor penalties' effects over time in order to better effectuate proportionate punishment in sentencing. Id. It is possible that such technologies could provide an objective means to discern between substantial and insubstantial forms of disutility for retributive sentencing.

(42) The focus of this Article is on the experience of prisoners with Axis I disorders. See DSM-IV-TR, supra note 7 (defining Axis I disorders). The studies detailed in this Article, however, did not employ uniform mental illness criteria and may be more or less inclusive in the diagnoses they accepted. For reasons explored later, see infra notes 236-239 and accompanying text, this paper is written as if prison were a unitary experience, which it is certainly not. This Article assumes that offenders with mental illnesses are housed with the general prison population because much of the research on mentally ill prisoners' experience is conducted in this setting. Many prisons lack specialized housing for inmates with stable mental illnesses. See ALLEN J. BECK & LAURA M. MARUSCHAK, BUREAU OF JUSTICE STATISTICS, DEP'T OF JUSTICE, MENTAL HEALTH TREATMENT IN STATE PRISONS 2000, at 4 (2001), available at http://www.bjs.gov/index-cfm?ty=pbdetail&iid=788 (identifying, from a 2000 census of state prison facilities, 155 facilities in forty-seven states specializing in mental health/psychiatric confinement, with twelve facilities reporting that their primary function is mental health confinement; this figure includes both facilities used temporarily to house inmates suffering from acute episodes and those used to house seriously mentally ill inmates separately from the general population for longer periods of time); 1 NAT'L COMM'N ON CORR. HEALTH CARE, THE HEALTH STATUS OF SOON-TO-BE-RELEASED INMATES: A REPORT TO CONGRESS, at xii (2002), available at http://www.ncchc.org/stbr/Volume1/ ExecutiveSummary.pdf (reporting that only 36% of prisons have specialized housing for inmates with stable mental health conditions). A minority of inmates with mental illnesses reside in segregated facilities. See BECK &; MARUSCHAK, supra, at 1 ("About two-thirds of all inmates receiving therapy/counseling or medications were in facilities that didn't specialize in providing mental health services in confinement."). I assume for purposes of this Article that prison facilities provide some minimal amount of mental health treatment in conformance with constitutional obligations. See infra note 282. But see BECK & MARUSCHAK, supra, at 2 (reporting that 13.2% of state minimum-security prison facilities reported providing no mental health screening or treatment). This is a simplification, but the simplified context probably suffices to establish the basic point that offenders with serious mental illnesses suffer disproportionately in this (common) setting.

(43) See infra Part II. For a discussion of limitations in this data, see infra notes 159-165.

(44) The degree to which sentencing judges should individualize punishments on grounds of offender hardship necessarily presents a line-drawing problem, both as to which offender characteristics to recognize, and as to the nature and degree of risks to consider. Given the strength of social science data outlined in Part II, this Article will focus on offenders with Axis I disorders. I leave to others the responsibility of debating the relevance of other offender characteristics to proportionate sentencing. As to which risks merit attention, I build on others' contributions, see infra notes 201-210, to argue that punishment, for purposes of sentencing, includes foreseeable, substantial risks of serious harm that are proximately caused by the state. Reasonable minds could, and have, disagreed on where to draw the lines on offender characteristics and risk, but I defend these lines for the reasons explained in this Article. See infra Part III.A 45 See infra notes 187, 283.

(46) See infra notes 188-191.

(47) See infra notes 192-194, 336-343 and accompanying text, as well as Part III.B.3.

(48) See infra notes 198-199 and accompanying text.

(49) See, e.g., 730 ILL. COMP. STAT. ANN. 5/5-3-2(a)(1) (West 2007) ("In felony cases, the presentence report shall set forth: the defendant's history of delinquency or criminality, physical and mental history and condition, family situation and background, economic status, education, occupation and personal habits...."); IOWA CODE ANN. [section] 901.3(8) (West 2003 & Supp. 2012) ("If a presentence investigation is ordered by the court, the investigator shall promptly inquire into ... [w]hether the defendant has a history of mental health or substance abuse problems."); N.Y. CRIM. PROC. LAW [section] 390.20 (McKinney 2005) (requiring presentence investigations and reports for all offenders convicted of felonies, and certain offenders convicted of misdemeanors); N.Y. CRIM. PROC. LAW [section] 390.30 (McKinney 2005) (requiring that the presentence report include information regarding defendant's mental health).

(50) See, e.g., FED. R. CRIM. P. 32(c)(1)(A) ("The probation officer must conduct a presentence investigation and submit a report to the court before it imposes sentence."); FED. R. CRIM. P. 32(d)(1)(D)(i) ("The presentence report must identify any factor relevant to the appropriate kind of sentence...."); MONT. CODE ANN. [section] 46-18-112(3) (2011) ("The court may, in its discretion, require that the presentence investigation report include a physical and mental examination of the defendant.").

(51) See ABA STANDARDS FOR CRIMINAL JUSTICE, DEFENSE FUNCTION (3d ed. 1993), available at http://www.americanbar.org/publications/criminal_iustice_section_archive/crim just_standards_dfunc_blkold.html (standard 4-8.1 for sentencing).

(52) While this Article focuses on prisoners with serious mental illnesses, evidence suggests that incarceration in jail also caries antitherapeutic consequences. See Diane S. Young, Jail Mental Health Services, in HANDBOOK OF FORENSIC MENTAL HEALTH WITH VICTIMS AND OFFENDERS: ASSESSMENT, TREATMENT, AND RESEARCH 425, 436 (David W. Springer & Albert R. Roberts eds., 2007) ("[J]ail settings are decidedly nontherapeutic environments. They have many environmental factors that contribute to poor physical and mental health."); cf E. FULLER TORREY ET AL., CRIMINALIZING THE SERIOUSLY MENTALLY ILL: THE ABUSE OF JAILS AS MENTAL HOSPITALS 58 (1992) ("A small number of families reported that jail had been a positive experience for their seriously mentally ill relative by being the only way the person had been able to get treatment.... For the vast majority of mentally ill persons who go to jail, however, the experience varies from being merely negative to being catastrophic.").

(53) See HUMAN RIGHTS WATCH, ILL EQUIPPED: U.S. PRISONS AND OFFENDERS WITH MENTAL ILLNESS 53 (2003), available at http://www.hrw.org/sites/default/files/reports/ usa 1003.pdf ("Mental health experts have described prisons as a 'toxic' environment for the seriously mentally ill."); Marshall T. Bewley & Robert D. Morgan, A National Survey of Mental Health Services Available to Offenders with Mental Illness: Who is Doing What?, 35 LAW & HUM. BEHAV. 351, 352 (2011) ("Simply stated, prison environments are not conducive to optimal mental health functioning."); Richard C. McCorkle, Gender, Psychopathology, and Institutional Behavior: A Comparison of Male and Female Mentally Ill Prison Inmates, 23 J. CRIM. JUST. 53, 54 (1995) ("For those with predispositions, the incarceration experience can actually trigger psychopathology.").

(54) This Article will present the magnitudes of risks of harm faced by prisoners with serious mental disorders as found in the scientific literature. This Article will assume, without providing a detailed argument in support of this position, that these risks are substantial. Whether a risk merits the label of "substantial" is a complicated question that this Article will not address. See, e.g., Sharon Dolovich, Cruelty, Prison Conditions, and the Eighth Amendment, 84 N.Y.U.L. REV. 881, 911-23 (2009) (discussing the objective measure of 'substantial risk of harm' due to prison conditions for purposes of the Eighth Amendment); Richard Siever, HMOs Behind Bars: Constitutional Implications of Managed Health Care in The Prison System, 58 VAND. L. REV. 1365, 1366-67 (2005) (discussing whether the denial of healthcare services presents a substantial risk to prisoners, thus implicating the Eighth Amendment). Future work will explore the extent to which the risks detailed in this section are tolerable or, alternatively, so great that housing offenders with serious mental illnesses with the general prison population should be considered inhumane and prohibited as anathema to retributive punishment.

(55) See infra notes 65-83 (physical assault), 91-103 (sexual assault) and accompanying text.

(56) See infra notes 120-123 and accompanying text.

(57) See infra notes 114-120, 124-128 and accompanying text.

(58) See infra notes 146-149 and accompanying text.

(59) See infra notes 152-155 and accompanying text.

(60) See infra notes 106-110 and accompanying text.

(61) See, e.g., HUMAN RIGHTS WATCH, supra note 53, at 94-127 (detailing the state of inadequate mental health treatment in prisons); JAMES & GLAZE, supra note 2, at 9 (finding that only 34% of state prisoners and 24% of federal prisoners who evidenced a mental health problem had received treatment since admission); Jeffrey L. Metzner & Jamie Fellner, Solitary Confinement and Mental Illness in U.S. Prisons: A Challenge for Medical Ethics, 38 J. AM. ACAD. PSYCHIATRY & L. 104, 105 (2010) ("Relative to the number of prisoners needing help, there is an insufficient number of qualified staff, too few specialized facilities, and few programs."); Nancy Wolff et al., Rates of Sexual Victimization in Prison for Inmates with and Without Mental Disorders, 58 PSYCHIATRIC SERVS. 1087, 1088 (2007) (stating that "underidentification and undertreatment of mental illness inside correctional settings are well established").

(62) See HUMAN RIGHTS WATCH, supra note 53, at 154-61 (documenting the lack of mental health care services available to inmates in solitary confinement); Craig Haney, Mental Health Issues in Long-Term Solitary and "Supermax" Confinement, 49 CRIME & DELINQ. 124, 143 (2003) (describing mental health care options often available to inmates with mental illnesses in segregation); Metzner & Fellner, supra note 61, at 105 ("Mental health professionals are often unable to mitigate fully the harm associated with isolation. Mental health services in segregation units are typically limited to psychotropic medication, a health care clinician stopping at the cell front to ask how the prisoner is doing (i.e., mental health rounds), and occasional meetings in private with a clinician."). For a list of characteristics of segregation facilities with inadequate mental health care, see Gary E. Beven, Offenders with Mental Illnesses in Maximum- and Supermaximum-Security Settings, in HANDBOOK OF CORRECTIONAL MENTAL HEALTH 209, 216 (Charles L. Scott & Joan B. Gerbasi eds., 2005) (Table 10-6).

(63) See infra notes 110 and 153.

(64) See, e.g., JAMES STEPHAN & JENNIFER KARBERG, BUREAU OF JUSTICE STATISTICS, U.S. DEP'T OF JUSTICE, CENSUS OF STATE AND FEDERAL CORRECTIONAL FACILITIES, 2000, at 9, 10 (2003), available at http://bjs.ojp.usdoj.gov/content/pub/pdf/csfcf00.pdf (reporting twenty-eight inmate-on-inmate assaults per 1,000 inmates in federal and state prisons in 2000); Nancy Wolff, Jing Shi & Ronet Bachman, Measuring Victimization Inside Prisons: Questioning the Questions, 23 J. INTERPERSONAL VIOLENCE 1343, 1344 (2008) ("Representative prevalence rates remain elusive, with ranges varying from ... 10% to 25% for physical victimization." (internal citation omitted)).

(65) JAMES & GLAZE, supra note 2, at 10. The study defined having a mental health problem as, in the twelve months preceding the interview, being told by a mental health professional that the individual had a mental disorder; staying overnight in a hospital because of a mental health problem; using prescribed medication to treat a mental health problem; receiving professional mental health therapy; or experiencing symptoms of major depression, mania, or psychotic disorder, Id. at 1-2.

66 PAULA M. DITTON, BUREAU OF JUSTICE STATISTICS, U.S. DEP'T OF JUSTICE, MENTAL HEALTH AND TREATMENT OF INMATES AND PROBATIONERS 9 (1999), available at http://bjs. ojp.usdoj.gov/content/pub/pdf/mhtip.pdf. This study identified a prisoner as mentally ill if he reported a current mental or emotional condition, an overnight stay in a mental hospital, or participation in a treatment program. See id. at 2. No data was reported on diagnosis.

(67) See Cynthia L. Blitz et al., Physical Victimization in Prison: The Role of Mental Illness, 31 INT'L J.L. & PSYCHIATRY 385, 389-90 (2008) (Tables 2 and 3) (showing that 20.5% of male prisoners had experienced physical victimization by inmates over the last six months, while 24.6% had been victimized by staff; figures for female inmates were 20.6% for victimization by inmates and 8.3% for victimization by staff). Reflecting on the difference in staff victimization of male and female inmates, Professor Nancy Wolff and her colleagues have suggested that the data reflect "gender-patterned interactions between inmate and staff in which (a) male inmates, compared to female inmates, are more aggressive against authority figures, resulting in physical altercations with staff; (b) staff is more willing to use physical force against male inmates than female inmates; or (c) some combination of both." Nancy Wolff, Jing Shi & Jane A. Siegel, Patterns of Victimization Among Male and Female Inmates: Evidence of an Enduring Legacy, 24 VIOLENCE & VICTIMS 469, 477 (2009).

(68) See, e.g., HUMAN RIGHTS WATCH, supra note 53, at 56-57 (suggesting that inmates with mental illnesses are particularly likely to be victimized by other inmates); TORREY ET AL., supra note 52, at 21 (finding that 51.8% of the 1,202 jails that reported housing inmates with serious mental illnesses reported that these offenders "increase the potential for outbreaks of violence").

(69) Blitz et al., supra note 67, at 385 ("[W]hether [people with mental disorders] are at elevated risk for victimization inside prison has not been shown empirically, although it has been suggested in numerous reports." (internal citations omitted)); Annette S. Crisanti & B. Christopher Frueh, Risk of Trauma Exposure Among Persons with Mental Illness in Jails and Prisons: What Do We Really Know?, 24 CURRENT OPINION IN PSYCHIATRY 431, 434 (2011) (presenting a review of the "scant" recent literature on this topic).

(70) See Blitz et al., supra note 67, at 385. This study was the first one to explore the rate of physical victimization within a state prison system, inclusive of male and female facilities, and to address inmates with mental illnesses in particular. See also Nancy Wolff & Jing Shi, Feelings of Safety Among Male Inmates: The Safety Paradox, 34 CRIM. JUST. REV. 404 (2009); Wolff et al., supra note 61; Nancy Wolff et al., Understanding Physical Victimization Inside Prisons: Factors that Predict Risk, 26 JUST. Q. 445 (2009) [hereinafter Wolff et al., Understanding Physical Victimization]; Nancy Wolff & Jing Shi, Victimisation and Feelings of Safety Among Male and Female Inmates with Behavioural Health Problems, 20 J. FORENSIC PSYCHIATRY & PSYCHOL. S56 (2009) [hereinafter Wolff & Shi, Victimisation and Feelings of Safety].

71 Blitz et al., supra note 67, at 389 (Table 2); see also Wolff & Shi, Victimisation and Feelings of Safety, supra note 70, at $67 (Table 3). Other studies, however, have found that schizophrenia and bipolar disorder do not function as risk factors for physical victimization. See Paul-Philippe Pare & Matthew W. Logan, Risks of Minor and Serious Violent Victimization in Prison: The Impact of Inmates" Mental Disorders, Physical Disabilities, and Physical Size, 1 SOC'Y & MENTAL HEALTH 106, 113, 116-17 (2001); Wolff et al., Understanding Physical Victimization, supra note 70, at 468 (finding that schizophrenia and bipolar disorder did not function as risk factors in a multilevel predictive model of victimization and explaining this counterintuitive finding by observing that, in the prison system studied, inmates with schizophrenia and bipolar disorder are partially protected from predation by being housed in separate residential units when they are actively psychotic or symptomatic).

72 Blitz et al., supra note 67, at 389 (Table 2).

73 Id. (42.8%).

74 Id. at 389-90. This study did not report on the severity of physical assaults that inmates experienced, although it did differentiate between assaults committed with and without weapons. Id. at 389.

(75) See Hung-En Sung, Nonfatal Violence-Related and Accident-Related Injuries Among Jail Inmates in the United States, 90 PRISON J. 353, 61 (2010) (finding, in a study of 6,982 inmates from 417 local jails, that a recent history of mental health treatment and symptoms of delusion or hallucination resulted in a statistically significant increase in the odds of violence-related injuries, and concluding that mental illness is a "powerful predictor" of violence-related injuries).

(76) See Pare & Logan, supra note 71, at 113, 116-17. The authors defined serious victimization as involving a stabbing, a gun wound, a broken bone or internal injury, unconsciousness, or sexual assault. Id at 112.

(77) For a description of the "prison code" and the difficulty inmates with mental illnesses have adapting to it, see KUPERS, supra note 9, at 18-19.

(78) See Merrill Rotter & Michael Steinbacher, The Clinical Impact of "Doing Time"--Mental Illness and Incarceration, in CIVIC RESEARCH INSTITUTE, FORENSIC MENTAL HEALTH: WORKING WITH OFFENDERS WITH MENTAL ILLNESS 16-1, 16-2 (Gerald Landsberg & Amy Smiley eds., 2001); Sung, supra note 75, at 364.

(79) See Kenneth Adams, Adjusting to Prison Life, 16 CRIME & JUST. 275, 310 (1992) (discussing research suggesting that "odd behavior" is a significant cause of violence between inmates); Wolff et al., Understanding Physical Victimization, supra note 70, at 448 ("Psychiatric or cognitive impairment may also increase victimization if impaired inmates are either easy marks or chronic violators of norms of behavior that are strictly enforced by captives and captors.").

(80) See Maureen L. O'Keefe & Marissa J. Schnell, Offenders with Mental Illness in the Correctional System, 45 J. OFFENDER REHAB. 81, 87 (2007).

(81) See Rotter & Steinbacher, supra note 78, at 16-5 ("Loners can be seen as weak or lacking protection and are often preyed upon. Individuals suffering from mental illness who withdraw, seeking to reduce stimulation or avoid others to meet safety or social distance needs, make the error of having no supports and put themselves at risk.").

(82) HUMAN RIGHTS WATCH, supra note 53, at 57 (quoting KUPERS, supra note 9, at 20).

(83) O'Keefe & Schnell, supra note 80, at 87.

(84) Wolff et al., supra note 61, at 1087 & n.17; see also GERALD G. GAES & ANDREW L. GOLDBERG, NAT'L INST. OF JUSTICE, PRISON RAPE: A CRITICAL REVIEW OF THE LITERATURE 2 (2004), available at https://www.ncjrs.gov/pdffilesl/nij/grants/213365.pdf (reporting prevalence estimates in existing studies ranging from 0% to 40%). This variance can be explained by methodological differences in sample size and location, definitions of victimization and perpetrator, the framing of questions, and the modes of questioning. See Wolff et al., supra note 61, at 1087 & n. 17 (providing examples of varying questions asked); Nancy Wolff et al., Understanding Sexual Victimization Inside Prisons: Factors that Predict Risk, 6 CRIMINOLOGY & PUB. POL'Y 535, 537 (2007).

