Group Violence and Group Vengeance: Toward a Retributivist Theory of International Criminal Law

HAQUEMACRO.DOC 11/23/2005 9:51 AM

Group Violence and Group Vengeance: Toward a Retributivist Theory of

International Criminal Law

Adil Ahmad Haque†

TABLE OF CONTENTS

Introduction.......................................................................274 I. A Relational Theory of Retributive Justice.............277 A. The Relational Structure of Retributive

Justice................................................................279 B. Retributivism and Consequentialism ...............283 C. Relational and Non-Relational Retributivisms..287 II. The Social Embodiment of Retributive Justice.......290 A. From Group Conflict to State Sovereignty .......291 B. From State Failure to International

Intervention.......................................................294 III. A Retributivist Theory of Crimes against

Humanity .................................................................297 A. Crimes against Humanity: Analysis.................299 B. Chander, Luban, and May on Group Violence...304 IV. An Expressivist Theory of Discriminatory Intent...308 A. The Concept of Expressive Harm......................311 B. The Expressive Harm of Group

Subordination....................................................315 C. Amann’s “Expressivist” Account .......................318 V. A Political Theory of Jurisdiction............................320 A. Universal Jurisdiction.......................................321 B. Passive and Active Jurisdiction ........................323 C. Complementary Jurisdiction.............................324 Conclusion .........................................................................326

† Law Clerk to the Honorable Jon O. Newman, United States Court of Appeals for the Second Circuit. J.D., Yale Law School, 2005. Many thanks to Diane Marie Amann, Anupam Chander, Mirjan Damaska, David Luban, and Larry May for their advice, assistance, and encouragement.


274 BUFFALO CRIMINAL LAW REVIEW [Vol. 9:273

INTRODUCTION

International criminal law is frequently portrayed as the strong arm of the international human rights regime, an instrument designed to safeguard the dignity of each human person. There is an important truth to this characterization: international crimes involve many of the most grotesque violations of individual rights human beings inflict and endure. Yet the areas of international criminal law that are the subject of this article apply to the doings and sufferings of individuals only indirectly. The law governing crimes against humanity and genocide frames the acts and fates of individuals against broader and darker patterns of group perpetration and group victimization. It is only within the context of group violence that international law attributes individual responsibility for wrongdoing and vindicates the rights of victims.

The conceptual mismatch between the moral individualism of human rights discourse and the collectivist structure of international criminal law requires theoretical resolution. The theory developed in part I locates the legitimacy of institutional coercion within a structure of rights and duties linking authors of wrongs, victims of crime, and agents of punishment. As part II explains, the need for international criminal law arises from the defective embodiment of this relational structure in social groups and failing states, defects which devolve retributive justice into cycles of escalating violence. The displacement of group vengeance by legal process is not the (broadly consequentialist) ground of the relational structure, but rather a reason for one set of social institutions rather than others to occupy a position of authority within that (broadly deontological) structure.

Although the relational theory is intended as an independent contribution to the philosophy of criminal law, its cash value lies in its power to illuminate the role of group perpetration and group victimization in justifying the displacement of domestic law by international law and


2005] GROUP VIOLENCE & GROUP VENGEANCE 275

the intervention of international tribunals into internal armed conflict (part III); the relative gravity of genocide and crimes against humanity as well as the grouping of persecution and apartheid with crimes whose constituent acts cause greater physical destruction (part IV); and the roots of state resistance to international tribunals and the role of complementary jurisdiction in rationing the latter’s political capital (part V).

The retributivist orientation of this article is bound to raise eyebrows. Retributivist approaches to international criminal law are often dismissed for lack of fit with current practice. International criminal prosecution seems too selective to satisfy the demands of retributive justice.1 Too many wrongdoers go unpunished; too many victims are forgotten or ignored. It is of course possible to argue that selective prosecution results from a compromise between the demands of retributive justice and institutional constraints such as lack of staff and funds, political constraints such as domestic resistance to international adjudication, as well as competing moral values such as group reconciliation. This article delves deeper, locating these seemingly disparate concerns within the dynamics of the retributivist problematic. International criminal law challenges the legitimacy of state coercive power, eliciting domestic opposition to the jurisdictional supremacy of its implementing institutions. The institutional fragility of international tribunals is the price they pay for the strength of their normative claims. Complementary jurisdiction emerges as a tolerable compromise that allows states with the capacity to discharge their duties the opportunity to do so, while empowering international

1. For a recent permutation of this concern see Mark A. Drumbl, Collective Violence and Individual Punishment: The Criminality of Mass Atrocity, 99 Nw. U. L. Rev. 539, 577 (2005). Drumbl also argues that the retributivists’ famed proportionality principle provides little useful guidance to sentencing judges, since those responsible for the deaths of many have but one life to lose in return. Id. The criticism has little independent force. If retributivists can offer a justification for concurrent sentences, then the objection dissolves. If they cannot, domestic criminal law presents sufficient obstacles to a satisfying retributivist theory of criminal sentencing.



institutions to displace states lacking that capacity and simultaneously ration their reserves of political capital. The social embodiment of retributive justice shifts between social groups, states, and international institutions, seeking to coordinate that value with the suppression of group violence rather than subordinate one value to the other. When criminal punishment in any form will provoke rather than prevent retaliatory violence there is nowhere left to go, no accommodation to reach but only conflict among values to resolve.

There is a second concern with retributivist approaches, one equally fundamental and equally illuminating. Legal proceedings begin after crimes occur, and glorifying retributive justice seems to celebrate the most disappointing feature of the current administration of international criminal law. Some fear that in the years to come the International Criminal Court will serve primarily as political cover for great powers unwilling to use their strength to defend the weak from the wicked, that the Security Council will increasingly refer humanitarian crises to the court rather than aggressively intervene to suppress ongoing violence.2 This concern is left for last because it challenges us to look beneath the structure of retributive justice to its ground, and we cannot interrogate the foundations of a theory without first glimpsing the edifice it is meant to support. The article concludes by suggesting that the duty to punish rests ultimately on the duty to protect, that invocation of the former implicitly admits failure to discharge the latter. The further the great powers run from their responsibilities the faster they will be overtaken.

This article is offered subject to the caveats indicated by its subtitle. The theory it sketches requires and will receive further elaboration and defense; the analyses it

2. For a clear and passionate statement, see W. Michael Reisman, The ICC and Congo: The Limits of International Law, Address to the Yale Journal of International Law’s Third Annual Young Scholars Conference, Toward a New World Order? The Role of International Law in the 21st Century, Yale Law School (April 15, 2005).



offers of existing law will require supplementation to keep pace with doctrinal evolution; and to fully redeem its promise of explanatory power it must prove an aid to understanding the laws of war. No doubt international adjudication protects war crimes trials from vices of victor’s and victim’s justice similar to those outlined in part V. No doubt the appeal of such a claim is owed in part to its breadth. The article contains a proposal, and hopefully a contribution, to the young and small but quickly growing literature on the philosophical foundations of international criminal law.

I. A RELATIONAL THEORY OF RETRIBUTIVE JUSTICE

Retributivism’s proponents and commentators tend to focus their attention on the content and consequences of the proportionality principle, which requires that the severity of criminal punishment must equal or at least not exceed some function of the harmfulness of the criminal act and the fault of the actor.3 Scholarly focus on the proportionality principle has lent an individualistic gloss to retributivist theory as a whole. Retributivist theory has

3. See, e.g., Hugo Adam Bedau, Classification-Based Sentencing: Some Conceptual and Ethical Problems, in Criminal Justice 89, 102 (J. Roland Pennock & John W. Chapman eds., 1985) (arguing that retributive criminal penalties “must be based on two basic retributive principles: (1) the severity of the punishment must be proportional to the gravity of the offense, and (2) the gravity of the offense must be a function of fault in the offender and harm caused the victim.”). Robert Nozick famously characterizes the retributive function of harm and culpability as P = r x H, where P is the severity of punishment justified, r the degree of culpability graded on a scale from 1 to 0 inclusive, and H the magnitude of the harm caused or risked. When r equals 0, as in cases of insanity or duress, P equals 0 as well, and so no punishment is called for. When H equals 0, as in the case of harmless acts, P again equals 0. In other words, in the absence of either harm or fault, no type or degree of punishment may be justified on retributive grounds. Robert Nozick, Anarchy, State, and Utopia 60 (1974). Cf. George P. Fletcher, The Fall and Rise of Criminal Theory, 1 Buff. Crim. L. Rev. 275, 289-90. Fletcher proposes the formula C = W x r, where W is the magnitude of wrongdoing, r the degree of personal responsibility, and C the level of overall culpability. W increases with the level of harm or proximity of the threat of harm while r increases with the nature of intentionality (less for negligence, more for recklessness, highest for purposefulness). In the case of excuses, r is reduced to 0.



long been identified with its theory of individual desert, of the proper treatment of individuals in light of their individual acts. Within international criminal law this individualistic focus has centered attention on the responsibility of superiors and subordinates, on principles of command responsibility4 and duress.5 Yet the proportionality principle itself cannot be understood unless one looks beyond features of individual acts and actors to a network of moral relations extending between the wrongdoer and the wronged, between the punisher and the punished, and between the agent of punishment and the victim of crime. This article seeks to recover certain background understandings of the constitutive role of retribution in marking the boundaries of moral communities and allocating the rights and responsibilities of individuals and groups within communities.

Section A elaborates the relational structure of retributive justice, a structure systematically unappreciated by retributivists and their opponents but deeply rooted in moral intuition. On this view, the offender’s violation of the victim’s right gives rise to a duty of the punishing agent, owed to the victim, to punish the offender. The legitimate authority of the punishing agent derives from its reciprocal relations with the victim to whom the duty to punish is owed and with the offender from whose wrong the right to punish derives.

4. See, e.g., Mirjan Damaska, The Shadow Side of Command Responsibility, 49 Am. J. Comp. L. 455 (2001); Allison Marston Danner & Jenny S. Martinez, Guilty Associations: Joint Criminal Enterprise, Command Responsibility, and the Development of International Criminal Law, 93 Cal. L. Rev. 75 (2005); Greg R. Vetter, Command Responsibility of Non-Military Superiors in the International Criminal Court (ICC), 25 Yale J. Int’l L. 89 (2000). 5. The leading case is Prosecutor v. Drazen Erdemovic, No. IT-96-22-Y (Mar. 5, 1998) (International Criminal Tribunal for the Former Yugoslavia [ICTY] Sentencing Judgment, Trial Chamber II), available at http://www.un.org/ icty/erdemovic/trialc/judgement/erd-tsj980305e.htm (last visited Sept. 14, 2005). For commentary see Rosa Ehrenreich Brooks, Law in the Heart of Darkness: Atrocity and Duress, 43 Va. J. Int’l L. 861 (2003); Valerie Epps, The Soldier’s Obligation to Die When Ordered to Shoot Civilians or Face Death Himself, 37 New Eng. L. Rev. 987 (2003).



Sections B and C show that earlier versions of retributivism share the instrumentalist structure of the very consequentialist theories they purport to challenge. On both consequentialist approaches and on all but one rival version of retributivism, the authority of the punishing agent is unilateral and managerial in nature, based on its ability to secure better results than alternative sources of administrative control, and subject to no personal moral grievance for its failures. The relational view, by contrast, rests the legitimacy of the state’s coercive power on reciprocity between offender, victim, and punisher. Rather than one acting upon the others for the sake of an impersonal goal, each makes moral claims on the others to respect the rights and discharge the duties that bind them together. The relational view therefore reveals the fundamental cleavage between retributivist and consequentialist approaches to criminal law, and shows why extant versions of retributivism fall on the wrong side. Most importantly for the purposes of this article, the relational theory shows that the asserted supremacy of international over domestic law threatens to replace not merely a tool, instrument, or mechanism for the pursuit of our goals, but the institutional framework that mediates our moral relations with others. International law challenges the legitimate authority of the state as the most basic moral community of which its citizens are a part, and as we shall see this challenge is not one states suffer lightly.