(85) Prison Rape Elimination Act of 2003, 42 U.S.C. [section] 15603 (2006).

(86) ALLEN J. BECK & PAIGE M. HARRISON, BUREAU OF JUSTICE STATISTICS, U.S. DEP'T OF JUSTICE, SEXUAL VICTIM1ZATION IN STATE AND FEDERAL PRISONS REPORTED BY INMATES, 2007, at 1 (April 2008), available at http://bjs.ojp.usdoj.gov/content/pub/pdf/svsfpri07.pdf.

(87) Id. at 2.

(88) ALLEN J. BECK & CANDACE JOHNSON, BUREAU OF JUSTICE STATISTICS, U.S. DEP'T OF JUSTICE, SEXUAL VICTIMIZATION REPORTED BY FORMER STATE PRISONERS, 2008, at 8 (May 2012), available at http://bjs.ojp.usdoj.gov/content/pub/pdf/svrfsp08.pdf.

(89) Id. (Table 1).

(90) Id. at 10.

(91) See Crisanti & Frueh, supra note 69; Wolffet al., supra note 61, at 1088; Wolff et al., supra note 84, at 540 (reporting data collected from 7,785 state prison inmates).

(92) See Wolffet al., supra note 61, at 1089-90 (Table 1).

(93) See id. at 1089 (Table 1); see also Wolff et al., supra note 84, at 546 ("Compared with inmates without mental disorders, reporting prior treatment for depression, anxiety, and posttraumatic stress disorder increases the likelihood of sexual victimization by 2.6 to 1.8 within a 6-month time period. Inmates with prior treatment for schizophrenia or (547) bipolar disorder were 50% more likely to report an abusive sexual contact within a 6-month time period compared with inmates without prior treatment for mental disorders.").

(94) Rates of sexual victimization by staff were also higher for disordered inmates than for those without a mental illness, but the variance was less dramatic. See Wolff et al., supra note 61, at 1090 ("Roughly one in ten male inmates with a mental disorder reported some form of sexual victimization by staff, compared with one in 14 male inmates without a mental health disorder."). Among inmates with a mental disorder, the rate of sexual victimization by any perpetrator was nearly twice as high among female inmates (27.2%) as among male inmates (15.1%). See id. at 1089-90 (Tables 1 and 2).

(95) Id. at 1089-90.

(96) See id. at 1089 (Table 1) (reporting sexual victimization rates of 10.1% for inmates with schizophrenia or bipolar disorder, compared to 8.3% for inmates with any mental disorder and 3.1% for non-disordered inmates).

(97) Id. at 1088.

(98) Id. at 1089 (Table 1).

(99) See id. (reporting rates of inmate-on-inmate nonconsensual sexual acts of 3.8% for inmates with schizophrenia or bipolar disorder, compared to 1.1% for non-disordered inmates).

(100) See JAMES AUSTIN ET AL., U.S. DEP'T OF JUSTICE, SEXUAL VIOLENCE IN THE TEXAS PRISON SYSTEM, at iv, 41 (2006), available at https://www.ncjrs.gov/pdffilesl/nij/ grants/215774.pdf (finding, in a review of nearly 2,000 officially reported sexual assaults in the Texas prison systems between 2002 and 2005, that prisoners classified as mentally ill were eight times more likely to be victimized than inmates not so classified).

(101) See, e.g., Prison Rape Elimination Act of 2003, 42 U.S.C. [section] 15601 (2006) ("Inmates with mental illness are at increased risk of sexual victimization."); HUMAN RIGHTS WATCH, supra note 53, at 58-59 (describing incidents of rape of individuals with mental illnesses in prison); TORREY ET AL., supra note 52, at 60 (sharing accounts from inmates with mental illnesses and their families of attempted and actual rape in jail).

(102) See NAT'L PRISON RAPE ELIMINATION COMM'N, NATIONAL PRISON RAPE ELIMINATION COMMISSION REPORT 72-73 (2009), available at https://www.ncjrs.gov/pdffilesl/226680.pdf [hereinafter PRISON RAPE ELIMINATION REPORT] ("For men, women, and juveniles coping with serious mental illness, both the disease itself and the treatment can render them extremely vulnerable. Symptoms ranging from hallucinations and paranoia to anxiety and depression may make it difficult to build the kind of supportive social networks that could protect prisoners from sexual abuse. Psychotropic medications often have side effects, such as sleepiness, slowed reactions, uncontrolled movements, and withdrawal that increase a person's vulnerability as well. Moreover, medications are often dispensed in open areas of the facility during peak traffic periods, such as around meal times, effectively 'outing' people with a mental illness."); TORREY ET AL., supra note 52 (observing that "mentally ill inmates who are confused and less able to defend themselves are more vulnerable" to attempted or actual rape).

(103) Wolff et al., supra note 84, at 539.

(104) See Beth Ribet, Naming Prison Rape as Disablement." A Critical Analysis of the Prison Litigation Reform Act, the Americans with Disabilities Act, and the Imperatives of Survivor-Oriented Advocacy, 17 VA. J. Soc. POL'Y & L. 281, 287-95 (2010) (exploring prison sexual assault as a process of disablement, observing the vulnerability of individuals with mental illnesses, and noting consequent long-term psychological and physical injuries, illnesses, and impairments).

(105) See Kim T. Mueser et al., Trauma, PTSD, and the Course of Severe Mental Illness." An Interactive Model, 53 SCHIZOPHRENIA RES. 123, 124 (2002).

(106) KUPERS, supra note 9, at 40; Nancy L. Wolff & Jing Shi, Trauma and Incarcerated Persons, in HANDBOOK OF CORRECTIONAL MENTAL HEALTH 277, 284 (Charles L. Scott ed., 2d ed. 2010); see also, e.g., Ellen Frank & Barbara Pazak Anderson, Psychiatric Disorders in Rape Victims: Past History. and Current Symptomatology, 28 COMPREHENSIVE PSYCHIATRY 77, 81 (1987) (concluding that the experience of rape "produces a period of acute psychological distress," and that individuals who have been raped later frequently meet the requirements for a diagnosis of major depressive disorder or generalized anxiety disorder); Mueser et al., supra note 105, at 126 (explaining that the trauma of sexual or physical abuse, in patients with severe mental illness, is related to increased severity of symptoms of depression, suicidality, anxiety, delusions, hallucinations, and dissociation); Mark Shevlin et al., Cumulative Traumas and Psychosis: An Analysis of the National Comorbidity Survey and the British Psychiatric Morbidity Survey, 34 SCHIZOPHRENIA BULL. 193, 197 (2008) (reporting that sexual abuse is one of the strongest correlates of psychosis). Dr. Terry Kupers has characterized the relationship between trauma and the development of mental illness in prison in the following way:

   Since schizophrenia and other major mental disorders usually
   surface during early adulthood, the age when most felons first
   enter prison, it is often difficult to discern whether a mentally
   disordered prisoner entered prison suffering from the disorder or
   the disorder was caused by harsh prison conditions and the massive
   traumas that regularly occur behind bars.... Many previously
   nondepressed people become severely depressed in jail and prison,
   and a significant proportion go on to commit suicide. Based on the
   large number of clinical cases I have reviewed and my interviews
   with prisoners and staff, I have come to the conclusion there is
   merit in both claims: A much greater number of mentally ill people
   are being sent to jails and prisons today, where their condition
   deteriorates on account of the harsh environment and inadequate
   mental health services; and the harsh conditions and brutality of
   life in prison are making previously very sane prisoners suffer
   psychiatric breakdowns.


KUPERS, supra note 9, at 17-18.

(107) See KUPERS, supra note 9, at 39, 44, 46; Charles B. Nemeroff & Pascal J. Goldschmidt-Clermont, In the Aftermath of Tragedy: Medical and Psychiatric Consequences, 35 ACAD. PSYCHIATRY 4, 5 (2011) (stating that "stress is known to precipitate episodes of, or exacerbate, a variety of severe psychiatric disorders including major depression, bipolar disorder, schizophrenia, and all of the major anxiety and substance abuse disorders").

(108) See Frank & Anderson, supra note 106, at 81 (finding that rape victims frequently develop posttraumatic stress disorder); Nemeroff & Goldschmidt-Clermont, supra note 107, at 5 (finding that syndromal posttraumatic stress disorder may develop in a significant minority of trauma victims); Barbara Olasov Rothbaum et al., A Prospective Examination of Post-Traumatic Stress Disorder in Rape Victims, 5 J. TRAUMATIC STRESS 455, 471 (1992) (reporting that 60% to 65% of rape victims satisfy the requirements for posttraumatic stress disorder one month after the incident).

(109) Blitz et al., supra note 67, at 393; see also Meaghan L. O'Donnell et al., Posttraumatic Disorders Following lnjury: An Empirical and Methodological Review, 23 CLINICAL PSYCHOL. REV. 587, 591 (2003) (reporting that between 80% and 85% of patients with posttraumatic stress disorder also satisfy requirements for at least one other condition, with depression being particularly common).

(110) See Maria A. Oquendo et al., Association of Comorbid Posttraumatic Stress Disorder and Major Depression with Greater Risk for Suicidal Behavior, 160 AM. J. PSYCHIATRY 580, 580 (2003); lldiko Suto & Genevieve L. Y. Arnaut, Suicide in Prison: A Qualitative Study, 90 PRISON J. 288, 294, 304 (2010). Studies indicate that mental disorder constitutes a salient risk factor for suicide in prison. See, e.g., Jacques Baillargeon et al., Psychiatric Disorders and Suicide in the Nation "s Largest State Prison System, 37 J. AM. ACAD. PSYCHIATRY & L. 188, 191 (2009) (finding, in a one-year study of 234,031 inmates incarcerated in the Texas Department of Criminal Justice, that inmates with major depressive disorder, bipolar disorder, schizophrenia, and nonschizophrenic psychotic disorders exhibited "strikingly elevated" risks of suicide); Anasseril E. Daniel & Jennifer Fleming, Suicides in a State Correctional System, 1992-2002: A Review, 12 J. CORR. HEALTH CARE 24, 27 (2006) (finding that 73% of prisoners committing suicide in Missouri had been diagnosed with an Axis I disorder); Raymond F. Patterson & Kerry Hughes, Review of Completed Suicides in the California Department of Corrections and Rehabilitation, 1999 to 2004, 59 PSYCHIATRIC SERVS. 676, 678 (2008) (finding that 73% of suicides in the California Department of Corrections and Rehabilitation--between 1999 and 2004--involved inmates with a history of psychiatric treatment, and 56% involved inmates currently on the mental health caseload).

(111) See Jamie Fellner, A Corrections Quandary: Mental Illness and Prison Rules, 41 HARV. C.R.-C.L.L. REV. 391,394 (2006) ("Apart from the mental health services that may or may not be provided, prisons typically treat prisoners with mental illness identically to all other inmates.").

(112) Ron Jemelka & David Lovell, When Inmates Misbehave: The Costs of Discipline, 76 PRISON J. 165, 167-68 (1996). Minor infractions include theft of food, horseplay, deceit, abusive language, and failure to comply with count procedures. Id. at 167. Major infractions encompass intoxication, assault, fighting, arson, and homicide. Id. at 168.

(113) See id at 166; O'Keefe & Schnell, supra note 80, at 97 (stating that mental health crises and disciplinary violations "can be thought of as behavioral reactions to the correctional environment"). For a thoughtful discussion of the relationship between severe mental illness, maladaptation, and disciplinary infractions, see T. Howard Stone, Therapeutic Implications of Incarceration for Persons with Severe Mental Disorders: Searching for Rational Health Policy, 24 AM. J. CRIM. L. 283, 300-02 (1997); see also Spearlt, Mental Illness in Prison: Inmate Rehabilitation & Correctional Officers in Crisis, 14 BERKELEY J. CRIM. L. 277, 280-93 (2010) (examining mentally ill prisoners' predisposition to break disciplinary rules, psychological deterioration resulting from segregation, exacerbation of mental illness due to inadequate mental health care, and deficiencies in mental health training for correctional officers).

(114) See TORREY ET At., supra note 52, at 58-59 ("Jails have rigid rules, both implicit and explicit, and the inmates who get along best in jails are those who can follow those rules. Inmates who are seriously mentally ill often can neither understand nor follow such rules and, consequently, may be very disruptive for other inmates and for the corrections officers."); Metzner & Fellner, supra note 61, at 105 ("Persons with mental illness are often impaired in their ability to handle the stresses of incarceration and to conform to a highly regimented routine. They may exhibit bizarre, annoying, or dangerous behavior and have higher rates of disciplinary infractions than other prisoners."); Young, supra note 52, at 429 ("Because of their illnesses and the corresponding confusion, suspicion, or fear, [severely mentally ill inmates] may have trouble understanding jail rules or following orders.").

(115) See, e.g., Craig Haney, The Psychological Impact of lncarceration: Implications for Post-Prison Adjustment ("From Prison to Home" Conference, The Urban Inst., U.S. Dep't of Health & Human Servs., Working Paper, Jan. 30-31, 2002), available at http://img2. tapuz.co.il/CommnnaFiles/19852476.pdf ("For mentally-ill and developmentally-disabled inmates, part of whose defining (but often undiagnosed) disability includes difficulties in maintaining close contact with reality, controlling and conforming one's emotional and behavioral reactions, and generally impaired comprehension and learning, the rule-bound nature of institutional life may have especially disastrous consequences.").

(116) See, e.g., HUMAN RIGHTS WATCH, supra note 53, at 53-54; O'Keefe & Schnell, supra note 80, at 86.

(117) HUMAN RIGHTS WATCH, supra note 53, at 54.

(118) See Adams, supra note 79, at 309 ("[I]t has been shown that seriously depressed inmates are more violent toward themselves, that highly confused or disoriented inmates are more violent toward others, and that inmates who are both depressed and confused are more destructive of property."); Beven, supra note 62, at 214.

(119) HUMAN RIGHTS WATCH, supra note 53, at 59 (quoting a letter from Keith R. Curry, Ph.D., to Donna Brorby, Mar. 19, 2002).

(120) JAMES & GLAZE, supra note 2, at 10 (Table 16).

(121) Id.

(122) See DITTON, supra note 66, at 9 (Table 13) (finding, for state prisoners, that 62.2% of inmates with a mental illness were charged with breaking rules, compared to 51.9% of non disordered inmates and, for federal prisoners, that 41.2% of inmates with a mental illness were charged with breaking rules, compared to 32.7% of non-disordered inmates); cf McCorkle, supra note 53, at 58-59 (analyzing data from the Bureau of Justice Statistics' 1986 Survey of Inmates at State Correctional Facilities and finding that prisoners utilizing past or present mental health services disproportionately committed disciplinary infractions in prison but, after accounting for race, concluding that only female prisoners currently on medication were significantly more likely to have experienced greater disciplinary problems).

(123) See, e.g., HUMAN RIGHTS WATCH, supra note 53, at 39 (reporting that, at the Bedford Hills Correctional Facility in New York, 80% of documented incidents involving serious threats to facility safety or security involved prisoners on the active mental health caseload); Kenneth Adams, The Disciplinary Experiences of Mentally Disordered Inmates, 13 CRIM. JUST. & BEHAV. 297, 304-05 (1986) [hereinafter Adams, Disciplinary Experiences] (finding that inmates referred to mental health units of two prisons in New York had significantly higher infraction rates than nonreferred inmates); Kenneth Adams, Former Mental Patients in a Prison and Parole System: A Study of Socially Disruptive Behavior, 10 CRIM. JUST. & BEHAV. 358, 366-68 (1983) [hereinafter Adams, Socially Disruptive Behavior] (finding that the annual rate of infractions for former mental patients was 21.6 per 100 inmates, whereas for the other inmates the annual rate was 14.0 infractions); O'Keefe & Schnell, supra note 80, at 97 (finding that offenders with mental illnesses committed 22% of the 23,852 disciplinary violations in Colorado prisons during fiscal year 2005, even though these inmates comprised only 25% of the offender population); Hans Toch & Kenneth Adams, Pathology and Disruptiveness Among Prison lnmates, 23 J. RES. CRIME & DELINQ. 7, 10-11 (1986) (finding that inmates who are mentally disturbed, as measured by a history of hospitalization or outpatient mental health treatment, have higher annual infraction rates for both violent and nonviolent infractions than inmates who are not mentally disturbed); see also Fellner, supra note 111, at 396 (discussing a subset of these studies).

(124) See Adams, Disciplinary Experiences, supra note 123, at 312-13; see also HANS TOCH, KENNETH ADAMS & J. DOUGLAS GRANT, COPING: MALADAPTATION IN PRISONS 63 (1989) (finding that hospitalized offenders are four times more likely than nonpatients to commit infractions suggesting unusual emotional states, such as throwing feces and setting fire to one's cell).

(125) Adams, Disciplinary Experiences, supra note 123, at 307; see also Adams, supra note 79, at 310 ("Disturbed inmates ... are more likely to engage in acts suggesting peculiar or extreme emotional states, such as self-injury, throwing feces, and arson.").

(126) Adams, Disciplinary Experiences, supra note 123, at 312-13.

(127) HANS TOCH & KENNETH ADAMS, ACTING OUT: MALADAPTIVE BEHAVIOR IN CONFINEMENT 107 (2002); see Adams, supra note 79 (discussing this phenomenon).

(128) TOCH & ADAMS, supra note 127, at 112.

(129) See HUMAN RIGHTS WATCH, supra note 53, at 65-68.

(130) Id. at 174.

(131) Id.

(132) Id.

(133) Fellner, supra note 111, at 397.

(134) Id.; see also KUPERS, supra note 9, at 31-32.

(135) See, e.g., HUMAN RIGHTS WATCH, supra note 53, at 68; Fellner, supra note 111, at 401-02.

(136) JAMES & GLAZE, supra note 2, at 8 (146 months compared to 141 months).

(137) For a fascinating list of activities or beliefs that have resulted in isolation, see Angela A. Allen-Bell, Perception Profiling and Prolonged Solitary Confinement Viewed Through the Lens of the Angola 3 Case: When Prison Officials Become Judges, Judges Become Visually Challenged, and Justice Becomes Legally Blind, 39 HASTINGS CONST. L.Q. 763, 772-76 (2012).

(138) These facilities may be denominated by various monikers, including "control units," "special management units," "security housing units," "high security units," "intensive management units," or "special controls units." Haney, supra note 62, at 151 n. 1.

(139) Though policies differ somewhat by institution, supermax facilities typically serve as a form of administrative segregation, housing inmates deemed to be dangerous or members of a disruptive group, such as a gang. See id. at 127. Thus, inmates are often kept in solitary confinement for an indefinite term "not specifically for what they have done but rather on the basis of who someone in authority has judged them to be." Id.

(140) For a history of supermax facilities and a description of the psychological problems that they may cause, see id A 2009 New Yorker article estimated the current population housed in supermax prisons at 25,000 or more inmates. See Atul Gawande, Hellhole, NEW YORKER, Mar. 30, 2009, at 42; cf. Heather Y. Bersot & Bruce A. Arrigo, Inmate Mental Health, Solitary Confinement, and Cruel and Unusual Punishment: An Ethical and Justice Policy Inquiry, 1 J. THEORETICAL & PHIL. CRIMINOLOGY 1, 12-13 (2010) (stating that, while the most frequently cited figure for the population in solitary confinement is 20,000 inmates, this estimate was obtained from reports compiled in the 1990s using dated findings).

(141) See Carl B. Clements et al., Systemic Issues and Correctional Outcomes." Expanding the Scope of Correctional Psychology, 34 CRIM. JUST. & BEHAV. 919, 926 (2007) (criticizing scholars' interchangeable use of such terms as "administrative detention," "solitary," "isolation," "super max," and "protective custody," and listing contextual variables that may vary and affect an individual's ultimate experience in segregation, including the layout of the cell, the size of the exercise yard, and access to recreational equipment, personal effects, and services).

(142) See Haney, supra note 62, at 125-27. For an in-depth look at solitary confinement in one prison system, see HUMAN RIGHTS WATCH, COLD STORAGE: SUPER-MAXIMUM SECURITY CONFINEMENT IN INDIANA (1997).

(143) Metzner & Fellner, supra note 61. For other descriptions of life in solitary confinement, see Beven, supra note 62, at 211-12; Haney, supra note 62, at 125-27; Peter Scharff Smith, The Effects of Solitary Confinement on Prison Inmates: A Brief History and Review of the Literature, 34 CRIME & JUST. 441, 448-49 (2006).

(144) Administrative segregation is often imposed for an indefinite time period. Disciplinary segregation, while ordered for a set term, can be extended if the offender commits new disciplinary infractions. Inmates with mental illnesses may be particularly likely to violate prison rules as their mental health deteriorates in isolation. See HUMAN RIGHTS WATCH, supra note 53, at 153 ("Achieving sufficient periods of good behavior to secure release from segregation is particularly difficult for mentally ill prisoners. The same inability to comply with the rules that got them placed in segregation originally then extends the time in isolated confinement."); Beven, supra note 62, at 214.

(145) See Haney, supra note 62, at 135.

(146) Bersot & Arrigo, supra note 140, at 7-8 (compiling sources); see also Stanley L. Brodsky & Forrest R. Scogin, Inmates in Protective Custody: First Data on Emotional Effects, 1 FORENSIC REP. 267, 269-70 (1988).

(147) See, e.g., MAUREEN L. O'KEEFE ET AL., ONE YEAR LONGITUDINAL STUDY OF THE PSYCHOLOGICAL EFFECTS OF ADMINISTRATIVE SEGREGATION, at iv (2010), available at https://www.ncjrs.gov/pdffilesl/nij/grants/232973.pdf (estimating that the rate of inmates with mental illnesses in administrative segregation is around 50% higher than the rate within the general prison population); Bersot & Arrigo, supra note 140, at 13 ("Current findings indicate that nearly a third (29%) [of segregated inmates] have been diagnosed with a psychiatric disorder. However, most researchers contend that the number of mentally ill incarcerates may be far greater." (internal citations omitted)); Haney, supra note 62, at 142 ("Research conducted over the past several decades suggests that somewhere between 10% to 20% of mainline prisoners in general in the United States suffer from some form of major mental illness. The percentages in supermax appear to be much higher. Although too few studies have been done to settle on the precise estimates of mentally ill supermax prisoners, and the numbers undoubtedly vary some from prison system to prison system, the percentages may be as much as twice as high as in the general prisoner population." (internal citations omitted)); cf. HUMAN RIGHTS WATCH, supra note 142, at 21 (stating that "even the staff acknowledges that somewhere between one-half and two-thirds of the inmates [in Secured Housing Unit] are mentally ill").

(148) See HUMAN RIGHTS WATCH, supra note 53, at 147-49; Fellner, supra note 111, at 403 n.54.

(149) O'KEEFE ET AL., supra note 147, at x (internal citations omitted).

(150) See 1 FRED COHEN, THE MENTALLY D|SORDERED INMATE AND THE LAW [paragraph] 11.1 (2d ed. 2008); Harley, supra note 62, at 143 ("Especially for prison systems that lack sufficient resources to adequately address the needs of their mentally ill mainline prisoners, disciplinary isolation and supermax confinement seems to offer a neat solution to an otherwise difficult dilemma. In such systems, supermax becomes the default placement for disruptive, troublesome, or inconvenient mentally ill prisoners.").

(151) See, e.g., KUPERS, supra note 9, at 53-58; Brodsky & Scogin, supra note 146, at 279 ("When inmates are subjected to extensive cell confinement and deprivation of activities and stimulation, a majority can be expected to report moderate to serious psychological symptoms."); Smart Grassian & Nancy Friedman, Effects of Sensory Deprivation in Psychiatric Seclusion and Solitary Confinement, 8 INT'L J.L. & PSYCHIATRY 49, 53-54 (1986); Haney, supra note 62, at 132 ("[T]here is not a single published study of solitary or supermax-like confinement in which nonvoluntary confinement lasting for longer than 10 days, where participants were unable to terminate their isolation at will, that failed to result in negative psychological effects."); Smith, supra note 143 (exhaustively reviewing historical and contemporary studies and concluding that "[r]esearch suggests that between one-third and more than 90 percent [of inmates] experience adverse symptoms in solitary confinement, and a significant amount of this suffering is caused or worsened by solitary confinement"). But see O'KEEFE ET AL., supra note 147, at 78 (finding that negative psychological and cognitive symptoms manifested by segregated offenders were not unique to the administrative segregation environment); Clements et al., supra note 141, at 925-26 (arguing that research on the effects of segregation is inconclusive, given the diversity among segregation experiences); Jeffrey Metzner & Joel Dvoskin, An Overview of Correctional Psychiatry, 29 PSYCHIATRIC CLINICS N. AM. 761, 765 (2006) (asserting that, "[d]espite claims to the contrary, it is not currently clear whether, how often, and under what circumstances such [long-term] confinement causes persons [without preexisting mental illness] to develop serious mental illness").

(152) See, e.g., Fellner, supra note 111, at 403 ("Prisoners with preexisting psychiatric disorders are at even greater risk of suffering psychological deterioration while in segregation. The stresses, social isolation, and restrictions of segregated confinement can exacerbate their illness or provoke a recurrence, immeasurably increasing their pain and suffering."); Metzner & Fellner, supra note 61, at 104-05. See generally HUMAN RIGHTS WATCH, supra note 142, at 39-40 (discussing expert and court opinions on vulnerability of offenders with mental illnesses to exacerbation and onset of new mental illnesses when housed in solitary confinement); HUMAN RIGHTS WATCH, supra note 53, at 149-53 (discussing the impact of isolation on inmates with mental illnesses); KUPERS, supra note 9, at 61-64; cf. O'KEEFE ET AL., supra note 147, at 82 (funding that "inmates with serious mental illness are less likely to improve in segregation and are less likely to get worse compared to mentally ill inmates in [the general population]"). But see Kenneth Adams & Joseph Ferrandino, Managing Mentally Ill Inmates' in Prisons, 35 CRIM. JUST. & BEHAV. 913, 921 (2008) (suggesting that research on the effects of segregation is inconclusive).

(153) See, e.g., HUMAN RIGHTS WATCH, supra note 147, at 38-39; Terry A. Kupers, What to Do with the Survivors? Coping with the Long-Term Effects of Isolated Confinement, 35 CRIM. JUST. & BEHAV. 1005, 1014 (2008). Studies have documented the link between isolation and increased risk of suicide. See CORR. ASS'N OF N.Y., MENTAL HEALTH IN THE HOUSE OF CORRECTIONS 57 (2004), available at www.correctionalassociation.org/wpcontent/uploads/2004/06/Mental- Health.pdf (finding that, between 1998 and April 2004, 34% of prison suicides occurred in disciplinary lockdown, although inmates in these units comprised less than 7% of the total prison population); Bersot & Arrigo, supra note 140, at 17 (reporting on a 2006 study of the California prison system that found that 70% of inmates who committed suicide over a one-year period had been housed in long-term solitary confinement); see also HUMAN RIGHTS WATCH, supra note 53, at 179-81 (discussing the increased risk of suicide in segregation units and providing examples of individual state prison systems).

(154) Madrid v. Gomez, 889 F. Supp. 1146, 1236 (N.D. Cal. 1995); see also I COHEN, supra note 150, [paragraph] 11.2, at 11-15; Grassian & Friedman, supra note 151, at 60.

(155) HUMAN RIGHTS WATCH, supra note 53, at 152 (quoting Testimony of Dr. Terry Kupers, Jones 'El v. Berge, 164 F. Supp. 2d 1096 (W.D. Wis. 2001)).

(156) See Madrid, 889 F. Supp. at 1279-80 ("With respect to the SHU [Special Housing Unit at California's Pelican Bay prison], defendants cross the constitutional line when they force certain subgroups of the prison population, including the mentally ill, to endure the conditions in the SHU, despite knowing that the likely consequence for such inmates is serious injury to their mental health, and despite the fact that certain conditions in the SHU have a relationship to legitimate security interests that is tangential at best."); Jones 'El, 164 F. Supp. 2d at 1122 (holding that imprisoning inmates with serious mental illnesses at Wisconsin's supermax facility has "more than a negligible chance" of constituting cruel and usual punishment). Largely in response to such litigation, some states--including Ohio, California, Illinois, Wisconsin, and recently New York--now exclude inmates with serious mental illnesses from confinement in supermax facilities. See Metzner & Dvoskin, supra note 151, at 763; Agreement Reached on SHU Bill--Will Help Enhance Quality of Life for Many New Yorkers with Psychiatric Disabilities in Prisons, MENTAL HEALTH ASS'N IN N.Y. STATE (Jul. 19, 2007), http://www.mhanys.org/publications/mhupdate/update070719.htm. For a discussion of these cases and the constitutionality of housing inmates with mental illnesses in solitary confinement, see 1 COHEN, supra note 150, at ch. 11; Haney, supra note 62, at 145-48.

(157) See Metzner & Fellner, supra note 61 (citing the Human Rights Committee Against Torture and the United Nations Special Rapporteur on Torture and stating: "Whatever one's views on supermax confinement in general, human rights experts agree that its use for inmates with serious mental illness violates their human rights").

(158) See supra note 54.

(159) See Nancy Wolff, Courting the Court: Courts as Agents for Treatment and Justice, in COMMUNITY-BASED INTERVENTIONS FOR CRIMINAL OFFENDERS WITH SEVERE MENTAL ILLNESS 143, 157-58 (William H. Fisher ed., 2003) (highlighting limitations in available data and arguing that unfounded generalizations about the distributional properties of the antitherapeutic impact of incarceration on offenders with mental illnesses should not be used to support differential treatment under the law).

(160) See supra notes 141, 146-149.

(161) See, e.g., Metzner & Dvoskin, supra note 151, at 763 ("There is general consensus among clinicians that placement of inmates with serious mental illnesses in these settings is contraindicated because many of these inmates' psychiatric conditions will clinically deteriorate or not improve.").

(162) See Adams & Ferrandino, supra note 152, at 92l; Clements et al., supra note 141, at 925-26 (2007).

(163) See PRISON RAPE ELIMINATION REPORT, supra note 102, at 73 ("[M]edications are often dispensed in open areas of the facility during peak traffic periods, such as around meal times, effectively 'outing' people with a mental illness.").

(164) See id. at 73.

(165) See id.

(166) See infra note 221 (drawing upon the work of Professor Uma Narayan to suggest that it might be morally defensible to limit sentencing accommodation for foreseeable harm to certain vulnerabilities).

(167) Cf. Kenneth W. Simons, Statistical Knowledge Deconstructed, 92 B.U.L. REV. 1, 6 (2012) (distinguishing between statistical and individualized knowledge for purposes of culpability determinations).

(168) See id. at 36-44 (setting forth and defending the "special stringency principle" for highly concentrated risks, which "explains the presumptive unjustifiability of acting with individualized knowledge" of serious harm, as opposed to mere statistical knowledge).

(169) Risk factors for sexual assault identified by the Attorney General in May 2012 in the National Standards to Prevent, Detect, and Respond to Prison Rape include mental disorder; physical or developmental disability; youth; diminutive stature; first incarceration in prison or jail; nonviolent history; sexual offender status; whether the inmate is likely to be perceived as gay, lesbian, bisexual, transgender, intersex, or gender nonconforming; previous sexual victimization; and the inmate's own perception of vulnerability. See National Standards to Prevent, Detect, and Respond to Prison Rape [section] 115.41, 76 Fed. Reg. 6248, 6280 (to be codified at 28 C.F.R. pt. 115) [hereinafter National Standards for Prison Rape], available at www.ojp.usdoj.gov/programs/pdfs/prea_final_rule.pdf; see also PRISON RAPE ELIMINATION REPORT, supra note 102, at 7-8, 69-74.

(170) See, e.g., JAMES AUSTIN & PATRICIA L. HARDYMAN, NAT'L INST. OF CORR., OBJECTIVE PRISON CLASSIFICATION: A GUIDE FOR CORRECTIONAL AGENCIES (2004), available at http://nicic.gov/Library/019319 (reviewing the current state of prison classification procedures and providing guidelines for implementing classification systems); PATRICIA L. HARDYMAN, ET AL., NAT'L INST. OF CORR., INTERNAL PRISON CLASSIFICATION SYSTEMS: CASE STUDIES IN THEIR DEVELOPMENT AND IMPLEMENTATION 2-4 (2002), available at http://nicic.gov/Library/017381 (describing the Adult Internal Management System (AIMS), whose purpose is "to reduce institutional predatory behavior by identifying predators and separating them from vulnerable inmates," and which is utilized by several facilities in the Federal Bureau of Prisons and by several state departments of corrections); id. at app. (including copies of the AIMS checklists used by the Missouri Department of Corrections); PRISON RAPE ELIMINATION REPORT, supra note 102, at 75-77 (recommending adoption of national standards for screening of all offenders for risk of victimization and abusiveness). Congress recently mandated that all federal confinement facilities assess inmates during intake and upon transfer for their risk of sexual victimization by other inmates. National Standards for Prison Rape, supra note 169, at [section] 115.41.

(171) See supra notes 21-23 and accompanying text (discussing states' abilities to depart from presumptive sentences on the basis of offender hardship and predicted vulnerability to harm).

(172) See, e.g., United States v. Graham, 83 F.3d 1466, 1481 (D.C. Cir. 1996) (stating that "extreme vulnerability to assault in prison may be a ground for departure" under 18 U.S.C. [section] 3553(b) and, "to qualify for a downward departure, a defendant's vulnerability must be so extreme as to substantially affect the severity of confinement, such as where only solitary confinement can protect the defendant from abuse"); United States v. Long, 977 F.2d 1264, 1277-78 (8th Cir. 1992) (upholding the lower court's downward departure under U.S. SENTENCING GUIDELINE [section] 5H1.4 based on reports that the defendant "would be exceedingly vulnerable to victimization and potentially fatal injuries"); United States v. Cotto, 793 F. Supp. 64, 65, 67 (E.D.N.Y. 1992) (granting a downward departure under U.S. SENTENCING GUIDELINE [section] 5K2.13 in part because the defendant's "dull mien, general slackness, and extreme passivity.., make it unlikely that he could resist attacks of predatory fellow inmates during a long prison term").

(173) See, e.g., United States v. Lara, 905 F.2d 599, 603 (2d Cir. 1990) (granting a downward departure under 18 U.S.C. [section] 3553(b) and U.S. SENTENCING GUIDELINE [section] 5H.13 based on the defendant's "particular vulnerability due to his immature appearance, sexual orientation and fragility," which created an extraordinary situation); United States v. Rausch, 570 F. Supp. 2d 1295, 1302, 1308 n.8 (D. Colo. 2008) (assigning a more lenient sentence in part because the defendant "may be easily taken advantage of by others, especially given his physical limitations and medical disabilities"); United States v. Ruff, 998 F. Supp. 1351, 1354-60 (M.D. Ala. 1998) (granting the defendant, who was effeminate, was slight of build, and had a history of sexual victimization, a downward departure because he "is vulnerable to sexual assault and victimization" and recognizing "not only the unconscionable commonality of sexual violence in this nation's prisons, but also the heightened risks facing an inmate with [the defendant's] personal characteristics").

(174) See, e.g., Lara, 905 F.2d at 603 (recognizing that the defendant's protective placement in solitary confinement "exacerbated" the "severity of [his] prison term" and justified a downward departure in sentencing under U.S. Sentencing Guidelines [section][section] 5H1.3 and 5H1.4); United States v. Noriega, 40 F. Supp. 2d 1378, 1379-80 (S.D. Fla. 1999) (reducing the defendant's sentence after finding that he had been confined to a type of "social isolation" that was considered "a more difficult ('harder') type of confinement than in general population," and claiming that, "if there was some divine way one could equate the nature of Defendant's confinement to that which would be more normal, I suppose we would find that he has in fact done more time now than the [actual] years which have passed"); United States v. Blarek, 7 F. Supp. 2d 192, 211 (E.D.N.Y. 1998) (granting the homosexual defendants a downward departure in part because "homosexual defendants may need to be removed from the general prison population for their own safety," which "would amount to a sentence of almost solitary confinement, a penalty more difficult to endure than any ordinary incarceration").