A. The Relational Structure of Retributive Justice

The first relation is the simplest: offender and victim stand as author and subject of the same wrong, as transgressor and holder of the same right. The proportionality principle tacitly incorporates a theory of crime, according to which no act may be criminalized—since no act may be punished—in the absence of harm or in the absence of fault.6 The fault element is often thought

6. Douglas N. Husak, Desert, Proportionality, and the Seriousness of Drug



the essence of retributive theories, but in fact the harm element is equally if not more central. Retribution, like the associated concepts of retaliation, reprisal, reciprocation, and revenge, involves the idea of a victim whose rights the legal system seeks to vindicate. The role of the victim gives sense to retributive metaphors of “payback,” “getting even,” and “balancing the scales.”7 Indeed, historically speaking the fault element is the latecomer.8

The moral relation between the punisher and the offender is in itself straightforward. This relation involves the authority to punish wrongdoing, as well as the duty to reflect justifications, excuses, and mitigating circumstances in the imposition and calibration of penalties. The former responds to the wrong suffered by the victim, the latter to the circumstances faced and choices made by the wrongdoer.9 The traditional proportionality principle incorporates both duties, stating that it is permissible (or

Offenses, in Fundamentals of Sentencing Theory: Essays in Honour of Andrew von Hirsch 187, 217-18 (Andrew Ashworth & Martin Wasik eds., 1998) (“If the seriousness of crime is dependent on harm and culpability, a criminal offence designed to prevent an obscure harm, and which involves little or no culpability, must have a degree of seriousness that approaches zero. . . . If a person does not deserve to be punished for what he has done, his conduct should not have been criminalized in the first place.”). 7. For an extended treatment of this point see 4 Joel Feinberg, The Moral Limits of the Criminal Law: Harmless Wrongdoing 159-65 (1990). Surprisingly, Feinberg does not explicitly link the role of the victim in retributivist theory to the harm principle he articulates. 8. See, e.g., Martin R. Gardner, The Mens Rea Enigma: Observations on the Role of Motive in the Criminal Law Past and Present, 1993 Utah L. Rev. 635, 651-52:

Early Anglo-Saxon law evidenced traces of the idea that some culpable state of mind must accompany harmful conduct in order for criminal liability to exist. Nevertheless, scholars routinely posit that primitive English law originated from an essentially strict liability basis. Because the early law focused so heavily on the perceived interests of outraged victims and would-be avengers in hopes of dissuading them from pursuing private vendettas, it likely paid little attention to niceties of culpability once it was shown that the accused caused harm to the victim. To the extent that the law was aimed at compensating and buying off the feud, it hardly mattered that the offender acted inadvertently or otherwise nonculpably.

9. Cf. John Gardner, Wrongs and Faults, in Appraising Strict Liability 51 (A.P. Simester ed., 2005).



required) to punish wrongdoers in proportion to their wrongdoing. But the permission (or requirement) at issue is not universal: modern states in particular claim exclusive authority to punish.10

The controversial relation is that between the agent of punishment and the victim of crime. How can the duty to punish the offender be owed to the victim given that (i) the state initiates and conducts prosecution; (ii) prosecution vindicates the state’s interest; (iii) the victim has no legal recourse against non-prosecution; and (iv) the victim may not waive the duty to punish and spare the offender? Each of these questions both essentializes and misunderstands legal practice.

First, it is not an essential feature of criminal law that the state initiates prosecution. In the early centuries of the common law the right (or duty) to exact retribution vested first in the victim herself, who initiated legal proceedings through the appeal of felony. Even when charges were brought by indictment the victim served as a private prosecutor in court; verbal altercation between victim and defendant replaced trial by battle between the two or their respective champions.11 In homicide cases the victim’s family served this role; only where no family was available did authority fall to a royal officer.12 Public prosecution was long restricted to state crimes such as treason, and was only gradually extended to ordinary crime.13

Second, the notion that criminal law vindicates the interests of the state rather than the interests of victims is

10. Hence a prison guard who kills a death row inmate in his cell is a murderer, not an executioner. 11. Sir Thomas Smith, De Republica Anglorum ch. 23 (Mary Dewar ed., 1982); Theodore F.T. Plucknett, A Concise History of the Common Law, 117-18 (5th ed. 1956). 12. See John H. Langbein, The Origins of Adversary Criminal Trial 10 (2003). No doubt this model exemplifies the English preference for decentralized administration of criminal justice. However, both the institutional arrangement and its motivational assumptions trade on the notion that the second relation—that of authority of the punisher over the punished—arises from the third—that between the agent of punishment and the victim of crime. 13. See John H. Langbein, The Origins of Public Prosecution at Common Law, 17 Am. J. Legal Hist. 313 (1973).



overbroad. It is true that criminal cases by and large bear a title of the form State v. D rather than P v. D as in tort suits. It is also true, however, that criminal law recognizes divisions between crimes against the state (such as treason) and crimes against persons, as well as divisions between crimes mala in se and crimes mala prohibita. Crimes mala prohibita depend for their wrongfulness on affirmative declarations by the state that the acts undermine government policies. Crimes mala in se are acts whose wrongfulness is independent of the state’s existence to say nothing of its interests. The distinctions among crimes seem far deeper features of our legal practice than the idea that criminal law vindicates the interests of the state, an idea that once again was late in its development. The legal fiction that crimes violate the King’s Peace developed gradually, extending first to crimes on highways and waterways and only later to acts of private violence.14 Once again, for much of the history of the common law, the functional distinction between tort and criminal law was that while tort suits offered victims the possibility of monetary gain, the only interest a victim advanced through criminal prosecution was vindication of her rights.

Third, the absence of legal recourse against non-prosecution is primarily a function of separation-of-powers concerns.15 Nonetheless, formal and informal mechanisms that allow victim input into the charging decision, plea bargaining, and punishment reflect the notion that the criminal law vindicates victims’ rights.16 When the state ignores violations of its citizens’ rights it is not uncommon to speak of a second crime inflicted by the state upon its neglected citizens. Finally, there are many duties in law

14. See, e.g., F.A. Inderwick, The King’s Peace: A Historical Sketch of the English Law Courts (2003); C.V. Wedgwood, The King’s Peace, 1637-1641 (1983); Frederick Pollock, The King’s Peace in the Middle Ages, 13 Harv. L. Rev. 177, 177 (1900). 15. See, e.g., United States v. Cox, 342 F.2d 167 (1965). 16. See, e.g., Mass. Gen. Laws Ann. ch. 258B, § 3 (West 2005) (“Rights afforded victims, witnesses or family members.”).



and morality that cannot be waived by their corresponding right-holder. One can neither consent to battery nor sell oneself into slavery. Not all rights and duties are grounded in the right-holder’s autonomy, or even in her welfare; some instead reflect the value morality attaches to our status as persons, as members of the moral community. On the relational view of retributive justice the duty to punish is of such a kind.

The state’s duty to punish offenders is owed to the victim of crime. Moreover, the duty to punish is grounded in the relationship between the three actors. That such a relationship must exist is reflected in the extension of active, passive, and territorial jurisdiction to ordinary crime, while universal and complementary jurisdiction are, for reasons we shall later investigate, generally limited to international crime involving group perpetration and group victimization. Active and passive jurisdiction reflect the principle that the state has the authority to right the wrongs committed both by and against its citizens. Territorial jurisdiction similarly extends to non-citizens a protected status in virtue of their relationship to (though not their membership in) the broader community.17 The dependence of the duty to punish on the relationship between the punisher and the victim can also be traced through the common law process described above: The authority to prosecute and seek punishment shifted from the smallest and most exclusive community to the largest and most inclusive community of which the victim was a member.

B. Retributivism and Consequentialism

The distinctive features of the relational theory will be illustrated first through a contrast with consequentialist approaches to criminal law, and in the next section through

17. For an early example of protected status see The Constitution of Medina arts. 14-16, 25-34, 40, in W. Montgomery Watt, Islamic Political Thought 130-34 (1968) (describing the “neighborly protection” offered to Jewish tribes by Arab patrons).



a contrast with extant versions of retributivism. Rather than enter into a long-running exchange of “fantastic examples” meant to drive a wedge between an opposing theory and moral intuition, the discussion below will bring out deeper structural differences among the different views.18 Ultimately the cleavage between the relational view and both consequentialist views and rival retributivist views lies in different conceptions of legitimate coercive authority, whether vested in a social group, a state, or an international legal system. While the relational view rests legitimacy on reciprocity between communities and their members, expressed through an interlocking network of rights and duties, consequentialist and rival retributivist views rest legitimacy on the unilateral assertion of administrative control by a managerial entity aimed at impersonal goals.19 The relative merits of the two ideals, as well as the conceptions of criminal law which depend upon them, must be adjudicated in the forum of political philosophy, and though that contest cannot be conducted in this article, the article reframes the terms of engagement.

First, on consequentialist approaches to criminal law the violation of a victim’s rights is not the moral ground of a duty to punish. Rather, the violation serves as evidence that the offender will commit more harmful acts in the future, and that punishment of the offender will deter others from committing similar harmful acts. The moral justification for punishment lies in the dangerousness of the offender and the opportunity to deter others by punishing him. It is the dangerousness of the offender that justifies removing him from society (incapacitation), imposing a cost on his conduct (specific deterrence), and preparing him for peaceful reentry into society

18. Cf. C.L. Ten, Fantastic Examples and Moral Principles, in Crime, Guilt, and Punishment 18 (1987). 19. Cf. Christopher Kutz, Responsibility, in The Oxford Handbook of Jurisprudence & Philosophy of Law 548, 569-70 (Jules Coleman & Scott Shapiro eds., 2002) (noting the managerial conception of the state associated with consequentialism and the absence of a competing conception associated with retributivism).



(rehabilitation). So the first structural difference between the relational view of retributive justice and the consequentialist view of criminal law is that on the former past wrongs generate practical reasons, reasons to act, while on the latter past harms generate only epistemic reasons, reasons to believe practical reasons for action (the dangerousness of the offender) exist.

Second, on the consequentialist view the victim of crime has no moral right (let alone a legal right) against the state to punish the offender. She has no personal moral grievance should her rights against the offender go without vindication. The state’s duty to punish is not owed to the victim; nor is the duty owed to the offender. Nor is the duty to punish a general duty to maximize the welfare of any individual members of the community. In its utilitarian manifestation, the goal of consequentialist social policy is the maximization of aggregate or average welfare. Pursuit of such a goal necessarily makes at least some members of the community better off, but does not necessarily make a majority of individuals better off. Indeed, consequentialism requires reducing the welfare of a majority of individuals when this decrease is outweighed by a large increase in the welfare of a minority of individuals. The point for the purposes of this article is that on the consequentialist view no individual has a personal moral grievance against the state when it fails to punish. All individuals have at most a shared, generic grievance when the state fails to maximize aggregate or average welfare. No wrong is done to them, but only to the impersonal cause of welfare maximization.

Third, on the consequentialist view the authority of the state to punish rests not on its relationship to the victim of crime but on its administrative competence. Individuals and social groups should refrain from punishment because consistent deference to the legal processes of the state will better maximize welfare than consistent reliance on individual and group judgment. The oft-noted instability in this view arises in cases in which it is clear that a deviation from legal processes will result in



better outcomes than blind obedience. For instance, consequentialists struggle to provide a sound reason why it would be wrong for state actors or private individuals to punish either an innocent person whose punishment would deter those disposed to actually commit crimes or to punish a clearly dangerous individual who has not yet committed a crime. Again, it is not the purpose of this article to rehash old arguments. The point here is that the very possibility of the described instability arises because consequentialism does not view the duty to punish as a duty owed to victims of crime, nor is the content of the duty to punish the vindication of rights against wrongdoers. To the extent that consequentialism recognizes a duty to punish at all, that duty is owed to no individual but is rather an impersonal duty to realize a desired state of affairs.