(175) See, e.g., United States v. Boutot, 480 F. Supp. 2d 413, 419, 421 (D. Me. 2007) (granting a downward departure because the court was "concerned about the impact that serving a prison term with the general inmate population would have" on the defendant, who was prone to mental decomposition when not adequately treated); United States v. Roach, No. 00 CR 411, 2005 WL 2035653, at *6-8 (N.D. Ill. Aug. 22, 2005) (finding "that[,] absent continuing and appropriate psychotherapy, [the defendant] will be placed at significant risk of a relapse to serious depression, placing her at risk of suicide ...," and thus reducing the defendant's sentence under 18 U.S.C. [section] 3553(a) to allow for treatment); United States v. Ribot, 97 F. Supp. 2d 74, 84 (D. Mass. 1999) (emphasizing the defendant's twenty-five year struggle with depression and prior suicide attempt and granting a downward departure under U.S. Sentencing Guidelines [section][section] 5H1.3 and 5K2.13 in part because "sending [the defendant] to prison would aggravate his major depressive disorder with potentially life threatening consequences"); People v. Jackson, Nos. 282708, 284430, 2009 WL 1361956, at *3 (Mich. Ct. App. May 14, 2009) (upholding the lower court's reduction of the defendant's sentence based in part on "her suicidal ideations"); People v. Zung, 531 N.Y.S.2d 615, 615-16 (App. Div. 1988) (substituting the punishment of community service for imprisonment in part because "uncontroverted medical documents indicate that a period of incarceration would be severely detrimental to this defendant's mental health, and could possibly exacerbate his suicidal tendencies").

(176) A rigorous analysis of each option is beyond the scope of this Article but will be explicated in future work.

(177) See Andrew von Hirsch, Proportionate Sentences: A Desert Perspective, in WHY PUNISH? HOW MUCH? A READER ON PUNISHMENT, infra note 191, at 207, 207 (describing the origins of desert theory and estimating that the movement toward proportionality-based sentencing began with the publication of AMERICAN FRIENDS SERVICE COMMITTEE, STRUGGLE FOR JUSTICE (1971)); Andrew von Hirsch, Proportionality in the Philosophy of Punishment: From "Why Punish?" to "How Much?," 1 CRIM. L.F. 259, 259-60 (1990) [hereinafter Proportionality] ("Recent philosophical writing about punishment has been devoted mainly to the 'why punish?' question.... The philosophical writings have paid comparatively little attention to the criteria for distributing punishments--particularly to the criteria for deciding how much to penalize convicted offenders.").

(178) See Andrew Ashworth & Elaine Player, Sentencing, Equal Treatment, and the Impact of Sanctions, in FUNDAMENTALS OF SENTENCING THEORY: ESSAYS IN HONOUR OF ANDREW VON HIRSCH 251, 252-53 (Andrew Ashworth & Martin Wasik eds., 1998) [hereinafter FUNDAMENTALS OF SENTENCING THEORY] (distinguishing between justifying the censure of wrongdoers and the imposition of hard treatment); H.L.A. HART, PUNISHMENT AND RESPONSIBILITY: ESSAYS IN THE PHILOSOPHY OF LAW 4 (1968) (stressing the importance of parsing questions concerning the general justifying aim of punishment and its distribution); Paul H. Robinson, The Ongoing Revolution in Punishment Theory: Doing Justice as Controlling Crime, 42 ARIZ. ST. L.J. 1089, 1089-90 (2010).

(179) See Robinson, supra note 178, at 1089-90.

(180) Retributivism defies easy definition. See Mitchell N. Berman, The Justification of Punishment, in THE ROUTLEDGE COMPANION TO PHILOSOPHY OF LAW (Andrei Marmor ed., 2012); John Cottingham, Varieties of Retribution, 29 PHIL. Q. 238 (1979) (delineating nine distinct retributivist theories); Mitchell N. Berman, Rehabilitating Retributivism 1 (July 19, 2012) (unpublished manuscript), available at http://ssrn.com/abstract=-2117619 ("Even if we limit consideration to those central or paradigmatic forms of retributivism that would justify punishment in terms of an offender's negative desert, particular accounts espouse different positions regarding, for example, just what it is that offender's [sic] deserve, in virtue of what they deserve it, and what justifies the state in endeavoring to realize those deserts."). A traditional variant of retributivism theorizes that deserved punishment, justified by the moral culpability and desert of the offender, is an intrinsic good. See G. W. F. HEGEL, ELEMENTS OF THE PHILOSOPHY OF RIGHT 127 (Allen W. Wood ed., H. B. Nisbet trans., 1991) (1821) ("[T]he universal feeling of peoples and individuals towards crime is, and always has been, that it deserves to be punished, and that what the criminal has done should also happen to him."); Michael Moore, The Moral Worth of Retribution, in RESPONSIBILITY, CHARACTER, AND THE EMOTIONS 179, 179 (Ferdinand Schoeman ed., 1987) ("Retributivism is the view that punishment is justified by the moral culpability of those who receive it. A retributivist punishes because, and only because, the offender deserves it."). Other retributivists, however, view punishment as an instrumental good and suggest that it may promote crime control or provide pleasure or utility. See Michael T. Cahill, Punishment Pluralism, in RETRIBUTIVISM: ESSAYS ON THEORY AND POLICY (Mark D. White ed., 2011). For recent scholarship complicating the dominant understanding of retributivism, and challenging the strict divide between retfibutivist and consequentialist theories of punishment, see Mitchell N. Berman, Two Kinds of Retributivism, in PHILOSOPHICAL FOUNDATIONS OF CRIMINAL LAW 433 (R.A. Duff& Smart P. Green eds., 2011).

(181) See BARBARA A. HUDSON, UNDERSTANDING JUSTICE: AN INTRODUCTION TO IDEAS, PERSPECTIVES AND CONTROVERSIES IN MODERN PENAL THEORY 47 (2d ed. 2003) ("Most retributivists in justification ... emphasize the denunciation aspect of punishment.... The degree of severity of a penalty announced marks the degree of disapproval of the crime. This censure is said by retributivists to be the core characteristic and function of punishment." (internal citations omitted)).

(182) Joel Feinberg, The Expressive Function of Punishment, 49 MONIST 400 (1965), reprinted in JOEL FEINBERG, DOING AND DESERVING 95, 98 (1970).

(183) See supra note 181. For examples of scholars who have defended expressive accounts of punishment, see, e.g., DUFF, supra note 30, at 233-66; ROBERT NOZICK, PHILOSOPHICAL EXPLANATIONS 363-97 (1981); Dan M. Kahan, What Do Alternative Sanctions Mean?, 63 U. CHI. L. REV. 591 (1996); Igor Primoratz, Punishment as Language, 64 PHIL. 187 (1989). Expressive accounts of punishment have generated intense criticism. See, e.g., Moore, supra note 180, at 181; Matthew D. Adler, Expressive Theories of Law: A Skeptical Overview, 148 U. PA. L. REV. 1363, 1414-27 (2000); Michael Davis, Punishment as Language: Misleading Analogy for Desert Theorists, 10 LAW & PHIL. 311, 312 (1991). Professor von Hirsch has embraced the view that communication of censure provides the dominant justifying aim for punishment, but that the hard treatment in punishment also "serves as a prudential reason for obedience to those insufficiently motivated by the penal censure's moral appeal." Andrew von Hirsch, Proportionate Sentences." A Desert Perspective, in PRINCIPLED SENTENCING: READINGS ON THEORY AND POLICY 115, 116-18 (Andrew von Hirsch et al. eds., 3d ed. 2009) [hereinafter PRINCIPLED SENTENCING]. Von Hirsch stresses that prudential reasons "should supplement rather than replace the normative reasons for desisting from crime conveyed by penal censure--that is, it provides an additional reason for compliance to those who are capable of recognizing the law's moral demands, but who are also tempted to disobey them." Id. at 118; see also R.A. Duff, Punishment, Retribution and Communication, in PRINCIPLED SENTENCING, supra, at 126, 128-29 [hereinafter Punishment] (characterizing von Hirsch's rationale for using hard treatment as the medium of communication of censure).

(184) See Ashworth & Player, supra note 178, at 252 (citing R.A. Duff, Andrew von Hirsch, and Uma Narayan).

(185) See R.A. DUFF, PUNISHMENT, COMMUNICATION, AND COMMUNITY 132 (2001); von Hirsch, supra note 177, at 277. This Article adopts a similar position.

(186) See Frederic R. Kellogg, From Retribution to "Desert," 15 CRIMINOLOGY 179 (1978) (describing the origins of the desert movement); Malcolm Thomburn & Allan Manson, The Sentencing Theory Debate: Convergence in Outcomes, Divergence in Reasoning, 10 NEW CRIM. L. REV. 278, 280-81 (2007) (reviewing VON HIRSCH & ASHWORTH, supra note 27).

(187) Von Hirsch developed and refined his sentencing theory over the course of four books, see infra note 283, and is widely regarded as a leading just desert theorist. See, e.g., MICHAEL TONRY, SENTENCING MATTERS 17 (1996) (describing von Hirsch as "the most influential modern proponent of just desert theories"); Thornburn & Manson, supra note 186, at 309 (reviewing VON HIRSCH & ASHWORTH, supra note 27) (describing VON HIRSCH & ASHWORTH, supra note 27, as "the best account yet of a theoretically-based sentencing model that fulfills the fundamental justice goal of providing a judge with a clear and principled basis to explain the imposition of state punishment").

(188) See VON HIRSCH & ASHWORTH, supra note 27, at 4 ("The desert rationale rests on the idea that the penal sanction should fairly reflect the degree of reprehensibleness (that is, the harmfulness and culpability) of the actor's conduct."); see infra note 289.

(189) See, e.g., von Hirsch, supra note 183, at 115-16 ("Desert theories for sentencing have had the attraction that they purport to be about just outcomes: the emphasis is on what the offender should fairly receive for his crime, rather than how his punishment might affect his future behaviour or that of others."). See generally infra notes 283-304 and accompanying text.

(190) See VON HIRSCH & ASHWORTH, supra note 27, at 4 ("Proportionalist sentencing is designed to avoid unjust results--through giving conceptions of justice a central role in sentencing policy. The desert rationale rests on the idea that the penal sanction should fairly reflect the degree of reprehensibleness (that is, the harmfulness and culpability) of the actor's conduct.").

(191) See infra Part III.B.3. Additional issues of concern include how to order offenses to reflect their relative seriousness, whether previous convictions should affect seriousness judgments, how to order penalties to reflect their relative severity, how to link particular penalties with particular crimes, how to anchor penalty systems, how to space penalties along a penalty scale, where to draw the line in the rank ordering of offenses between incarceration and other penal options, how to determine if penalties are too severe or too lenient to perform the justifying aim of punishment, and how much (if at all) to take into account utilitarian concerns. For sources addressing some of these topics, see generally VON HIRSCH & ASHWORTH, supra note 27; FUNDAMENTALS OF SENTENCING THEORY, supra note 178; WHY PUNISH? HOW MUCH? A READER ON PUNISHMENT (Michael Tonry ed., 2010).

(192) See infra notes 336-343 and accompanying text.

(193) Andrew von Hirsch has defined punishment without limiting it to deprivations or hard treatment intended by the state: "Punishment (for our purposes) means the infliction by the state of consequences normally considered unpleasant, on a person in response to his having been convicted of a crime." ANDREW VON HIRSCH, DOING JUSTICE: THE CHOICE OF PUNISHMENTS 35 (1976). He derived this definition from the iconic and narrower definition crafted by H.L.A. Hart, so von Hirsch's decision to allow punishment to extend beyond intended hardships appears to have been purposeful. See id. at n. 1 (citing HART, supra note 178); see also infra note 197.

(194) See infra Part III.B.3; VON HIRSCH & ASHWORTH, supra note 27, at 176 (explaining that the aim of the equal-impact principle, when applied in the case of an offender with a physical handicap, is to make "adjustments in sentence to deal with certain foreseeable differential impacts").

(195) This Article largely avoids exploring the important remedial implications of embracing a definition of punishment that includes foreseeable, substantial risks of serious harm.

(196) See JESPER RYBERG, THE ETHICS OF PROPORTIONATE PUNISHMENT: A CRITICAL INVESTIGATION 111-13 (2004) (arguing that punishment should not be confined to intentional consequences); Bronsteen et al., supra note 34, at 1482-95 (arguing that punishment includes reasonably foreseeable acts proximately caused by the state); Gray, supra note 32, at 1622 (arguing that "punishment should be described, accounted for, and justified on objective grounds without reference to the subjective experiences of particular offenders"); Kolber, Subjective Experience, supra note 35, at 185-86 (arguing that sentencing should reflect offenders' subjective experiences of punishment); Kolber, Unintentional Punishment, supra note 36, at 2-3 (arguing that retributivism must measure and take account of unintentional harms associated with punishment to ensure that the punishment is just); Markel & Flanders, supra note 8, at 961 (arguing that "if the ancillary burden the inmate experiences during his imprisonment lacks authorization, then we cannot equate that burden with justified, authorized punishment; thus, it does not necessarily warrant relief from otherwise justified and authorized punishment"). These issues have also been debated within the context of the Eighth Amendment. See, e.g., Dolovich, supra note 54, at 897-908 (arguing that all state-created prison conditions constitute punishment for Eighth Amendment purposes); Thomas K. Landry, "Punishment" and the Eighth Amendment, 57 OHIO ST. L.J. 1607, 1607 (1996) (advancing a governmentalist definition of punishment, which includes the express terms of the penal statute and sentence, and those conditions and events in prison that are attributable to the punitive intent of the government in its role in controlling the machinery of punishment); Alice Ristroph, Sexual Punishments, 15 COLUM. J. GENDER & L. 139, 167-71 (2006) (arguing that, "when a person is sentenced to prison as criminal punishment, the standard and foreseeable conditions of incarceration," including sexual violence, "are part of that punishment"); Ristroph, supra note 30, at 1391-94 (discussing the work of David Enoch, who has argued that the state's reliance on the intention-foresight distinction is often an attempt to evade moral responsibility and that state actors should assume special responsibility for the foreseen effects of their actions). My ultimate conclusion largely coheres, at least with respect to foreseeable risk of serious harm, with those reached by Professors Kolber, Ristroph, Dolovich, Enoch, Bronsteen, Buccafusco, and Masur, though for reasons slightly different from the ones they express.

(197) See, e.g., DUFF, supra note 185, at xiv-xv (observing that "punishment is, typically, something intended to be burdensome or painful, imposed on a (supposed) offender for a (supposed) offense by someone with (supposedly) the authority to do so"); HART, supra note 178, at 4-5 (defining punishment as the imposition of something unpleasant for a legal offense on a supposed offender by a person who intends to administer such punishment within the framework of a legitimate legal authority); VON HIRSCH, supra note 193, at 35 ("Punishment ... means the infliction by the state of consequences normally considered unpleasant, on a person in response to his having been convicted of a crime."). Scholars disagree as to whether the harm of punishment should be confined to intentional deprivations of liberty or whether the harm may also include intentional suffering. See Kolber, Unintentional Punishment, supra note 36, at 20 nn.41-43 (making this observation and collecting sources).

(198) See supra note 197; Richard A. Wasserstrom, Punishment, in PHILOSOPHY AND SOCIAL ISSUES 112, 112 (1980) ("Punishment, whatever else may be said of it, involves the intentional infliction of pain and suffering upon human beings by other human beings."); Hugo Adam Bedau, Feinberg's Liberal Theory of Punishment, 5 BUFF. CRIM. L. REV. 103, 111-12 (2001) (observing that the definitions of punishment offered by Joel Feinberg, Stanley Benn, Antony Flew, H.L.A. Hart, and John Rawls all specify that deprivations or suffering imposed on a person for a legal wrong must be "intended" by a recognized legal authority); Richard W. Burgh, Do the Guilty Deserve Punishment?, 79 J. PHIL. 193, 193, 194 n.1 (1982) (adopting Anthony Flew's definition of punishment and stating that punishment "involves the intentional infliction of suffering"); Anthony Flew, The Justification of Punishment, 29 PHIL. 291, 293-95 (1954) (defining punishment as the suffering of an offender for his offense imposed intentionally by human agencies in connection with a system of laws); see also Kolber, Unintentional Punishment, supra note 36, at 5 & n.9 ("Criminal law scholars widely agree that in order for some conduct to constitute punishment, it must be intentionally imposed."). As early at the 1930s, philosophers distinguished "punishment" as an intentional deprivation of liberty from "its accessories," including both the foreseeable and unforeseeable consequences of such deprivation. See J.D. Mabbott, Punishment, 48 MIND 152, 165 (1939) ("When a man is sentenced to imprisonment he is not sentenced also to partial starvation, to physical brutality, to pneumonia from damp cells and so on. And any movement which makes his food sufficient to sustain health, which counters the permanent tendency to brutality on the part of his warders, which gives him a dry or even a light and well-aired cell, is pure gain and does not touch the theory of punishment.").

(199) See Gray, supra note 32, at 1653 (arguing that, "because it is incidental, objectivist forms of retributivism ... bear no burden to justify ... additional or surplus suffering because it is not 'punishment,' and therefore is not justified"); Markel & Flanders, supra note 8, at 959-64 (challenging Kolber's argument that a just punishment system calls for contemplating both intended liberty deprivation and unintended ancillary distress); Dan Markel, Chad Flanders & David Gray, Beyond Experience: Getting Retributive Justice Right, 99 CALIF. L. REV. 605, 618 (2011) (asserting that, "[i]f the hardship endured by the offender is not authorized, intentionally imposed, and proximately caused by the state, then it is a conceptual error to call it 'punishment'").

(200) See Gray, supra note 32, at 1648 ("If it is true that some suffering is incidental and some not, then it may simply be the case that all the subjective inequalities Kolber and Bronsteen, Buccafusco, and Masur are concerned with, whether measured subjectively, comparatively, or diachronically, are incidental to punishment and therefore impose no duties of accommodation or accounting on theories of punishment."). But see RYBERG, supra note 196, at 113 ("What I am claiming, of course, is not that possible differences in sensibility or in prison conditions would be irrelevant to the proportionalist sentencer, who believes that it is intended severity that counts; if, for instance, there is an intention to punish two persons equally severely and the sentencer knows that there are such differences then they should be accounted for when the sentencer seeks to carry out the intention. The problem rather is that, if the sentencer is misinformed or lacks information on these matters, then there will not be reasons concerning justice to object to the resulting punishments as long as what was intended did not violate justice."). Although Professor Gray appears to believe that sentencing judges are not obligated to consider the risk of prison violence in sentencing as a matter of proportionate punishment, he argues that "prudence" or mercy may call for action by other institutional players. See, e.g., Gray, supra note 32, at 1692-93. In particular, because such violence is not punishment, it may motivate prison officials to modify penal circumstances, supply grounds for a tort claim or criminal action, or, if pervasive, require reform of punishment practices. See id. at 1627, 1630 n.46, 1653, 1670; see also Markel & Flanders, supra note 8, at 961 ("If an unconstitutional tort occurs during the punitive encounter, the state's obligation may reasonably take the form of compensation, apology, injunctive relief, or administrative reform. Such harm to the offender does not necessitate the remission of the offender's balance of punishment; there are other currencies the state can use."). Professor Gray also suggests that "excessive suffering at the hands of other prisoners ... may well provide good reason for early release from a justly imposed term of imprisonment," and observes with approval that "judges and executive-branch officials routinely entertain pleas for mercy from prisoners who have suffered inordinately during their incarceration." Gray, supra note 32, at 1692.

(201) These arguments have been made both as a matter of moral theory and within the context of the Eighth Amendment.

(202) Ristroph, supra note 196, at 168.

(203) See Bronsteen et al., supra note 34, at 1488-89; Ristroph, supra note 30, at 1399-1400 (observing that "[r]arely can a single coherent intent be attributed to the entire institutional apparatus that imposes punishment" and detailing "all the state actions that must occur in order for a person to be punished with a prison sentence"); supra note 30.

(204) See Kolber, Unintentional Punishment, supra note 36, at 6-7.

(205) Id. at 7; Ristroph, supra note 30, at 1393 (discussing the work of David Enoch).

(206) David Enoch, Intending, Foreseeing, and the State, 13 LEGAL THEORY 69, 91 (2007); see also id. at 82 ("A responsible agent, it can be argued, accepts responsibility for all (foreseen) consequences of her actions, both intended and unintended. This suspicion--that hiding behind the intending-foreseeing distinction is really just evading responsibility--is arguably at least a part of the rationale for the entrenched doctrine of the criminal law, according to which under certain circumstances foresight can substitute for intention."); Ristroph, supra note 30, at 1391-94 (discussing the work of David Enoch).

(207) See RYBERG, supra note 196, at 112-13.

(208) See Kolber, Unintentional Punishment, supra note 36, at 14-15. Though Kolber argues that the state must take responsibility for the foreseeable results of its actions, he does not challenge the "technical" definition of punishment. See also id. at 2 ("[E]ven if the unintended side effects of punishment are technically not punishment, the state has a moral obligation to take account of the actual or expected ways in which punishment affects inmates' lives.").

(209) Id. at 3-4.

(210) See Bronsteen et al., supra note 34, at 1487; see also Kolber, Subjective Experience, supra note 35, at 208-10 (exploring why offenders' subjective experiences should matter to expressive views of retributivism).

(211) See supra notes 181-184 and accompanying text.

(212) See Duff, Punishment, Retribution and Communication, supra note 183, at 126, 132 (explaining that his expressive account of punishment "requires us to attend not just to the general meaning of punishment as a mode of censure, but to the distinctive meanings of different modes of punishment"); Dan M. Kahan, What's Really Wrong with Shaming Sanctions, 84 TEX. L. REV. 2075, 2086 (2006) ("[C]itizens will expect punishments not only to express condemnation but also to express condemnation in a way that coheres with ... their more basic cultural commitments."); Primoratz, supra note 183, at 201 ("[The punishment] has to be appropriate as moral condemnation. It has to be truthful, just and deserved, and to be seen as such, by everyone involved: by those conveying it, and by all those to whom it is being conveyed.").

(213) See, e.g., Markel et al., supra note 199, at 624 ("[T]he polity need only be constrained by the reasonable interpretation of the sentence imposed, and this will largely follow the polity's perspective since it is the polity that is creating and reflecting the social meaning involved here."); Simons, supra note 40, at 3 (arguing that, to expressive retributivists who believe that punishment expresses the community's resentment, "it is absolutely crucial that the public view the conditions of the offender's punishment as proportionate to the initial blaming judgment"). But see Bronsteen et al., supra note 34, at 1478 (arguing that a communicative theory of punishment that values only the perception of how a typical offender experiences punishment (as opposed to evidence of the actual experience of a typical offender) is unappealing and incredible).

(214) See Kolber, Subjective Experience, supra note 35, at 209 ("[P]eople may not investigate the more detailed facts about an objectively defined punishment so as to know its true severity. But surely an offender cannot be said to deserve the vague punishment given by ill-informed societal condemnation any more than an innocent person deserves the culpability judgment of an ill-informed factfinder."); Kolber, Unintentional Punishment, supra note 36, at 2; cf. Markel & Flanders, supra note 8, at 954-55 ("To be sure, awareness of hedonic adaptation or expected subjective preference patterns of the public at large may conceivably inform the ex ante selection of sentencing ranges or penal techniques approved by legislatures.").

(215) For a discussion of limitations in data and how they may affect particularized risk assessments, see supra notes 159-169 and accompanying text. Of course, psychiatrists may also disagree about the existence of a disorder in a particular offender.

(216) The differential effect of incarceration on offenders with mental illnesses--especially the effect of prolonged solitary confinement--has been the subject of recent media attention. See, e.g., Vince Beiser, A Necessary Evil?, L.A. TIMES MAG., Oct. 19, 2003, at 14; Colin Dayan, Barbarous Confinement, N.Y. TIMES, July 17, 2011, at A19; Editorial, Prison Suicides, N.Y. TIMES, Jan. 15, 2011, at A22; Andrew Cohen, An American Gulag: Descending into Madness at Supermax, THEATLANTIC.COM (June 18, 2012, 2:10 PM), http://www.theatlantic.com/national/archive/2012/06/an-american-gulag-descending-into-madness-at-supermax/258323/; Andrew Cohen, Supermax: The Faces of a Prison's Mentally Ill, THEATLANTIC.COM (June 19, 2012, 6:00 AM), http://www.theatlantic.com/ national/print/2012/06/supermax-the-faces-of-a-prisonsmentally-ill/258429/; Frontline: The New Asylums, PBS (May 10, 2005), http://www.pbs.org/wgbh/pages/frontline/shows/ asylums/inmates/; Laura Sullivan, At Pelican Bay Prison, a Life in Solitary, NAT'L PUB. RADIO (Jul. 26, 2006), http://www.npr.org/templates/story/story-php?storyId=5584254.

(217) Judges and legal commentators have cited the differential suffering of prisoners with mental illnesses as a reason supporting the creation of mental health courts. See, e.g., RICHARD D. SCHNEIDER ET AL., MENTAL HEALTH COURTS: DECRIMINALIZING THE MENTALLY ILL 2 (2007); Robert Bernstein & Tammy Seltzer, Criminalization of People with Mental Illness: The Role of Mental Health Courts in System Reform, 7 U.D.C.L. Rev. 143, 143 (2003); James D. Cayce & Kari Burrell, King County's Mental Health Court: An Innovative Approach for Coordinating Justice Services, 53 WASH. ST. B. NEWS 19 (1999); Susan Stefan & Bruce J. Winick, A Dialogue on Mental Health Courts, 11 PSYCHOL. PUB. POL'Y & L. 507, 510 (2005). Mental health courts typically allow participants to avoid incarceration in exchange for court-supervised treatment. See E. Lea Johnston, Theorizing Mental Health Courts, 89 WASH. U. L. REV. 519, 521 (2012). Though intuitions regarding offenders with major mental illnesses were not specifically tested, recent empirical work suggests that a substantial minority of the public would support mitigation of punishment for offenders who would suffer undue hardship. See Paul H. Robinson, Scan E. Jackowitz & Daniel M. Bartels, Extralegal Punishment Factors: A Study of Forgiveness, Hardship, Good Deeds, Apology, Remorse, and Other Such Discretionary Factors in Assessing Criminal Punishment, 65 VAND. L. REV. 737, 782, 824 (2012) (reporting that 28% of study respondents believed that mitigation of punishment was justified when the punishment would have an undue hardship on the offender); id. at 782, 824-25 (reporting that 22% of study respondents believed that advanced age would warrant mitigation).

(218) See Alice Ristroph, Desert, Democracy, and Sentencing Reform, 96 J. CRIM. L. & CRIMINOLOGY 1293, 1309 (2006).

(219) ANDREW ASHWORTH, SENTENCING AND PENAL POLICY 277 (1983) ("The argument, then, is that whilst it is just to impose the same sentence on two equally culpable offenders for two equally grave offences, it is unjust to do so if the two offenders have such differing 'sensibilities' that the sentence would have a significantly different effect on each of them. The sentencer should take account of any relevant and significant differences, and should strive to achieve equality of impact."); Ashworth & Player, supra note 178, at 253 (advocating "a general principle of equal treatment, by which we mean that a sentencing system should strive to avoid its punishments having an unequal impact on different offenders or groups of offenders"); Kolber, Subjective Experience, supra note 35, at 199-210 (arguing that various versions of retributivism must factor subjective experience into sentencing in order to fulfill the proportionality requirement). I explore the concept of "equal impact" in more depth at Parts III.B.2-3.

(220) See Dolovich, supra note 54, at 939 (observing in the context of incarceration that "there will likely be few cases in which harm to prisoners is not traceable to official conduct").

(221) If penalties that pose a foreseeable, substantial risk of serious harm are not inhumane, it may be that a sentencing accommodation is only warranted for offenders with certain vulnerabilities. See VON HIRSCH & ASHWORTH, supra note 27, at 173 & n.f (discussing, in the context of application of the equal-impact principle, when a living standard analysis should be tailored to members of "nonstandard" groups for whom "imprisonment typically becomes more burdensome" and referencing the dissertation of Professor Uma Narayan, infra, for elucidating these special cases). For instance, Professor Uma Narayan has argued, in the context of regulating offensive conduct, that nonstandard interests held by individuals with "special natural vulnerabilities"--those stemming from relatively natural causes such as physical or mental disability, illness, youth, or advanced age--are particularly worthy of protection "because these vulnerabilities are not even remotely matters of choice, often pose serious risk of harm or offense, and are potentially likely to affect any currently 'standard' person. Also, people in this category are usually unable to provide the requisite special protection for themselves." Uma Narayan, Offensive Conduct: What Is It and When May We Legally Regulate It? 212-13, 223 (1990) (unpublished Ph.D. thesis, Philosophy Dept., Rutgers University). To the extent that Narayan has identified factors sufficient to distinguish morally cognizable vulnerabilities from those without similar salience, a number of classes beyond the seriously mentally ill may be worthy of recognition in the sentencing context. In particular, cognizable vulnerabilities, if verified, could include those stemming from advanced age, youth, diminutive stature, effeminate appearance, gay and transgender orientation, mental retardation, and physical disability. This Article leaves to others the important work of making that case and defending the use of Narayan's criteria--or another set--to distinguish groups of offenders whose foreseeable harm should factor into sentencing.

(222) See Kolber, supra note 41, at 635-40.

(223) See Kolber, Subjective Experience, supra note 35, at 208-10; Kolber, supra note 36, at 28-29.

(224) I argue here that, under an expressive theory of retributive punishment, the offender should be the primary intended audience for society's message of condemnation. This understanding mutes any difference, for my purposes, between expressive and communicative theories of punishment (though adherents to communicative theories of punishment may be more receptive to the arguments advanced in this Part of the Article). Cf. Elizabeth S. Anderson & Richard H. Pildes, Expressive Theories of Law: A General Restatement, 148 U. PA. L. REV. 1503, 1508 (2000) ("To express a mental state requires only that one manifest it in speech or action. To communicate a mental state requires that one express it with the intent that others recognize that state by recognizing that very communicative intention."); Markel & Flanders, supra note 8, at 929 & n.89 (differentiating "communicative" action directed to a designated recipient "in a way the sender of the message thinks will make sense to the recipient, and is performed in a way that the thought conveyed can be made sense of, or effectuated, through the free will of the recipient" from "expressive" action that "emit[s] certain views or attitudes but does not require that a particular member of the audience for the action understands the basis for or purpose behind the action").

(225) See supra note 212.

(226) See supra note 183; see also DUFF, supra note 185, at 75-130.

(227) R.A. Duff, Penal Communications: Recent Work in the Philosophy of Punishment, 20 CRIME & JUST. 1, 61-62 (1996).

(228) See infra notes 305-312 and accompanying text.

(229) See DUFF, supra note 185, at 129-30 ("[Punishment] addresses offenders, not as outlaws who have forfeited their standing as citizens, but as full members of the normative political community; it is inclusionary rather than exclusionary. It treats them as citizens who are both bound and protected by the central liberal values of autonomy, freedom, and privacy. It holds them answerable, as responsible moral agents, for the public wrongs they commit. But it also respects their own autonomy (since it seeks to persuade rather than merely to coerce), their freedom (since it constitutes a legitimate response to their wrongdoing and leaves them free to remain unpersuaded), and their privacy (since it addresses only those aspects of their lives and actions that properly fall within the public sphere)."); Bronsteen et al., supra note 34, at 1487 (arguing that "the state cannot define the content of its messages by authorial fiat" and that, "[t]o the extent that the state ignores an offender's reasonable interpretation of the message, it fails to treat her fully as moral agent"); von Hirsch, Proportionality, supra note 177, at 273-74 ("Such communication of judgment and feeling is the essence of moral discourse among rational agents.").

(230) See generally Kolber, Subjective Experience, supra note 35, at 208 ("If the severity of punishment depends on how the condemnatory message is understood by offenders, then it is easy to see why offenders' punishment experiences matter."); Primoratz, supra note 183, at 200 ("So if society's condemnation of their misdeeds is really to reach [offenders], if they are really to understand how wrong their actions are, it will have to be translated into the one language they are sure to understand: the language of self-interest. This translation is accomplished by punishment.").

(231) See DUFF, supra note 185, at 87 ("Punishment as censure gives offenders the opportunity to listen to the law's moral voice and so to repent their crimes and seek their own moral reform. But it does not find its justifying purpose in an attempt to elicit (or coerce) such moral responses." (citations omitted)); von Hirsch, supra note 183, at 116-17 ("A response to criminal wrongdoing that conveys blame gives the individual the opportunity to respond in ways that are typically those of an agent capable of moral deliberation: to recognise the wrongfulness of action; feel remorse; to make efforts to desist in future--or to try to give reasons why the conduct was not actually wrong."); Markel & Flanders, supra note 8, at 933 ("[T]hough the offender must be able to rationally understand the communication, he need not be persuaded by it.").

(232) See supra note 212 and accompanying text (stressing that, at base, society must view a penalty as censorious and roughly proportionate to a crime).

(233) See supra notes 49-51 and accompanying text.

(234) See Ashworth & Player, supra note 178, at 260; Bronsteen et al., supra note 34, at 1482-95; cf. Kolber, supra note 36, at 15-16.

(235) See VON HIRSCH & ASHWORTH, supra note 27, at 156 ("The criterion for substitutions among penalties should, on a desert model, be that of comparable severity: approximate equivalence in penal bite. The principle of proportionality addresses the severity of penalties, not their particular mode."); NORVAL MORRIS & MICHAEL TONRY, BETWEEN PRISON AND PROBATION, at ch. 4 (1990) (setting forth principles of interchangeability of punishments to provide for the principled distribution of punishments with rough equivalence of punitive bite on utilitarian grounds); id. at 93 ("[F]rom a moral perspective, the measure of punishment is not its objective appearance but its subjective impact. Our goal is to achieve a system of interchangeable punishments that the state and the offender would regard as comparable in their punitive effects on him."); Robinson, supra note 26, at 151-52 (arguing that, "[a]s long as the total punitive 'bite' of the punishment achieves [ordinal] ranking, [deontological and empirical] conceptions of desert have little reason to care about the method by which that amount of punitive 'bite' is imposed" and suggesting the adoption of an equivalency table for alternative sanctions). It is possible, however, that only a showing of individualized risk warrants a sentencing adjustment. See supra notes 167-170 and accompanying text.

(236) See supra note 30. It is for this reason that offender vulnerability is a proper consideration of sentencing, as opposed to an issue solely of penal administration.

(237) Judges typically lack the authority to select the institution to which an individual will be assigned to serve his term of imprisonment. See, e.g., 18 U.S.C. [section] 3621(b) (2006) ("The Bureau of Prisons shall designate the place of the prisoner's imprisonment."); FLA. STAT. ANN. [section] 20.315(7) (West 2009) ("The department [of corrections] shall place each offender in the program or facility most appropriate to the offender's needs, subject to budgetary limitations and the availability of space."); R.I. GEN. LAWS [section] 42-56-10(13) (2006) (granting the director of the department of corrections the power to "assign or transfer those persons [committed to the custody of the department] to appropriate facilities and programs"). Judges may, however, recommend particular housing assignments. See 18 U.S.C [section] 3621(b)(4) (stating that, in making its determination under the statute, the Bureau of Prisons may consider any statement by the sentencing court concerning the purposes of a term of imprisonment or recommending a certain type of correctional facility). The degree to which these recommendations are honored varies by jurisdiction.

(238) Violence levels and victimization rates, in reality, will vary by facility and security level. See ALLEN J. BECK & CANDACE JOHNSON, BUREAU OF JUSTICE STATISTICS, DEP'T OF JUSTICE, SEXUAL VICTIMIZATION REPORTED BY FORMER STATE PRISONERS, 2008, at 22-23 (May 2012), available at http://bjs.ojp.usdoj.gov/content/pub/pdf/svrfsp08.pdf (showing prevalence rates of sexual victimization, for male former prisoners, that are 1.7 to 3.0 times higher in maximum-security prisons than in minimum-security prisons and comparing victimization rates for prisons by type of facility); GAES & GOLDBERG, supra note 84, at 50 (suggesting that sexual victimization rates are highest at higher-security-level prisons). In addition, treatment opportunities and program options vary by facility and prison security level. See, e.g., JAMES AUSTIN & KENNETH McGINNIS, NAT'L INST. OF CORR., CLASSIFICATION OF HIGH-RISK AND SPECIAL MANAGEMENT PRISONERS 31, 33 (2004), available at http://www.nicic.org.

(239) See supra notes 49-51.

(240) See infra notes 387-393.

(241) See Bronsteen et al., supra note 34, at 1466 ("The state is responsible for the foreseeable, proximately caused effects of punishment--effects that the typical offender will understand to be part of her punishment--and this responsibility should influence the legislative crafting of punishments.").