Finally, the consequentialist view presupposes an independently defined community whose welfare the state seeks to maximize through the punishment of harmful behavior. The principle of utility cannot determine whose welfare the state should take into account when crafting criminal justice policy. The principle logically extends beyond borders20 and (on some views) even species21 to all sentient beings whose welfare can be enhanced by our actions. Some independent membership criteria must be introduced to limit the scope of the principle to citizens and protected persons (immigrants and foreign visitors, most prominently). The relational view, by contrast, treats the duty to punish as partly constitutive of the community itself. As we will see again in part II, practices such as outlawry and exile manifest the internal relationship between community membership and the right to be avenged. The principle of utility can be applied to any group of individuals whatsoever precisely because its application does not presuppose that the individuals whose doings and sufferings provide the occasion for punishment stand in some important relation to one another or to the

20. Peter Singer, Famine, Affluence, and Morality, 1 Phil. & Pub. Aff. 229 (1972). 21. See Peter Singer, Animal Liberation (2002).



agent of punishment. On the relational view the duty to punish is not impersonal or free-floating in this way; the state’s duty to punish one person is owed to another in virtue of the reciprocal moral relations that extend between the three parties. To include or exclude an individual from the network of rights and duties is to broaden or widen the community itself, and must be justified on the basis of a revised understanding of the nature of the community. Rather than presupposing an independently defined community, our criminal law practices are in part a forum for negotiation over the boundaries of our communities and the scope of our rights and duties.

C. Relational and Non-Relational Retributivisms

Retributivist theories can be plotted along two axes: evaluative theories refer to value created or disvalue eliminated through punishment; deontic theories refer to requirements or permissions to punish. Positive theories refer to values and requirements, negative theories to disvalues and permissions. Positive evaluative retributivism is the view that deserved punishment is intrinsically valuable in its own right.22 Negative evaluative retributivism is the view that deserved punishment merely eliminates an intrinsic disvalue, such as impunity.23 Negative deontic retributivism is the view that it is permissible to punish individuals less than they deserve, but impermissible to punish them more than they deserve.24 Consequentialist

22. See, e.g., Michael Moore, Placing Blame: A General Theory of the Criminal Law 87-88, 156-58 (1997). Moore claims that deserved punishment is “an intrinsic value that we each categorically are enjoined to realize in our actions,” and we therefore “each have an agent-relative obligation to punish the guilty . . . .” Id. at 163. He therefore asserts both a form of positive evaluative retributivism and a form of positive deontic retributivism. See infra note 26 and accompanying text. 23. See, e.g., George P. Fletcher, The Place of Victims in the Theory of Retribution, 3 Buff. Crim. L. Rev. 51, 60-63 (1999). Fletcher’s remarks on domination, though illuminating, do not have this structure. Punishment is a means to eliminate domination, not its absence or logical opposite. 24. For discussion see J.L. Mackie, Morality and the Retributive Emotions, Crim. Just. Ethics, Winter/Spring 1982, at 3; Jeffrie G. Murphy, Retributivism



considerations determine whether and how much punishment is inflicted within the permitted range. Finally, positive deontic retributivism is the view that there is a duty to punish individuals in proportion to their desert. The relational view is a form of positive deontic retributivism that claims the duty to punish is owed to the victim of crime, in contrast to Herbert Morris’s view that the duty to punish is owed to the offender,25 as well as Michael Moore’s view that the duty to punish is an impersonal duty owed to no one in particular.26

Since the violation of a victim’s rights is part of the basis of the offender’s desert, only negative deontic retributivism fails to treat such violations as part of the moral justification for punishment. On that view the desert of the offender is not a reason to punish, but merely an enabling condition that empowers consequentialist reasons to justify punishment.27 Desert serves only as a constraint on the amount of punishment that may be inflicted for other—generally consequentialist—reasons. On the other rival views, deserved punishment is alternately the good to be pursued through punishment, the conceptual opposite of the evil to be avoided by punishment, or the ground of the duty to punish. On each view, that the offender deserves to be punished for violating the victim’s rights is what makes it the case—not merely evidence of what makes it the case—that the state ought to punish.

and the State’s Interest in Punishment, in Criminal Justice: Nomos XXVII 156, 159 (J. Roland Pennock & John W. Chapman eds., 1985). 25. Herbert Morris, Persons and Punishment, in Punishment 74 (Joel Feinberg & Hyman Gross eds., 1975). This is not to deny that the state is under a duty to choose methods of punishment, if any, that are neither degrading nor dehumanizing. 26. Moore, supra note 22, at 91, 163; Michael Moore, Victims and Retribution: A Reply to Professor Fletcher, 3 Buff. Crim. L. Rev. 65 (1999) (arguing that victims are central to retributivist theory only insofar as the norms whose violation renders the offender deserving of punishment are rights-based). 27. On the distinction between considerations that favor action and enabling conditions that permit other considerations to favor action, see Jonathan Dancy, Ethics Without Principles (2004).



However, each of the alternative theories shares the other structural features the previous section identifies with consequentialism. First, on none of the alternatives does the failure to punish ground a personal moral grievance on the part of the victim. On Morris’s view offenders have a right to demand their own punishment, but victims have no moral basis to demand vindication of their rights. On the other alternatives no one has grounds for personal complaint; everyone shares a generic complaint that the punishing agent is not discharging its impersonal obligations or pursuing the best results—alternately understood in terms of the pursuit of value, the reduction of disvalue, or the pursuit of maximum welfare within deontological side-constraints.

Second, like rule-consequentialist approaches to punishment, each of the alternatives rests the authority of the state to punish on its ability to secure better results over time than individuals acting on their own contextual judgment. The good results the state achieves are on these views not the maximization of welfare but the infliction of deserved punishment, the elimination of impunity, and the discharge of duties to punish owed to offenders or to no one in particular. In each case the state is an instrument for the achievement of certain ends rather than a collective agent standing in a network of moral relations with both victims and offenders.

Finally, the rival versions of retributivism are unrestricted in scope, acting on independently defined communities rather than participating in their constitution. The values and duties adduced may be pursued through any number of institutional mechanisms, instruments which may be discarded when they cease to be useful. Such an instrumentalist conception of social institutions occludes the true nature of conflicts over institutional legitimacy. International law seeks to replace the nation with humanity as the foundational moral community of which we are part, and replace the state with a new set of basic institutions that mediate our moral relations with one another. To acknowledge the supremacy



of international law is not merely to lay down one tool and pick up another; it is to radically revise our understandings of ourselves as persons and the state as a collective moral agent.

II. THE SOCIAL EMBODIMENT OF RETRIBUTIVE JUSTICE

Specification of the formal features of retributive justice is the first step toward revealing its capacity to justify and explain central features of international criminal law. This part examines the embodiment of the relational structure of retributive justice in human practices and institutions. For while the inner logic of retributivism is illuminated by the moral relations of authors of wrongs, victims of crime, and agents of punishment, the capacity of this relational structure to illuminate international criminal law depends critically on its social and historical manifestations. While retributive justice is claimed by many to have its origins in emotion28 or in reason,29 this part highlights its place in human experience and its constitutive role in the formation and regulation of moral communities and their members. The narrative focuses on the interaction between three agents of retribution: social groups, state actors, and international institutions. Each agent suffers from its own deficiencies: social groups from the excesses of group vengeance; states from their vulnerability to indifference toward minorities and co-option by majorities; and international institutions from the political instability of the inspiring but elusive ideal of a common humanity. These deficiencies prevent the fair and effective discharge of each agent’s duties and the authority to punish shifts from one to the next as each is successively stripped of its legitimacy.

28. 2 James Fitzjames Stephen, A History of the Criminal Law of England 81-82 (1883); J.L. Mackie, Retributivism: A Test Case for Moral Objectvity, in Philosophy of Law 677 (Joel Feinberg & Hyman Gross eds., 4th ed. 1991). 29. See, e.g., Immanuel Kant, The Metaphysics of Morals 140-43 (Mary Gregor trans., 1991) (describing the principle of retributive justice as a categorical imperative).



A. From Group Conflict to State Sovereignty

The narrative traced in the following sections is not intended as a history of criminal law, a history too vast and varied to be treated let alone captured in this article. What follows is offered in the mode of an origin myth, a collection of familiar tales of the sources and development of criminal law and, with it, human society.30 If it is sufficiently familiar to resonate and sufficiently rich to illuminate then it will serve its limited purpose.

The founding myth of criminal justice—indeed of law itself—begins with familial and tribal conflict.31 “In the beginning,” the story goes, a wrong inflicted by one individual upon another was punished by the victim’s in-group: a family, a clan, a tribe, or some such collectivity. Punishment was both a right and a duty of community membership, which not only shielded its members from abuse but also played an expressive and constitutive role in marking the contours of the group and the relations of its members. These functions of retribution were performed not only through punishment but also through its denial.

30. See, e.g., Jerome Bruner, Making Stories: Law, Literature, Life 45-46 (2002) (describing criminal law’s aim, as reflected in its most particular doctrines, as “to stem cycles of revenge”). Cf. James Q. Whitman, At the Origins of Law and the State: Supervision of Violence, Mutiliation of Bodies, or Setting of Prices?, 71 Chi.-Kent L. Rev. 41, 46 (1995) (calling for refinement of partial truth of “self-help” model of state formation):

Stage one is the stage of the state of nature. This is a stage of ordered vengeance and vendetta. In this first stage, clans and/or individuals exact vengeance, in a systematic and rule-governed way, when injured by other clans and/or individuals; in particular, they exact talionic vengeance, seeking, in the famous biblical phrase, “an eye for an eye, a tooth for a tooth.” In stage two, the early state emerges. This early state does not, however, attempt to prevent violence. Rather, it sets out to supervise the existing system of vengeance. Thus, the early state assumes a kind of licensing power over acts of talionic vengeance, requiring that injured parties seek formal state sanction before avenging themselves. In stage three, the early state itself begins to function as enforcer, taking vengeance on behalf of injured clans; in Weber’s phrase, the early state of stage three monopolizes the legitimate use of violence. Only in stage four does the early state at last move to eliminate private violence.

31. The canonical literary treatment remains Aeschylus, Oresteia (David Grene & Richmond Lattimore eds., 1969).



Indeed, exile was counted among the most severe punishments due in large part to the refusal to avenge wrongs subsequently committed against former group members.32

Despite its constitutive and regulative role in human communities, group vengeance suffers from a well-known pathology. When doer and sufferer are of different collectivities the duty of retribution, coupled with imperfect information and in-group bias, gives rise to spiraling cycles of revenge, community rivalry, and civil unrest. Partiality to group members leads to disproportionate responses while unresolved factual disagreements make it difficult for one group to accept another group’s acts as acceptable retribution rather than further (punishable) wrongdoing. Finally, and most destructively, against a background of hostility and distrust the wrongs of individuals are imputed to the groups of which they are members; widespread violence may therefore arise from isolated incidents.33 As Robert Jackson noted in his opening arguments before the first Nuremberg Tribunal, it can be “hard to distinguish between the demand for a just and measured retribution,” sanctioned by the principle of retributive justice, and “the unthinking cry for vengeance . . . .”34 When groups judge their members’ cases, this distinction becomes impossible to draw.