(242) The decision to limit the risks that should factor into sentencing to substantial risks of serious harms is supported by several moral and prudential considerations. First, when sentences ignore substantial risks of serious harm, the ultimate penalty experienced is likely to be vastly disproportionate to the penalty deserved. Sentences cannot practically contemplate all risks of harm, and those that are insubstantial are far less likely to impose vastly disproportionate punishments. See ASHWORTH, supra note 219, at 274 (arguing that "there are degrees of injustice, and that a slightly inaccurate estimation of desert is strongly preferable to an approach to sentencing which in no way aspired to proportionality and departed widely from it"). Second, although accounting for all nonserious harms might require great variation in length or severity of punishment, see Simons, supra note 40, at 5, this should be less true when accounting for only truly serious harms for which a foreseeable, substantial risk exists at the moment of sentencing. Third, the state's moral obligation to consider foreseeable risks of harms--and act to prevent their realization--is at its apex when those risks are substantial and the harms are of a serious nature. Fourth, while sentencing two equally blameworthy offenders differently on the basis of vulnerability--assuming the basis of the distinction is inadequately explained or publicized--may result in the appearance of a lack of uniformity and unfairness, this cost is arguably dwarfed by that of imposing vastly disproportionate and inhumane sentences on morally significant, vulnerable populations, which is a likely result if judges ignore foreseeable, substantial risks of serious harm at sentencing. The extent to which the definition of punishment should turn on prudential, as opposed to moral, reasoning is questionable, however. After all, proponents of restricting punishment to intended harm can call upon the prudential reasons of ease of administration and commensurability to support their position as well. Future work will explore in more depth the existence of moral grounds for limiting the scope of punishment to substantial risks of serious harms proximately caused by the state.

(243) Farmer v. Brennan, 511 U.S. 825, 834 (1994) (internal quotation omitted). One important issue, not addressed in this Article, is the extent to which prison violence effectuated by inmates--as opposed to prison guards--may constitute punishment even though private actors, who are not authorized to inflict sanctions, carry it out. See, e.g., Alon Harel, Why Only the State May Inflict Criminal Sanctions: The Case Against Privately Inflicted Sanctions, 14 LEGAL THEORY 113, 114 (2008) ("Insofar as the state is the source of criminal prohibitions, it should also determine the nature and the severity of the sanctions that follow their violation and should inflict these sanctions."). This Article assumes, but does not defend the notion, that prisoner-on-prisoner violence constitutes punishment when the state has created the conditions of confinement in which prison violence is likely to occur, such that acts of prisoner-on-prisoner violence foreseeably arise from those conditions and may be considered proximately caused by the state. See Bronsteen et al., supra note 34, at 1485 n.86. See generally Mary Sigler, Just Deserts, Prison Rape, and the Pleasing Fiction of Guideline Sentencing, 38 ARIZ. ST. L.J. 561, 568 (2006) (describing when the Eighth Amendment protects individuals from victimization by other inmates). With that said, it is worth emphasizing that prison violence certainly can never be considered just punishment.

(244) See Sigler, supra note 243; supra note 199 and accompanying text. For arguments that foreseeable prison violence should be conceived as part of an offender's punishment for purposes of the Eighth Amendment, see Dolovich, supra note 54 at 906-07; Alexander A. Reinert, Release as Remedy for Excessive Punishment, 53 WM. & MARY L. REV. 1575, 1611-22 (2012); and Ristroph, supra note 196, at 167-70.

(245) See Gray, supra note 32, at 1649-53 (characterizing prison violence as "crime," not "punishment"); Markel & Flanders, supra note 8, at 960-61.

(246) See Sigler, supra note 243, at 562 (arguing that "the doctrine of downward departures on the basis of extreme vulnerability is problematic in political, moral, and practical terms").

(247) See id. at 573 (observing that the common result of granting a downward departure for extreme vulnerability, under the U.S. Sentencing Guidelines, is not total relief from incarceration, but rather, a reduced carceral term "at rape"); see also Gray, supra note 32, at 1650 ("[I]f the suffering occasioned by prisoner-on-prisoner violence is 'punishment,' and 'punishment' is the suffering which offenders deserve as a consequence of their crimes, then the perpetrators of sexual assault in prison are by definition immune from prosecution because the suffering they inflict is 'punishment.'").

(248) See Gray, supra note 32, at 1649-50. For an effective response to this point, see Reinert, supra note 244, at 1619.

(249) See E-mail from Dan Markel to author (Feb. 6, 2012) (on file with author).

(250) Cf. Reinert, supra note 244, at 1619.

(251) See infra Part III.B.2.

(252) See infra Part III.B.3.

(253) See Dolovich, supra note 54, at 915-16 ("To force prisoners to live in constant fear of violent assault, under conditions in which many of the most vulnerable among them can expect that fear to be realized, is to inflict a form of physical and psychological suffering akin to torture. It is plainly cruel to punish criminal offenders with the strap, with rape, or with any other form of brutal corporal treatment. And for the same reason, the state may not place incarcerated offenders in a position of ongoing vulnerability to assault by predatory prisoners, thus creating conditions that would amount to the same thing.").

(254) See, e.g., Jones v. Diamond, 636 F.2d 1364, 1373 (5th Cir. 1981) ("Confinement in a prison where terror reigns is cruel and unusual punishment. A prisoner has a right to be protected from the constant threat of violence and from sexual assault."), overruled on other grounds by Int'l Woodworkers of Am. v. Champion Int'l Corp., 790 F.2d 1174 (5th Cir. 1986); Ashworth & Player, supra note 178, at 262-64 (arguing that "prison conditions which fall below the minimum and which constitute inhuman or degrading treatment are simply not acceptable"). For a discussion of inhumanity in the context of retributivism, see Part III.B.1.

(255) See supra Part III.B.2.

(256) See Ashworth & Player, supra note 178, at 262 (observing that, after one determines that prison conditions are not inhumane but vary, one must ask "whether, and, if so, how, to take account of those variations when calculating sentencing length"). I am grateful to Professor Michael Seigel for suggesting the "collateral consequences" terminology.

(257) Haney, supra note 62, at 135 ("Conditions of confinement for protective custody prisoners are in many ways similar to those in supermax confinement. That is, they are typically segregated from the rest of the prison population, restricted or prohibited from participating in prison programs and activities, and often housed indefinitely under what amount to oppressive and isolated conditions."); James E. Robertson, The Constitution in Protective Custody: An Analysis of the Rights of Protective Custody Inmates, 56 U. CIN. L. REV. 91, 91 (1987) ("In exchange for security, protection inmates often experience conditions of confinement similar to those imposed as punishment for disciplinary infractions.").

(258) See Sharon Dolovich, Strategic Segregation in the Modern Prison, 48 AM. CRIM. L. REV. 1, 3-4 (2011) (stating that, due to the fact that gay men and transgendered women are "almost automatically" targets for sexual abuse, many carceral facilities routinely house these inmates in protective custody--"a classification that typically involves isolation in 'a tiny cell for twenty-one to twenty-four hours a day,' the loss of access to any kind of programming (school, drug, treatment, etc.), and even deprivation of basics like 'phone calls, showers, group religious worship, and visitation'") (internal citations omitted).

(259) It is possible that confinement in isolation may, given its likely effects on an individual, be inhumane. For purposes of this discussion, I am assuming that conditions in protective custody, while harsh, would not be inhumane for the individual in question. In addition, to reduce an offender's prison sentence based on the foreseeable, substantial risk of psychological damage from protective custody, this harm must be capable of justification. See infra note 262.

(260) See ASHWORTH, supra note 219, at 277 (observing that some courts in England grant sentencing discounts to offenders likely to serve their sentences in protective segregation in order to reflect "the greater pains of imprisonment" to be experienced by these offenders). Indeed, some courts that have departed downwards under the U.S. Sentencing Guidelines on the basis of extreme vulnerability to abuse in prison have done so on this basis. See United States v. Lara, 905 F.2d 599, 602 (2d Cir. 1990) (observing that "the only means for prison officials to protect Morales was to place him in solitary confinement"); see also Sigler, supra note 243, at 571 (arguing, in reference to Lara, that "[i]n effect, the court found that a shorter period of time in protective custody is equivalent to a lengthier sentence in the general population"). For purposes of clarity, it is important to highlight that the sentencing discount addressed in general terms in this section does not equate substantively or procedurally to the downward departure currently available under the U.S. Sentencing Guidelines. See 18 U.S.C.A. [section] 3553(b)(1); U.S. SENTENCING GUIDELINES MANUAL [section] 5K2.0(a)(2).

(261) In addition, by granting a sentencing discount to a vulnerable inmate, a judge will make sexual assault less likely to occur and, if it does occur, the discount should make a remedy more likely. By highlighting an offender's vulnerability in a sentencing order, a judge will put prison officials on notice of the offender's susceptibility to harm. Such notice will be relevant to prison officials' liability under the Eighth Amendment. See Farmer v. Brennan, 511 U.S. 825, 828 (1994) (holding that "a prison official's 'deliberate indifference' to a substantial risk of serious harm to an inmate violates the Eighth Amendment").

(262) This Article does not address the extent to which foreseeable harms flowing from incarceration can be justified by the state's legitimate retributive aims, beyond the limited circumstance of prison violence, which can never be justified. Such foreseeable harm might--and might not--be justified through application of the doctrine of double effect, originally attributable to St. Thomas Aquinas's Summa Theologiae and summarized by Professor Warren S. Quinn as involving:

   [A] set of necessary conditions on morally permissible agency in
   which a morally questionable bad upshot is foreseen: (a) the
   intended final end must be good, (b) the intended means to it must
   be morally acceptable, (c) the foreseen bad upshot must not itself
   be willed (that is, must not be, in some sense, intended), and (d)
   the good end must be proportionate to the bad upshot (that is, must
   be important enough to justify the bad upshot).


Warren S. Quinn, Actions, Intentions, and Consequences: The Doctrine of Double Effect, 18 PHIL. & PUB. AFF. 334, 334 n.3 (1989). See generally THE DOCTRINE OF DOUBLE EFFECT: PHILOSOPHERS DEBATE A CONTROVERSIAL MORAL PRINCIPLE (P.A. Woodward ed., 2001) (presenting arguments for and against the principle of double effect). Professor Adam Kolber has argued that, while "retributivists might argue [that] many of the unintended harms of incarceration can be justified by the state's legitimate retributive intentions ... the job of justifying the side-effect harms of incarceration ... is hardly so easy." Kolber, supra note 36, at 22-23. Kolber observes, for instance, that "there are limits on the magnitude of foreseen harm that can be justified by an intended positive aim" and that, if an alternative means of achieving a legitimate objective exists that is like the original means in all respects except that it would pose a lower risk of harm to others, the person seeking the objective is morally obligated to select the alternative. See id. at 23. If, considering evidence of statistical and individualized risk, see Simons, supra note 167, a serious, foreseeable harm stemming from incarceration cannot be justified, a judge should not order a carceral penalty in a given instance, but should look to potential alternative sanctions. See Part III.B.2.

(263) Cf. Dolovich, supra note 54, at 916 (characterizing as "excruciating" the psychological harm of "being forced to live for extended periods in dread of attack," which could leave individuals "desperate to protect themselves at all costs and rob them of the ability to function in any reasoned or self-possessed way").

(264) This example serves the very limited purpose outlined in the text. Otherwise, employment by the U.S. Department of State bears little similarity to confinement in a state or federal prison. For instance, an employee of the State Department can resign at will; a prisoner, of course, has no such option.

(265) See Danger Pay Allowance (DSSR 650), Percentage of Basic Compensation, U.S. DEP'T OF STATE, http://aoprals.state.gov/Web920/danger_pay_all.asp (rates effective Feb. 24, 2013).

(266) See Frequently Asked Questions About Danger Pay, U.S. DEP'T OF STATE, http://aoprals.state.gov/content.asp?content_id=177&menu_id=78#01 (last visited Mar. 7, 2013).

(267) See id.

(268) See Overseas Compensation and Medical Benefits for Department of Justice Employees, U.S. DEP'T OF JUSTICE, http://www.justice.gov/jmd/ps/docs/overseas-comp.pdf (last visited Mar. 2, 2012).

(269) The Federal Tort Claims Act allows civil actions on claims against the United States for money damages for:

   [P]ersonal injury or death caused by the negligent or wrongful act
   or omission of any employee of the Government while acting within
   the scope of his office or employment, under circumstances where
   the United States, if a private person, would be liable to the
   claimant in accordance with the law of the place where the act or
   omission occurred.


28 U.S.C. [section] 1346(b)(1) (2006). The Act bars claims against the United States based on injuries suffered in a foreign country, however. See id. [section] 2680(k); Sosa v. Alvarez-Machain, 542 U.S. 692 (2004).

(270) See supra note 262.

(271) In some cases, judges may use "vulnerability to abuse" as a proxy for harsher prison conditions of varying sort, which may sometimes include, unfortunately and unintentionally, prison violence. Probably in light of their inability to control housing assignments and conditions of incarceration--and because sentencing is at base a predictive enterprise--sentencing judges often do not go to great lengths to delineate the many forms that harsher conditions in prison might take for an individual vulnerable to abuse. In these cases, it may be impossible to divine the predicted source of harm that motivated a sentencing discount.

(272) See Gray, supra note 32, at 1649-50; Sigler, supra note 243, at 573-74.

(273) See Reinert, supra note 244, at 1619.

(274) See Farmer v. Brennan, 511 U.S. 825, 837 (1994); see also supra note 261.

(275) See supra note 262. In addition, prison conditions must not be so degrading or dehumanizing that the penalty is rendered inhumane. See infra Part III.B. 1.

(276) See supra notes 114-134 (difficulty coping and rule violation as result of mental illness), notes 146-157 (incidence and consequences of confinement in isolation) and accompanying text. Some prisons, often in response to court orders or settlements, exclude prisoners with Axis I disorders from prolonged confinement in isolation. See supra note 156.

(277) Prisons are not unitary institutions that offer uniform experiences but, since judges have little control over placement, they can only make educated guesses about how an offender will actually experience his punishment. See supra notes 237-238.

(278) See infra note 371.

(279) See infra note 370.

(280) See infra note 369.

(281) See infra note 368.

(282) See, e.g., DeShaney v. Winnebago Dep't of Social Servs., 489 U.S. 189, 200 (1989) ("[W]hen the State by the affirmative exercise of its power so restrains an individual's liberty that it renders him unable to care for himself, and at the same time fails to provide for his basic human needs--e.g., food, clothing, shelter, medical care, and reasonable safety--it transgresses the substantive limits on state action set by the Eighth Amendment and the Due Process Clause."); Estelle v. Gamble, 429 U.S. 97, 104 (1976) ("[D]eliberate indifference to serious medical needs of prisoners constitutes the 'unnecessary and wanton infliction of pain,' proscribed by the Eighth Amendment." (internal citation omitted)); Ramos v. Lamm, 639 F.2d 559, 575 (10th Cir. 1980) ("A medical need is serious if it is 'one that has been diagnosed by a physician as mandating treatment or one that is so obvious that even a lay person would easily recognize the necessity for a doctor's attention.'" (internal citation omitted)); Bowring v. Godwin, 551 F.2d 44, 47 (4th Cir. 1977) (holding that an inmate is "entitled to psychological or psychiatric treatment if a physician or other health care provider, exercising ordinary skill and care at the time of observation, concludes with reasonable medical certainty (1) that the prisoner's symptoms evidence a serious disease or injury; (2) that such disease or injury is curable or may be substantially alleviated; and (3) that the potential for harm to the prisoner by reason of delay or the denial of care would be substantial").

(283) Professor von Hirsch has developed and refined his sentencing theory, within an expressive framework, over the course of four books: DOING JUSTICE, supra note 193, PAST OR FUTURE CRIMES: DESERVEDNESS AND DANGEROUSNESS IN THE SENTENCING OF CRIMINALS (1985) [hereinafter PAST OR FUTURE CRIMES], CENSURE AND SANCTIONS (1993) [hereinafter CENSURE], and, with Andrew Ashworth, PROPORTIONATE SENTENCING, supra note 27.

(284) See von Hirsch, Proportionality, supra note 177, at 278-79 (outlining three implications and concluding that they are supported by multiple versions of expressive theory).

(285) VON HIRSCH & ASHWORTH, supra note 27, at 135.

(286) Id.

(287) See id. at 138-41. Ordinal proportionality is sometimes referred to as "proportionality simpliciter" or "relative proportionality." See Greg Roebuck & David Wood, A Retributive Argument Against Punishment, 5 CRIM. L. & PHIL. 73, 76-77 (2011).

(288) VON HIRSCH & ASHWORTH, supra note 27, at 135.

(289) See id at 4, 186.

(290) Id. at 139-40; see also von Hirsch, Proportionality, supra note 177, at 282.

(291) VON HIRSCH & ASHWORTH, supra note 27, at 139-40.

(292) Id. at 140; see also von Hirsch, Proportionality, supra note 177, at 282.

(293) See VON HIRSCH, supra note 193, at 90; see also Roebuck & Wood, supra note 287, at 76-77.

(294) See JEFFRIE G. MURPHY, RETRIBUTION, JUSTICE, AND THERAPY 235 (1979) ("A theory of justice ... can at least demand the following: that everyone has the right to have offenses graded in terms of individual fault or blameworthiness (i.e. desert) and not mere social utility, that other even substantive bases for grading be reasonable, that punishments be graded on a comparable basis, and that there be a matching between seriousness of punishment and seriousness of offense."); von Hirsch, Proportionality, supra note 177, at 282 ("These requirements of ordinal proportionality are not mere limits, and they are infringed when persons found guilty of equally reprehensible conduct receive unequal sanctions on crime preventive grounds."). Limiting retributivists, on the other hand, view ordinal proportionality as a "weak" or "limiting" principle. See, e.g., JOHN BRAITHWAITE & PHILIP PETTIT, NOT JUST DESERTS: A REPUBLICAN THEORY OF CRIMINAL JUSTICE 101-02, 126 (1990); NICOLA LACEY, STATE PUNISHMENT: POLITICAL PRINCIPLES AND COMMUNITY VALUES 193-95 (1988); NORVAL MORRIS, MADNESS AND THE CRIMINAL LAW 151 (1982); Michael Tonry, Obsolescence and Immanence in Penal Theory and Policy, 105 COLUM. L. REV. 1233, 1263-68 (2005); Michael Tonry, Purposes & Functions of Sentencing, 34 CRIME & JUST. 1, 19-20 (2006) (identifying three difficulties with the concept of ordinal proportionality). For a detailed description and critique of limiting retributivism's treatment of ordinal proportionality, see VON HIRSCH & ASHWORTH, supra note 27, at 180-85 (comparing and contrasting their modified desert proposal to Morris's limiting retributivism).

(295) See Roebuck & Wood, supra note 287, at 77.

(296) See VON HIRSCH, PAST OR FUTURE CRIMES, supra note 283, at 39 (defining cardinal magnitudes as "what absolute levels of severity should be chosen to anchor the penalty scale"); id. at 43-46 (discussing the role of desert in determining cardinal magnitudes); see also Hugo Adam Bedau, Classification-Based Sentencing: Some Conceptual and Ethical Problems, 10 NEW ENG. J. ON CRIM. & CIV. CONFINEMENT 1, 13 (1984) (exploring the distinction between ordinal and cardinal punishment).

(297) See, e.g., Ristroph, supra note 218, at 1308-09 (arguing that notions of desert are indeterminate and highly elastic); Robinson, supra note 26, at 164-67 (observing that moral philosophers lack agreement as to how to translate an offender's moral blameworthiness into a specific punishment).

(298) See von Hirsch, Proportionality, supra note 177, at 282.

(299) See Markel & Flanders, supra note 8, at 954-58 (identifying as an "island of

agreement" with subjectivists the idea that retributive justice must be sensitive to the experience of punishment in ensuring that punishments sufficiently communicate condemnation and are not excessive or cruel).

(300) See VON HIRSCH, supra note 193, at 111 n. * (rejecting corporal punishment as a permissible form of punishment because it "evokes in its victim intense feelings of humiliation and terror," and asking whether there is "a right to the integrity of one's own body, that not even the state's interests in punishing may override"); JOHN KLEINIG, PUNISHMENT AND DESERT 123 (1973) ("[T]here is a limit to the severity of the punishment which can be humanely inflicted upon a wrongdoer. What these limits are is of course a matter for debate, to be decided partly by recourse to normative considerations."); MURPHY, supra note 294, at 233 ("A punishment will be unjust (and thus banned on principle) if it is of such a nature as to be degrading or dehumanizing (inconsistent with human dignity). The values of justice, rights and desert make sense, after all, only on the assumption that we are dealing with creatures who are autonomous, responsible, and deserving of the special kind of treatment due that status."); infra notes 307-312.

(301) See VON HIRSCH, PAST OR FUTURE CRIMES, supra note 283, at 53 ("If the state is to carry out the authoritative response to [wrongful] conduct--as it must if it visits any kind of sanction upon its perpetrators--then it should do so in a manner that testifies to the recognition that the conduct is wrong.").

(302) See VON HIRSCH, CENSURE, supra note 283, at 37.

(303) Id.

(304) See Anthony Bottoms, Five Puzzles in von Hirsch's Theory of Punishment, in FUNDAMENTALS OF SENTENCING THEORY, supra note 178, at 53, 61 n.26; VON HIRSCH, PAST OR FUTURE CRIMES, supra note 283, at 39.

(305) See, e.g., HUDSON, supra note 181, at 51 (discussing the moral theory of Immanuel Kant and characterizing it as resting "on a model of the human as someone whose actions are the result of moral choices"); Jeffrie G. Murphy, Marxism and Retribution, 2 PHIL. & PUB. AFF. 217, 217, 231 (1973) (outlining Immanuel Kant's theory of punishment with an emphasis on its manifestation of respect for dignity, autonomy, rationality, and rights).

(306) WON HIRSCH & ASHWORTH, supra note 27, at 76.

(307) See Margaret Jane Radin, The Jurisprudence of Death: Evolving Standards for the Cruel and Unusual Punishments Clause, 126 U. PA. L. REV. 989, 1054 (1978) ("Because the value underlying modern retributivism is to treat people with the concern and respect due persons, a punishment that violated our current conception of human dignity could not be justified on retributivist grounds.").

(308) See MURPHY, supra note 294, at 233 (decrying "a punishment which is in itself degrading, which treats the prisoner as an animal instead of a human being, which perhaps even is an attempt to reduce him to an animal or a mere thing" as inconsistent with human dignity); Herbert Morris, Persons and Punishment, 52 MONIST 475, 490 (1968) ("When we treat a human being merely as an animal or some inanimate object our responses to the human being are determined, not by his choices, but ours in disregard of or with indifference to his. And when we 'look upon' a person as less than a person or not a person, we consider the person as incapable of rational choice."). (309) Morris, supra note 308, at 488.

(310) See Richard L. Lippke, Arguing Against Inhumane and Degrading Punishment, 17 CRIM. JUST. ETHICS 29, 36-37 (1998) ("The state is not permitted to attempt to seize control of the moral personalities of offenders, nor to manipulate their personalities in ways that preclude them from forming and acting on their own judgments about the sanctions being inflicted on them.... These forms of treatment would interfere with the important retributive requirement that offenders be able to comprehend their punishments as justifiable losses or deprivations imposed on them for their past misconduct."); Herbert Morris, A Paternalistic Theory of Punishment, 18 AM. PHIL. Q. 263, 270 (1981) ("Punishments will not be permitted that destroy in some substantial way one's character as an autonomous creature. Certain cruel punishments, then, may be ruled out, not merely because they are conducive to hardening the heart but, more importantly, because they destroy a good that can never rightly be destroyed.").

(311) See Markel & Flanders, supra note 8, at 957-58.

(312) See Radin, supra note 307, at 1047 ("[I]t appears that retributivist systems define dignity coextensively with permissible punishment, with the result that all violations of human dignity are inherently excessive."). Under a lex talionis perspective, however, some would argue that punishments can be proportionate yet inhumane. See Bedau, supra note 296, at 17-18. In this situation, moral principles of humane treatment would serve as an external constraint on proportionality. See id.

(313) As previously noted, offenders' experiences in prison will vary by the facilities to which they are assigned and their levels of security. See supra note 238 and accompanying text.

(314) When prison conditions present a risk of harm great enough to warrant a designation as "inhumane" is a tricky issue, and any assumptions here are controversial and will be difficult to unpack and defend. See, e.g., KLENIG, supra note 300, at 123 (observing that limits of humanity are "of course a matter for debate, to be decided partly by recourse to normative considerations"); Arnold S. Kaufman, The Reform Theory of Punishment, 71 ETHICS 49, 52 (1960) (arguing that "[t]he conditions that exist in many of the prisons of even the most civilized nations are degrading and barbaric" and that "such conditions violate those moral rights a criminal retains even inside prison walls"); Rod Morgan, Not Just Prisons: Reflections on Prison Disturbances, 13 POL'Y STUD. 4, 6 (1992) (observing that the Council of Europe Committee for the Prevention of Torture has adopted "a cumulative view of adverse prison conditions," finding that the combination of overcrowding, lack of integral sanitation, and lack of out-of-cell activities results in inhumane and degrading treatment). In the context of the Eighth Amendment's Cruel and Unusual Punishment Clause, to warrant relief, "the inmate must show that he is incarcerated under conditions posing a substantial risk of serious harm." Farmer v. Brennan, 511 U.S. 825, 834 & n.3 (1994); see also id. at 834 & n.3 (stating that "the deprivation alleged must be, objectively, 'sufficiently serious'; a prison official's act or omission must result in the denial of 'the minimal civilized measure of life's necessities,'" but leaving for another day the question of "[a]t what point a risk of inmate assault becomes sufficiently substantial for Eighth Amendment purposes" (internal citation omitted)). In addition, the inmate must establish that a prison official was subjectively aware of the risk and failed to take reasonable measures to abate it. See id. at 828 (holding that "a prison official's 'deliberate indifference' to a substantial risk of serious harm to an inmate violates the Eighth Amendment" and defining "deliberate indifference" as "requiring a showing that the official was subjectively aware of the risk"). Arguably, retributive theory, with its focus on justice, morality, the dignity of the offender, and proportionate punishment, can and should be more sensitive to risk of physical and psychological harm than current Eighth Amendment jurisprudence. See Andrew von Hirsch & Uma Narayan, Degradingness and Intrusiveness, in CENSURE, supra note 283, at 80, 81 ("[W]hereas the Eighth Amendment to the US Constitution (as now construed) scarcely may outlaw even the most grossly disproportionate punishments, a fair system of punishment should observe more stringent proportionality requirements. The same point should hold for the present issue of degrading punishments."); Barry Pollack, Deserts and Death: Limits on Maximum Punishment, 44 RUTGERS L. REV. 985, 988 n.17 (1992) (stating that, while "some desert arguments no doubt equate to constitutional arguments," desert and analysis under the Eighth Amendment must be kept distinct because "the desert argument is a matter of policy, not constitutionality").

(315) See Dolovich, supra note 54, at 3-4; Robertson, supra note 257, at 126 (noting that inmates in protective custody spend twenty-two hours in isolation per day).

(316) Assault of inmates while housed in protective custody is not unheard of. See Little v. Walker, 552 F.2d 193, 197 (7th Cir. 1977); PRISON RAPE ELIMINATION REPORT, supra note 102, at 79 (detailing the case of an inmate who was double-celled with, then raped by, a convicted sex offender while in protective custody).

(317) In reality, prisons utilize, to varying extent, a range of short-term and long-term housing options for offenders with serious mental illness, especially those in crisis. See NAT'L INST. OF CORR., PROVISION OF MENTAL HEALTH CARE IN PRISONS 5-6 (2001), available at http://nicic.gov/Library/016724. One option employed by many prisons consists of transferring seriously mentally ill prisoners to psychiatric hospitals during acute episodes. See id.

(318) See supra Part III.B.2.

(319) Some jurisdictions have created segregated residential facilities for vulnerable populations, apparently with some success. See Dolovich, supra note 258, at 44. In these jurisdictions, prison officials make housing determinations. I am grateful to Professor Jerold Israel for suggesting that a reconfigured "guilty but mentally ill" verdict could serve as a segregating mechanism and possible means to eliminate the disproportionate portions of an offender's sentence.

(320) See Robinson, supra note 26, at 151 ("If the endpoint of the punishment continuum changes, the amount of punishment that an offender deserves under [deontological and empirical] conceptions of justice also changes, to the amount of punishment necessary to keep it in its proper ordinal rank."). This would be the case if incarceration under currently existing conditions were not an inhumane penalty for other offenders. However, it may well be that the risk of physical and psychological harm to offenders with major mental illnesses does not differ substantially from (or may even be less than) that experienced by other vulnerable prison populations, such as gay or transgendered inmates, those of diminutive stature, or those with mental retardation or physical disabilities. See supra note 169 (detailing risk factors for sexual violence identified by the Attorney General in May 2012). There is a shortage of data on how the risks of physical or psychological harm for offenders with serious mental illnesses differ from the risks faced by these subgroups. See Wolff, supra note 159, at 158. It also could be persuasively argued that long-term incarceration--with widespread risk of physical and psychological harm--is currently morally intolerable for any offender. See MURPHY, supra note 294, at 239-40 ("Studies on the effects of long-term term incarceration in 'total institutions' indicate that long-term confinement develops in persons an 'institutional personality'--i.e, a personality with diminished affect, neurotic dependencies, loss of autonomy and mental competence generally: in short, a kind of death (of personhood). If these studies are correct, then long-term incarceration will be a kind of slow torture and psychic mutilation and should no doubt be banned on Eighth Amendment grounds...."); Melvin Gutterman, The Contours of the Eighth Amendment Prison Jurisprudence: Conditions of Confinement, 48 SMU L. REV. 373, 404-07 (1995) (considering whether the cumulative effect of physical and psychological health risks posed by contemporary prisons may rise to the level of an Eighth Amendment violation); cf. Eva S. Nilsen, Decency, Dignity, and Desert: Restoring Ideals of Humane Punishment to Constitutional Discourse, 41 U.C. DAVIS L. REV. 111, 113-34, 157-74 (2007) (arguing that prison sentence length, prison conditions, and poor postrelease reintegration should have greater constitutional significance and advocating for a more robust concept of human dignity in the context of the Eighth Amendment). If this is the case, then the long-term incarceration of any offender should be disallowed until these risks are eliminated or brought within acceptable levels.

(321) Professor John Kleinig has suggested that the most severe penalty deemed morally permissible should anchor a penalty scale. See KLEINIG, supra note 300, at 124. For critiques of this proposal, see VON HIRSCH, PAST OR FUTURE CRIMES, supra note 283, at 44 n. *; Don E. Scheid, Theories of Legal Punishment 173-82 (1977) (Ph.D. dissertation, New York University).

(322) See VON HIRSCH, PAST OR FUTURE CRIMES, supra note 283, at 92 ("Anchoring points are needed that begin to establish the levels of severity appropriate for given degrees of blameworthiness. Otherwise, the crime-seriousness rankings and the punishment scale will 'float' independently of each other.").

(323) See id. at 44 ("Once ... the magnitude and anchoring points of the scale have been chosen (with whatever uncertainties this choice involves), then the internal scaling requirements of proportionality--the ordinal requirements--become binding.").

(324) This would be true assuming that the jurisdiction decided not to impose greater maximum penalties--exceeding in severity incarceration in the segregated facility for forty years--through the use of alternative sanctions. See supra Part III.B.2.

(325) Again, this is assuming that no terms of incarceration are inhumane, under currently existing conditions, for other offender populations. See supra note 320.

(326) Cf infra note 352 (describing the broader potential of this analysis for the sentencing of offenders with serious mental illnesses). Scholars have questioned whether noncarceral sanctions are capable of communicating the necessary censure for the most serious crimes. See, e.g., VON HIRSCH, supra note 193, at 111 ("One reason for preferring incarceration is simply that we have not found another satisfactory severe punishment.").

(327) See, e.g., id. at 118-23; MORRIS & TONRY, supra note 235, at 11-12.

(328) See Andrew von Hirsch & Nils Jareborg, Gauging Criminal Harm: A Living Standard Analysis, 11 OXFORD J. LEGAL STUD. 1, 17-23, 28-32 (1991); see also VON HIRSCH & ASHWORTH, supra note 27, at app. 3 (offering a revised and expanded version of the living standard analysis in the context of offense seriousness).

(329) VON HIRSCH & ASHWORTH, supra note 27, at 194.

(330) See Andrew von Hirsch, Seriousness, Severity and the Living Standard, in PRINCIPLED SENTENCING, supra note 183, at 143, 146 ("What makes punishments more or less onerous is not any identifiable sensation; rather, it is the degree to which those sanctions interfere with people's interests. The unpleasantness of intensive probation supervision, for example, depends not on its 'feeling bad' in some immediate sense, but on its interfering with such important interests as being in charge of one's own life or moving about as one chooses.").

(331) See id.

(332) See id.

(333) See Andrew von Hirsch, Reduced Penalties for Juveniles: The Normative Dimension, in PRINCIPLED SENTENCING, supra note 183, at 323, 327.

(334) See VON HIRSCH & ASHWORTH, supra note 27, at 147; von Hirsch, supra note 330, at 144 ("Most victimizing offences involve one or more of the following interest-dimensions: (i) physical integrity; (ii) material support and amenity; (iii) freedom from humiliation; and (iv) privacy."); id. at 147 n.4 ("To apply the living-standard idea to penalties, there would have to be modifications in the analysis. When evaluating harms, the main interests are ... those of physical integrity, material amenity and so forth. For punishments, other interests also need to be taken into account: for example, the interest in freedom of movement that is affected by incarceration, home detention and intensive probation supervision."); Thornburn & Manson, supra note 186, at 284-85 (reviewing VON HIRSCH & ASHWORTH, supra note 27). This list was not derived from "deep theory", but rather, from "impressions of the main kinds of concerns that seem typically involved in victimizing crimes" and penalties, VON HIRSCH & ASHWORTH, supra note 27, at 147, 205, and scholars have criticized the list as arbitrary. See JACEK CZBANSKI, ESTIMATES OF COST OF CRIME 67 (2007); Mirko Bagaric & James McConvill, Giving Content to the Principle of Proportionality: Happiness and Pain as the Universal Currency for Matching Offense Seriousness and Penalty Severity, 69 J. CRIM. L. 50, 59 (2005) ("[T]he problem with their ranking system is that despite the fact that they concede that their analysis is normative, since it is a theory on how harms ought to be rated, it is devoid of an underlying rationale or an empirical or scientific foundation--it is built on armchair speculation.").

(335) See von Hirsch, supra note 330, at 145 ("One might use three living-standard levels: (i) subsistence; (ii) minimal well-being; and (iii) 'adequate' well-being. The first, subsistence, refers to survival, but with maintenance of no more than elementary human capacities to function--in other words, barely getting by. The remaining levels refer to various degrees of life quality above that of mere subsistence.").

(336) VON HIRSCH, CENSURE, supra note 283, at 34 (emphasis added).

(337) Id. (emphasis added); see also id. at 35 n.* ("What matters is how much a given penalty does affect the ordinary person's living standard. This is a matter, not of surveying what people think, but of analysing the effects of the penalty on the quality of persons' lives."). To this end, von Hirsch suggests that research be conducted into "what interests penalties intrude upon, how those intrusions would affect the quality of life, and why so." Id. If cognizable harm were circumscribed by the intent of the sentencing judge or legislature, then such an inquiry into actual effect would be unnecessary.

(338) Id. at 35 n.*.

(339) See VON HIRSCH & ASHWORTH, supra note 27, at app. 3, 186, 201-08, 212-15.

(340) See id. at 194-96.

(341) See id. at 205 (instructing that, to perform a living standard analysis in the context of a criminal offense, "first identify and separate out the interest dimensions involved in an offense"). For illustrations of the living standard analysis in the context of several crimes and to see how harm is evaluated through the perspective of the typical victim, see id. at 208-12.

(342) See id. at 186-87, 215. But see id. at 206 n.g (noting a degree of overlap between harm and culpability, in the interest of freedom from humiliation, because humiliation normally presupposes at least apparent intent).

(343) See VON HIRSCH, CENSURE, supra note 283, at 34-35.

(344) See id. at 35; see also VON HIRSCH & ASHWORTH, supra note 27, at 147 (rejecting a "subjectivist conception of severity" that would "depend[] upon how disagreeable the sanction typically is experienced as being" and asserting that "[w]hat makes punishments more or less onerous is not any identifiable sensation; rather, it is the degree to which those sanctions interfere with people's legitimate interests").

(345) See VON HIRSCH & ASHWORTH, supra note 27, at 188.

(346) Id. at 189.

(347) Of course, the empirical evidence at this point may not suffice to identify precisely which offenders with serious mental illnesses are most likely to suffer serious harms in prison. See supra text accompanying notes 159-168 (sketching limitations in available data).

(348) See supra Part III.B.3 (explaining how a living standard analysis may be used to effectuate proportionate punishment for offenders with serious mental illnesses under the equal-impact principle).