Into this arena of group conflict comes the law. Enlightened and neutral, the state displaces the authority of other collectivities and takes upon itself the duty of retribution, absolving social groups of this burden and its potential to provoke endless and destructive tit-for-tat

32. Albert M. Rosenblatt, The Fifty-Fifth Annual Cardozo Memorial Lecture: The Law’s Evolution: Long Night’s Journey into Day, 24 Cardozo L. Rev. 2119, 2123 (2003) (“Banishment was among the most severe punishments imaginable . . . . Banishment meant isolation, the loss of the protection of the tribe, and in all probability death by starvation or violence.”). 33. See the discussion of Gujarat massacre infra. 34. Robert H. Jackson, Opening Speech for the Prosecution at Nuremberg (Nov. 21, 1945), in 2 Trial of the Major War Criminals before the International Military Tribunal, Nuremberg: 14 November 1945–1 October 1946, at 98, 101 (1947).



reprisals.35 As Martha Minow has argued, the task of the law is

to avoid such escalating violence [by] transfer[ing] the responsibilities for apportioning blame and punishment from victims to public bodies acting according to the rule of law. This is an attempt to remove personal animus, though not necessarily to excise vengeance. Tame it, balance it, recast it as the retributive dimension of public punishment.36

Though the modern liberal state is understood to stand in an unmediated relationship with the individual, understood as the basic unit of moral concern, the state assumes the responsibility to protect individuals and punish their persecutors in large part because other social groupings prove incapable of discharging that responsibility.37

Although state punishment coexists for a time with private vengeance,38 ultimately it demands total control over the dispensation of retributive justice. The state claims to replace all other collectivities as the foundational moral community to which individuals owe primary allegiance and from which they demand justice. Retribution does not lose its expressive and constitutive

35. Robert Postawko, Towards an Islamic Critique of Capital Punishment, 1 UCLA J. Islamic & Near E. L. 269, 304 (2002) (arguing that “the clear impetus of Islamic law is to break the cycle of revenge in murder cases . . .”). 36. Martha Minow, Between Vengeance and Forgiveness: Facing History after Genocide and Mass Violence 11-12 (1998). 37. See John Gardner, Crime: In Proportion and in Perspective, in Fundamentals of Sentencing Theory: Essays in Honour of Andrew von Hirsch, supra note 6, at 31 (“The blood feud, the vendetta, the duel, the revenge, the lynching: for the elimination of these modes of retaliation, more than anything else, the criminal law as we know it came into existence.”); Jeremy Horder, The Duel and the English Law of Homicide, 12 Oxford J. Legal Stud. 419 (1992). Cf. V.F. Nourse, Reconceptualizing Criminal Law Defenses, 151 U. Pa. L. Rev. 1691 (2003) (arguing that defenses in domestic criminal law can be understood to control individual vengeance). 38. See 2 Sir Frederick Pollock & Frederic William Maitland, The History of English Law 449-51 (2d ed. 1968) (describing predominant responses to wrongdoing: outlawry, blood-feud, tarrifs and fines, and corporeal and capital punishment).



functions upon its incorporation into the rights and duties of the state. Well into the Middle Ages the practice of outlawry confirms that the right to be avenged remains the mark of community membership.39 Outlawry has long since passed from the scene, but even now “[t]he precondition of criminal punishment is the severing of bonds with the wrongdoer—the withdrawal of social protection that opens him to shunning, imprisonment, or, in the extreme case, execution.”40

B. From State Failure to International Intervention

Like familial and tribal retribution, state punishment suffers from certain pathologies. But while families, tribes, and clans err through excess of passion, states fail primarily through indifference. The risk of such indifference is particularly acute when the state has been co-opted by one social group to the detriment of others. Often in cases of cooption the state not only fails to mediate group conflict, but itself becomes a weapon of war.

For example, between February 28 and March 2, 2002, over 2000 people were killed in an outbreak of ethnic violence in the Indian state of Gujarat. The Hindu nationalist government failed to intervene then, and even now refuses to address the majority of the crimes committed because their victims were Muslim civilians.

39. See, e.g., id. at 449: He who breaks the law has gone to war with the community; the community goes to war with him. It is the right and duty of every man to pursue him, to ravage his land, to burn his house, to hunt him down like a wild beast and slay him; for a wild beast he is; not merely is he a “friendless man,” he is a wolf.

See also Mirjan R. Damaska, Adverse Legal Consequences of Conviction and Their Removal: A Comparative Study, 59 J. Crim. L., Criminology & Police Sci. 347, 350 (1968) (“The outlaw’s children were considered as orphans and his wife a widow. Besides losing his family rights, he also lost all his possessions and even his right to life (if we can use that expression), for anyone could kill him with impunity.”). 40. David Luban, A Theory of Crimes against Humanity, 29 Yale J. Int’l. L. 85, 140 (2004).



In almost all of the incidents documented by Human Rights Watch the police were directly implicated in the attacks. At best they were passive observers, and at worse they acted in concert with murderous mobs and participated directly in the burning and looting of Muslim shops and homes and the killing and mutilation of Muslims. In many cases, under the guise of offering assistance, the police led the victims directly into the hands of their killers.41

According to Smita Narula, a senior researcher for Human Rights Watch,

[s]ixteen months after the beginning of the violence, not a single person has been convicted. . . . Hindu nationalist groups continue to arm civilians in Gujarat and many other Indian states. Instead of cracking down on these groups, the Gujarat state Bharatiya Janata Party (BJP) government has included the distribution of arms as part of its election manifesto.42

A similar pattern played out in the context of the Rwandan genocide, where the murder of nearly one million Tutsi and moderate Hutu was organized through the Hutu-controlled Presidential Guard and armed militias.43 Both the Gujarat massacres and the Rwandan genocide were understood as retaliation for the acts of the victim group: a train fire that

41. Human Rights Watch, “We Have No Orders to Save You”: State Participation and Complicity in Communal Violence in Gujarat, April 2002, at 5, available at http://hrw.org/reports/2002/india/gujarat.pdf (last visited Sept. 13, 2005):

Many witnesses testified that their calls [to police] either went unanswered or that they were met with responses such as: “We don’t have any orders to save you”; “We cannot help you, we have orders from above”; “If you wish to live in Hindustan, learn to protect yourself”; “How come you are alive? You should have died too”; “Whose house is on fire? Hindus’ or Muslims’?” In some cases phone lines were eventually cut to make it impossible to call for help.

Id. at 5-6. 42. Human Rights Watch, India: Gujarat Massacre Cases Sabotaged (July 1, 2003), at http://www.hrw.org/press/2003/06/india070103.htm (last visited Sept. 13, 2005) (internal quotation marks omitted). 43. See Gérard Prunier, The Rwanda Crisis: History of a Genocide (rev. ed. 1997).



killed Hindu tourists and BJP activists in the one case, the assassination of Rwandan President Juvenal Habyarimana in the other.44 Far from replacing destructive cycles of group retaliation with informed and impartial justice, the co-opted states were indifferent to and complicit in the suffering inflicted on the minority group.

When the state is either a tool of group conflict or a passive observer to such conflict it fails to discharge its role as a neutral source of retributive justice with the ability to prevent cycles of group retaliation for past violence. This failure alone would not provide a justification for international intervention. The justificatory gap must be filled by a sociopolitical premise, namely that just as the state supplanted the clan as the foundational moral community, the state has been supplanted in turn. The claim must be that following the rise of the human rights movement after World War II, we inhabit a world in which the moral status of individuals is seen to rest not on religion, nationality, or other social group membership, but on a common humanity. Human dignity, the equal moral worth of every human person, is now accepted as the precondition of all other legal and political values. Humanity itself, the story goes, has overthrown the state as the foundational moral community, and international criminal law exists to vindicate the rights of its members when states fail due to co-option by social groups.45

The temptation to make such sweeping claims without qualification must be resisted. The claim that the international community is our fundamental moral and

44. There is considerable evidence that Habyarimana was assassinated by extremist Hutu within his own political party. See id. at 213-21 (arguing that most probable assassins of Habyarimana were desperate members of his own circle who feared dilution of Hutu power); Human Rights Watch, Playing the “Communal Card”: Communal Violence and Human Rights 1, 9 (Cynthia G. Brown & Farhad Karim eds., 1995) (finding that Habyarimana’s airplane was downed by ground fire from positions near airport, and that this region was controlled by Rwandan army). 45. For a thoughtful attempt to flesh out such a view, see James Bohman, Punishment As a Political Obligation: Crimes against Humanity and the Enforceable Right to Membership, 5 Buff. Crim. L. Rev. 551 (2002).



political unit remains an aspiration rather than a social fact. Critically, while the state can overwhelm social groups with pure power, international institutions rely on states themselves for their survival. At the same time, it would be a mistake to conclude that because the international community is at least in part a legal fiction the idea of a common humanity holds no theoretical power. Moral universalism may fall short of global consensus, but it remains the guiding vision of those individuals and groups most responsible for the establishment of international tribunals and the elaboration of international criminal law. Similarly, the tenuous grip of this vision on the minds of world leaders does not entail that international institutions have no reason to exact retribution in the face of state failure. The state did not supplant the social group before assuming the responsibility to punish wrongs. Rather, the state achieved its foundational status by making good on its promise to effectively discharge the duty to punish. International tribunals may therefore strive to constitute an international moral community rather than reflect one that already exists. It is because international tribunals seek to change rather than maintain existing social realities that the political will supporting international institutions will likely remain limited and subject to fluctuation for the foreseeable future. As we shall see, complementary jurisdiction allows tribunals to ration their political support while balancing their universalistic aspirations with political realities.

III. A RETRIBUTIVIST THEORY OF CRIMES AGAINST HUMANITY

The narrative sketched above incorporates the goal of forestalling victim self-help into the broader project of instituting an effective and impartial regime of retributive justice. Similarly, the adjudication of individual guilt as a mechanism for the absolution of others cools intergroup hostility and reduces the likelihood of cycles of retaliation.



Rather than independent objectives these goals are aspects of the retributive problematic. This part interprets the evolving law governing crimes against humanity in terms of the dynamics of group conflict, state co-option, and international intervention. Although the constituent acts of crimes against humanity violate core human rights, theories that focus on the dignity of the individual cannot explain why such acts only constitute crimes against humanity when committed as part of “a widespread or systematic attack directed against any civilian population . . . . pursuant to or in furtherance of a State or organizational policy . . . .”46 The moral individualism at the heart of the human rights tradition entails that the killing of a human person is equally wrong whether committed by an individual or by a group and equally unfortunate whether or not the victim is a member of a group under attack. Nor can the scale of victimization alone justify the intervention of international law into internal conflict. International law makes group perpetration and group victimization—not numbers of victims—the central features of crimes against humanity.

The first function of international criminal law is to provide a body of law applicable both in domestic and international courts that prohibits serious wrongdoing permitted by domestic law at the time of the offense. This function is explained by showing how acts mala in se constitute a minimum content of retributive justice, such that the punishment of these acts can be justified in the absence of domestic legislation. The second function of international criminal law is to provide a legal basis for prosecution by international tribunals. This function is explained and justified in terms of the ability of social groups to co-opt state power, the propensity of states to remain indifferent to the suffering of minority groups, and

46. Rome Statute of the International Criminal Court, United Nations Diplomatic Conference of Plenipotentiaries on the Establishment of an International Criminal Court, July 17, 1998, pt. 2, arts. 7(1), 7(2)(a), U.N. Doc. A/ CONF.183/9, at http://www.un.org/law/icc/statute/romefra.htm [hereinafter Rome Statute] (last visited Sept. 13, 2005).



by the tendency of group violence to result in cycles of retaliation.

Anupam Chander, David Luban, and Larry May have recognized the salience of groups in international criminal law and have proposed interesting rationales for this distinctive object of concern. Chander defends the democratic credentials of international criminal law by showing how it protects minorities from majority will.47 Luban argues that crimes against humanity assault both human individuality and human community and manifest political pathology.48 May argues that non-individualized treatment harms humanity as a whole, and that this general harm provides an affirmative justification for the enforcement of international law by international tribunals.49 The powerful insights of these authors are integrated into and illuminated by the retributivist framework.

A. Crimes against Humanity: Analysis

Crimes against humanity are marked by five especially salient features:

Crimes against humanity are [1] international crimes [2] committed by politically organized groups acting under color of policy, [3] consisting of the most severe and abominable acts of violence and persecution, and inflicted on victims [[4] who are typically fellow nationals] [5] because of their membership in a population or group rather than their individual characteristics.50

It is important to make explicit the logical relationship between these five features: the second, third, and fifth features must justify and explain the first and fourth features. This logical relationship is by no means obvious.

47. Anupam Chander, Globalization and Distrust, 114 Yale L.J. 1193, 1232-35 (2005). 48. Luban, supra note 40. 49. Larry May, Crimes against Humanity: A Normative Account (2005). 50. Luban, supra note 40, at 108.



International documents impose the requirements of group perpetration and group victimization and list the constituent acts of violence and persecution, but do not limit crimes against humanity to fellow nationals or explain their status as international crimes.51 Yet the implicit features are theoretically more central than the explicit features.

The centrality of intra-national violence to the law of crimes against humanity follows from its design and logical structure. The law in this area was developed specifically to fill the legal void left open by the laws of war, which govern only violence between states.52 The law of crimes against humanity was developed specifically to override state sovereignty, justify prosecution in international tribunals, and criminalize acts permitted under domestic law. No injustice would result if a state fairly prosecuted and punished under its domestic law acts which would otherwise qualify as crimes against humanity.53 International criminal law is designed to provide an

51. See, e.g., Rome Statute, supra note 46, pt. 2, art. 7. See also Statute of the International Tribunal for the Prosecution of Persons Responsible for Serious Violations of International Humanitarian Law Committed in the Territory of the Former Yugoslavia since 1991, S.C. Res. 827, U.N. SCOR, 48th Sess., 3217th mtg., art. 5, U.N. Doc. S/RES/827 (1993), amended by S.C. Res. 1166, U.N. SCOR, 53rd Sess., 3878th mtg., U.N. Doc. S/RES/1166 (1998) [hereinafter ICTY Statute]; Statute of the International Criminal Tribunal for the Prosecution of Persons Responsible for Genocide and Other Serious Violations of International Humanitarian Law Committed in the Territory of Rwanda and Rwandan Citizens Responsible for Genocide and Other Such Violations Committed in the Territory of Neighbouring States, between 1 January 1994 and 31 December 1994, S.C. Res. 955, U.N. SCOR, 49th Sess., 3453d mtg., art. 3, U.N. Doc. S/RES/955 (1994), amended by S.C. Res. 1165, U.N. SCOR, 53rd Sess., 3877th mtg., U.N. Doc. S/RES/1165 (1998) [hereinafter ICTR Statute]. 52. M. Cherif Bassiouni, Crimes against Humanity in International Criminal Law 72 (2d ed. 1999). Crimes committed against fellow nationals represent the “pure case” of crimes against humanity that are not also war crimes, and therefore provide a focal point for theory construction. Luban, supra note 40, at 94. 53. Concurrent charges of crimes against humanity may communicate the state’s understanding of the true nature and full gravity of the underlying crimes, but this is a consideration of political craft, not of criminal justice. Note that the special intent element of genocide discussed in part V grounds a form of culpability unknown to domestic law.



international forum when fair domestic prosecution is impossible and to provide governing law when domestic law falls short.

The law of crimes against humanity is at its core designed to govern intra-national violence and justify international intervention. The task for the retributivist theory is to provide a compelling and unified account of how group perpetration, group victimization, and severe violence and persecution explain and justify the status of crimes against humanity as (primarily) international crimes committed against fellow nationals.

Crimes against humanity include only the most severe and abominable acts of violence and persecution. International law exists in large part to counterbalance group-based discrimination in domestic legislation. No society can survive without laws generally prohibiting acts of private violence, but when organized groups co-opt the state their first order of business is often to insulate their future crimes from domestic prosecution. When the state has been transformed into an instrument for group conflict, its authorization of group victimization by state officials and private individuals must be overridden by international law.

The constituent acts of crimes against humanity are limited to violent acts which are mala in se rather than mala prohibita. The wrongfulness of the latter crimes depends on the content of discretionary legislation that responds to the circumstances of particular countries and the demands of affected communities. It is both unnecessary and illegitimate for an international body to promulgate global rules for localities they do not understand and to whose populace they remain unaccountable. In the absence of such legislation, the acts are not wrongful and on the retributivist theory cannot be punished.

The wrongfulness of acts mala in se, by contrast, depends not on the content of legislation, but on the independently specifiable moral rights of victims. These rights fill out what we may call “the minimum content of retributive justice”: In the absence of any positive



legislation, harm to interests protected by moral rights may justify punishment when committed with awareness of the wrong-making features of the transgression. The requisite elements of harm and fault do not require specification in domestic legislation to provide the basis of just punishment. Institutional fairness concerns may nonetheless require prospective international law governing acts mala in se, but such law will merely codify preexisting moral wrongs rather than prohibit acts that would otherwise be morally permissible.

Crimes against humanity are committed by politically organized groups acting under color of policy. Outbreaks of widespread violence are almost invariably provoked and guided by organized groups exploiting popular unrest to achieve their own ends.54 Politically organized groups are appropriate targets for international law because they are either identical with the state, among the agents of the state, or are uniquely capable of co-opting the state to either assist or ignore group violence. As such, the state is incapable of dispensing just retribution or stemming cycles of group vengeance. If the state is not directing these groups it is assisting or acquiescing to their activities.

54. The Gujurat massacres, for instance, “were planned, well in advance of the Godhra incident, and organized with extensive police participation and in close cooperation with officials of the Bharatiya Janata Party (Indian People’s Party, BJP) state government.” Human Rights Watch, supra note 41, at 4.

Between February 28 and March 2 the attackers descended with militia -like precision on Ahmedabad by the thousands, arriving in trucks and clad in saffron scarves and khaki shorts, the signature uniform of Hindu nationalist—Hindutva—groups. Chanting slogans of incitement to kill, they came armed with swords, trishuls (three-pronged spears associated with Hindu mythology), sophisticated explosives, and gas cylinders. They were guided by computer printouts listing the addresses of Muslim families and their properties, information obtained from the Ahmedabad municipal corporation among other sources, and embarked on a murderous rampage confident that the police was with them. In many cases, the police led the charge, using gunfire to kill Muslims who got in the mobs’ way. A key BJP state minister is reported to have taken over police control rooms in Ahmedabad on the first day of the carnage, issuing orders to disregard pleas for assistance from Muslims. Portions of the Gujarati language press meanwhile printed fabricated stories and statements openly calling on Hindus to avenge the Godhra attacks.

Id. at 5.



Even after the fact, these groups are well positioned to threaten, bribe, or otherwise co-opt investigators, witnesses, and officials, further undermining the possibility of a just response to past crimes.

We generally expect states to punish private violence committed by individuals or even unstructured groups such as mobs or rioters. But politically organized groups pose special dangers to the impartiality and effectiveness of the state. Victim groups who have entrusted the authority to exact retribution to the state are likely to reclaim that authority when the state has been co-opted in the ways described. Seeking justice for their members, victim groups continue the cycle of group vengeance, and the possibility of a just, measured, and final retribution is lost to all.

Crimes against humanity are inflicted on victims based on their group membership rather than their individual characteristics. Although crimes against humanity lack a special intent requirement of discriminatory purpose, they require knowledge of a widespread or systematic attack against a civilian population. Though this population need not be defined by nationality, race, ethnicity, or religion, the chapeau indicates that the victims of a crime against humanity form a recognizable group, while the knowledge requirement implies that group membership is not an accidental feature of the victims but explains why they were targeted for violence and persecution. Crimes against humanity may not share the specific intent requirement of genocidal acts, but they are not random or indiscriminate. They are committed in furtherance of an organizational policy and against a background of systematic violence and persecution, and this implies criteria for the selection of victims. These criteria define the target group, membership in which is the basis for victimization.55

International law’s concern with group victimization flows from the failure of states to prevent group violence,

55. Cf. Luban, supra note 40, at 103-04. As we shall see, the special intent requirement for genocidal acts functions primarily as an aggravating factor.



which leads in turn to group retaliation. Groups transfer to the state their right and responsibility to exact retribution, expecting that the state will discharge this duty impartially, effectively, and in a manner acceptable to all the groups in society. Minority groups in particular rely on the state to vindicate rights they could not effectively champion themselves. The domestic emphasis of crimes against humanity reflects the priority of state responsibility for and to its own citizens, and the special significance of state failure to prevent and punish internal violence. Crimes against humanity committed against fellow citizens represent a fundamental collapse of state authority in the face of group conflict, a complete failure of the state to discharge its most basic obligations to contain group violence through impartial and effective retribution. States are either the agents of group victimization or are unwilling or unable to prevent and punish the extreme and systematic violence of organized groups against their fellow citizens. Once the moral authority of the state is so thoroughly compromised, the task of meting out retributive justice falls to the international community.

B. Chander, Luban, and May on Group Violence

In Globalization and Distrust, Anupam Chander argues that international criminal law should not be criticized as antidemocratic because when applied to democratic governments it protects minorities from majority abuse.56 Chander argues that international actors need not weigh demands of substantive justice against respect for democratic self-rule, since on the conception of democracy he inherits from John Hart Ely there is nothing to place on the right-hand side of the scales. Chander’s negative claim is well-taken. Chander’s account does not, however, support the positive claim that international criminal law is best understood as a mechanism of

56. Chander, supra note 47, at 1232-35. For a contrary view see Madeline Morris, The Democratic Dilemma of the International Criminal Court, 5 Buff. Crim. L. Rev. 591 (2002).



democratic representation reinforcement, a claim difficult to square with the law’s application to non-democracies and to non-state actors within democracies. Although international criminal law protects minorities from extreme abuse by democratic majorities, democratic principles cannot fully explain its structure and content.

Chander’s Elysian account balances a commitment to democratic governance against a commitment to the equal treatment of minorities. The retributivist theory, by contrast, disentangles egalitarianism (which it requires) from democracy (which it does not). Even where neither majorities nor minorities have institutional access to state power, the state bears a duty to vindicate the rights of all its citizens to live free from public and private violence. The legitimacy of the state’s coercive power depends in part on its displacement of group vengeance as the operative social embodiment of retributive justice and its integration of all citizens into a shared political community. If the state is to stem destructive cycles of group vengeance and secure all citizens full community membership then all groups must be treated equally under its laws. To leave some citizens outside the law’s protections would be to choose sides in group violence rather than replace it with the conditions for peace and equal justice. Democracies and non-democracies alike bear the same duty of equal justice to their citizens, and this duty requires the prosecution and punishment of non-state actors.

David Luban writes that a theory of crimes against humanity must identify both the distinctive human attribute targeted by these crimes and the universal human interest in suppressing them. Luban argues that crimes against humanity simultaneously transgress human individuality by targeting victims on the basis of group membership rather than individual characteristics as well as the human need for community by targeting such communities for destruction. Individuality and the need for community are two sides of our nature as political animals: we require political organization to survive and flourish yet fear that politics will turn against us. Our



nature as political animals in turn explains the universal human interest in suppressing what Luban describes as “Politics Gone Cancerous,” the capacity of the political organization we cannot live without to become a source of violence and oppression rather than protection and cooperation.57

It is easy to exaggerate the limited role of Luban’s account of individuality and community in his broader theory. Violence committed by one individual upon others on the basis of the latter’s group membership may be a crime against individuality and a crime against community but it will not constitute a crime against humanity.58 “Rather, what makes crimes against humanity distinctive lies in the fact that they are atrocities committed by governments and government-like organizations toward civilian groups under their jurisdiction and control.”59 Luban accounts for the requirement of political organization in terms of the universal interest in suppressing political pathology. The retributivist theory accounts for the same requirement in terms of the danger of state indifference or co-option posed by organized groups and the resulting need for a legal basis for international intervention. Interestingly, both the interest in suppressing political pathology and the danger of state indifference and co-option serve as grounds of jurisdiction rather than bases of culpability. On neither account does group criminality make the constituent acts deserving of greater punishment (though either theory could conceivably support that further position as well). Rather, both treat group criminality primarily as a reason for foreign and international courts to exercise jurisdiction over their authors. Luban argues that this latter interest supports universal jurisdiction, a claim explored and challenged in part V.

57. Luban, supra note 40, at 116-17. 58. For instance, a private individual who sets off an explosive during a gathering of some group could be guilty of mass murder but not of a crime against humanity. 59. Luban, supra note 40, at 120.



In his recent volume on the foundations of international criminal law, Larry May agrees that state complicity or acquiescence in group violence is necessary to overcome or (perhaps more accurately) undermine the presumption that states can be trusted to respond appropriately to internal armed conflict.60 With the presumption lifted, legal intervention remains in need of affirmative justification. May introduces his “international harm principle” with the familiar but misleading thought that, like domestic crime, international crime seeks to prevent and punish harms not to individual victims but to collectivities, to the state or to humanity as a whole.61 In passages that parallel Luban’s argument, May argues that crimes against humanity may be distinguished from their domestic analogues by the fact that the former target victims on the basis of their group membership rather than their individual characteristics.62 Non-individualized treatment of this kind constitutes an additional harm to all of humanity, and provides affirmative justification for legal intervention.

As we have seen, the extent to which criminal law enforcement vindicates the interests of political entities rather than the rights of victims can easily be overstated. Displacement of domestic by international legal norms is legitimate only because the norms of the latter refer to acts

60. May, supra note 49, at 68. May’s “security principle” emphasizes the failure of the state to protect its citizens, while the retributivist theory emphasizes the failure of the state to vindicate their rights. In this way the retributivist theory maps more closely on to the Rome Statute’s jurisdictional requirement that a state prove unwilling or unable to prosecute offenders. 61. Id. at 80, 82. It is not always clear whether May regards non-individualized treatment as a free-standing harm to the human community, or as a proxy for the likely number of victims and the danger of regional conflict. See, e.g., id. at 83. (“Group-based harms are of interest to the international community because they are more likely to assault the common humanity of the victims and to risk crossing borders and damaging the broader international community.”) The latter line of argument is the less interesting of the two: a free-standing requirement of widespread or systematic attack would respond to the risk of mass violence without targeting group violence as such, and the laws of war presumably place independent limitations on cross-border conflict. 62. Id. at 85-86. Luban and May appear to have read each other’s manuscripts. See Luban, supra note 40, at 85, author’s note.



mala in se whose wrongfulness does not derive from their impact on government policies, to crimes against persons rather than crimes against states.63 It is therefore not clear that we must look beyond the impact of the constituent acts of murder, rape, and enslavement on their individual victims to justify the application of international law. It is in fact quite unlikely that the international community could justifiably remain indifferent to mass violence were it not for the special harm involved in non-individualized treatment. Luban and May’s emphasis on non-individualized treatment is nonetheless instructive. In part IV of this article, it is argued that discriminatory intent inflicts an expressive harm on its victims that justifies punishing genocide more severely than crimes against humanity involving similar constituent acts as well as grouping persecution and apartheid with crimes against humanity whose constituent acts appear more physically destructive. Non-individualized treatment is not a normative tipping point necessary to justify the application of international law; such treatment constitutes instead an aggravating factor that justifiably figures into the elements of crimes and the gradation of punishments.

IV. AN EXPRESSIVIST THEORY OF DISCRIMINATORY INTENT

Within the framework developed above genocide is, as it is generally thought to be, the paradigmatic international crime.64 Genocidal acts, because they are

63. See supra part III.A. 64. See, e.g., Prosecutor v. Musema, Case No. ICTR-96-13-I, para. 981 (Trial Chamber, International Criminal Tribunal for Rwanda [ICTR], Jan. 27, 2000), at http://www.ictr.org/ENGLISH/cases/ Musema/index.htm (describing genocide as “the crime of crimes”) (last visited Sept. 12, 2005); Prosecutor v. Rutaganda, Case No. ICTR-96-3, para. 451 (Trial Chamber, ICTR, Dec. 6, 1999), at http://www.ictr.org/ENGLISH/cases/Rutaganda/index.htm (same) (last visited Sept. 12, 2005); Prosecutor v. Kayishema, Case No. ICTR-95-1-T, para. 9 (Trial Chamber, ICTR, May 21, 1999), at http://www.ictr.org/ENGLISH/cases/KayRuz/ index.htm (same) (last visited Sept. 12, 2005); Prosecutor v. Serushago, Case No. ICTR-98-39-S, para. 4 (Trial Chamber, ICTR, Feb. 5, 1999), at http://www.ictr.org/ENGLISH/cases/Serushago/judgement/os1.htm (same) (last visited Sept. 12, 2005).



directed at groups as such, are the most likely to foster an “us vs. them” attitude and ultimately provoke group retaliation. At the same time, the aim of such acts is to eliminate the very community which, in the absence of a neutral and effective state, would seek retribution on behalf of those wronged. The intent of genocidal acts is to leave no one left to avenge the victims, while their effect is often to set off a cycle of reprisals. These considerations support special attention to such acts by international institutions, which may both preempt in-group retaliation with its attendant evils and provide the retribution victims deserve but would otherwise not receive.

But it is not enough to justify the special attention international law pays to genocidal acts; we must also decide whether genocidal acts should be punished more severely than other international crimes involving similar underlying conduct.65 Perhaps the most straightforward justification for such differential response derives from the special value of social groups, a value that supervenes over and is not reducible to the value of their members. On this account, genocide would constitute a crime against a group “as such,” while persecution and apartheid would remain crimes against individuals perpetrated because of their group membership.66 Despite the many attractions of this account, it is not entirely satisfying. The greater wrongfulness of genocide cannot reside in its impact on the survival of the target group, for these effects are indifferent

65. The case law remains divided on this issue. Compare Prosecutor v. Jelisic, Case No. IT-95-10, para. 13 (Trial Chamber I, Partial Dissenting Opinion of Judge Wald, Dec. 14, 1999), available at http://www.un.org/icty/jelisic/trialc1/ judgement/index.htm (“Indisputably, genocide is at the apex” of “a hierarchy of ‘seriousness’ among the crimes”), with Prosecutor v. Kayishema, Case No. ICTR-95-1-A (Appeals Chamber, June 1, 2001), available at http://www.ictr.org/ ENGLISH/cases/KayRuz/appeal/index.htm (finding that “there is no hierarchy among crimes in the Statute and that all the crimes within it constitute ‘serious violations of international humanitarian law’ punishable by the same sentence”). 66. Cf. May, supra note 49, at 168. See also Report of the International Law Commission on the Work of its 48th Session, U.N.GAOR, 51st Sess., Supp. No. 10, at 87, 88, UN Doc. A/51/10 (1996) (genocidal intent “must be [the intent] to destroy the group ‘as such’ meaning as a separate and distinct entity, and not merely some individuals because of their membership in a particular group”).



to the mental state of the perpetrators. Yet what distinguishes genocide from extermination is that the former crime involves the purpose to partially or wholly destroy a group, while the latter involves only the knowledge that such will be the likely consequence of the constituent acts. It is not clear whether the goal of protecting social groups is better served by one provision than by the other.

One might attempt to internalize the wrongfulness of genocide within the mind of the actor, observing that though genocide and extermination are wrong for the same reason (they both involve the partial or complete destruction of a social group), those who commit genocide do so for the very reason that makes their acts wrong, while those who engage in extermination, though aware of the reason that makes their acts wrong, do so for other reasons. Such an argument may succeed in explaining why those who commit genocide are worse criminals than those who participate in extermination, but it will not succeed in explaining why genocide is a worse crime than extermination, and it is the latter claim and not the former we set out to establish.

This part argues that the gravity of genocide derives from the relationship between what it means—the evaluations genocidal acts express—and what it does—the social structures genocidal acts seek to bring into being. Genocidal acts do not merely deny the equal moral worth of victims; they seek to create a social world in which that denial is an operating principle in society. Under such conditions the expressive harm inflicted on victims does not end with the criminal acts but persists in the social reality created by those acts. The expressivist theory explains not only why genocidal acts may be punished more severely than crimes against humanity involving similar constituent acts, but also why acts of persecution and apartheid are grouped with crimes against humanity whose constituent acts appear far more physically destructive. The expressivist theory therefore extends beyond genocide and provides a broader account of the full ordering of international criminal offenses.



A. The Concept of Expressive Harm

Genocidal acts are generally distinguished from crimes against humanity not by their conduct elements, which often overlap, or by their results, which may vary, but rather by the specific intent “to destroy, in whole or in part, a national, ethnical, racial, or religious group, as such.”67 As with domestic legislation targeting hate crimes, it is the specific intent element of genocide that is considered by many to make acts carrying out that intent more serious and deserving of greater punishment, all other things being equal.

Hate crime laws remain controversial among commentators, who describe their enforcement as tantamount to punishing thought.68 Dan Kahan has recently attacked this line of argument, observing that criminal law regularly adjusts the severity of punishment to match the deviation of an offender’s values from those the law endorses.69 Kahan defends this pervasive practice through an appeal to an expressive theory of punishment, according to which the purpose of criminal sanctions is to express certain values by punishing those who reject them. Allison Danner has incorporated Kahan’s arguments into her own proposal for a hierarchy of international crimes, arguing both that “the perpetrator of a bias crime deserves greater punishment because his motivation offends important societal norms” and that “[p]unishing bias crimes more severely than other crimes expresses the international community’s condemnation of such crimes . . . .”70

67. Convention on the Prevention and Punishment of the Crime of Genocide, Dec. 9, 1948, art. 2, 78 U.N.T.S. 277. 68. See, e.g., Anthony M. Dillof, Punishing Bias: An Examination of the Theoretical Foundations of Bias Crime Statutes, 91 Nw. L. Rev. 1015, 1017 (1997). 69. Dan M. Kahan, Two Liberal Fallacies in the Hate Crimes Debate, 20 Law & Phil. 175 (2001). 70. Allison Marston Danner, Constructing a Hierarchy of Crimes in International Criminal Law Sentencing, 87 Va. L. Rev. 415, 480 (2001). See also id. at 481, 483 (relating discriminatory motivation to genocide). Danner’s dense article also argues that bias crimes inflict greater harm on individual victims as well as the groups of which they are members, id. at 479-80, and that genocide



Expressive theories of punishment have long been viewed with great skepticism. It is difficult to believe that the intrinsic value of expressing certain evaluations can justify the willful creation of human suffering.71 Yet if such expression finds its value in its educative function, in the role of legal punishment in the dissemination and internalization of public values, then the expressivist theory collapses into a form of consequentialism. More significantly for the purposes of this article, punishment that targets deviant beliefs cannot be understood as retributive. The apportionment of happiness according to virtue may be a desirable goal of distributive justice, but it is not a goal of retributive justice.72

In this part it is argued that independently harmful acts manifesting group-based animus impose an added expressive harm on their victims, and that an increase in the severity of punishment appropriate to the underlying harm is therefore justified on retributivist grounds. In this way the concept of expressive harm cleaves the expressivist theory of crime from the expressivist theory of punishment

adds to persecution the danger of group extermination, id. at 482-83. 71. As Hart famously argued, it is not clear that the infliction of suffering is either necessary to express our disavowal of a harmless act or sufficient to justify the willful infliction of human suffering when doing so increases the net amount of suffering in the world without vindicating the rights of victims of past violations. H.L.A. Hart, Law, Liberty, and Morality 65-66 (1963). Especially where there is no victim with whom to show solidarity or whose disrespected value calls for affirmation it seems that vocal disapproval and other diffuse social sanctions are a more humane method of communication. In the absence of both retributive and consequentialist justifications, the creation of human suffering for the sole purpose of public condemnation “is uncomfortably close to human sacrifice as an expression of religious worship.” Id. at 66. 72. Compare Immanuel Kant, Critique of Practical Reason 135 (T.K. Abbott trans., 1996) (arguing that “the distribution of happiness in exact proportion to morality (which is the worth of the person, and his worthiness to be happy) constitutes the summum bonum of a possible world”) with Kant, The Metaphysics of Morals 105 (Mary Gregor ed. & trans., 1996):

Punishment by a court (poena forensis)—that is distinct from natural punishment, in which vice punishes itself and which the legislator does not take into account—can never be inflicted merely as a means to promote some other good for the criminal himself or for civil society. It must always be inflicted upon him only because he has committed a crime.



and leaves an enriched and refined retributivist theory in its place.

According to Anderson and Pildes, “[a] person suffers expressive harm when she is treated according to principles that express negative or inappropriate attitudes toward her.”73 Central to the concept of expressive harm is the idea that such treatment really does harm its target, that such treatment wrongs her in such a way as to give rise to a personal grievance. Expressive harm supervenes over the physical and psychological damage caused or threatened by the offender; it is the proverbial insult added to injury. To be the victim of an expressive harm is to be in a position from which an offender’s inappropriate attitudes are not mere curiosities, or nuisances, or private vices, or impersonal evils, but rather direct attacks upon one’s own dignity.74 Acts inflicting expressive harm may be distinguished from other acts motivated by improper attitudes by the personal nature of their normative results, that is, by the actions and attitudes they warrant in response.

To make these general statements more concrete, let us briefly examine the paradigmatic modern case of expressive harm. The wrongfulness of racial segregation did not lie entirely in the gross material inequality of social conditions maintained for whites and for persons of color, but also in a pervasive message of racial hierarchy and white purity.75 To say that “separate . . . [is] inherently unequal”76 is in part to say that when segregation expresses

73. Elizabeth S. Anderson & Richard H. Pildes, Expressive Theories of Law: A General Restatement, 148 U. Pa. L. Rev. 1503, 1527 (2000). 74. Cf. Jean Hampton, Correcting Harms Versus Righting Wrongs: The Goal of Retribution, 39 UCLA L. Rev. 1659, 1666 (1992) (“All wrongful actions are actions that violate a moral standard applicable in the circumstances. However, I will say that some moral actions violate those standards in a particular way insofar as they are also an affront to the victim’s value or dignity. I call such an affront a moral injury.”). 75. For leading expressive accounts of segregation see Anderson & Pildes, supra note 73, at 1542-45; Charles L. Black, Jr., The Lawfulness of the Segregation Decisions, 69 Yale. L.J. 421 (1960); Deborah Hellman, The Expressive Dimension of Equal Protection, 85 Minn. L. Rev. 1 (2000). 76. Brown v. Bd. of Educ., 347 U.S. 483, 495 (1954). The Brown Court



negative evaluations of one group this is itself wrongful.77 Yet the wrongfulness of racial segregation, even under fictitious conditions of material equality, is not “free-floating.”78 Though all would be justified in expressing outrage and disgust, only one group could properly express resentment. There was only one party in interest, only one group with its own grievance to litigate, and only members of one group had standing to bring suit. Racial segregation was not merely wrongful in the abstract; it was a wrong inflicted upon African-Americans. In any other context this would be a truism; in this discussion it clarifies the personal nature of expressive harm.

With the concept of expressive harm thus demystified, we are in a better position from which to interpret Kahan’s claim that the reason why deliberate acts are punished more severely than negligent acts, rape more than battery, and crimes motivated by hate more than crimes motivated by greed is not that the former are more physically damaging than the latter, but that the former are seen to express morally worse attitudes. The concept of expressive harm allows us to see that the former acts are indeed morally worse than the latter, but that they are not (merely) morally worse for the offender, not merely indications of an inferior moral character and indictments of his evaluative framework. They are morally worse for the victim, deeper violations of her rights to equal concern and respect. It is for this reason that the proper response to expressive harm is not Platonic pity for the offender, who

equivocated between an expressive and a psychological account of the wrongfulness of segregation, writing, for instance, that segregating black school children “generates a feeling of inferiority as to their status in the community that may affect their hearts and minds in a way unlikely ever to be undone.” Quoting a lower court ruling, the Court added that “‘[t]he impact is greater when it has the sanction of the law; for the policy of separating the races is usually interpreted as denoting the inferiority of the negro group.’” Id. at 494. 77. For instance, it is because of the different evaluations expressed that race-segregated and gender-segregated bathrooms do not pose comparable equality concerns. 78. Feinberg uses the phrase “free-floating evil” to pick out those morally objectionable actions or states of affairs whose immorality does not derive from a wrong or other harm to any person. Feinberg, supra note 7, at 4.



dwells in moral ignorance, but resentment and indignation on behalf of victims, who have been both damaged and devalued. The occasion for punishment lies not in the offender—his improper attitudes—but in the normative effect on the victim of the externalization of these attitudes through injurious acts.

B. The Expressive Harm of Group Subordination

With these theoretical foundations established we can examine the expressive harm of genocidal intent that warrants enhancing the punishment appropriate to the underlying conduct alone. To inflict violence upon an individual because of her group membership is to express the view that her moral status, her claim to respectful treatment, rests not upon her status as a human person but upon fortuities of her birth and upbringing. Genocidal intent reverses the moral priorities asserted by human rights norms. Rather than the moral significance of one’s group membership resting on one’s status as a human person, the moral significance of one’s status as a person is seen to rest on one’s group membership. Acts in conformity with that intent reassert precisely the normative universe international law is meant to replace.79

Genocidal intent realized in genocidal acts imposes a supervening expressive harm, for the acts manifest a denial of the victims’ dignity as persons, their equal moral worth founded on their shared humanity. This expressive

79. Indeed, to affirm the moral significance of group membership for all people and, conversely, to denounce the destruction of groups as distinct cultural entities is already to credit the notion that the moral significance of group life, though very great, is derivative rather than foundational. Group life has value because it has value for persons. Cf. David Luban, The Self: Metaphysical Not Political, 1 Legal Theory 401 (1995).

Larry May has asked why “the intent behind ethnic cleansing [cannot] be to go back to a time of separate spheres rather than a world in which one side is denied its humanity.” Email from Larry May, Professor of Philosophy, Washington University (Aug. 8, 2005) (on file with author). Though logically possible, it is psychologically implausible that many who forcibly expel entire populations from their homes sincerely adhere to the principle of “separate but equal.”



assault on the victim’s moral status may call for additional punishment to match the additional harm inflicted. Rather than targeting the attitude (genocidal intent) itself, however, the retributive view targets the added expressive harm inflicted on the victims of genocidal acts through the realization in and through those acts of a denial of the victim’s equal moral worth.

What distinguishes genocidal acts from individual hate crimes is that these acts seek to violently reorder the social world to realize and reflect the underlying discriminatory attitude. Genocidal acts attempt to create a social hierarchy to mirror the perpetrators’ distorted moral hierarchy. Through extermination, enslavement, expulsion, and other methods, they seek to make the inferiority of the target group a social reality. The expressive harm inflicted through genocidal acts is thus distinguished from other, more common forms of violence not so much by intensity as by duration, and from crimes against humanity not by the impact on those who perish but by the impact on those who survive. The expressive harm of genocidal acts does not end with the acts themselves. It persists in the social structures those acts create.

The manifestation and perpetuation of expressive harm justifies greater punishment for genocidal acts than for crimes against humanity involving similar constituent acts, as well as grouping together acts of persecution and apartheid with other crimes against humanity that involve physical violence.80 While crimes against humanity involving physical violence require only knowledge of a widespread or systematic attack against a civilian population, persecution and apartheid necessarily involve discriminatory intent.81 Although the definitions of persecution and apartheid incorporate by reference the acts of violence constitutive of other crimes against humanity, the former categories would be redundant if their content were exhausted by constituent acts of violence. When one

80. Rome Statute, supra note 46, pt. 2, art. 7(1). 81. Id. art. 7(1)-(2).



combines discriminatory intent with the known background of a widespread or systematic attack against the persecuted population as part of an organizational policy, these acts of violence constitute genocide. An organizational policy of inflicting widespread and systematic violence on a group “by reason of the identity of the group or collectivity”82 necessarily intends to destroy, at least in part, that same group, as such.83 The same could be said of acts of violence “committed with the intention of maintaining [an institutionalized] regime” of racial oppression and domination.84 It is only with respect to nonviolent but nonetheless “intentional and severe deprivation of fundamental rights”85 that the prohibitions of persecution and apartheid do independent work.

The question, then, is why nonviolent forms of persecution and apartheid should be grouped with the violence that characterizes every other constituent act of crimes against humanity. The expressivist theory suggests that nonviolent acts of persecution and apartheid are comparable in gravity to violent acts because they create and perpetuate institutions and practices that inflict widespread and systematic expressive harm on an ongoing basis. These acts create a social world that reflects discriminatory attitudes in every aspect of its structure, “an institutionalized regime of systematic oppression and domination by one . . . group over any other . . . groups . . . .”86 Persecution and apartheid generate a hierarchy of social groups whose very existence expresses a

82. Id. art. 7(2)(g). 83. The crime of persecution extends beyond race (the sole concern of apartheid), ethnicity, religion, and nationality (the concerns of genocide), to gender, politics, and culture. Id. art. 7(1)(h). It is unclear whether anything significant follows. It is difficult to imagine that a distinct cultural group targeted for extermination would not be considered an ethnic group, and it is difficult to imagine an organizational policy of attacking women as such as opposed to the women of a particular social group. Political affiliation might not invoke the protections of genocide law, but would fall within the “civilian population” requirement of crimes against humanity. 84. Id. art. 7(2)(h). 85. Id. art. 7(2)(g). 86. Id. art. 7(2)(h).



denial of the equal moral worth of human persons that its victims must experience every day of their lives. It is this systemic expressive injury that ranks persecution and apartheid among crimes against humanity.

C. Amann’s “Expressivist” Account

In a recent article Professor Amann articulates an approach to the law of genocide she describes as “expressivist.” The “expressive” act at the center of her theory is not the criminal act punished, but rather the judicial act of determining that a victim group is protected under the law of genocide.

The requirement that a perpetrator acted with “intent to destroy, in whole or in part, a national, ethnical, racial or religious group, as such” sets genocide apart from all other offenses. This group mentality element has fostered the understanding, shared by jurists and lay public alike, that genocide is the most awful crime. As expressivist writings demonstrate, that social meaning constrains judges to exercise care in determining whether a victim group fell within the proscription of genocide . . . . On the one hand, too loosely declaring groups protected could undercut the understanding that genocide is unique. On the other hand, too rigidly withholding protection from deserving groups could give rise to perceptions that law is unfairly selective, or that it fails to comprehend the true nature of today’s tragedies.87

Strictly speaking, Amann invokes communication rather than expression, looking beyond the attitudes which motivate an act to the intended effect of the act on the beliefs of an audience.88 The administrator who behind closed doors denies a qualified individual a desired position on the basis of her race expresses but does not communicate an attitude of racial prejudice; she does not

87. Diane Marie Amann, Group Mentality, Expressivism, and Genocide, 2 Int’l Crim. L. Rev. 93, 142-43 (2002). 88. See Anderson & Pildes, supra note 73, at 1508.



intend that anyone recognize her motivating attitude and indeed seeks to avoid such recognition. Amann does not question the moral appropriateness of judges’ motivations; rather, she urges judges to consider the effect of their decisions on a global audience. Her concern is not with the attitudes their decisions express, but with the attitudes they may engender in others.

Nonetheless, there is no logical conflict between the retributivist theory’s claim that the expressive harm inflicted by genocidal acts justifies enhanced punishment and Amann’s claim that judicial determination of which groups are protected by the law of genocide should be sensitive to the communicative dimension of their decisions. Amann is concerned with the identification of genocidal acts, while the retributivist theory provides an account of the wrongfulness of such acts. True, the values asserted by the retributivist theory may be implicated in acts Amann would not characterize as genocidal for fear of adverse audience reaction. To determine whether such divergence exists would require more searching sociological inquiry than can be conducted here. But one could in principle retain both Amann’s account and the retributivist account, hoping for broad convergence and preparing for occasional conflict between the values at stake.

Yet consilience is a theoretical virtue, and the retributivist theory suggests its own approach to the identification of protected groups. The retributivist justification of enhanced punishment for genocidal acts follows from the expressivist analysis of the greater wrongfulness of those acts. Genocidal acts express a denial of the victim’s equal moral worth and seek to realize a social world in which members of the victim group occupy a subordinate position in society. Rather than identify protected groups on the basis of physical characteristics or social history, judges applying the law of genocide should ask whether the perpetrator perceives the victim group as possessing features which mark them as social inferiors. The existence of these features as a matter of biological or historical fact should not control the decision, both because



the independent existence of these features is irrelevant to the perpetrator’s mental state and because the subordination of individuals on the basis of perceived differences creates real differences. Genocidal acts may create distinct social groups, marked by their shared subordination, where no such groups existed before.

The retributivist account of enhanced punishment for genocidal acts rests on the expressive harm inflicted by acts motivated by genocidal intent and perpetuated by the social reality these acts seek to create. International criminal law intervenes into the process of social construction to reassert the primacy of shared humanity over group difference and dismantle regimes founded on the premise of unequal moral worth. Yet, as with crimes against humanity, international prosecution of genocide trades on an unstable reserve of political capital. The final part of this article argues that political will supporting international institutions must be rationed through complementary jurisdiction.

V. A POLITICAL THEORY OF JURISDICTION

Discussions of jurisdiction over international crimes often proceed against a background of debates over the nature and boundaries of state sovereignty. In the context of the current discussion it is useful to begin with a different challenge: How is international law to protect legal proceedings from the very group loyalties and animosities that defeat the aspirations of retributive justice in the context of group violence? How can we ensure that the forms of law are not mere masks for partisan bias, for absolution of one’s own and persecution of one’s enemies, for continuation of group conflict by other means? Once again it falls to international law to ensure retributive justice in the face of group bias on one side and state indifference, co-option, and impotence on the other. The complementary jurisdiction of the International Criminal Court reflects, though it does not fully satisfy, the concurrent needs to respect the deep and persistent



relationship between states and their citizens, to ensure the integrity of legal processes, and to ration the reserve of political will that supports international institutions.

A. Universal Jurisdiction

Let us return to David Luban’s argument that universal jurisdiction is justified by the interest of all individual human beings in the suppression of pathological politics, constrained by the independent demand for due process in the determination of guilt and the imposition of punishment.89 Though the crimes alleged are committed by and upon foreign nationals on foreign soil, Luban argues that individual states should bring criminal charges on behalf of their own citizens rather than on behalf of the victims of crime. Luban denies a universal moral interest in the prosecution and punishment of international crimes, and denies that humanity forms a political community with the power to enact laws or an interest in preventing and punishing violations of such laws.90 Universal jurisdiction rests on individual self-interest: Since we are all vulnerable to group violence we all have an interest in its prevention and punishment.

It is not clear whether Luban’s account succeeds on its own terms. First, universal jurisdiction is a poor deterrent, since its invocation can seldom be predicted before the fact. If we seek to deter political violence then we should value certainty of prosecution and punishment, which is best achieved through a centralized institutional process. Second, nowadays the state rather than individual citizens brings criminal charges, and few nations recognize claims against the government for failure to prosecute. So it is not clear whether citizens will be able to vindicate their interest. Third, if a state prosecutes a foreign national for international crimes then this likely indicates that the state already recognizes the dangers of pathological

89. Luban, supra note 40, at 137-46. 90. Id. at 125-27.



politics. A state itself beset by pathological politics would likely decline to punish others for behavior contemplated by its own officials. So the states which bring charges are the ones their citizens need not worry about, and the states citizens should worry about will not bring charges.

Its internal structure to one side, Luban’s account provides an instructive contrast to the retributivist theory developed above. On the retributivist theory, the state’s authority to punish an offender derives from the reciprocal rights and duties that bind the state to the victim of crime. On Luban’s view the state’s authority to punish derives from its managerial role in promoting the interests of its citizens. On the retributivist theory the wrong inflicted on the victim is a moral ground for punishment. On Luban’s view the wrong inflicted on the victim is only a reason to believe that such a reason exists, namely that public punishment of the offender will deter domestic political groups from committing similar offenses in the future. The state prosecutes and punishes not to vindicate the victim’s rights but to benefit its own citizens.

On the retributivist theory universal jurisdiction must rest on a conception of human beings as members of a single moral community in virtue of their shared status as persons. Only such a conception can ground the reciprocal rights and duties that on the retributivist theory bind together authors of wrongs, victims of crime, and agents of punishment. Conceived as self-contained political communities, states lack the authority to punish crimes against humanity committed by and upon non-citizens on foreign soil, since they do not stand in the right sort of relation to the offenders and victims. It is only as representatives of a broader moral community founded not on political ties but on shared humanity that individual states may have the authority, absent international institutions, to prosecute and punish. Yet the instability of such a conception in the beliefs and practices of states creates uncertainty over the prudence and propriety of regularly extending universal jurisdiction to international crimes involving non-citizens. As we shall see, the



complementary jurisdiction of the International Criminal Court seeks to moderate this same instability while promoting certainty that international criminals will face justice in one forum or another. The exercise of the right to punish on behalf of the human community is conditioned on the inability or unwillingness of the state in which a crime occurs to discharge its duty to deliver retributive justice to its own citizens.

B. Passive and Active Jurisdiction

From the perspective of the retributivist theory sketched above, passive jurisdiction seems to provide a compelling basis for state processes. A community’s vindication of its members’ rights is the paradigm application of the collective duty of retribution. But the possibility of bias against outsiders who have wronged nationals should occasion concern. As Gary Bass has argued,

If the international community will not step in, then vengeance will be left in the hands of the victims. They may not be legalistic; they may have highly politicized trials (as in Croatia) or show trials (as in Republika Srpska and Serbia); they may find that the pressure of trying war crimes overwhelms their judicial system (as in Rwanda); or they may prefer revenge to legalism.91

When the scale of victimization is such that the whole society is overborne with grief and anger, the possibility of fair and impartial determinations of individual guilt is undermined.

Assertions of active jurisdiction raise the opposite concern, of unwillingness to aggressively investigate and prosecute citizens accused of international crimes. These concerns are particularly acute where, as is often the case, the state claiming jurisdiction is complicit in the original

91. Gary Jonathan Bass, Stay the Hand of Vengeance: The Politics of War Crimes Tribunals 310 (2000).



crimes or appears implicated in the group antagonisms that led to them. On the other hand, active jurisdiction facilitates authoritative disavowal of the individual and his actions by the collectivity. By distancing themselves from the offender and punishing on their own initiative, a group may better maintain peace and foster reconciliation with the victim’s group. By holding the offender responsible for his own acts, the group also absolves its other members from responsibility, decreasing the likelihood of group retaliation. This process of disavowal and absolution is essential to prevent the duty of retribution from degenerating into further group conflict.92

Ironically, the need for states to accept responsibility for the crimes of their citizens is most acute where the state was complicit in them to begin with. Thus the need for active jurisdiction varies directly with the skepticism with which its assertion will be met. Allowing a national to be prosecuted elsewhere can also be an important gesture of disavowal of the offender, but a state which disputes the underlying allegations or the fairness of the victims’ state’s legal system would have reason to control prosecution in order to guarantee its citizens a just result. When offenders and victims are of different states, duties of retributive justice and concerns with fair processes often cancel each other out. When both states have valid moral claims but neither state can guarantee an impartial trial, international assistance is justified if it is possible.

C. Complementary Jurisdiction

A state’s claim to prosecute violations of international criminal law may seem at its height in cases of internal conflict in which both the victims and the violators are citizens. In these cases the state may assert both passive and active jurisdiction, and bears the duty to give justice to victims while disavowing offenders and their wrongful acts.

92. For more on punishment as a method of disavowal and absolution, see Joel Feinberg, The Expressive Function of Punishment, 49 Monist 397 (1965).



Yet, as we have seen, internal conflict bespeaks of a greater failure of the state to stem group violence. Though states may wish to redeem themselves, it will not always be obvious whether the legal system can resist corruption by surviving elements of the criminal groups responsible for the acts condemned.

The Rome Statute addresses these concerns only partially, by creating a standing international tribunal to prosecute international crimes when state parties are unwilling or unable to do so.93 When states find themselves deadlocked with opposing and equally valid claims of active and passive jurisdiction and concerns with the fairness of the other state’s processes, they may appeal to the ICC to do justice to all concerned. When states devastated by internal conflict find themselves incapable of effectively prosecuting those most responsible for the violence, they may ask the ICC to bear that burden and discharge their shared responsibilities. When a state harbors a citizen accused of international crimes, the Security Council or another state party may ask the chief prosecutor to initiate an independent investigation. In this way the risks of state acquiescence and impotence in the face of past crimes are minimized. The opposite vice, however, of group bias leading to unfair trials, unsupported convictions, or unjust sentences is left unchecked.

The limited and conditional jurisdiction of the ICC reflects its unstable foundation in the political will of states. The ICC threatens not merely to intrude upon sovereignty in the abstract, but to relieve states of fundamental duties with respect to their citizens, denying states their role as guarantor of fair processes and vindicator of victims’ rights. On the relational conception, the ICC calls into question the foundation of the state’s claim to legitimate coercive authority. The ICC’s own claim to authority ultimately rests on the fairness and transparency of its processes and its claim to represent the moral interests of the international community. Yet, as we

93. See Rome Statute, supra note 46, pt. 2, art. 17.



have seen, the concept of the international community remains largely an aspiration of activists and scholars rather than a political reality. States remain unwilling to allow the ICC to sit as a court of appeals when domestic prosecution of international criminals is allegedly biased against defendants. While states generally have the power to overwhelm social groups, international institutions rely on states for their support. The incomplete and consensual nature of ICC jurisdiction remains a necessary imperfection in a legal regime struggling to deal with the deficiencies of groups and states as agents of retributive justice.

CONCLUSION

This article has examined the claims to legitimate coercive authority made by social groups, nation states, and international institutions, claims redeemed by the ability of these collectivities to incorporate the relational structure of retributive justice into institutions and practices that fairly and impartially vindicate the rights of victims. Perhaps inevitably, then, this article has taken the form of a study in comparative incompetence. Each collectivity examined suffers from its own pathologies, for which the others can only partially compensate. The biases of social groups—their favoritism toward their own, their wrath toward outsiders—enables real or perceived wrongs to provoke escalating cycles of retaliation and lasting hostility. The impartiality of the state may be so thoroughly overborne by dominant groups that the state ceases to be an agent of just and measured retribution and becomes an instrument of group violence. The relational structure of retributive justice must be given substance by politically stable institutions, yet support for the International Criminal Court remains uncertain and conditional.

These admittedly pessimistic implications should not obscure the light shed by the theory developed above. Group perpetration and group victimization challenge the legitimacy of the state, and thereby explain and justify



international intervention. The need for substantive law independent of domestic legislation is met by recurring to the formal structure and minimal contents of retributive justice. Genocide, like persecution and apartheid, seeks to reconstruct social reality to systematically express attitudes of contempt for certain groups and deny the moral equality of their members. Jurisdictional principles ration political support for international institutions by limiting their reach to cases in which indifference or impotence have stripped states of legitimate coercive authority and with it the right and duty to secure retributive justice Finally, the international system has largely failed to control the opposite vice, of persecution by a state dominated by victim groups, whose biases may distort legal processes and defeat the aspiration to deliver justice.

In the introduction, a concern was raised that international criminal law frequently operates in the wake of avoidable tragedies, that nations with the power but without the courage to prevent atrocities will welcome the political cover available through referrals to the ICC, and that retributivist approaches seem to provide a philosophical justification for moral cowardice. This concern is both legitimate and welcome, for it turns our attention from the relational structure to the moral foundations of duties of retributive justice. Part I argued that the violation of moral rights grounds the duty of the community, owed to the victim, to punish those responsible. On one interpretation, the community’s duty to punish arises from the wrongdoer’s breach of a duty of restraint. On this interpretation, so far as retributive justice is concerned, the community has no duty to exercise its coercive power on behalf of its members until after their rights have been violated. It is this account that the concern described above rightly attacks.

There is another interpretation, however, according to which the community’s duty to punish arises from its own unwillingness or inability to protect its members from the wrongdoing of others. On the latter interpretation,



punishment is not a substitute for protection but rather a response to a failure to protect. The duties of protection and punishment are related not as alternative grounds for action but as ground to consequent, as primary to remedial obligation. Nations with the means to prevent mass violence cannot escape their responsibility by appealing to the backward-looking nature of retributive justice, for when we look back for a ground of the duty to punish we will find that ground in an earlier dereliction of duty. To invoke the duty to punish atrocity is both to assert the duty to prevent atrocity in the first instance and to admit breach of that prior obligation.

Does such a model of retributive justice, one that grounds backward-looking duties of punishment on forward-looking duties of prevention, cede too much to consequentialist models of crime control? No more, perhaps, than models of corrective justice, those that ground retrospective duties of repair on prospective duties of reasonable care, cede to consequentialist models of cost-avoidance.94 It is not compromise but only candor to admit that justice, for all its majesty, is an often feeble response to our countless failures to force a recalcitrant world into conformity with our highest ideals. In an area of law so fraught with death and suffering, such candor is the very least we owe those whose violations we fail to prevent and instead merely punish.

94. See, e.g., Jules L. Coleman, The Practice of Principle: In Defence of a Pragmatist Approach to Legal Theory (2003).

Group Violence and Group Vengeance: Toward a Retributivist Theory of International Criminal Law

Documents