(349) WON HIRSCH & ASHWORTH, supra note 27, at 173 n.f ("Certain kinds of non-standard cases ... [in an interest analysis for gauging severity of punishments] can themselves be analysed in terms of typical impacts for the members of that group. One thus could analyse how imprisonment typically becomes more burdensome for those with specified physical disabilities.").

(350) See id. at 41-42 (defining "punitive bite" as "the extent to which those sanctions interfere with important interests that people have"); supra note 235. For a discussion of the potential scope of the living standard analysis in application, see infra note 353.

(351) See, e.g., ALA. CODE [section] 13A-12-213 (LexisNexis 2005) (first degree possession of marijuana); Id. [section] 13A-5-6 (sentences for felony convictions).

(352) A living standard analysis holds great potential for the sentencing of offenders with serious mental illnesses beyond this limited scenario. For instance, if a legislature used a living standard analysis to identify a slate of penalties of roughly equivalent punitive bite for most combinations of criminal offenses and criminal history scores (or adopted similar recommendations offered by a sentencing commission), and granted judges the discretion to select among these sentencing options for offenders with serious mental illnesses (in light of the empirical risks of serious harm posed by incarceration detailed in Part II), then these offenders could receive noncarceral sanctions of equivalent punitive bite even if incarceration would not reach the threshold of inhumanity. Cf infra text accompanying note 354 (observing that the sanction of incarceration may be reserved for the most serious offenses). Indeed some states, in their sentencing guidelines grids, authorize community sanctions or stayed sentences of confinement for the least serious offenses, for all offenders. See, e.g., MINNESOTA SENTENCING GUIDELINES, pt. 4.A (2012), available at http://www.msge.state.mn.us/guidelines/grids/2012%20MN%20Sentencing%20Guidelines %20Grid.pdf (indicating combinations of offenses and criminal history scores with presumptive stayed sentences).

(353) For a discussion of scaling principles and grids of sanctions of comparable severity see, e.g., MORRIS & TONRY, supra note 235 (relying on limiting retributivism to propose the establishment of "exchange rates" to achieve interchangeability between punishments); Andrew von Hirsch et al., Punishments in the Community and the Principles of Desert, 20 RUTGERS L.J. 595 (1989) (applying principles of desert to the choice among noncustodial penalties, and advocating for limited substitutability and the ranking of penalties based on degree of intrusion on offenders' interests); Paul Robinson, A Sentencing System for the 21st Century?, 66 TEX. L. REV. 1, 55 (1987) (providing a sentencing grid that prescribes "sanction units" of comparative punitive bite); Martin Wasik & Andrew von Hirsch, Non-Custodial Penalties and the Principles of Desert, 1988 CRIM. L. REV. 555 (applying principles of desert to the choice among noncustodial penalties and advocating for limited substitution).

(354) See supra note 326.

(355) Cf Part III.B.1.

(356) VON HIRSCH & ASHWORTH, supra note 27, at 131.

(357) Id. at 172; see also Ashworth & Player, supra note 178, at 253 ("We would argue for a general principle of equal treatment, by which we mean that a sentencing system should strive to avoid its punishments having an unequal impact on different offenders or groups of offenders. It is a principle with similar roots to proportionality, in that it seeks to respect individuals by ensuring fair treatment."). The roots of the equal-impact principle can be found in the writing of Jeremy Bentham. See JEREMY BENTHAM, PRINCIPLES OF MORALS AND LEGISLATION, ch. XIV, para. 14 (1789) (articulating a principle of equal impact); see also id. at ch. VI, para. 6 (delineating "circumstances influencing sensibility"). The principle of equal impact is one of several (sometimes competing) values that, as a normative matter, arguably should inform general sentencing policy or individual sentencing decisions. See ANDREW ASHWORTH, SENTENCING AND CRIMINAL JUSTICE 95-100 (5th ed. 2010); ASHWORTH, supra note 219, at 277-78. Other principles with normative claims in sentencing include rule-of-law values, parsimony and restraint in the use of custody, economy, and equality before the law. See ASHWORTH, supra, at 95-100 (listing and briefly describing these principles).

(358) See Ashworth & Player, supra note 178, at 253 (equal treatment, nondiscrimination); id. at 255 (fairness); Andrew Ashworth & Julian Roberts, Doing Justice to Difference: Diversity and Sentencing, in PRINCIPLED SENTENCING, supra note 183, at 342, 343-44 (equality, nondiscrimination).

(359) See ASHWORTH, SENTENCING, supra note 357, at 295 ("When the principle of equal impact is invoked, the decision is often based on assumptions about the typical reaction to the sentence of persons in a class to which the offender belongs (e.g. the elderly or the young) or of persons placed in the situation in which this offender finds himself (e.g. segregated in prison under [rules concerning protective custody]...."); VON HIRSCH & ASHWORTH, supra note 27, at 172.

(360) See VON HIRSCH & ASHWORTH, supra note 27, at 173 ("The equal-impact principle does not actually function as true equity mitigation, because it actually does not call for qualifying defendants to suffer less punishment: it merely is a way of avoiding that such defendant be made to suffer more."); von Hirsch, supra note 333, at 323, 327; supra note 235.

(361) See ASHWORTH, supra note 219, at 275 (arguing that, to avoid "the result that objectively similar sentences will have a subjectively variable effect on offenders[,] ... it would be necessary to travel outside proportionality and to adopt equality of impact as a principle, so as to take account of manifest differences between offenders which affect the degree of pain and deprivation caused to them by particular sentences").

(362) See Ashworth & Roberts, supra note 358, at 342, 345-46 ("Desert theories can plausibly claim that the principle of equal treatment forms part of their rationale: sentences should be determined chiefly by the seriousness of the offence...."); Ashworth & Player, supra note 178, at 255 (arguing that "those theories which have some requirement of proportionality of sentence to the seriousness of the offence must surely concern themselves with this problem [of equal impact]"); see also von Hirsch, supra note 333, at 328 (arguing that juveniles should be punished less in part because punishments are more onerous for them, but stating: "This still assumes, however, that the conventions linking severity and seriousness are unchanged. Where the crimes (adjusting for culpability factors) have similar seriousness-ratings, and where the penalties (adjusting for juveniles' greater vulnerability) should have the same severity-ratings, then juveniles and adults would receive equivalent punishments."); cf. VON HIRSCH & ASHWORTH, supra note 27, at 172 (asserting that, although the "'equal impact' principle is connected with the proportionalist sentencing model, [it] is not part of it in standard cases" and that its use should be reserved for "unusual cases that diverge significantly from the norm").

(363) See supra notes 336-343 and accompanying text.

(364) See VON HIRSCH & ASHWORTH, supra note 27, at 176 (considering "certain foreseeable differential impacts" that incarceration poses to offenders with particular handicaps).

(365) Id. at 172, 176 (explaining that the aim of the equal-impact principle, when applied in the case of an offender with a physical handicap, is to make "adjustments in sentence to deal with certain foreseeable differential impacts").

(366) Id. at 172-73.

(367) Id. at 173.

(368) See ASHWORTH, supra note 357, at 100; VON HIRSCH & ASHWORTH, supra note 27, at 173 ("[The equal-impact principle] would work for old, ill, or disabled defendants, where the sanction appears to be altered in its impact by the person's disability...."); Ashworth & Player, supra note 178, at 255 ("Many mentally disordered offenders may find the experience of imprisonment significantly more painful than others.... We would argue that fairness requires a recognition that the same sentence may have a disproportionately severe impact on certain offenders, and that only if one adopts a principle of equal impact can this problem be minimized.").

(369) See VON HIRSCH & ASHWORTH, supra note 27, at 172-73, 176.

(370) See id. at 173, 176; Ashworth & Player, supra note 178, at 259-60.

(371) See VON HIRSCH & ASHWORTH, supra note 27, at 42-43; Andrew Ashworth, Sentencing Young Offenders, in PRINCIPLED SENTENCING, supra note 183, at 294, 300 (asserting that "we should recognise that punishments are generally more onerous for the young because they impinge on important developmental interests, in terms of education and socialisation, for which the teenage years are a crucial phase" and that "[p]roportionality theory thus requires that sentence levels be significantly lower than those for adults...."); von Hirsch, supra note 333, at 323 (arguing that, "in applying a policy of proportionate sentencing to juveniles, substantial overall penalty reductions are called for" in part due to "criminal sanctions' greater 'punitive bite' when applied to juveniles"); see also Barry C. Feld, A Century of Juvenile Justice: A Work in Progress or a Revolution that Failed?, 34 N. KY. L. REV. 189, 247-48 (2007) (concluding that "[s]tates should adopt a categorical 'youth discount' and sentence youths based on a sliding scale of criminal responsibility"). Scholars have also argued that youthful offenders deserve categorical age-related discounts because of their diminished capacity to assess and appreciate the harmful consequences of their actions and their reduced volitional controls. See VON HIRSCH & ASHWORTH, supra note 27, at 36-40.

(372) My analysis depends upon subscription to a broad understanding of punishment that includes foreseeable, substantial risks of suffering serious harms, proximately caused by the state in the context of imprisonment. See supra III.A.

(373) See VON HIRSCH & ASHWORTH, supra note 27, at 203-04 (identifying "preservation of one's major physical and cognitive functions, and preservation of a minimal capacity of social functioning" as necessary for minimal subsistence and "[p]rotection against grossly demeaning or insulting treatment" as inherent to minimal well-being). Some have argued that sentence mitigation is particularly appropriate when incarceration exacerbates a current medical condition. See Ashworth & Player, supra note 178, at 259 ("A distinction can be drawn between the effect of a medical condition on the experience of imprisonment, and the effect of imprisonment on the medical condition.... The justification for reducing the length of a custodial sentence because of an offender's medical state is arguably more compelling if ... there is evidence that imprisonment has a deleterious effect on that condition."). In addition, an offender with a major mental disorder may be more likely to experience violations of his physical integrity, including health, safety, and the avoidance of pain. However, for reasons explained in notes 243-247 and 271-272, and accompanying text, an offender's increased likelihood of physical assault should not be used as the basis of a sentencing discount. This argument applies within an equal-impact framework as well.

(374) See supra note 282.

(375) See von Hirsch, supra note 333, at 327 (arguing that young people "have certain development interests"--"critical opportunities and experiences that need be provided between the ages of 14 and 18," such as adequate schooling in a reasonably nurturing atmosphere with exposure to adequate role models "in order to mature adequately"--and that incarceration is more onerous for juveniles because of its intrusion upon these interests) (emphasis omitted).

(376) The fact that community mental health care is inadequate does not detract from the interest offenders with serious mental disorders have in receiving such treatment when confined. It should be noted, however, that for many offenders, prison offers an opportunity to receive mental health care that they were not receiving in the community. See H. Richard Lamb & Linda E. Weinberger, The Shift of Psychiatric Inpatient Care from Hospitals to Jails and Prisons, 33 J. AM. ACAD. PSYCHIATRY & L. 529, 531 (2005); E. Fuller Torrey, Editorial, Jails and Prisons--America's New Mental Hospitals, 85 AM. J. PUB. HEALTH 1611, 1611 (1995). Other sentencing options--including those provided by mental health courts--may be more treatment-friendly than jails and prisons. See Jacques Baillargeon et al., Psychiatric Disorders and Repeat Incarcerations: The Revolving Prison Door, 166 AM. J. PSYCHIATRY 103, 107-08 (2009) (reviewing prebooking and postbooking diversion programs).

(377) See supra notes 61-62.

(378) Cf von Hirsch, supra note 333, at 328 ("If punishments are thus more onerous when undergone by juveniles, proportionality would require that they be reduced.").

(379) See supra Part III.B.2. To the extent that noncarceral sanctions pose foreseeable, substantial risks of serious harm, proximately caused by the state to offenders with serious mental illnesses, this foreseeable harm would be factored into the living standard analysis through the process outlined above. The extent to which any sanction other than imprisonment poses foreseeable harm to offenders with serious mental illnesses is an empirical question beyond the scope of this Article.

(380) See supra notes 167-170 (discussing the fact that a showing of individualized risk of serious harm, in addition to mere statistical risk, may be necessary before a sentencing accommodation is warranted). An important additional caveat is that incarcerating the individual must not be inhumane.

(381) See supra note 326.

(382) See VON HIRSCH & ASHWORTH, supra note 27, at 172; see also Ashworth & Player, supra note 178, at 253.

(383) At least two scholars have proposed that vulnerability to "significant physical and mental abuse from which the prison authorities fail to provide adequate protection" merits consideration under the equal-impact principle. See Ashworth & Player, supra note 178, at 260.

(384) See supra notes 243-249, 271-274 and accompanying text.

(385) See supra notes 262-269 and accompanying text. A living standard analysis would include anxiety justifiably arising from threats to personal safety. See VON HIRSCH & ASHWORTH, supra note 27, at 208.

(386) See supra note 262 and accompanying text. Again, this Article does not address the question of whether foreseeable harms stemming from prison conditions, beyond physical and sexual violence, may ever be justified by valid retributive aims. See id.

(387) Just desert theorists have not provided extensive detail on how to calculate sentencing discounts. See Ashworth & Player, supra note 178, at 270-71 (anticipating some problems with the implementation of sentencing discounts); Austin Lovegrove, Proportionality Theory, Personal Mitigation, and the People's Sense of Justice, 69 CAMBRIDGE L.J. 321, 324 (2010) ("The drawing of conclusions about the potential effect of personal mitigation is not easy, since von Hirsch and Ashworth speak in only general terms about it.").

(388) VON HIRSCH &; ASHWORTH, supra note 27, at 172.

(389) For instance, Professors von Hirsch, Ashworth, and Jareborg have identified "preservation of one's major physical and cognitive functions," "preservation of a minimal capacity of social functioning," and "avoidance of intense pain" as necessary for minimal subsistence. See id. at 203. Impairment of these interests would pose a grave threat to an individual's well-being and should be reflected in a living standard analysis. See id. In addition, necessary to individuals' "minimal wellbeing" are "a minimum opportunity for self-respect" and "[p]rotection against grossly demeaning or insulting treatment." Id. at 204. Thus, at least to some extent, just desert theorists' gaze appears focused on accounting for risks of serious harm.

(390) See, e.g., Ashworth & Player, supra note 178, at 255 ("Many mentally disordered offenders may find the experience of imprisonment significantly more painful than others.").

(391) This Article assumes that the substantiality of the risk of harm is not in question for offenders with serious mental illnesses. See supra note 54.

(392) VON HIRSCH & ASHWORTH, supra note 27, at 176.

(393) Id. The lack of clarity as to whether punishment should contemplate nonserious harms probably stems from the origin of the living standard analysis. In the context of gauging the harms exacted by criminal offenses, it is fully justifiable--and useful--to compare the harms effected by criminal offenses by what interests, critical to individuals' quality of life, they disturb. But punishment, simply put, may not be concerned with all the ways in which hard treatment foreseeably affects offenders' means or capabilities for achieving a good life.

(394) See, e.g., Stephen P. Garvey, Questions of Mercy, 4 OHIO ST. J. CRIM. L. 321,321-23 (2007) (outlining four major positions on the relationship of mercy to justice); Mary Sigler, Contradiction, Coherence, and Guided Discretion in the Supreme Court's Capital Sentencing Jurisprudence, 40 AM. CRIM. L. REV. 1151, 1153 n.14 (2003) ("Although 'mercy' is commonly invoked in the sentencing context to refer simply to leniency in punishment, its role in legal justice is actually ambiguous and complicated. One issue is whether mercy is a feature of justice or an independent value that potentially conflicts with justice.").

(395) Lucia Zedner, Sentencing Young Offenders, in FUNDAMENTALS OF SENTENCING THEORY, supra note 178, at 165, 173, 174. Scholars have differed in how they conceive of mercy as responding to likely offender hardship. See, e.g., Claudia Card, On Mercy, 81 PHIL. REV. 182, 184 (1972) (arguing that an offender deserves mercy, but has no right to it, when otherwise (l) he would suffer unusually more than he deserves, and (2) he would be worse off than those who would benefit from the punishment); David Dolinko, Some Naive Thoughts About Justice and Mercy, 4 OHIO ST. J. CRIM. L. 349, 354 (2007) (suggesting that "a judge exercises mercy when she imposes a sentence that is: (1) more lenient than what would normally be expected in a case of this sort; (2) yet just, based on consideration of a range of mitigating factors broader than what would be standard in sentencing a criminal like this one for the same crime").

(396) See VON HIRSCH & ASHWORTH, supra note 27, at 172-78 (discussing equity mitigation and outlining when it might apply).

(397) See David Garland, Sociological Perspectives on Punishment, 14 CRIME & JUST. 115, 118-19 (1991) (observing that "difficult issues tend to escape detailed moral scrutiny because they do not feature in the oversimplified conception of 'punishment' that philosophers conventionally use: they are not part of the problem that this tradition has set out for itself").

E. LEA JOHNSTON, Associate Professor of Law, University of Florida, Levin College of Law. I wish to thank Kevin Cieply, Mark Fenster, Erica Hashimoto, Jerold Israel, Adam Kolber, Michelle Jacobs, Elta Johnston, Wes Jones, Lyrissa Lidsky, Tom Lin, Dan Markel, William Page, Rachel Rebouche, Alice Ristroph, Elizabeth Rowe, Peter B. Rutlidge, Michael Seigel, Christopher Slobogin, Daniel Sokol, John Stinneford, Tonia Werner, Michael Wolf, and Stewart Young for their valuable comments and suggestions. I also appreciate feedback from participants at the Southeastern Association of Law Schools 2012 Annual Meeting, the Arizona State University Legal Scholars Conference, and the faculty workshops at the University of Georgia School of Law and the University of Florida Levin College of Law. I am grateful for the summer grant provided by the Levin College of Law. Finally, I thank Joseph Eagleton, Jessie Ervolino, Kelly Gay, Sarah Jeck, Caitlin Kovacs, Matt Michel, Patricia Morgan, Janet Owens, Erica Perdomo, Kimberly Tolland, and Daniel Tullidge for their outstanding research assistance.
COPYRIGHT 2013 Northwestern University, School of Law
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2013 Gale, Cengage Learning. All rights reserved.

 
Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:III. How Vulnerability May Affect the Distribution of Deserved Punishment through IV. Concluding Thoughts, with footnotes, p. 183-229
Author:Johnston, E. Lea
Publication:Journal of Criminal Law and Criminology
Date:Jan 1, 2013
Words:37175
Previous Article:Vulnerability and just desert: a theory of sentencing and mental illness.
Next Article:An empirical assessment of corporate environmental crime-control strategies.
Topics:

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters