DATE DOWNLOADED: Wed Oct 7 10:18:01 2020 SOURCE: Content Downloaded from HeinOnline Citations: Bluebook 21st ed. Miriam J. Aukerman, Extraordinary Evil, Ordinary Crime: A Framework for Understanding Transitional Justice, 15 HARV. HUM. Rts. J. 39 (2002). ALWD 6th ed. Aukerman, M. J., Extraordinary Evil, Ordinary Crime: A Framework for Understanding Transitional Justice, 15 Harv. Hum. Rts. J. 39 (2002). APA 7th ed. Aukerman, M. J. (2002). Extraordinary Evil, Ordinary Crime: Framework for Understanding Transitional Justice. Harvard Human Rights Journal, 15, 39-98. Chicago 7th ed. Miriam J. Aukerman, "Extraordinary Evil, Ordinary Crime: A Framework for Understanding Transitional Justice," Harvard Human Rights Journal 15 (2002): 39-98 McGill Guide 9th ed. Miriam J Aukerman, "Extraordinary Evil, Ordinary Crime: A Framework for Understanding Transitional Justice" (2002) 15 Harv Hum Rts J 39. MLA 8th ed. Aukerman, Miriam J. "Extraordinary Evil, Ordinary Crime: A Framework for Understanding Transitional Justice." Harvard Human Rights Journal, 15, 2002, p. 39-98. HeinOnline. OSCOLA 4th ed. Miriam J Aukerman, 'Extraordinary Evil, Ordinary Crime: A Framework for Understanding Transitional Justice' (2002) 15 Harv Hum Rts J 39 -- Your use of this HeinOnline PDF indicates your acceptance of HeinOnline's Terms and Conditions of the license agreement available at https://heinonline.org/HOL/License -- The search text of this PDF is generated from uncorrected OCR text. -- To obtain permission to use this article beyond the scope of your license, please use: Copyright Information Extraordinary Evil, Ordinary Crime: A Framework for Understanding Transitional Justice Miriam J. Aukerman ° I. INTRODUCTION A. The ProsecutionPreference The international debate about the use of prosecutions in transitional justice has focused on the conditions that permit prosecuting those who commit human rights violations. Some critics suggest that international law imposes a duty to prosecute a former regime's atrocious crimes, and contend that states have overstated claims that prosecutions are impossible.' These critics contend that while governments should not press prosecutions when there is "a genuine and serious threat to national life,"2 states should assume "reasonable risks" associated with prosecution, including military discon- tent.3 Other scholars, however, criticize the proponents of prosecutions for assuming that prosecutions will be possible in the wake of human rights disasters. Not only is an amnesty for human rights abusers often a precondition for securing a smooth political transition,4 they argue, but "tm]any fledgling democracies have simply not had the power, popular support, legal tools, or conditions necessary to prosecute effectively."' * Miriam J. Aukerman is currently a staff attorney for Westem Michigan Legal Services. She has clerked forJudge Pierre N. Leval of the United States Court ofAppeals for the Second Circuit, worked forthe New York and Moscow offices of the Ford Foundation, and served as the United States' rcepmntative forthe Gulag Museum, based on the site of the Perm-36 Gulag camp inPerm, Russia. The author would like to thank Alex Boraine, David Garland, Priscilla Hayner Charles Pazdernk, John \Witt. and the editors ofthe HARVARD HumtN RIGHTS JOURNAL for their thoughtful comments on this Article. 1. See Diane E Orenlicher, Setting Accounts:Tke Duty toPrcsrtaeHvt.n RIgtk V:latirs' cfa Pn.rRegime, 100 YALE LJ.2537, 2548 (1991) (arguing that international law requires pmeecution of human rights violations). 2. Id 3. Id.at 2549. 4. See Paul van Zyl, Dilemmas of TranritionalJustixTk. Casecf Sc:th Afitwuos Trmtb a=JRerm.thlzatt= Commission,52J. Ia'r'LAFF. 647,651(1999). 5. Stephan Landsman, AlternativeResponsestoSerious HuanRisks Awes: OfPrc masttarJTruth C .-- missions, 59 LAW & CONTMP. PsoBs. 81, 84 (1996); See alsoCharles Villa-Vicencio, Why Perparatt:m Should Not Always Be Punishuk Where the IntromticnalCrziinal Ccirt arJ Th'ih Cc-zzm:r.s ,if-s.49 EMORY I.J.205, 220 (2000) (describing the challenge ofdeciding whether or not to pro.ecute as Mdiang between Scylla and Charybdis: "t(rhe duty to prosecute ...can shipwreck non-prosecutonal iruttaives; HarvardHuman RightsJournal/ Vol. 15 While participants in this debate disagree as to when trials are possible in practice, they generally share a basic assumption: prosecuting perpetrators of injustice is the optimal method for dealing with past atrocities. 6 The assumption that prosecutions are preferable, while perhaps not always feasible, has fostered a belief that alternative approaches, such as truth commissions, are an inferior substitute for prosecution. Predictably, strong proponents of prosecution argue that "[wihatever salutary effects it can produce, an official truthtelling process is no substitute for enforcement ofcriminal law through prosecutions." 7 But even those who advocate non-prosecution alternatives generally concede the desirability of prosecution.3 Such critics tend to justify non-prosecution alternatives by referring to the difficulty of bringing perpetrators to trial due to inadequate legal systems, the staggering number of potential defendants, or the political consequences of trials.9 What accounts for the widespread assumption that the best way to deal with perpetrators of serious human rights atrocities is to prosecute them? by nations seeking seriously to move away from past gross violations of human rights. The unbridled affirmation of national sovereignty, which may allow nations to devise a form of amnesty that bypasses the demands ofinternational human rights, has, in turn, the capacity to negate the important advances made in the affirmation of human rights [through prosecution]"). 6 See Mary Margaret Penrose, Lest We Pail: The Importanceof Enforcement in InternationalCriminal Lw', 15 Am.U. INT'L L. Rav. 321, 373 (1999) ("Punishment, via criminal prosecutions, is perceived as the most favored method of combating impunity."). 7. Orentlicher, supranote 1, at 2546 n.32; see alsoJuan E.Mendez, In Defense of TransitionalJustice, in TRANSITIONAL JUSTICE AND THE RULE OF LAW INNEw DEMOctCIEs 1, 15 (A. James McAdams ed,, 1997) (arguing that using a truth commission as an alternative to criminal prosecution is "f[t]he most extreme form oftokenism," and suggesting that "[s]ocieties that are in a position to provide both truth and justice to the victims of human rights violations should be encouraged to pursue both objectives as much as they can"). 8. See, eg., Landsman, supra note 5, at 83 (arguing that the best response is usually the "vigorous prosecution of perpetrators"); MARTHA MNOW, BETWEEN VENGEANCE AND FORGIVENESS: FACING HISTORY AFTER GENOCIDE AND Mf.ASSVIOLENCE 58 (1998) (noting that most commentators believe prosecution is the best option and truth commissions should be used only when prosecution is impossi- ble). 9. Susan Dwyer, for example, argues that truth commissions are a morally inferior option only if there is some other strategy a nation could undertake that would be better. For example, if justice, in the sense offairand comprehensive trials and punishment, could be effected, reconciliation would rightly be judged morally inferior. But the availability of realistic alternatives is precisely what is in question in most of the situations in which reconciliation is being recommended. Whether the establishment of truth commissions and efforts at reconciliation are morally inferior responses to violent pasts depends on the availability of other morally acceptable options. Where no such options exist, calls for reconciliation need not be impugned. Susan Dwyer, Reconciliationfor Realists, 13 ETHICS & INT'L AFF. 81, 97 (1999). Seealso Luc Huyse, To Punishorto Pardon.A Devil'sChoice, in REINING IN IMPUNITY FOR INTERNATIONAL CRIMES AND SEIOUS VIOLATIONS OF FUNDAMENTAL HumAN RIGHTS: PROCEEDINGS OF THE SIRACUSA CONFERENCE, 17- 21 SEPTEMBER, 1998 79, 89 (Christopher C.Joyner & M. Cherif Bassiouni eds., 1998) ('[]f the balance of forces at the time of the transition makes a negotiated mildness inevitable, a truth-telling operation with fullexposure of the crimes of the former regime is the least unsatisfactory option."); Landsman, supra note 5, at 83 (arguing that certain violations should always be prosecuted and that, while truth commissions may sometimes be acceptable when "for practical reasons or on the basis of sound policy, prosecution is inappropriate," trials are generally preferable); ARYEI NEIER, WAR CRIMES: BRUTALITY, GENOCIDE, 'TERROR, AND THE STRUGGLE FOR JUSTICE 104 (1998) (describing the amnesty that accompanied the South African Truth and Reconciliation Commission as the "price the black majority needed to pay for a more or less peaceful transition"). 2002 / A Franeworkfor UnderstandingTransitiona/Justice Scholars ascribe a wide range of values to such trials. Douglass Cassel, for example, argues that the International Criminal Courtl u would contribute to justice through the "identification, exposure, condemnation and proportionate punishment of individuals who violated fundamental norms recognized internationally as crimes, and ...reparations to their victims, by means of fair investigations and fair trials by an authorized judicial body."" Stephan Landsman suggests that prosecution "makes possible the sort of retribution seen by most societies as an appropriate communal response to criminal conduct."' 2 Moreover, he argues, prosecution can educate and deter, provide a predicate for compensating victims, enhance the rule of law, and help to heal a society's wounds. 13 Similarly, David Crocker concludes that "(e]thically defensible treatment of past wrongs requires that those individuals and groups responsible for past crimes be held accountable and receive appropriate sanctions or punishment."' 14 Underlying such justifications is the assumption that atrocious human rights violations are in fact crimes. The instruments of systemic savagery can take many forms, from industrialized gas chambers to machete-wielding mobs, from sophisticated torture chambers to steel-toed boots. But the actions of individual human rights violators, such as murder, rape, assault, and torture, are prohibited by almost every domestic criminal justice system.1s Since the Western conceptual framework for dealing with ordinary crime revolves around prosecutors, judges, and trials, it is easy to assume that the same mechanisms should be used to deal with genocide and similar horrors. Is ordinary crime really an appropriate analogy for massive human rights atrocities, what Kant has called "radical evil?" 16 Ordinary crime-individual conduct that violates domestic criminal law and is undertaken for nonpolitical purposes-concerns individual criminals. Extraordinary evil-massive or systematic human rights violations prohibited by international 10. In 1992 the UN General Assembly instructed die International Law Commission to draft a statute for an international criminal court. In 1998 a conference among states was held inRomre. at %htdi the Statute forthe International Criminal Court was adopted. This statute will enter into force aiter ratification by sixty states. Sez HENRY STEINER & PHILiP A sasoN. INTERNATIONAL HuI2A.N RoIGnIMIN CoNTExT 1192 (2d ed. 2000). 11. Douglass Cassel, Why We Nizd tbeInr'itsnar~lCrittzalCc.I, 116 TtE CAIISTIAN C.,TsrVY 532,533 (1999). 12. Landsman, supranote 5,at 84. 13. Id at 83-84. 14. David A. Crocker, Reckoning uwtb Past Wrongs: A Ncrratte Frar.ucrk, 13 Etics & 1sT'L Att. 43, 53 (1999). 15. The author recognizes that this discussion of domestic criminal justice systems and -ordinry crime" reflects the American framework for criminal prosecutions. Because of the power of the Unted States, along with other Western countries, in the international debate about pre!_--uton of human rights atrocities, the criminal justice analogy used in that debate largely relies on Vestern assumptions about ordinary crime. One of the points ofthis Article, however. isthat different socites have ditfternt goals and methods flrdealing with ordinary crime. If we apply an ordinary crime analogy to tranittonal justice, we should recognize these differences. 16. See CARLos SANTIAGO NINO, RADICAL EVIL ON TRIAL vii 11996). HarvardHuman RightsJournal/ VoL 15 law' -involves individuals committing many of the same actions, such as the unjustified intentional taking of human life, that constitute ordinary crimes. The ordinary crime analogy presumes, as Martha Minow puts it, that"even ... massive horrors can and should be treated as punishable criminal offenses perpetrated by identifiable individuals."1 Yet are genocide and ethnic cleansing really just more egregious versions of premeditated murder? Are they merely much larger conspiracies with many more victims? Or are such atrocities qualitatively different from ordinary crime because of the numbers of victims involved and because they are typically undertaken or at least countenanced by state or quasi-state actors for political reasons? Moreover, would such qualitative differences require different responses and remedies? Assuming that the analogy of extraordinary evil to ordinary crime is an appropriate framework within which to examine transitional justice, one must ask, why do societies prosecute ordinary crimes? The short answer is that different societies have different goals for criminal justice. Even within societies there are often fundamental disagreements about the purposes of domestic criminal justice. For example, in the United States, there is a de- 17. For the purposes of this paper, terms such as "extraordinary evil," "radical evil," "massive human rights abuses," or "mass atrocities" are intended as a short-hand for a wide range of conduct, such as that specified inArticle 7 of the Rome Statute of the International Criminal Court: For the purpose of this Statute, "crime against humanity" means any of the following acts when committed as part of a widespread or systematic attack directed against any civilian population, with knowledge ofthe attack: (a) Murder; (b) Extermination; (c) Enslavement; (d) Deportation or forcible transfer of population; (e) Imprisonment or other severe deprivation of physical liberty in violation of fundamental rules of international law; (W Torture; (g) Rape, sexual slavery, enforced prostitution, forced pregnancy, enforced sterilization, or any other form of sexual violence ofcomparable gravity; (h) Persecution against any identifiable group or collectivity on political, racial, national, ethnic, cultural, religious, gender... , or other grounds that are universally recognized as impermissible under international law, in connection with any act referred to in this paragraph or any crime within the jurisdiction of this Court; (i) Enforced disappearance ofpersons; (j) The crime ofapartheid; (k) Other inhumane acts of a similar character intentionally causing great suffering, or serious injury to body or mental or physical health. Rome Statute of the International Criminal Court, 37 I.L.M. 999 (1998), reprintedin part in STEIN8r, supra note 10, at 1192, 1193. See alsoJohn E Murphy, Civil Liabilityfor the Commission of International Crimesasan Alternativeto CriminalProsecution, 12 HARV. Hus. R'rs. J. 1, 19-20 (1999) (distinguishing crimes against humanity from ordinary crime on the grounds that the former "involve either atrocities on a large scale or a policy of acting in a preconceived or systemic way," and noting that there has been a shift away from requiring state action to transform ordinary crime into crimes against humanity); Kristin Henrard, The Viability ofNationalAmnesties in View ofthe IncreasingRecognitionof IndividualCriminalResponsibilityat InternationalLaw, 8 MVcH. ST. U.-DETROIT C. L. J. INT'L L. 595, 605-09 (discussing definitions ofinternational crime). 18. MINOW, supranote 8, at 25. 2002 / A Frameworkfor UnderstandingTransitionalJustice bate about whether offenders should be punished in prisons or rehabilitated in half-way houses; whether or not harsh sanctions have an appreciable deterrent effect; and whether or not reconciliation programs can be attached effectively to prosecutions. Despite such disagreements, however, in the West the determination of individual culpability through prosecution is commonly regarded as necessary to redress criminal actions. The widely held faith in prosecution-whether the goal is punishment or deterrence, condemnation or rehabilitation-disguises disputes about the underlying purposes ofcriminal justice. This Article, while discussing and employing the ordinary crime analogy, ultimately questions the utility of this analogy. The aim here is to examine the reasons why societies prosecute and punish "common criminals," and to question whether prosecutions are the best method of redressing criminal actions in the context of transitional justice. Specifically, this Article will challenge the assumption that prosecutions are always the best way to pursue justice in societies in transition by arguing that the choice between prosecution and non-prosecution alternatives should depend on what one is seeking to achieve. The best known alternative to prosecution is the truthcommission. However, truth commissions are not the only alternative; other options include reports by international delegations, lustration, civil liability, reparations, and historical inquiry.19 Additionally, each of these alternatives may take many different forms.20 The purpose of this Article is not to champion any specific alternative, although contrasts will be drawn betveen prosecution and alternative mechanisms-particularly truth commissions-in order to highlight the distinctive features of prosecutions. Rather, this Article will challenge the primacy of prosecution and will argue for goal- and culturespecific responses to mass atrocity.As the term "non-prosecution alternative" suggests, prosecution21 and forgetting are nor the only options available in the context of transitional justice. 19. See, eg., ML CherifBassiouni, Searchingfor Pce andAdin ng jettca TcNhiviAccngt..c dtt 59 LAw & CONTEMP. PROBS. 9, 20-22 (1996) (listing a variety of accountability mechanisms; Murphy supra note 17, at 47-55 (arguing fbr the viability of civil liability as an alternative 1o prvAccutons), Roman Boed, An Evaluation f the Legality andEf/jcai ofLatratica as a Tcl cfTrantt:alJuttte, 37 CoLIi1. J. TRANSNAT'L L 357 (1999) (questioning the ability of lustration to achieve the goals of transitional justice); Graciela Fernandez Meijide er al., The Role of Hiocriwalluquiry inCratir AecezalthtyjJr Human Rights Abuses,12 B.C. THIRD WoRLD UJ.269 (1992). 20. See,eg., Priscilla B. Hayner, Fifteen Truth C':,fissins-1974 to 1994: A Cc=parattwStcrty,16 HUm. RTs. Q. 597 (1994) (describing a wide variety of truth commissions). 21. In this Article the term "prosecution" is used to refer to both domestic and international efforts to bring perpetrators to trial. Some statesmen and scholars have argued that international prosecutions are necessary because domestic trials often do not meet due process standards. S, eg.. Seth Mydans, T, Shape ofJustice Is Undefin d in Cambodia, N.Y. TorEs, Dec. 5, 1999. at A4 (quoting Kofi Annan as demanding international control over efforts to prosecute Khmer Rouge leaders on the grounds chat "crimes of such a magnitude demand a judicial process that answers to the highest norms of integrity and law"). Although there may be differences in the extent to which domestic versus international prosecutions can meet the traditional goals of criminal justice while simultaneously respecting defendants' rights, such a comparison is beyond the scope of this Article. See Neil J. Knz. Cc--ng n, Tra th 44 HarvardHumanRightsJournal / Vol, 15 The social and political realities of a particular transitional context will affect the kind of justice that can be pursued. Before one determines whether or not prosecution is feasible, however, one must ask if it is even desirable. Prosecutions are better designed to achieve some goals than others. Nonprosecution alternatives are indeed a second-rate option when prosecution, though politically difficult, would best serve the goals of transitional justice, But one cannot presume the inferiority of non-prosecution alternatives without first articulating the desired goals of transitional justice. Some goals can be best achieved through non-prosecution alternatives, regardless of whether prosecutions are politically feasible. As Argentinian philosopher and human rights activist Carlos Nino has suggested, "the extent of the duty of a government.., to prosecute past human rights abuses depends... on the theory that underlies the justification of punishment."22 In other words, we must decide what we want from transitional justice before we can decide ifprosecutions are the best way to achieve it. B. The GoalsofCriminalJustice:An Overview In the context of domestic crime, penologists have provided a variety of theoretical frameworks for justifying punishment and for dealing with offenders and the crimes they commit.23 These approaches may be classified generally as: desert/retribution/vengeance, deterrence, rehabilitation, restorative justice, and communication/condemnation/social solidarity.2 4 This Atrocities: A Review ofAccountability AMechanismsforMass Violations ofHuman Rights, 59 LAw & CONTIMP. PRODS. 127, 129-33 (1996) (discussing the advantages and disadvantages ofinternational and domestic prosecutions). 22. Carlos S. Nino, The Duty to PunishPastAbuses ofHuman RightsPut into Context: The Caseof Argcttina, 100 YALE L.J.2619, 2620 (1991). 23. See generally PRINCIPLED SENTENCING: READINGS ON THEORY & POLICY (Andrew von Hirsch & Andrew Ashworth eds., 2d ed. 1998); CONTEMPORARY PUNISHMIENT: VIEWS, EXPLANATIONS, AND JUSTIFICATIONS (Rudolph Gerber &Patrick McAnany eds., 1972); THEORIES OF PUNISHAtENT (Stanley E. Grupp ed., 1971); ANDREW VON HIRSCH, CENSURE AND SANCTIONS (1993); NIGEL WALKER, WiY PUNISH? (1991). 24. Notably absent from this list is incapacitation, rendering the offender incapable of reoffending through physical restraint. Unlike rehabilitation, incapacitation does not aim to change or improve the offender in order to make her less likely to offend; rather, incapacitation seeks to impede the person from carrying out whatever criminal inclinations she might have. See generallyAndrew von Hirsch, Incapacitation, in PRINCIPLED SENTENCING: READINGS ON THEORY & POLICY supra note 23, at 88, Because the concept of incapacitation has been almost utterly absent from the debate about how and why to punish perpetrators of mass atrocities, a detailed comparison of the concept in the contexts of criminal and transitional justice has been omitted. Several points are nonetheless worth making. First, prosecutions for human rights atrocities are generally not undertaken where the perpetrators remain an immediate threat, or where there is a realistic danger that the forces responsible for past abuses will remain in power or attempt to regain it. As Van Zyl notes, in the South African case, "[t]he former government and its security forces never would have allowed the transition to a democratic order had its members, supporters or operatives been exposed toarrest, prosecution and imprisonment." Van Zyl, supra note 4, at 650. Even after a transition has taken place, potential defendants may be "an integral part ofan intact military establishment fully capable of bringing down the government if too directly threatened," Landsman, supra note 5, at 84. Thus, in the transitional justice context, prosecutions are rarely used to incapacitate those persons likely to reoffend. In fact, prosecutions may be avoided precisely because they may make particular perpetrators who remain dangerous even more dangerous. Second, to the extent that 2002 / A Framwworkfor UnderstandingTransitionalJustice Article will argue that although each of these themes is relevant to transitional justice, none can blindly be transposed from the domestic context. Each of the five main Sections below corresponds to one approach, and will begin with a brief outline of the theoretical underpinnings of that approach as applied in the domestic context. Each Section then considers the implications of that approach for transitional justice, highlighting the ways in which the scale and nature of "radical evil" complicate not only the pursuit of the five possible goals of punishment, but also the very analogy of extraordinary evil to ordinary crime. Prosecution can certainly satisfy a society's demand for retribution in reaction to massive human rights violations. However, if in reaction to massive human rights violations a society seeks to achieve another of the possible goals of punishment, a nuanced inquiry is required to determine which mechanism would be best suited to that goal. In order to decide whether prosecution or non-prosecution alternatives are more likely to promote deterrence, restorative justice, rehabilitation, or condemnation, we need a clearer understanding of what these concepts mean in the context of transitional justice, as well as how they might apply in a country undergoing transition. If the appropriate mechanism for confronting grave human rights violations depends upon specified goals, who has the authority to set these goals? If transitional societies themselves have the right to decide, then we must recognize that different societies will have differing goals.2 Some societies emerging from mass trauma will demand retribution, while others will focus on compensation. Still others may concentrate on rebuilding a shattered economy or on strengthening democratic institutions. If different societies want different things, and if prosecution is a more effective tool for achieving some goals than others, we cannot presuppose that all societies in transition should choose prosecution. exile offbrmer leaders serves to insulate post-transition societies from powerful former leaders, itcan be a form of incapacitation. When faced with the threat of prosecution, abusive rulers may be less willing to give up power, making it more difficult to incapacitate them. Thus prosecutions, because [hey make it harder to find a "safe"country of exile forformer dictators, can run counter to the goal of incapa tation. Third, it is usually impossible to incapacitate all of those individuals who have been involved in past abuses. Yet it is unclear whether incapacitation of a few key leaders will successfully inoculate society against the danger that such crimes will be repeated. Moreover, while incapa.itation is justified by the individual's predisposition towards crime, in cases of mass violence the link between individual dangerousness and the likelihood ofinvolvement in future atrocities is less clear.Finally; one might think in terms ofincapacitating a system ofabuse, rather than incapacitating individual offenders. The tmrtment of Germany after World War I is the classic example of the failure of soetal mcal"actation unposed from the outside. Disarmament and heavy reparations were supposed to prevent Germany. beltved to be inherently aggressive, from acting on these intentions. Where incapacitation reflects a society's decision to check its own worst impulses, however, it need not be so monumentally unsuccessful. One could, for example, envision a post-transition commitment to civilian control of the armed forces, independence of the judiciary, and the rule of law as an effort to restrain tendencies towards military coups, victor's justice, and unbridled vengeance. 25. SeeMAtNow, supra note 8, at 4 (describing differing survivor responses to collective violence). HarvardHumanRightsJournal / Vol. 15 If, on the other hand, the international community has the right to set goals for transitional justice, we must ask whether or not the international community even agrees about what the most important goals should be.26 An indiscriminate duty to prosecute 7 assumes that the international community shares a fixed hierarchy of goals, agrees that these goals are best served by prosecution, and feels comfortable imposing such a vision on the society in question. Certainly there is at least superficial agreement that human rights atrocities should be prevented and condemned. But this limited consensus provides little aid in deciding how to prioritize these two goals as against others, nor does it signify complete agreement that prosecution is the best way to prevent future atrocities or communicate shared outrage. Moreover, even if the international community could make a clear choice as to its goals and the means of achieving them, it is not obvious that such an international strategy should trump the wishes of the local society.28 Those who have not suffered cannot presume to determine for those who have what should be attempted through transitional justice.29 Domestic criminal justice systems have sometimes been criticized for ignoring the needs and desires of victims, for "stealing conflicts" from those involved.30 Similar problems will arise if transitional justice reflects only the priorities of the international community (or of powerful states within it), and not those of the affected country. Of course, the international community does have both a role and an interest in transitional justice. Just as ordinary crime is not simply an offense against the individual victim but against the entire society, so extraordinary evil is not merely an assault on the particular traumatized society but on humanity as a whole. As a result, the choice of retribution or deterrence, reconciliation or condemnation cannot be left solely to either the international community or the local society. Transitional justice must reflect the needs, desires, and political realities of the victimized society, while at the same time recognizing the international community's right and responsibil- 26. While particular societies may also suffer from disagreements about appropriate goals and meth. ods, states-unlike the international community-have governments and political processes for resolving such conflicts. 27. See, e.g., Orentlicher, supranote I passim (arguing that international law imposes a duty to prosecute massive human rights violations). 28. An example of the tension between local and international control can be seen in the case of Rwanda. While the Rwandan government initially requested that the UN Security Council establish an international tribunal, Rwanda became dissatisfied with various aspects ofthe resulting plan, and cast the only vote against the establishment of the tribunal. See Penrose, supranote 6, at 342-43. 29. See Paul van Zyl,Justice Without Punishment: GuaranteeingHuman Rights in TransitionallSodcej, In LOOKING BACK/REACHING FORWARD: REFLECrIoNs ON THE TRUTH AND RECONCILATION CoMiisSION OF SOUTH AFRICA 42, 52 (Charles Villa-Vicencio & Wilhelm Verwoerd eds., 2000) (suggesting that one ofthe factors in determining whether a state should be able to derogate from the obligation toprosecute is whether or not there is convincing evidence that the majority of citizens have freely chosen the transitional justice policy endorsed by the state). 30. See Nils Christie, Conflicts as Property, in PRINCIPLED SENTENCING: READINGS ON THEORY & POLICY supra note 23, at 312, 312 (arguing that conflicts can be seen as a form of property the state appropriates through the criminal justicesystem). 2002 / A Frameworkfor UnderstandingTransitionalJustice ity to intervene. In practice, the balance between domestic and international control over transitional justice will likely be determined in large part by the political realities ofa particular transition.3 1 Nonetheless, we must think more carefully about the extent to which transitional justice should reflect local rather than international choices. This Article does not suggest that any one of the goals discussed below should be given priority, either in general or with regard to any particular society considering how best to come to grips with its past. Indeed, since the choice ofgoals will affect the choice of mechanisms, and since the choice of goals and effectiveness of mechanisms will depend at least in part on cultural factors, it would be far from realistic to suggest that any one means of dealing with the past will always be superior. The main thrust of this Article, then, is that greater honesty in explaining why we punish those who commit horrific abuses will clarify what transitional justice can achieve, and increase the likelihood of achieving it. C. Assumptions Before moving to a discussion of each of the different approaches, it is necessary to identify three assumptions implicit in this Article. First, forgetting is unacceptable. In part this isbecause victims of horrible atrocities are simply unable to forget. Without some form of accounting, past atrocities inevitably fuel future ones.32 A nation's unity, explains Chilean human rights advocate Jos6 Zalaquetr, "depends on a shared identity, which in turn depends largely on a shared memory. The truth also brings a measure of healthy social catharsis and helps to prevent the past from reoccurring." -3 Truth and accountability are essential if traumatized societies are to begin resolving their political, ethnic, racial, and religious conflicts through democratic processes, rather than through torture, rape, and genocide. The recent horrors in Rwanda and the former Yugoslavia can be attributed in part to the failure to confront earlier ones. By contrast, denazification contributed to the stability of Germany's post-war democracy.34 As Richard Goldstone, the former chief prosecutor for the International Criminal Tribunals for the former Yugoslavia and Rwanda, has noted, "Itheonly hope of breaking cycles of violence is by public acknowledgement of such violence and the exposure 31. The physical location of human rights abusers, and the willingness of local authonres to hand them over, are significant Factors in determining the extent of local control over transitional justice. Far example, the international effiirts to prosecute Augusto Pinochet were possible because he was abroad. and hence could bearrested. Where international prosecutions depend more heavily on local cooperation, local communities will have more say in the form oftransitional justice adopted. 32. Kritz, supranote 21, at 127 ('[to suggest] that individuals or groups who have been the victims of hideous atrocities will simply forget about them or expunge their feelings without wime form of accounting, some semblance ofjustice, is to leave in place the seeds offuture conflict-). 33. Jos6 Zalaquett, BalancingEthiralImprtites and PclitizalCcwstraints:Tk Dz4=c-a cf Neu Dt=erades ConfrentingPastHuman RightsVilations,43 HAsTnGs LJ.1425. 1433 (1992). 34. See KRK OSIEL, SS ArRocrn; COLLECrivE MIEMORy, AND THE Lw 193 (1997) (noung dur trials ofNazi officials allowed for the creation ofa new German identity). HarvardHuman RightsJournal / Vol. 15 of those responsible for it." 3 5 Victims have a deep need for acknowledgment, he continues, and they "cannot get on with building the future until their calls for justice have been answered. '36 Amnesia or blanket amnesties without any recognition of the past are simply not an option.37 Second, if prosecutions are undertaken, they must comport with accepted standards of due process. While this may seem obvious, in fact procedural abnormalities are common in trials undertaken in transitional contexts,3 8 Jon Elster has cataloged the use in such prosecutions of exceptional measures, including illegal internments, the presumption of guilt, biased selection of judges and jurors, lack of appeals mechanisms, collective guilt, and retroactive legislation.39 Moreover, it is not self-evident that punishment undertaken in the name of transitional justice necessarily requires trials. For example, as World War II drew to a close, Britain and Russia argued for the summary execution of their enemies: to them, the crimes were so apparent that no trial was necessary.4 0 Although a number of German war criminals were eventually tried at Nuremberg, elsewhere summary executions did take place; in post-World War II France, between 20,000 and 50,000 alleged collaborators were assassinated. 41 When trials are eschewed, the criminal justice analogy is typically replaced by an alternative conceptual framework. For example, political leaders who have insisted that evidence against those who orchestrated the recent World Trade Center and Pentagon attacks need not be sufficient for an indictment in a court of law have done so within the framework of a "war on terrorism."4 2Within this framework, terrorists may be punished without trials or due process. However, if the conceptual framework were to shift back to criminal justice, and suspected terrorists were to be brought to trial, due process safeguards should still apply, just as they have in other terrorism-related trials.43 35. Richard Goldstone, Exposing Human Rights Abuses-A Help or Hindrance to Reconciliation?, 22 HASTINGS CONST. L.Q. 607,615 (1995). 36. Id. 37. See Nino, supranote 22, at 2630. 38. SeeJon Elster, Coming to Terms with the Past:A Frameworkfor the Study ofJustice in the Transition to Democracy, 39 EUR. J. SoC. 7, 24 (1998) (arguing that procedures used in dealing with the past tend to be exceptional). 39. Id.at 24-26. 40. See MINOW, supra note 8, at 29. 41. Roy C. Macridis, France: From Vichy to the FourthRepublic, in FROM DICTAToRsIP TO DutMocRACY: COPING WITH THE LEGACIES OF AUTHORITARIANISI AND TOTALITARIANISM 161, 171 (John ff. Herz ed., 1982). 42. See, e.g., Suzanne Daley, NATO Says U.S. HasProofAgainst bin Laden Group,N.Y. TIMEs, Oct. 3, 2001, at Al (quoting U.S. Secretary of State Colin Powell as stating that although the U.S. had "pretty good information" establishing the guilt ofOsama bin Laden, "it is not evidence in the form of acourt case"); Elizabeth Bumiller, Bush Pledges Attack on Afghanistan Unless It Surrenders bin Laden Now, N.Y. TIMES, Sept. 21, 2001, at Al (quoting U.S. President George Bush as framing the American response to the September 11, 2001 attacks as a "war on terrorism"). 43. For example, the United States postponed the execution of Timothy McVeigh, convicted of the bombing of the Federal Building in Oklahoma City, until the judiciary could review the government's failure to turn over materials to which the defendant was entitled as amatter of law. 2002 / A Frameworkfor UnderstandingTransitionaIJustice While certain procedural deviations may be acceptable-a relaxed application of the concept of nulla poena sine lege, for example-basic due process rights should be respected, even when those on trial are accused of genocide.4 There are ethical as well as practical reasons to respect defendants' rights. By upholding standards of fairness, one can, in Vaclav Havel's famous words, show that "we are not like them."4 ' Due process is designed in part to protect the innocent from punishment and prevent excessive punishment of the less guilty. But due process is also what gives legitimacy to trials and convictions. Much of what prosecutions can achieve, from communicating reprobation to legitimating societal demands for revenge, depends for its success on a public belief in the fairness of trials. The failure to adhere scrupulously to fair trial standards and apply laws equally may lead to the perception that trials are merely exercises in partisan politics or victor's jus- tice. 46 Of course, due process requirements must also apply to non-prosecution alternatives. For example, the ongoing debate about whether truth commissions should make public the names of offenders reflects the concern that those individuals have not had the chance adequately to defend themselves.4 7 Arguably, however, there is a critical distinction between due process in the non-prosecution and prosecution contexts. When applied to non-prosecution alternatives, due process is an inherently flexible concept. 48 For example, due process might require more procedural safeguards if a commission can strip offenders of their jobs than if it can merely publish reports about their conduct. In criminal prosecutions, by contrast, there is a fairly fixed idea of what due process means, reflecting in part the severe consequences of a 44. Se, eg., fNow, supra note 8, at 50 (arguing that because one important goal of trals is to promote human rights, they should be used only where there is a chance for or the percepton of firnts). 45. Quotedin The Center far Security Policy, "Fra'Czzloiaki.&a?Shadzis crr i: Trawtt:, at http:j/ www.security-policyorglpapersl19909O-55.html (last visited Nov. 7, 2001). 46. See Kritz, supra note 21, at 137. 47. For example, Jos6 Zalaquert, a commissioner on the Chilean Truth Commission, has argued that "[io name culprits who had not defended themselves and were not obliged to do so would have been the moral equivalent to convicting someone without due process." RrPoRT OF THE CHLEAN NAxtom. COmmISION oN TatrTII AND REcoILATiON .xtxii (Phillip E. Berryman trans., 1993), Su-ah, Hayner, supra note 20, at 647-50 (examining the debate about whether or not truth commissions should name names, and reporting on the practices ofindividual truth commissions). 48. The American jurisprudential approach to due process is helpful in understanding that due process can mean different things in different contexts. As the U.S. Supreme Court noted in Xtatkt,,i c El. dridge, due process generally requires consideration ofthree distinct factors: First, thce prvate interest that will be affected by the official action; second, the risk of an erroneous dep~nitatn of such interest through the procedures used, and the probable value, ifany. ofadditional or substitute procedural safeguards; and finally, the Government's interest, including the functon involved and the fiscal and administrative burdens that the additional or substitute procedural requirement would entail.... 424 U.S. 319, 335 (1976). In considering 'what process is due," U.S. courts havegenerally attempted to determine which of the procedural formalities of a trial should be incorporated into the ministrative proceeding. See Judge Henry Friendly, Soe Kird of Hcarmg, 123 U. PA. L Rrv 1265 (1975) (lsting possible ingredients ofdue process). HarvardHumanRightsJournal / Vol. 15 criminal conviction. Despite considerable variation among legal systems, "fair trials" are generally understood to require a set of basic rights, such as the right to an impartial tribunal, the right to counsel, the right to crossexamine one's accuser, and the right to present evidence in one's own de- fense.49 Respecting due process requirements certainly makes it harder to convict human rights violators, just as it makes it more difficult to convict common criminals. While we must be willing to pay this price, it can be dear indeed. In the case of Chilean ex-dictator Augusto Pinochet,' 0 a Chilean appeals court suspended his prosecution on the grounds that he was not mentally fit to stand trial."1 Human rights activists argued that Pinochet's illness was exaggerated and that he "could [have] befen] tried without a violation of due process."52 But given the court's finding of incompetence, due process under Chilean law prohibited Pinochet's prosecution.53 Less well-known is the case of Jean-Bosco Barayagwiza, a Rwandan media leader who incited Hutus to take up arms against Tutsis and who has been called the "number 49. For example, Article 21 of the Statute of the International Tribunal for the Former Yugoslavia provides: 1. All persons shall be equal before the International Tribunal. 2. In the determination of charges against him, the accused shall be entitled to a fair and public hearing, subject to article 22 ofthe Statute. 3. The accused shall be presumed innocent until proved guilty according to the provisions of the present Statute. 4. In the determination of any charge against the accused pursuant to the present Statute, the accused shall be entitled to the following minimum guarantees, in full equality: (a) to be informed promptly and in detail in a language which he understands of the nature and cause of the charge against him; (b) to have adequate time and facilities for the preparation of his defence and to communicate with counsel ofhis own choosing; (c) to be tried without undue delay; (d) to be tried in his presence, and to defend himself in person or through legal assistance of his own choosing; to be informed, ifhe does not have legal assistance, of this right; and to have legal assistance assigned to him, in any case where the interests ofjustice so require, and without payment by him in any such case if he does not have sufficient means topay for it; (e) to examine, or have examined, the witnesses against him and to obtain the attendance and examination ofwitnesses on his behalf under the same conditions as witnesses against him; (0 to have the free assistance of an interpreter if he cannot understand or speak the language used in the International Tribunal; (g) not to be compelled to testify against himself or to confess guilt. Statute of the International Tribunal for the Former Yugoslavia, Article 21, at http://www.un.org/ictyl basicstatut/stat2000.htm#21 (last visited Oct. 10, 2001). Virtually identical language is contained In Article 20 ofthe Statute of the International Tribunal for Rwanda. See Statute of the International Tribunal for Rwanda, at http://www.ictr.org (last visited Oct. 10, 2001). 50. Augusto Pinochet ruled Chile from 1973 until 1990, after taking power in a violent coup. An estimated 3,200 people were executed or "disappeared" during General Pinochet's rule, and tens of thou. sands more were tortured. Clifford Krauss, Chile's Effort to Try PinochetIs Running Out of Steam, NY. TmsaS, June 25, 2001, at A3. 51. See Human Rights Watch, Pinochet Decisions Lamented, at http://www.hrw.org/pressl2001/07/ pino0709.htm (last visited Oct. 4, 2001);seealsoKrauss,supranote 50, at A3. 52. Human Rights Watch, supranote 51. 53. Id. 2002 / A Frameworkfor UnderstandingTransitionalJustice one" culprit in the Rwandan genocide.54 In November 1999, the International Criminal Tribunal for Rwanda ("CICTR") dismissed the case against him on the grounds that Mr. Barayagwiza's fundamental rights had been violated by prolonged detention without trial.5 Respecting the speedy trial rights of a perpetrator responsible for the deaths of thousands is one cost of using a prosecution-based approach premised on the rule of law.56 Due process requirements also affect who is prosecuted. For example, in deciding how to focus the limited resources of the Office of the Prosecutor of the International Criminal Tribunal for the Former Yugoslavia (ICTY"), Richard Goldstone and his colleagues had hoped to target those who planned and ordered the commission of crimes, rather than those who simply undertook them. But because the Tribunal requires sufficient evidence for guilt to be proven beyond a reasonable doubt, and because there was often not enough evidence to establish who ordered the commission of the atrocities, the Office of the Prosecutor was forced to focus on lower-level cases where a longer paper trail and more witness testimony were available.17 Thus, the evidentiary requirements of due process may make it not only generally more difficult to convict, but also particularly difficult to convict those who are most culpable. The third assumption implicit in this Article is that prosecutions are necessarily selective. Often entire societies are implicated in atrociries.5s As Jos6 Alvarez notes, massive human rights violations "usually involve massive complicity by large numbers of perpetrators, at all levels of domestic and international society, and not merely by a select group of government elites."'59 Most commentators recognize that in the wake ofsuch widespread guilt, only a small number of even the worst perpetrators will ever stand trial. Because mass atrocities are generally perpetrated by a large number of 54. Christopher S.Wren, U.N. TriunalWrong to FretTcp Susa. Ruz.mda Says, N.Y TWES, No-. 12. 1999, at A10. 55. I 56. One can disagree about whether or not due process requires a speedy tral for Sgcrdaffa or about how long a person may legitimately be detained without trial. For example, some jurisdictions have created exceptions tostandard speedy trial requirements in murder cases. St,eg., N.Y. Cwtsa. Paoc- Lime, § 30.30(3Xa) (2001) (establishing that statutory speedy trial requirements are inapplicable for defendanis charged with murder). Even in the absence of any right to a speedy trial forEtrodaers,however, the basic problem remains. Due process will sometimes require placing a higher prionty on protecting the rights ofdefendants than on establishing their guilt. As a result, some of the guilty will go free. 57. See Goldstone, supranote 35, at 617. 58. For example, ofthe Rwandan genocide, Bernard Mfuna has said, mhe genocide in Rwanda was five times faster dan de one in Germany. even though the German genocide had gas chambers. If you take the lower figure of 500.000 people killed you are looking at 5,000 people a day. If you take the higher figure of one million people killed, you are looking at 10,000 people killed a day without guillotines or gas chambers. Instead, most of the killings were done with match heads [sic] and spears. This meant dt a large proportion ofthe population were (sic] implicated for this to succeed. Bernard Muna, Conferae en War Crime Triunals: Tk Ruardan Tritsal arJis Rdlatsz".skp tiNaut=l Trialsin Rwanda, 13 AL U. L'n'L L REV. 1469, 1480 (1998). 59. Jos6 E. Alvarez, CrimesofStareslCrimes ofHate Lcscsufr_ Ruar.a,24 YALE J. IN-r*L L 365. 467 (1999). HarvardHumanRightsJournal/ Vol 15 people, "[pirosecution of every single participant in the planning, ordering, or implementation of the atrocities in question-not to mention all those who collaborated with them-would be politically destabilizing, socially divisive, and logistically and economically untenable." 60 Indeed, even the strongest advocates of prosecution do not believe that it is possible to prosecute every perpetrator.61The fact that the ICTY has only two trial chambers demonstrates that, in Goldstone's words, it only has the capacity to try "a relatively small number of cases. '6 As of October 2001, sixty-one accused had appeared in proceedings before the ICTY. Only twenty-three defendants had been tried, of whom two had been acquitted, and twenty-one convicted.63 As of September 2001, there had been fifty-two ICTR detainees, of whom one had been acquitted, three had pled guilty, and five had been con- victed."1 Prosecutions, in part because of the due process requirements involved, are also expensive. 6' Costs escalate if trials are held outside of the transitional country, as witnesses and evidence must be transported to the distant forum. Criminal justice resources-not only funds, but staff and expertise as wellmay be sufficient to prosecute "only a small fraction of those responsible for gross violations of human rights. '66 In its efforts to prosecute widely, the Rwandan government has confronted the problem of inadequate resources. While at least 1420 people have been tried, Amnesty International has reported that some trials have been unfair, that tens of thousands of genocide suspects have been detained for years under horrible conditions, and that due to the backlog the government plans to introduce a system for handling lower-level offenders which "do[es] not conform to basic international standards for fair trials."16 As Paul van Zyl notes, "[clriminal justice systems... are designed for societies in which the violation of law is the exception and 60. Kritz, supra note 21, at 138-39. See also Report to the United Nations Secretary-General, 16 March 1999, Human RightsQuestions,Including AlternativeApproachesfor Improvingthe Effcctive Enjoyncnt of Human RightsandFundamentalFreedomsin Cambodia,U.N. Doc. A/5318501/19991231 106 (heteinfter CambodiaReport] (arguing that prosecutions in Cambodia should be limited to top leaders because an effort to try all perpetrators would be "logistically and financially impossible for any sort oftribunal that respects the due process rights ofdefendants"). 61. See, e.g., Orentlicher, supranote 1,at 2601 (arguing that international law does not require prosecution of all offenders). 62. Goldstone, supranote 35, at 617. 63. See Fact Sheet on ICTY Proceedings, at hrtp://www.un.org/iccy/glance/procfact-e.htm (last updated Oct. 10, 2001). See alsoPenrose, supranote 6, at 368 (noting that during the first six years ofopera. ion, the ICTY completed only six trials involving nine separate defendants). 64. See ICTR Detainees--Status on 10 October 2001, at http:l/www.ictr.org/ENGLISH/factshects/ deainee.hrm (last visited Oct. 10, 2001). 65. Penrose, writing in 1999, estimated that the ICTR and ICTY had spent approximately $400 million to obtain a handful ofconvictions. See Penrose, supranote 6, at 391. 66. Van Zyl, supranote 4, at 651. 67. Amnesty International, Rwanda: The Troubled Course ofJnstie, at http:l/www.amnesty-usa.org/ news/2000/14701500.hrm (last visited Oct. 9, 2001). See also Amnesty International, Annual Report 2000: Rwanda, at http:l/www.web.amnesty.org/web/ar2000web.nsf1785alacb99d4e7d4802568f8OO3cb 05917d4208bcf86ce147802568f200552963 (lastvisited Oct. 9, 2001). 2002 / A Frameworkfor UnderstandingTransitionalJustice not the rule. Once the violation of the law becomes the rule, criminal justice systems simply cannot cope."68 The result, he concludes, is a defactoamnesty for many.69 While selective prosecution is used in virtually every legal system " u and prosecutors everywhere are faced with the need to concentrate their limited resources on the most culpable offenders, selective prosecution after genocide or massive human rights abuses presents a qualitatively different challenge. In a functioning criminal justice system, a prosecutor's decision not to proceed typically reflects a lack of evidence (so that a conviction is unlikely), an understanding that the defendant's actions were justifiable (so that a defense to the crime exists), or a recognition that the crime was relatively insignificant (so that obtaining a conviction is not an efficient use of resources). Selective prosecution in the transitional justice context, on the other hand, takes place despite compelling evidence that the perpetrators have committed the most heinous ofcrimes, and have done so without justification. The inevitable selectivity of prosecution does not render it an unacceptable approach. Despite this limitation, trials can achieve important criminal justice goals. Nevertheless, the narrow scope of prosecution--due in part to the requirements of due process-is central to an assessment of how well prosecution can fulfill each of the five possible goals of criminal justice. H. CRMIINAL JUSTICE As TRANSITIONAL JUSTICE A. Desert/RetributionVengeance 1. Theory Desert theory is premised on the relationship between punishment and culpability. "Retributivism is a very straightforward theory of punishment," writes penologist Michael Moore. "W]e are justified in punishing because and only because offenders deserve it."' Unlike other theories of criminal justice, "just deserts" is thus explicitly backward-looking. For a pure retributivist, there is a Kantian categorical imperative to punish, whether or not punishment will prevent future crime.72 While desert theory is distinctive in its concern for the past rather than the future, it is like most other approaches to criminal justice in its focus on the offender. Victims may re- 68. Van Zyl, supranote 4, at 661. 69. Id. 70. See Orendicher,supranote 1,at 2601-02 (noting that selectivity in prosecutns is a common teature ofalmost all legal systems). 71. Michael S. Moore, The Moral Wch of Ranrduttn. in PRINCtPLrD Sa'N1-NcIu Rt.Di.tNS O.% THEORY AND PoLicy supra note 23, at 150, 150. 72. SeeWALKER, supranote 23, at 77 (discussing Kant's view that society has an ebligation to exe-ute the "last murderer," even if society is about to come to an end). Some scholars are pcrmmtse retnbutists" who hold that criminals lose their immunity from punishment, but beiese that punshment is not mandatory and should be justified by other goals. See Nino, iara note 22, at 2620 (d.nbing the "putmissive retributivist" position). This Article concentrates on pure retrnbuttvtsm, since such a vew t ms to be the predominant reributivist approach in the context of massive human rights atrcittes HarvardHumanRightsJournal / Vol. 15 ceive satisfaction from the knowledge that the perpetrator is punished.73 But because crime is an offense against society as a whole, it is society, and not the victim, that determines the perpetrator's guilt and the appropriate level of punishment. Moreover, decisions about the seriousness of the offense are based on the offender's moral culpability, not on the degree of harm suffered by the victim. Three basic questions usually arise in considering approaches to punishment: who should be punished, how much should they be punished, and why should they be punished? Desert theorists generally agree on the answers to the first two questions.7 4 Regarding the first question, desert theorists believe that punishment should be restricted to those who have committed crimes.75 As for how much to punish, desert theorists rely on the principle of proportionality, the notion that the severity of a sentence should be proportional to the seriousness of the criminal conduct.7 6 The related requirement of ordinal proportionality determines the degree to which a crime should be punished relative to other crimes. 7 The real debate in desert theory concerns the third question: why punish? The tension among desert theorists-like the skepticism of critics of retribution-is grounded in a deep discomfort with vengeance. Revenge, some fear, will lead to a downward spiral of violence and recrimination.78 Moreover, explains scholar Jeffrie Murphy, one tends to think "that it is only primitives who would actually hate criminals and want them to suffer to appease an anger or outrage 73. A theory of mandatory retribution, unlike most other approaches to criminal justice, may even imply that victims have a right that their abusers be punished. SeeNino, itipranote 22, at 2621: Almost all views ofpunishment, with the exception of mandatory retribution, deny that anybody has a right that someone else be punished for a past crime. Punishing those who have relinquished their right not to be punished [by committing a crime] is not due to the rccogni tion that the victims or their relatives have a right to that punishment. It is a consequence of a collective goal imposed by the policy ofprotecting human rights for the future. 74. Andrew Ashworth, Desert, in PRINCIPLED SENTENCING: READINGS ON THEORY AND POLICY Ilpra note 23, at 141, 141. 75. See id at 143. Although this limitation may seem obvious, desert theorists attack approaches like deterrence and incapacitation on the grounds that they would logically allow for punishment that is not based on the commission ofa crime. See, eg., Andrew Ashworth, Deterrence,in PRINCIPLED SENTENCING: READINGS ON THEORY AND POLICY supra note 23, at 44, 46 (noting that deterrence theory may require punishment of the innocent for the greater social good); Alan H. Goldman, DeterrenceTheory: Its Moral Problems, in PRINCIPLED SENTENCING: READINGS ON THEORY AND POLICY supra note 23, at 80, 81 (arguing that deterrence theory cannot explain why the innocent should not be punished without importing retributivist ideas); Von Hirsch, supranote 24, at 89 (explaining that in incapacitation theory the justification for sentencing is based on the likelihood of future offending, not on any past criminal be- havior). 76. Ashworth, supra note 74, at 143. 77. See id.; VON HIRsCH, CENSURE AND SANCTIONS, supra note 23, at 18. Ordinal proportionality requires parity (crimes of similar seriousness deserve comparable punishments), rank-ordering (the relative severity of punishments should represent the level of social disapproval for different crimes), and spacing (the severity of punishment should reflect not only whether the first crime is more serious than the second, but by how much). 78. MINOW, supranote 8, at 10. 2002 / A FrameworkforUnderstandingTransitionaIJustice that is felt toward them. Good people are above such passions or at least they try to be."79 Desert theorists have responded to the "vengeance problem" in several ways. Some have sought to prove that retribution and vengeance are not the same thing, or that retribution can be justified without reliance on the concept of revenge.80 Other theorists, notably Susan Jacoby, have sought to rehabilitate the concept of revenge itself.81 Jacoby argues that we should recognize that criminal justice is based on revenge82 and stop pretending that "justice and vengeance have nothing, perish the uncivilized thought, to do with each other."8 3 Law serves to channel vengeance, thereby both discouraging less controlled forms of victims' justice, such as vigilantism,' and restoring the moral and social equilibrium that was violently disturbed by the offender.8 5 Some desert theorists, while avoiding the term "vengeance," rely on the similar concept that "[m]ost people react to [atrocious crimes) with an intuitive judgment that punishment (at least of some kind and to some degree) is warranted."8 6 Under this theory, qualms about the legitimacy of retribution blind one to the fact that the belief in punishment is actually moti- 79. Jeffrie G. Murphy, RaributimHairdAn Essa) cn Crinisrcal L rad -aml tk Ew;:.-i, m LiAu.m" AND RESPONSIBILITY 351, 353 (R. G. Frey & Christopher W Morris eds., 1991). 80. Jean Hampton, fbr example, argues that retribution is justified as a way to reassert the victim's value as a human being by negating the evidence of superiority implied by the wrongdoer's offense. How a society reacts to an individual's victimization is a reflection of how valuable society thinks that individual is. Thus the failure to punish denies the human value of victims. So Jean Hampton, A Nro Tk.ry q Retribution,in LIABuar- AND REsPONsiBIIsu ; Supra note 79, at 396-412. Similarly. Martha Minow seeks to distinguish retribution and vengeance by arguing that [r]etribution can be understood as vengeance curbed by the intersentuon of someone other than the victim and by principles of proportionality and individual rights. Retribution motivates punishment out of fairness to those who have been wronged and reflects a belief that wrongdoers deserve blame and punishment in direct proportion to the harm inflicted. MiNow, supra note 8, at 12. Such victim-based desert theories are attractive, and may provide a helpful way to distinguish between retribution and vengeance. In the popular imagination, retribution is intv-d often understood as a way of providing relief to victims. Certainly some victims do find satisf ation in knowing that their abusers will pay. It is critical to remember, however, that ultimately retribution is about the offender's culpability, nor the victim's suffering. A rerribunvist punishes attempts, though they cause no harm, but not accidents, though the harm caused may be severe. Moreover, retributivists determine how much to punish depending on how much the offender deserves it, not on how much the victim wants the offender to pay. The benefits for victims are thus better understood as side-effects of, rather than justifications foir,a retributive approach to punishment. 81. SusAN JAcOB's WiLD JusTc: THE EVOLLMoN OF REVENGE 6 (1983) (the question is "not whether retribution perse is a 'firbidden' objective of criminal justice but which forms of revenge are consistent with the aims ofa just society"). 82. Id. 83. Id at 8. 84. SeeidL at 10. 85. Seeia at 291. 86. Moore, supra note 71, at 152. Moore continues: Many will quickly add, however, that what accounts for their intuitive judgment is the netd fir deterrence, or the need to incapacitate such a dangerous person, or the need to reform the person. My own view is that these addenda are just "bad reasons for what we believe on instinct anyway." ... HarvardHumanRightsJournal / Vol. 15 vated by the desire for retribution.87 The problem with such intuition-based arguments for retribution is that not everyone shares the desire to punish;88 in fact, some victims plead for clemency for their tormentors. Nor does sympathy for the retributive victim's desire that her wrongdoer suffer necessarily mean that such suffering is justified, or even morally right. As Nino argues, retributivism "presupposes that it is sometimes appropriate to redress one evil with another evil. However, when I add the evil of the crime to the evil of the punishment ... my moral arithmetic leads me consistently believe that we have 'two evils' rather than 'one good."'89 2. Desert/Retribution/Vengeancein TransitionalJustice How effective is prosecution in achieving the goal of retribution against those who commit massive human rights abuses? Because it provides a legitimate way in which to impose severe punishment, prosecution is better suited to retribution than other forms of transitional justice. In the face of atrocity, both individuals and societies have a powerful need to call those who caused the suffering to account. It is difficult to accept that the worst perpetrators of genocide and war crimes should escape responsibility. In fact, advocates of prosecution emphasize its retributive qualities. Some openly use the image of "getting even." Aryeh Neier, for example, expresses the wish that in establishing the ICTY, the international community had simply stated, "[flrom now on, those who commit great crimes will pay,"90 Generally, however, retribution is more politely described in terms of combating impunity or bringing perpetrators to justice. For example, Diane Orentlicher writes that the world community "has resolved emphatically that it will not countenance impunity for massive atrocities against persecuted groups."91 Similarly, the argument for prosecutions made in a recent report to the United Nations by a group of experts assigned to assess possibilities for transitional justice in Cambodia is that "crimes such as those of the Khmer Rouge deserve punishment as a matter of morality and fundamental considerations of justice. '92 The UN Security Council resolution that established the ICTY likewise speaks of the world's "determintation] ...to bring to justice" those responsible for the atrocities.93 Although not framed 87. Jeffrie Murphy, for example, argues that due to an unwillingness to acknowledge retributive emotions, many people prefer to believe that punishment is grounded in social utility or the demands of justice; in fact, however, the criminal law "institutionalizes certain feelings of anger, resentment, and even hatred that we typically (and perhaps properly) direct toward wrongdoers." Murphy, supra note 79, at 352. 88. See WALKER, supra note 23, at 72. 89. NINo, supranote 16, at 137. 90. Neier argues, however, that the actual motivations for the establishment of the tribunal were more complex, and included reasons of political expediency. See NmER, supranote 9, at 111-33. 91. Orentlicher, supranote 1,at 2595. 92. CambodiaReport,supranote 60, 99. 93. Reprintedin STEINER, supra note 10, at 1146. 2002 / A Frameworkfor UnderstandingTramitionaljustice in the language of vengeance or retribution, the underlying assumption of such statements is that such horrible crimes should not go unpunished. Prosecution is not, ofcourse, the only way to exact retribution. Lustration can strip perpetrators of their jobs or remove them from elective office."' Through civil suits, victims and their heirs can exact financial penalties from those who committed abuses.9 ' National truth commissions can "generate social opprobrium," 96 turning perpetrators into social outcasts and forcing them to face victims on television. However, the sanctions imposed through alternative mechanisms-sanctions such as social opprobrium, ostracism, money judgments, or loss of jobs and privileges-simply are not proportional to the crimes committed by human rights violators.' As we have seen, because courts afford defendants far more due process protections than truth commissions or lustration committees, prosecutions can legitimately sanction behavior more severely. As a result, "It]he greater the felt need for punishment, the more seriously the prosecution option must be consid- ered."98 In other words, if one adopts a retributive theory of transitional justice, prosecutions have significant advantages over other accountability mechanisms. While prosecutions may be more effective than other approaches in achieving retributive goals, true retributive justice is almost always unachievable in the wake of radical evil. This is true for several reasons. First, it is often impossible even in prosecutions to impose a punishment that is proportional to the crime. As Minow argues, massive human rights atrocities "call for more severe responses than would any ordinary criminal conduct, even the murder of an individual .... And yet, there is no punishment that could express the proper scale of outrage."9 "" What punishment would be proportional to such heinous offenses? Theoretically, one could torture torturers and rape rapists. But societies committed to human rights and individual dignity must not take this path. Killing killers remains a popular form of retribution in many domestic criminal justice systems. However, while certain courts of transitional justice, such as that established at Nuremberg after World War II, have condemned human rights abusers to death, the death penalty now clearly violates international human rights norms.",, Even if capital punishment is used, it cannot satisfy the requirements of or- 94. For a description oflustration, see, for example, Bassiouni, !s'ranote 19, at 21; Bcd, Lp'ra note 19. 95. See Bassiouni, supra note 19, at 22. 96. Landsman, supranote 5, at 89. 97. See id. at 88 ("The truth commission's one real shortcoming is lhat it cannot match pro-an-uctons with respect to the fulfillment of the important policy goals regarding punishment,-. 98. Ld at 88. 99. MsNow, supranote 8, at 121. 100. See, eg., Second Optional Protocol to the International Covenant on Civil and Political Rights. Aiming at the Abolition of the Death Penalty, prcdfcr gratureDec. 15, 1989, rt rtidin CLNTL ion THE STuDY oF HumttAN RIGHTS, WE'arI'rrY-FIVE HutmAN RIGHTS Docutrms 33 (1991). Both thr ICT'Y and ICTR prohibit the death penalty.Ste Penrose, supranote 6. at 374. HarvardHumanRightsJournal / Vol. 15 dinal proportionality because it fails to distinguish the punishment of radical evil from the price paid by an ordinary murderer.10 1 Moreover, where a perpetrator's actions may have caused the deaths of thousands, the loss ofone life is not a punishment that adequately reflects the gravity of the crime. Adequate retribution is impossible unless those inflicting punishment violate the rights of human rights violators, and even then the punishment often fails to approach the horror of the crime. "U[ustice," explains theologian Donald Shriver, "falls limp before monster-sized evil."'1 2 In the final analysis, however, the ultimate futility of retribution should not be a reason to discount the utility of prosecution. Even if human rights violators can never be punished enough, they can still be punished severely. And if the desired goal for transitional justice is retribution, inadequate penal sanctions imposed after trial are still preferable to grossly inadequate civil liability or public shaming. The second and more fundamental critique of a retributive approach to transitional justice is that it depends on the concept of blame, which requires character evaluation.10 3 As Nino has argued, the viability of character evaluation in the context of radical evil is unclear. 1 Noting Hannah Arendt's puzzlement that Adolf Eichmann organized mass murder in order to advance his career rather than to cause harm to others,' 0 5 Nino concludes, Of course, banal evil is still evil. But are we prepared to blame a character which we evaluate as banal rather than full of burning hatred, sadistic inclinations, and cruelty? Even if retributive punishment were justified in general.., it may be unsuitable for radical evil.10 6 Yet perhaps it is easier to blame the banal than Arendt and Nino recognize: we might find a technocrat who believes that managing the details of geno- 101. Ofcourse, if a society has banned capital punishment for ordinary crime and then introduces it to punish human rights atrocities, such a distinction might be achieved. In Peru, for example, the government ofAlberto Fujinori tried retroactively to create an exception to that country's prohibition on capital punishment in order to punish Abimael Guzman, leader of the leftist guerrilla movement, the Shining Path. See NINO, supranote 16, at 142. Allowing exceptions to human rights principles can be dangerous, however. "Should we also make an exception," asks Nino, "to abhorrence of prisoner mistreatment, imitating what we condemn the prisoners for doing?" Id. 102. DONALD SHRIVER, AN ETHIC FOR ENEMIES: FORGIVENESS IN POLITICs 82 (1995), 103. Character evaluation may not be required for determinations of criminal guilt pcr se, since it is the individual's mens rea to commit the prohibited act, rather than the motivation of the offender, which is critical. Still, the structure of the penal law-which builds in defenses like duress or self-defense, and does not punish accidents that cause serious harm-suggests that the motivation and character of the offender are central to an understanding of individual culpability. Under a retributive understanding, punishment is justified not simply because an offense was committed, but also because the offender is blameworthy and should get what she deserves. 104. NINo, sapranote 16, at 142. 105. Eichmann, who organized the mass murder ofJews in Nazi Germany, was tried and convicted in Israel in 1961. See STEINER, Supra note 10, at1138. 106. NINO, supra note 16 at 142. 2002 / A Frameworkfor UnderstandingTransitionalJustice cide will further his career more reprehensible than a hate-filled murderer. Moreover, societies regularly punish crimes that are undertaken in pursuit of commonplace goals. While the burglar's incentive may be a bigger paycheck and the drug pusher's motive may be a promotion in the drug ring, retribution may still be appropriate. Nevertheless, Arendt's and Nino's insights are powerful. Radical evil involves horrific acts that even ordinary criminals would find appalling. 0 7 It is often committed by average people who would never commit ordinary crime. How can one understand intent in such circumstances? Perhaps, as Nino suggests, the proper response is to suspend reactive judgments (much as one might do with the insane), because the perpetrators of mass atrocity "have gone beyond the pale of humanity by rejecting the framework of interactions that blame presupposes.""" Character evaluation in the aftermath of mass human rights violations is further complicated by the difficulty of assigning individual responsibility, especially in the case of lower-level participants. Minow argues that [the central premise of individual responsibility portrays defendants as separate people capable of autonomous choice-when the phenomena of mass atrocities render that assumption at best problematic. Those who make the propaganda but wield no physical weapons influence those with the weapons who in turn claim to have been swept up, threatened, fearful, mobilized. 0 9 Of course, individuals always have choices. As the German philosopher Karl Jaspers wrote shortly after World War II, "I, who cannot act otherwise than as an individual, am morally responsible for all my deeds, including the execution ofpolitical and military orders. It is never simply true that 'orders are orders."'110 But individual autonomy in the context of dictatorship or mass violence may not be the same in the context of ordinary crime. Unlike ordinary criminals, who violate social norms by committing crimes, individuals who are swept up in mass violence do not step outside the prevailing moral framework. Rather, they succumb to intense social pressure. While this does not relieve such individuals of moral agency or responsibility, it does make them more difficult to judge. Former German President Richard von Weizsacker has noted a widespread and unfortunate tendency among young Germans "to believe that then people were evil but today they are good."'I As Reinhold Neibuhr has quipped, the universality of sin is the only concept in 107. The revulsion ordinary criminals feel for homfic crimes is well-demonstrated by threk. tht prisoners convicted of especially egregious acts, like child molestation, frequently must be held in pra:"rive custody to prevent assaults by those who are detained for lesser offenses. 108. N No, supranote 16, at 141. 109. INow, supranote 8, at 46. 110. KARL JASPERS, THE QUESTION oF GERAN GUILT (1948), ct,,is rt',4d t:I Tng smO.,,L JUsTICE 157, 159 (Neil Kritz ed., 1995). 111. Quotedin SHRIVER, supra note 102, at 84. HarvardHumanRightsJournal / Vol 15 the Judeo-Christian tradition that is empirically verifiable.112 Those who have never faced such choices should not be too quick to assume that they would have acted differently. In addition to the conceptual obstacles to a retributive understanding of transitional justice, prosecutions themselves complicate the retributive framework. The first difficulty lies in the relationship between retribution and vengeance. Ethnic conflicts around the world, in which each side justifies the atrocities it inflicts by referring to the wrongs it has suffered, demonstrate the tendency of vengeance to lead to a downward spiral of violence. Prosecutions are designed to channel these demands for vengeance and to break the cycle of personal revenge. As Minow explains, in order to avoid such escalating violence ... (one must) transfer 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.11 3 Whether or not we equate retribution and vengeance,114 it is clear that in the context of transitional justice prosecutions founded on a desire for retribution will have many of the drawbacks of vengeance. Where prosecutions are publicly perceived as a form of victor's justice, they will be unlikely to break the cycle of violence. As Jos6 Alvarez notes, "[the majority of the thousands detained in Rwanda's jails today report, and perhaps genuinely feel, that 'they have done nothing wrong' and are being victimized merely because they were on the 'wrong side of the war."'115 Even if prosecutions satisfy demands for revenge from Tutsis, they may stoke the desire for vengeance by Hutus. Nino describes a related problem in Argentina, where human rights groups, by adopting an "all-out retributive" approach and demanding that all the guilty be punished, ended up undermining their own credibility, fanning a backlash by military and government forces against the trials, and ultimately weakening the impact of those trials that did take place.'16 Nino suggests that although many people approach the issue of human rights violations with a strong retributive impulse, almost all who think momentarily about the issue are not prepared to defend a policy of punishing 112. Citedinid.at 83. 113. MINOW, sipra note 8, at 11-12. 114. See stpranote 80. 115. Alvarez, supra note 59, at 468. 116. Nino, supranote 22, at 2635. 2002 / A Frameworkfor UnderstandingTy-amitionalJustice those abuses once it becomes clear that such a policy would probably provoke, by a causal chain, similar or even worse abuses.'1 ' The second difficulty is that prosecutions, as we have seen, are necessarily limited and selective. Retribution theory, by contrast, is predicated on the notion that everyone should get their just deserts. In the wake of genocide or other mass violence, many people in society will "deserve" to be punished. Yet prosecutions, even at their most extensive, will only reach a few of the culpable. Many offenders who raped, tortured and killed will never be tried. As will be argued below, a deterrence-based rationale for prosecutions can account for exemplary prosecutions ofgenocidaires.A retributive one, by contrast, cannot. 118 Selective prosecution further undermines retributive goals because prosecutors rarely succeed in targeting only the most culpable. "IThe actual set of individuals who face prosecution," notes Minov, "is likely to reflect factors far removed from considered judgments about who deserves prosecution and punishment."'1 9 The failure to prosecute all equally culpable individuals, however, violates the principle of proportionality, which dictates that like crimes should be treated alike. 2 0 Proportionality is further undermined when prosecutions target lower-level offenders while ignoring more blameworthy ones, since such decisions do not reflect the relative level of social disapproval accorded to different crimes.121 Yet accepted legal principles may make it extremely difficult to convict those who orchestrated abuses rather than simply carrying them out. For example, reunified Germany found it difficult to hold East German leaders adequately accountable for ordering that fleeing citizens be shot, and decided instead to prosecute several young East German border guards.'2 2 Such selective, limited prosecu- 117. Id at 2620. 118. Proponents of prosecuions frequently elide the two rationales. Orentlicher. for example. aes a retributive line in arguing that "atrocious acts committed on a mass scale against racial. religious, or political groups must be punished." Orenlicher, .upra note 1, at 2594. Yet in defending selective prosecution, Orentlicher argues that fill prosecution is unnecessary for effective deterrence. Id at 2601. 119. AftNow,.supra note8,at31. 120. See, eg., Nino,srpranote 22, at 2620 (arguing that because retributive theory values punishmmt above all else,and because selective prosecution may be necessary to protect democra retributive theory raises the question of equality before the law). 121. Neier provides an example of how easily lower-level offenders can be equated with thow who plan and organize atrocities. In 1992, Lawrence Eagleburger, then acting U.S. Secretary ofStare, included a Bosnian Serb soldier, Borislav Herak, alongside Yugoslav President Slobodan ,alilevd and Bomian Serb leaders Radovan Karadik and Ratko Mladi4 on a list of ten potential .r crimes defenants. The twenty-one-year-old Herak, who had committed rwenty-nine murders, imas clearly a foot soldier i the conflict, bur Bagleburger felt compelled to add his name to the list of those who masterminded the ethnic cleansing because Herak had been profiled in Tki New )kr VTTh-a. For a detailed acount of the= developments, see NwER, supra note 9, at 125-27. 122. For a description of these prosecutions, see 'DNA RossNawiRG, Tnt HMu.roD LA.ND; F^ciNG EuRoP'es GHOSTS AFTER ComiuNisM 261-305, 340-51 (1995). Rosenberg argues that the rder guard prosecutions did not comport with public notions ofculpability. (Miost East Germans were repelled. They felt the border guards had taken the hit for a chain of superiors leading at least to Honecker and possibly to the Kremlin-most of whom would HarvardHuman RightsJournal / Vol, 15 tions-the only kind possible in transitional justice-fail to meet the basic requirements of retributive justice. The third complication is the fact that transitional societies must continue to deal with ordinary crime even as they confront the extraordinary evil of the past. In many societies transitional periods have been marked by soaring crime rates and accompanying efforts to prosecute and punish offenders.123 In light of this reality, a retributivist must ask who, for example, is more guilty: the young, uneducated, unemployed black carjacker from a Johannesburg township, or the white police officer who killed the boy's father during a political rally several years before? Due to the inevitable limitations on prosecutions in the transitional context, even those prosecutors who believe that the youth is less culpable will concede that they will be more likely to prosecute him than his father's murderer.124 It may well be true, as Paul van Zyl has argued, that "[if the police and prosecuting authorities were to devote a significant share of their resources to dealing with human rights violations, many ofwhich occurred a decade or more ago, the country would almost certainly lose the current battle against ongoing crime."125 But the need to control crime is no answer to the basic retributivist objection: the greater the culpability, the more severe the punishment should be. Retributive theory cannot support a form of prosecutorial selectivity that excuses more serious past crimes in order to address less serious current ones. Furthermore, amnesties for past offenses typically do not apply to ordinary crime. South Africans, for example, could only seek amnesty if their offenses were "associated with political objectives and committed in the course of conflicts of the past." 26 From a retributive point of view, it is not immediately clear why a murderer who kills for political reasons should be entitled to amnesty in return for the truth, while one who kills out of passion or greed should not. Society's willingness to forgo punishment for some political offenses (whether through an express amnesty or a defacto one resulting from selective prosecution), but not for common crimes, suggests never face justice. One of the most distasteful aspects of the trial was seeing the guards' superior officers ... most ofthem now working as unified German border guards or policemen, arriving to testify and leaving as free men. They were even paid for missing work in order to be at court. Id.at 344. 123. See Villa-Vicencio, supranote 5, at 214 (noting the challenges posed by the escalation ofordinary crime in countries moving away from repressive rule and the way in which the line between ordinary and political crime can become blurred). 124. Paul van Zyl argues that "even if the South African transition could have occurred without any form of amnesty agreement, thus leaving open the possibility of large-scale prosecutions, only a small fraction of those responsible for gross violations of human rights could have been prosecuted successfully." Van Zyl, supranote 4, at 651. Van Zyl's assessment ofthe South African situation mirrors a widespread understanding among scholars of transitional justice that it is impossible or undesirable to prosecute all those who were involved in atrocities. See Section I. C. 125. Van Zyl, supranote 4, at 652. 126. S. AFR. CONST. of1993 (amended 1994), postamble. 2002 / A Frameworkfor UnderstandingTransitionalJustice that individual culpability is understood very differently in the contexts of extraordinary evil and ordinary crime. To summarize, in the face of horrific crimes, many people, both in the affected society and the international community, share a powerful sense that such atrocities should not go unpunished. Prosecution, which provides a legitimate way to impose severe punishment, is the most effective method of pursuing retributive goals in transitional societies. But it has limits: one cannot punish fairly, and one cannot punish enough. As an Indonesian victim and prosecution proponent explains, "(it] will have to be a very selective justice and, of course, selective justice is not exactly justice."'12 ' An advocate for retributive prosecution might answer these criticisms by simply arguing for more prosecution. But unless everyone can be prosecuted, which is impossible, the objections remain. Although the desire for retribution is a justifiable response to extraordinary evil, retribution itself is only partially achievable. B. Deterrence 1. Theory The term "deterrence" is often used interchangeably with "prevention." In fact, deterrence is only one way to prevent crime. Under deterrence theory, potential offenders may still be capable of committing crimes (since they are not incapacitated) and may still desire to commit crimes (since they are not rehabilitated). But despite their capacity and desire, potential offenders are inhibited by the "intimidation or terror of the law."'128 There are two main types of deterrence: individual deterrence and general deterrence. Individual deterrence seeks to prevent future crime by setting sentences that are strict enough to ensure that a particular offender will not reoffend.12 9 General deterrence, on the other hand, attempts to prevent crime by "induc~ing) other citizens who might be tempted to commit crime to desist out of fear of the penalty."130 Notably, deterrence theory does not allow, much less require, the punishment ofall who are guilty. Moreover, general deterrence does not even require punishment of all those who might be dererrable as individuals. If 127. Seth Mydans, IndonesiansDiffer on PenalhiesfarthrPast,N.Y.T. tes, Aug. 27, 2000, at A14. 128. Jeremy Bentham, Punishentand DaerrXce,in PRINCIPLIED SNr.NciNG: READINGS ON THEORY AND Poucy supra note 23, at 53, 54. Bentham argues that there are three principal ways to prevent crime: take away the physical power to violate the law, take away the desire to offend, or make the individual afraid ofoffending. "In the first case," Bentham explains, the individual can no more commit the offense; in the second, he no longer desires to commit ir, in the third, he may still wish to commit it, but he no longer dares to do it. In the first case, there is a physical incapacity;, in the second, a moral reformation; in the third, there is intimi. dation or terror ofthe law. id 129. For example, a sentencing scheme which increases penalties based on prior convictions may reflect the belief that, while the previous, lower sentence was not sufficient to derter cnmiality by this offender, a higher sentence might be. SeeAshworth, svpra note 75, at 44. 130. eL HarvardHumanRightsJournal / Vol. 15 exemplary punishments adequately prevent future crime, they are suffi- cient.1 31 Deterrence occurs, writes Nigel Walker, when people refrain from certain actions because they fear the possible consequences of those actions.13 2 In other words, the potential benefit of committing crime is outweighed by the risk of sanctions. Deterrence thus assumes that, were it not for the possibility of adverse consequences, people would engage in crime. "A person is not deterred," writes Walker, "if he refrains because he is not tempted, or is tempted but restrained by his code ofmanners or morals." 133 Deterrence also assumes that the potential offender will undertake a two-part calculation, assessing both the gravity of the consequences and the likelihood of getting caught. This calculation is based not on the objective severity of sanctions or the real risk of apprehension, but on the potential offender's subjective assessment of these factors. Thus, the effectiveness ofany deterrent depends on the potential offender's perception of possible sanctions, and on her assessment of her ability to evade law enforcement.134 The actual severity or certainty of punishment is less important than its perceived severity or cer- tainty.' 35 While the logic of deterrence is intuitively appealing, the available empirical evidence regarding the effectiveness of deterrence in domestic criminal justice systems is inconclusive. 136 There are several possible reasons for this.137 First, it is difficult to prove that threats of legal sanctions, rather than other motivations, have prevented people from offending. 38 "Marginal deterrence," the amount by which deterrence increases or decreases based on changes in the severity of sanctions, is particularly difficult to demon- strate.139 The problem is not in showing that deterrence can occur when punishment is certain, swift, and severe, but in determining when and to what extent it occurs under real-world conditions, under which punishment is never certain, rarely swift, and only sometimes severe.140 Second, effec- 131. Thus, pure deterrence theory can be deeply unsatisfying: why should the perpetrators of horrific but undeterrable crimes go unpunished? As penologist Alan Goldman argues, "surely the fact that a crime is unplanned, that the criminal does not contemplate the consequences to him of his action"--in other words, that the crime is undeterrable--"should not entirely exempt it from punishment." Goldman, supranote 75, at 81. 132. WALKER, supra note 23, at 13. 133. Id. at 14. 134. Some penologists believe that, in practice, the only factor affecting deterrence is the likelihood of getting caught, judged from the viewpoint of the potential offender, See id. at 17; SANrORD H. KADIsH & STEPHEN J. SCHUL-OFER, CRIBINAL LAW AND ITS PROCESSES: CASES AND MATERIALS 117 (Richard A. Epstein et al. eds., 7th ed. Aspen Law & Bus. 2001) (arguing that certainty of punishment is important only ifit contributes to the perception of certainty). 135. Seeid. 136. See, e.g., Ashworth, supranote 75, at 48; Deryck Beyleveld, DeterrenceRescareb andDeterrencePoliCieS,in PRINCIPLED SENTENCING: READINGS ON THEORY AND POLICY supranote 23, at 66, 72. 137. This discussion offactors relies heavily on the analysis inAshworth, supranote 75, at 48-50. 138. Seeid.at48. 139. See id. 140. The criminologist Deryck Beyleveld has noted that deterrence is possible if one is willing to 2002 / A Frameworkfor UnderstandingTransitionalJustice tiveness depends on context.141 The risk of detection and punishment must not be so low as to be readily discounted, and the penalty must be adequately publicized. 14 2 Moreover, the offender and the crime must be deterrable; that is, they must be rational.' 43 Heightened enforcement and increased sanctions may reduce crimes that reflect rational choices, but they are unlikely to have an impact on irrational offenders. Deterrence, then, only works in relation to some crimes and some offenders. 1"4 4 2. Deterrencein TransitionalJustice Scholars and human rights activists have trumpeted deterrence as perhaps the most important justification for prosecution in transitional justice. Orentlicher, for example, writes that "[t]he fulcrum of the case for criminal punishment is that it is the most effective insurance against future repression." 145 Neil Kritz shares Orentlicher's confidence in the effectiveness of trials as a deterrent, noting that the failure to prosecute at least key figures "can be expected not only to encourage new rounds of mass abuses in the country in question but also to embolden the instigators of crimes against humanity elsewhere."'146 M. Cherif Bassiouni adds that "[t]he relevance of prosecution and other accountability measures to the pursuit of peace is that through their effective application they serve as deterrence, and thus prevent future victimization.' ' 47 Given the manner in which deterrence functions in the domestic criminal justice context, how tenable is the widespread deterrence justification for prosecution in the wake of massive human rights violations, war crimes, or genocide?148 Deterrence theory is quite useful as a justification for selective adopt extreme measures: "Ifthe choice is a clear one between compliance and, for example, certain death, then it is a good bet that a deterrable individual will comply" Beyleveld, nTpa note 136. at 77. Hluotwer, such swift and certain punishment is not only difficultto achieve in practice, but may alw interfere with principles of fairness,such as the right of the accused to a trial before an impartial tribunal. While 'Ye have good reason to believe that immoral, unimplementable policies would *work,- concludes Beyleveld. "[there are] rarely reason[s] to believe that more sane and realistic policies will achieve anything.' Id 141. SeeAshworth, supranote 75, at 49. 142. Seeid 143. For example, if, as some scholars have argued, "Ih]omacide is a crime usually committed in undrterrable states ofmind," WALKER, supranote 23, at 16, one cannot expect deterrence measures to afft-t homicide rates significantly, nor can we use deterrence as a justification for punishing such "crimes of passion." 144. See Ashworth, supra note 75, at 50. 145. Orentlicher, supranote 1, at 2542 (footnotes omitted). Orentlicher argues that ([b]ylaying bare the truth about violations ofthe past and condemning them, prosecutions can deter potential lawbreakers and inoculate the public against future temptation to be complicit in state-sponsored violence.- Sr alb Landsman, supra note 5, at 84 (identifying deterrence offuture violations as one of prvecution's pnmary benefits); Goldstone, supra note 35, at 619 (emphasizing the deterrent value of prosecutions). 146. Kritz, supranote 21, at 129. 147. Bassiouni, supranote 19, at 18. 148. As a preliminary matter it is worth pointing out that in the transitional justice context deterrence" almost always refers to "general deterrence." Prosecutions generally cannot take place before highlevel leaders responsible forthe atrocities have been removed from office. Nor are prosecutions likely to be undertaken where deposed leaders have sufficient political capital to gain political power once again. HarvardHuman RightsJournal/ Vol. 15 prosecution. 149 In particular, because deterrence does not require punishment of all the guilty, it does a better job than desert theory of explaining why exemplary punishments of a few offenders may be acceptable. Although the failure to prosecute heinous abuses is impermissible on retributive grounds, deterrence theory can accommodate selective prosecution. The explicitly utilitarian goals of deterrence also provide a justification for abandoning post-transition prosecutions when "the mischief [they] would produce would be greater than what [they] prevented." 150 When ousted leaders retain significant power in a country in transition, harsh punishment could lead them to seize power again. Thus, criminal punishment may have a positive or negative overall effect on human rights.151 A deterrence rationale allows for a cost/benefit analysis in which one assesses whether the advantages ofpreventing crime through prosecution outweigh the costs to democracy and human rights that might result if trials lead to political instabil- ity.152 Deterrence permits a decision not to prosecute where trials would cause more problems than they would solve. If deterrence is the justification for prosecution, one must determine if prosecutions actually prevent human rights abuses. It is even more difficult to show the effectiveness of prosecution as a method of deterring mass atrocity than it is to demonstrate its effectiveness in the context of ordinary crime. In fact, it is virtually impossible to assess whether or not the threat of prosecution has ever prevented genocide and war crimes. As Minow notes, "[nlo one really knows how to deter those individuals who become potential dictators or leaders of mass destruction ....One hopes that current-day prosecutions would make a future Hitler, or Pol Pot, or [Bosnian Serb leader] Radovan Karadi6 change course, but we have no evidence of this." 53 Given the unyielding stream of atrocities the world has witnessed since Nuremberg, it is difficult to argue that these trials had any discernable effect." 4 Similarly, many of the worst atrocities in the former Yugoslavia took place after the ICTY was established.15 ' It could be that in the absence Prosecutions are not intended to teach individual leaders like Pinochet a lesson in the hopes that they will behave better next time. Rather, the goal is to encourage other political leaders to follow international human rights standards. 149. See Orentlicher, supra note 1,at 2601 (defending selective prosecution on the grounds that full prosecution is unnecessary for effective deterrence). 150. Bentham, supra note 128, at 57. 151. See id 152. As Jon Elster notes, "if the threat of harsh punishment is in fact credible, it may cut both ways. Although it will make it less likely (but not impossible) that coups will occur in the future, it will also make coup-makers more reluctant to step down." Elster,supranote 38, at 37. 153. MINOW, supranote 8,at 146. 154. See, eg., id. at 27. 155. Aryeh Neier, The QuestforJustie,N.Y. REv.BOOKS, Mar. 8, 2001, at 31, 32. Noting that the genocide in Rwanda also occurred after the establishment of the ICTY,Neier argues that the tribunal's creation "[clertainly... did not make the authors of grave crimes in other parts of the world worry about being called to account." Id.See alsoPenrose, supranote 6, at 326 (expressing concern that "ifthe Interna. tional Criminal Tribunals at Nuremberg, Tokyo and the recent additions at the Hague and Arusha are used as a gauge for deterring future violence, the international community must admit failure") (footnote 2002 / A Frameworkfor UnderstandingTransitionalJustice of these prosecutions, many more such atrocities would have taken place. But those who point to the deterrent effect of prosecutions bear a heavy burden ofproof indeed. Any deterrent effect that prosecutions might have will depend on context, including the risk of "getting caught," the severity of penalties, the extent of public knowledge about such sanctions, and the degree to which the crime and the offender are deterrable. In domestic criminal justice, the risk of "getting caught" typically concerns the risk of detection, since in a functioning criminal justice system, offenders whose guilt can be proven will generally be punished if they can be apprehended." 6 In the transitional justice context, on the other hand, "getting caught" usually has little to do with the risk of detection; indeed, many atrocities are committed in plain view.157 Rather, "getting caught" primarily concerns the chance of being punished. Thus, the fact that in the wake of mass atrocities only a small number of those implicated will ever be prosecuted undermines the logic of the deterrence argument. 158 Those who "merely" kill, rape, and plunder, but do not mastermind the carnage, have little to fear from prosecution. Even for those who orchestrate human rights abuses, the risk of "getting caught" is low. "[I]t is not irrational," writes Minow, "to ignore the improbable prospect of punishment given the track record of international law thus far."' 59 Ofcourse, in the rare cases where perpetrators do "get caught," the sanctions may be considerable. However, it seems doubtfiil that severe penalties for massive human rights violations will have much deterrent value when they are so heavily discounted by the negligible likelihood of prosecution."tO Nor is it clear that the foot soldiers of atrocity will even be aware of the heavy sanctions imposed on a few high-level perpetrators in some far-off land.161 omitted). 156. Potential offenders may further discount the risk of detection based on the passibility that the case will not actually result in a conviction due, fbr example, to procedural safeguards. protrcutonal sloppiness, or lack ofevidence. 157. Prosecutions are frequently undertaken when there is compelling, publicly ackowledged evidence of an individual's guilt. Although prosecutions may bring out additional details and although prosecutors may have difficulty finding sufficient legally admissible evidence, prosecutions in the context of transitional justice are generally not used to determine who is responsible. Some forms of transitional justice do increase the risk ofdetection. Revelations to truth commissions or lustration laws which expose an individual's prior contacts with the secret police can be used to identify previously unknown pcqxpusrots. However, such individuals are rarely prosecuted. 158. See, eg., Van Zyl, supranote 4, at 658 (questioning the validity of the deterrence rationale in the context ofSouth Africa's limited ability to prosecute). 159- Mrow, supranote 8, at 50. 160. In the wake ofthe near-extradition ofPinochet toSpain, the risk ofprosecution was high enough that it had some effect on international travel by high-profile statesmen and politicians whose current or past activities could conceivably land them behind bars. See Jayson Blair, Ptrwtdj RCmS=S OlntvrNcvd', You'd Better Watch Out, N.Y.T1mzEs, Mar. 26, 2000, Section 4, at 5. However, at seems implausible thut limitations on foreign travel can serve as an effective deterrent against human rights abuse. 161. It is more reasonable to assume that "lower-level" perpetrators would be aware of previous domestic prosecutions in their owncountry. HarvardHuman RightsJournal/ Vol. 15 In response, one might argue that the international community should substantially increase the likelihood that human rights abusers will face criminal prosecution. But it is not clear how much of a deterrent effect increased prosecutions would have on genocidaires. Potential war criminals may underestimate the actual risks. Robert Jackson, the lead prosecutor at Nuremberg, questioned the degree to which that tribunal could serve as a deterrent, given that wars are almost always started in anticipation that they will be won. "Personal punishment, to be suffered only in the event the war is lost," he argued, "is probably not [enough] to be a sufficient deterrent to prevent a war where the war-makers feel the chances of defeat to be negligible. ' 162 Moreover, as Jon Elster notes, "even if violations are harshly punished now, how can future would-be violators know that they, if overthrown, will be treated in the same way? Incentive effects presuppose stable institutions, which almost by assumption do not exist. '163 Scholars have also debated whether or not massive human rights violations involve crimes or criminals that are deterrable. Are such crimes subject to a rational assessment of costs and benefits? One former prosecutor of the ICTY has claimed that "deterrence has a better chance ofworking with these kinds of crimes [war crimes, genocide, crimes against humanity] than it does with ordinary domestic crimes because the people who commit these acts are not hardened criminals; they're politicians or leaders of the community that have up until now been law abiding people. ' 164 Such analysis seems fundamentally misguided. When "ordinary" people commit horrible crimes, it suggests that the normal restraints of law and deterrence are not working, or that these people are no longer functioning rationally. 165 At the same time, some individuals do make rational choices when committing horrific crimes, and would therefore potentially be deterrable. As Douglass Cassel notes, certain dictators like former Yugoslav President Slobodan Milogevi6 are manipulators, not fanatics, and might be restrained by credible threats. 166 Other perpetrators, such as Hitler, however, are probably undeterrable. Moreover, some atrocities are carefully planned and staged, while others are "crimes ofpassion" or crimes of hate. Just as in the domestic context, deterrence, if it works at all, will only work against some offenders and some crimes.' 67 162. MiNow, supranote 8, at 50 (quoting FACING HISTORY AND OURSELVES NAT'L FOUND,, TfIg NEW ENGLAND HoLocAusT MEMORIAL STUDY GUIDE 12 (1996)). 163. Elster, supra note 38, at 37. 164. Michael P. Scharf, The Casefor a PermanentInternationalTruth Commission, 7 DUKIB J. COMP. & INT'L L. 375, 398 n.12 4 (1997) (quoting interview by Michael P Scharf with Grant Niemainn, Prosecutor, International Criminal Tribunal for the Former Yugoslavia, the Hague, Neth, (July 25, 1996)). 165. MINOW, supranote 8, at 50 ("Individuals who commit atrocities on the scale of genocide ateunlikely to behave as 'rational actors,' deterred by the risk of punishment."). See dso Villa-Vicenclo, stpra note 5,at 210 ("More is required than the heavy hand of the law ... to deter those driven by the ideological factors that have either destroyed ... nations or threatened to bring them to their knees."). 166. Cassel, supra note 11, at 534. 167. Timothy Garton Ash suggests that prosecution is most likely to occur where deterrence is least 2002 / A Frameworkfor UnderstandingTransitionalJustice It is worth remembering that deterrence does not work when perpetrators, while recognizing and fearing the possibility of punishment, nevertheless decide to engage in crime because the potential rewards outweigh the risks. Dictators who knowingly accept the risk of prosecution, but nevertheless commit atrocities in order to retain political and economic control, may rationally believe that crime does pay. In the domestic context, a deterrence theorist's response to this problem would be to continually increase both the penalties and the likelihood of apprehension. In the transitional context, however, penalties could not be much more severe,1 " and the chances of punishment, while they could be increased, will likely never be extremely high. Given that prosecution is not a particularly effective deterrent against gross human rights violations, are alternative mechanisms any better? In comparing prosecutions and other sanctions-such as civil liability, lustration, or public shaming-we must examine their relative effectiveness in terms of marginal deterrence, which as we have seen is difficult to assess. If deterrence depends on "intimidation or terror of the law," it is likely that a potential human rights violator will fear incarceration more than a money judgment, the loss of his job, or the shame of a public confession. Moreover, even dictators who never expect to stand trial may fear the consequences of an indictment. Once an international warrant of arrest has been issued against a suspected war criminal, that person is liable to arrest in virtually every country in the world, making it difficult for her to hold high public office in her own country and virtually impossible for her to participate in international negotiations. 169 But even though trials threaten more severe punishment than alternative mechanisms, it is not clear what effect this increased potential sanction has on deterrence. Deterrence depends not only on the severity of the sanction but also the certainty of punishment. While the likelihood that human rights violators will face any sort of sanction remains small, some accountability mechanisms are better equipped than others to handle large numbers of offenders. Prosecutions may yield severe punishments, but they are rare. Will a highly uncertain but severe punishment have a greater deterrent effect than a lesser but more likely sanction? Perhaps the deterrent effect differs depending on the person being deterred; would-be dictators might be dissuaded by the fear of prosecution, while necessary: "[Where [such] deterrence might still be important (as inRussia) there hase 14en no suh trials, and where there have been trials (as in Germany) the deterrence is hardly needed." Timothy Garton Ash, The Trutb About Ditatorship,N.Y.REv. BooKs, Feb. 19, 1998, at 35. 168. Although some states, like Rwanda, have imposed the death penalty on perpetrators, the oppaittion of the international community to the death penalty suggests that this sanction is unlikely to be imposed by international tribunals. 169. Goldstone, supra note 35, at 620. Sez also Neier. supra note 155, at 32 (noting that the Indiaments of Radovan Karad=1i and Ratko Mladi6 'sidelined both men politically' and resulted in their exclusion from the Dayton peace negotiations). HarvardHumanRightsJournal / Vol. 15 low-level functionaries, well-aware that they are unlikely ever to be tried, might regard lustration or truth commissions as a more credible threat. Prosecutions may deter some future human rights abusers, and prosecutions may even have a greater deterrent value than alternative post-transition mechanisms. However, it is unlikely that post-atrocity prosecution is the most effective way to prevent future atrocities. As penologists have noted in the context of ordinary crime, the severity and likelihood of a legally imposed penalty are not "the only or necessarily the most powerful influence[s] on a person's behavior."'170 An individual's actions are often affected to a greater degree by moral norms than by fear of punishment. The reason most people do not murder is not because they are afraid of getting caught, but because they believe that murder is wrong. Similarly, the constraints a society imposes on itselfmay have more to do with its political culture and form of government than with concern about the possible consequences of misbe- havior.171 Even where a person does refrain from taking a desired action out of fear of the possible consequences, legal sanctions may play only a minimal role. Frequently, non-legal deterrents, a pervasive fact of human life72 , are much more powerful than legal ones. An individual's decision not to assault someone after an insult in a bar may have more to do with a fear of being beaten up than with any worry about what the police might do. Similarly, while the threat of prosecution may deter some leaders contemplating atrocities, such persons are probably more likely to hold back out of fear of vigorous public criticism, political pressure, diplomatic isolation, economic sanctions, or even military intervention. To summarize, if deterrence is our goal, our underlying concern will be the prevention of future crimes. It is by no means clear that prosecution is the most effective mechanism for preventing atrocities. If the international community sits by and watches while atrocities occur, demanding prosecution only after the violence has stopped, arguments about the deterrent effect of such trials will ring hollow.173 Alvarez is right to warn that the international community will lose its credibility unless "tijnternational efforts to prevent the continuation ofgenocidal acts and other acts ofviolence ...pre- 170. Ashworth, supranote 75, at 50. 171. Note also that, as with domestic crime, punishment and censure can shape public values and help enshrine moral standards. In transitional justice literature, the language ofdeterrence is often used to refer to such moral education. See, ag., Kritz, supranote 21, at 129 (suggesting that international tribunals are better positioned than local courts to convey the message that the international community will not stand for atrocities); Cassel, supra note 11 (arguing that the International Criminal Court will reinforce moral norms). While punishment may indeed serve an important communicative function and help to reinforce values that will reduce future commission of atrocities, it is not clear that prevention occurs through fear ofpunishment. The communicative function ofpunishment is addressed in a separate Section below. 172. Beyleveld, supranote 136, at 78 ("We tend to associate deterrence with a legal context, but we need not do so, and doing so does not alter the phenomenon, only the context."), 173. The establishment of the ICTY has "been widely understood as (a] symbolic international effort ... undertaken after no nation indicated a willingness to risk the loss of its own soldiers to stop the massacres." MINOW, supra note 8, at 37. 2002 / A Frameworkfor UnderstandingTransitionalJustice cede attempts at criminal accountability." 74 This is not to say that prosecutions, or other transitional justice mechanisms, have no deterrent value. But if the goal is to deter future human rights abuses by making potential abusers afraid to act, the international community has bigger sticks to shake than the threat of trial. C. Rehabilitation 1. Theory Rehabilitation seeks to prevent the future commission of crime by curing previous offenders of their criminal tendencies. 1 7 Thus, the success of rehabilitation is measured by recidivism rates, rather than by changes in the aggregate incidence of crime.176 There are two possible reasons for wanting to "cure" the offender. First, one might argue that society will be safer once the offender is rehabilitated and is no longer committing crimes.' 7 Second, one could believe that offenders should be given the opportunity to have productive lives for their own sake. 78 Rehabilitation de-emphasizes the link between the gravity of the crime and the severity of the sentence. Appropriate rehabilitative sentences reflect the measures necessary to reintegrate a particular offender into the community. Once an offender has been rehabilitated, further punishment is unnecessary. Strict proportionality is not required; among offenders who commit the same crime, some will take longer than others to rehabilitate.1 7 9 The primary objection raised by critics of rehabilitation is that it simply does not work.18 0 Even researchers who favor rehabilitation believe that treatment programs are effective only with certain types of offenders.' 8 ' Other critics argue that rehabilitation denigrates human dignity.1s2 Still 174. Alvare;, supranote 59, at 458. 175. For an overview of rehabilitation theory, see generally Andrew yon Hirsch, R xtIza'at:, M PRINCIPLED SENTENcING: READINGS ON THEORY & POLICY supra note 23, at 1. Rehabilitation tradidonally involves counseling, training, psychological assistance, and other support ervices. Ser r. However, rehabilitation need not be pleasant;, it can take place in prisons, boot camps, or mental hospicals. and may involve aversive therapies, like shock treatment of sex offenders. 176. See i. 177. See M'fichael Moore, Law and Psychiatry. Rethinking the Relationship 234-35 (1934). rn&trJm SANFORD H. KADISH & STEPHENJ. SCHULHOFER, .srnanote 134, at 123. 178. Sec id. 179. SeeVon Hirsch, supranote 175, at 2. 180. See Francis A. Allen, The Dcdine of tke RchahitatitrIdeal, in PRINCIPLED Sr'M,crNco: READINGS ON THEORY & PouLCYspranote 23, at 14, 14-15 (discussing reasons for growing public skcpucism about penal rehabilitation); Stephen Brody,How EffraireAre PemalTrexart.s?,in PruNciFL.D SEN'TENcING: READINGS ON THEORY & POLICY supra note 23, at 9, 9 (describing a 'widespread conviction that rehabilitation is not effective in preventing recidivism). 181. See Brody,supranote 180, at 10. 182. Comparing rehabilitation and retribution, C.S.Lewis wrote: To be "cured" against one's will and cured ofstares which we may not regard as di.-i'is is to be put on a level with those who have not yet reached the age of reason .... But to be punished. however severely, because we have deserved it, because we "ought to have known better.- is to be treated as a human person made in God's image. HarvardHumanRightsJournal/ Vol. 15 others attack the allegedly benevolent purpose of rehabilitation, since rehabilitative regimes may in fact inflict a greater deprivation of liberty on their subjects than avowedly punitive programs. 183 2. Rehabilitationin TransitionalJustice Advocates for prosecutions as the optimal form of transitional justice frequently use the language of rehabilitation, but their focus is on societal, not individual, rehabilitation.'14 Few indeed would think of prosecution and punishment as a way to redeem despots like Pol Pot'8 5 and Pinochet. By contrast, support for societal rehabilitation-the idea that prosecutions can change a society's moral values by "foster[ing] respect for democratic institutions and thereby deepen[ing] a society's democratic culture" 186 -is wide- spread. Prosecutions are believed to have at least three curative powers. First, prosecutions help to establish the truth. Most scholars of transitional justice agree that exposure and acknowledgement of the past is a prerequisite for future social stability. 8 7 Prosecutions educate the public about the nature and extent of prior wrongdoing 188 and contribute to a shared historical understanding. Through this educational process, writes Stephan Landsman, prosecutions "may serve both to inoculate the populace against lapses into oppressive behavior and as a means of establishing an accurate account of what actually transpired before the democratic regime came to power."18 9 Second, prosecutions help to establish the rule of law. "Holding violators accountable for their misdeeds," explains Landsman, demonstrates to "all members of society that the law's authority is superior to that of individu- als."'190 By contrast, failure to enforce the law undermines its authority.19 ' C. S. Lewis, The HumanitarianTheory OfPunishment, 6 RES JUDICATAE: J.L. STUDENTS' SOC'Y VICTr. 2 (1953), quotedin WALKER, supra note 23, at 61. 183. Allen, supra note 180, at 16. 184. For a rare example of efforts to rehabilitate individuals previously involved in human rights abuses, see Ian Fisher, AfandheeraJournal:Somali Militiar Now Undergoing 'Rebumanization,' N.Y. Tinms, Dec. 4, 1999, at A4 (describing efforts to retrain former Somali militia members as policemen by "mak[ing] them into humans again"). To the extent that the transitional justice community has sought to respond to the needs of offenders, it has been in the context ofnon-prosecution alternatives, such as truth commissions. Although such efforts have a rehabilitative aspect, they are arguably more closely related to restorative justice, which will be discussed below. 185. Pol Pot led the Khmer Rouge regime, which held power in Cambodia from 1975 to 1979. Pol Pot's policies of imprisonment, torture, starvation, overwork, and execution resulted in the death of an estimated two million Cambodians. See generally CNN, CNN Newsmaker-Profiles:Pol Pot, at http:ilwww. cnn.com/resources/newsmakers/world/asia/pol.html (last visited Oct. 4, 2001). 186. Orentlicher, supranote 1, at 2543. 187. See Section I. C. 188. See Landsmansupra note 5, at 83. 189. Id. 190. Seeid. 191. See Orentlicher, supra note 1, at 2542 (when "atrocious crimes fare] committed by a prior regime on a sweeping scale ... (this] failure of enforcement vitiates the authority of the law itself') (footnotes omitted). 2002 / A Frameworkfor UnderstandingTransitionalJustice Third, prosecutions reinforce moral norms and contribute to a shared understanding that certain behavior is wrong. In order to prevent future atrocities, one must establish not only the truth about past abuses, but also a national and international consensus that such acts are unacceptable."' 2 One important way to communicate such a moral consensus is through the criminal law. 193 Punishment thus not only reflects but also shapes moral values. The fact that prosecutions can promote societal rehabilitation does not necessarily mean, however, that trials are more effective than nonprosecution alternatives at curing societies of their evil tendencies. The superiority of prosecution as a means of accomplishing the first goal--establishing the truth, educating the public, and forming a shared historical understanding-is dubious. Some scholars have suggested that trials provide a "higher quality" truth than alternative mechanisms because they are more narrative and dramatic.194 While trials may have moments of high drama, their formalism and rigidity can also make them excruciatingly boring.'" This is particularly true ifdue process standards are respected, since "[w]hat makes for a good 'morality play' tends not to make for a fair trial."", Other forms of dramatic truth-telling, such as the televised confrontations between victims and perpetrators before the South African Truth and Reconciliation Commission, may be as good or better at capturing the public imagination. Some scholars have suggested that, because of the higher evidenciary standards imposed on trials, the truth produced through prosecution is more accurate than that established through alternative mechanisms.19' In fact, 192. See, eg., Van Zyl, supra note 4,at 658 ("Knowing the truth about violations of human rights and building a national consensus that such acts are illegitimate are essential in preventing their recur- rence."). 193. The Pinochec case demonstrates the pedagogical benefits of prosecution: 'a generation of youth has been taught that his alleged crimes.., are so unconscionable that he is pursued forthem een today.Cassel, supranote 11, at 533-34. 194. See Osis, supra note 34, at 15-16 (arguing that trials are preferable to truth commissions in shaping public memory because they are more dramatic); NiN . svpra note 16, at 146 (suggesting that truth "is much more precise and much more dramatic when done through trial'. 195. Telford Taylor, in his memoir about the Nuremberg trials, describes a scene of incredible tedium. But fine weather, even forthose able to enjoy its amenities, could not dtspel the terrible boredom. The Tribunal, like the Apocalypse, was supposed to dnve out evil and enthrone good, but the goal was not atrained on four horses. For nearly a year the inmates and workers of the courthouse had been fairly drowned in documents, arguments, speeches, witnesses, translators, reporters, and other judicial whatnor. In the nature of things, thee ingredients of the proceedings became more and more repetitive as time went on. In August the trial of the organizations had been especially wearisome, enlivened only occasionally by surprises sud as the ghastly Sievers explosion orJudge Biddle's unintentional public lecture on brothels. TELFORD TAYLOR, THE ANATOMY OFTHE NURaMBERG TRiALs: A PERsoNu. .MLsoin 546 (19'92). 196. OsLEL, supra note 34, at 59. Osiel is restating a common critique here. He heiesS in the F05sbility and desirability of orchestrating trials of human rights abusers in a dramatic. pedagogcal l1aalon, so long as the lesson taught is a liberal one. ld at 65. Yet the examples he gives to show that lair trials can be conducted to maximize their dramatic effect (such as de decision to conduct the Argentinian trials in a single oral proceeding rather than in sequential proceedings with primarily 'ntten submissions) do not undermine the basic point that fair trials can, at times, be very dull indeed. Sc: i. at 77 197. See, eg., Mendez, supra note 7, at 16 (arguing that society has more faith in truth produced at trial since courts are stricter than truth commissions in examining evidence)- HarvardHumanRightsJournal/ Vol. 15 prosecution is at best an imperfect means to develop a complete record of the past.198 First, rules of evidence typically reflect not only a desire to ascertain the truth, but also competing public policy or constitutional concerns. For example, in the United States relevant and probative evidence is routinely excluded from criminal prosecutions when it is considered excessively prejudicial, denies the defendant her right of confrontation, or forces a defendant to incriminate herself. Yet in the wake of gross human rights violations, gruesome photographs, flagrant hearsay and perpetrator confessions are essential to developing an accurate picture of the past.199Moreover, while the threat of prosecution can be an important tool in forcing perpetrators to participate in other truth-seeking mechanisms,200 prosecutions themselves are ill-suited to eliciting testimony from perpetrators, the very people who know the most about the atrocities. "The primary sources of information concerning those infamies, the perpetrators themselves," notes the late South African Constitutional Court Justice John Didcott, "would hardly be willing to divulge it voluntarily, honestly, and candidly without the protection of exemptions from liability."201 The most serious deficiency of prosecutions when it comes to truthseeking is that trials focus on select individuals and thus do not account, in Minow's words, for "the complex connections among people that make massacres and genocides possible."20 2 The history produced by judges is "the byproduct of particular moments of examining and cross-examining witnesses and reviewing evidence about the responsibility of particular individuals. 203 Moreover, international tribunals are often located far from the affected society in transition, making the process less accessible to victims, witnesses, and the public. In sum, "if the goal to be served is establishing consensus and memorializing controversial, complex events, trials are not ideal." 2 1 Alternative mechanisms may be better suited than prosecution to developing full records of the past. Some have argued that due to the lower standard of proof and evidentiary and discovery advantages of civil proceedings, 198. See MiNOW, supranote 8, at 60. 199. Ofcourse, more lax evidentiary standards can be adopted, at least where perpetrators are not being prosecuted through domestic justice systems. However, those standards must still comply with basic due process, and due process does not permit the introduction of all relevant evidence, 200. See Villa-Vicencio, supranote 5, at 209 (noting the ways in which prosecutorial justice and truth commissions reinforced one another in South Africa: "[flew would have applied for amnesty if not faced with the threat ofprosecution"). 201. QuotedinNIER, supranote 9, at 105. 202. MNOW, supranote 8, at 47. For example, Babacar Sine, a Senegalese intellectual, has said of the prosecution of Hiss~ne Habr6, the former president ofChad, This case is much more complex than the role of Habt. There is the role ofFrance that supported him. There is the role of the United States that supported him. If we are to judge Hiss~ne Habr6, we have to also judge those who supported him. Quoted in Norimitsu Onishi, An African DictatorFacesTrial in His Place of Refuge, N.Y. Th~is, Mar.1, 2000, atA3. 203. MINoW, supranote 8, at 60. 204. Id. at 47. 2002 / A Frameworkfor UnderstandingTransitionalJustice civil liability is better than criminal prosecution at establishing "a definitive, historically accurate account of the atrocities."2- 5 This may be true, but civil trials, just like criminal trials, focus on individual responsibility and shift public focus away from systemic, shared culpability. By contrast, truth commissions are not limited to the facts of individual cases, but highlight the vast scope of and widespread complicity in the human rights violations. In addition, the history produced through truth commissions focuses not just on perpetrators but also "on victims, including forgotten victims in forgotten places."206 The merits of prosecution in achieving the second goal of rehabilitation, establishing the rule of law, are also unclear. Trials conducted before impartial courts that scrupulously observe due process requirements may showcase the benefits of the rule of law and contrast favorably with the lawless behavior ofthe defendanrs.2° Yet when those same due process protections free those who are perceived to be guilty, fair trials may well inspire contempt for the rule of law.208 On the other hand, trials which are conducted unfairly or which offend legal principles will undermine the "spirit of legality"&) that such trials are supposed to inculcate. For example, human rights activists initially cheered the prosecution in Senegal of former Chadian dictator Hiss~ne Habr6' 0 as demonstrating that even such an "ultimate untouchable"211 may not violate the law. However, after a change of government in Senegal, the case against Habr6 was dismissed,212 reinforcing the notion that the powerful are untouchable.213 A similar problem arises when the "small fry" are prosecuted instead of high level officials, fostering a perception that prosecutions are a form of scapegoating, not a means ofachieving justice. Perhaps the most distinctive contribution prosecution can make to societal rehabilitation is in establishing the wrongfulness of past atrocities. The prosecution and punishment of atrocities is a forceful way to disavow such conduct. A process of censure not only expresses disapproval of those who violate international human rights norms, but also serves to define and strengthen the norms themselves. While prosecutions help to establish 205. Murphy, supranote 17, at 47-48. 206. MiNow, supranote 8, at 60. 207. See, eg., NINo, supranote 16, at 146. 208. See Wren, supranote 54. Following the ICTR's decision, Rwanda suspended cooperattion with the tribunal. Joseph Mutaboba, Rwandan representative to the UN, described the ICTR as negligent in its duties, and said that the hardship suffered in jail by Jean-Bosco Barayagwiza 'is negligible compared to the suffering his victims endured." Id 209. Nuno, supranote 16, at 148. 210. Hisslne Habr6 ruled Chad from 1982 to 1990. His regime has been accused of mass murder and systematic torture. SeegenerallyHuman Rights Watch, The CastAganat Huir.e Habr, athttp-I11%%'w.hr%% orgljusticelhabrefintro_web2.htm (last visited Oct. 4., 2001); se alo Reed Brody Tk Prmciatrn61 HisuseHabri--An"African Piccuhet," 35 NEw ENG. L REV. 321 (2001). 211. Onishi,supranote 202, atA3. 212. SeegenerallyHuman Rights Watch, The CaseAgainst Hisub. Ha-Irl, qa note 210 213. Human rights activists are continuing their efforts to bring Habre to trial. It is still poible that this will occur, which would obviously change the lesson to be drawn from his care. HarvardHuman RightsJournal / Vol, 15 moral norms, one cannot assume that in every case "once accusations are leveled and indictments and judgments are issued, all will come to acknowledge the barbarous evils committed by genocidaires. 214 Rather, one must identify those situations in which prosecutions will help to forge common moral values.21 5 While prosecutions can produce moral consensus, they can also create scapegoats and feelings of bitterness, particularly where the selection of defendants appears to be politically motivated or where there is a perception that the trial represents victor's justice. One must be wary of a vicious circle: where prosecution is not grounded in moral consensus it will be seen as victor's justice. And if it is seen as victor's justice, it will lose its effectiveness as a tool for creating moral consensus. A society cannot be cured ofa condition it does not regard as a disease. Finally, is prosecution the most effective way to foster societal rehabilitation? Certainly prosecutions shape norms by condemning undesirable conduct. But trials also express the belief that "specific individuals-not entire ethnic or religious or political groups-committed atrocities."216 While individual crime is a component ofmass atrocity, "radical evil," as Carlos Nino notes, also requires an evil political and legal framework in which to flourish, Without that framework, it is unlikely that massive, statesponsored human rights violations will ensue regardless of whether punishment for previous violations takes place. With that framework in place and given certain antecedent circumstances, however, violations are highly likely even with previous convictions and punishment for human rights violations.217 Perhaps funds spent on prosecutions would have a stronger rehabilitative effect if spent on reforming a society's political and legal framework. Prosecutions may well be less successful in rehabilitating a society than a concerted effort to reduce inequalities in wealth, provide basic public education, create functioning courts, establish civilian control over the military, ensure the independence of the press, or hold free and fair elections. 214. Alvarez, supranote 59, at 468, 215. Since rehabilitation justifies "punishment" only to the extent that it contributes to moral reeducation, proportionality is not required. Thus, if there are other ways to rehabilitate societies which are as or more effective than prosecuting and punishing human rights abusers, then those methods may be preferable. 216. Krirz,supranote 21, at 128. 217. NINo, supranote 16, at 145. 2002 / A Framework/orUnderstandingTransitionalJustice D. Restorativejustice 1. Theory Two major paradigms fall within the rubric of restorative justice.21 , The first focuses on compensating victims219 and views crime as a harm that criminal justice should seek to undo.220 Under this view, the purpose of punishment is to repair injuries to victims,22 and thus the goal of criminal justice is for the offender to provide restitution to the victim. The second paradigm envisions crime as conflict and criminal justice as a form of conflict resolution.2 22 The basic assumptions of this approach are: (1) Crime is primarily a conflict among individuals resulting in injuries to victims, communities and the offenders themselves; only secondarily is it lawbreaking. (2) The overarching aim of the criminal justice process should be to reconcile parties while repairing the injuries caused by crime. (3) The criminal justice process should facilitate active participation by victims, offenders and their communities. It should nor be dominated by the government to the exclusion ofothers.Y" Under this view, the goal of criminal justice is the reconciliation of the offender, victim, and community.224 These two paradigms, compensation and conflict resolution, are often linked. An apology without restitution may mean little; if a friend apologizes for raking a pen but does not return it, her statement is worthless.225 "Apologies set the record straight; restitution sets out to make a new rec- 218. This discussion of these two paradigms draws heavily on Andrew Ashworth, RtawsijnrJwir::, ms PRINCIPLED SENTENCING: READINGS ON THEORY & Poicyz sspranote 23, at 300, 219. Se id 220. Dean E. Peachey, Rertitution,Rceacliatien,Retrituticn:Idarim tk-- Fcr,-i cfJXszae Perc,!r Dalr, in RESTORATrIVE JUSTICE ON TRIAL-PITFALLS AND POTENTIALS OF VICTIM.OFFr-NDER NfrDoATIONINTERNATIONAL RESEARCH PERSPECTIVES 551, 552 (Heinz Messmer & Hans-Uwe Otto eds., 1991). 221. For example, Daniel W.Van Ness explains that punishment does not help repair the injuries caused by crime (but] simply creates new injuries; now both the victim and the offender are injured. Recompense, on the other hand,issomething given or done to make upfor an injury. This underscores hat the offender who caused the injury should be the active party, and that the purpose ofpunishment should be to repairas much as possible the injury caused by the crime. Daniel W.Van Ness, RestoratitvJusticeandIntrnationalHtt-w.rnRShts, in REs'roRAr JustIcE INRLIINAIONAL PERSPECTIVES 17, 27 (Burr Galaway &Joe Hudson eds., 1996). 222. SeeAshworth, spra note 218, at 301. 223. Van Ness, supra note 221, at 23. 224. Strictly speaking, restorative justice, under either paradigm, is nor a theory ofpunishmnt. The purpose of requiring an offender to pay her victim's hospital bills or to perform community service is to provide compensation fbror to encourage reconciliation with the victim; punishment is a side-effecr, S: Ashworth, supra note 218, at 301. Retribution, deterrence and rehabilitation may all be sidrcffects of restorative justice. While some argue that restorative justice is better able to adieve the goals than more traditional forms ofcriminal justice, others believe that restorativism is its own end, and s jusfied even ifit has no preventive benefits. Sa id at 304. 225. SeeSHRIVER, spranote 102, at 224 (providing this example). HarvardHumanRightsJournal / Vol. 15 ord," explains theologian Donald Shriver.226 Yet conflict resolution cannot rest on compensation alone, particularly since it is often impossible to restore to the victim what she has lost or repair the harm that she has suffered. While a stolen pen can be replaced, a murdered child cannot. Restorative justice differs from retribution, deterrence and rehabilitation in its focus on the victim. 227 Restorative justice raises questions about the identities of the parties to the conflict. In other words, it asks whose interests are relevant to the case.228 In the West, crimes are defined as offenses against the state; in this way, as Nils Christie argues, the state has stolen conflicts from victims, 229 communities and offenders. "Virtually every facet of the criminal justice system works to reduce victims, offenders and communities to passive participants," notes Daniel Van Ness.230 Restorative justice aims to return conflicts to the parties to the conflict. Yet the idea that crimes are offenses against society at large should not be ignored, for offenders harm not only specific victims but also entire societies. Thus, restorative justice must consider the interests of both the individual victim and the wider community. Theoretically, the compensatory paradigm of restorative justice is somewhat disconnected from culpability, since the degree of harm caused may not reflect the blameworthiness of the offender.231 Proportionality in this context is based on the amount of harm inflicted on the victim, not the maliciousness of the offender's intent.23 2 The conflict resolution paradigm, on the other hand, is more likely adequately to address culpability; a victim will be angrier at someone who hits her intentionally than at someone who does so accidentally, and will require more ofthe offender in order to resolve the conflict.233 2. RestorativeJusticein TransitionalJustice In the context of transitional justice, reparations can be seen as a form of compensation, while approaches seeking to heal society's wounds can be un- 226. Id. 227. Ashworth notes, however, that not all restorative theories are necessarily victm-centered, and not all victim-centered approaches are necessarily restorative. Ashworth, supra note 218, at 300. Nevertheless, under both a compensatory and a conflict resolution model of restorative justice, punishment, deterrence and rehabilitation of offenders are insufficient. The victim must also be considered. 228. See id. at 300-02. 229. See Christie, supra note 30, at 312. 230. Van Ness, supra note 221, at 24. 231. See Ashworth, supra note 218, at 305. 232. See Lucia Zedner, ReparationandRetribution:Are They Rcconcilable?, in PRINCIPLED SENTENCINO: READINGS ON THEORY & POLtcYsupra note 23, at 336, 344. 233. Some scholars have argued that restorative justice and retribution are not as antithetical as is sometimes assumed. Lucia Zedner, for example, has noted that both reparation and retribution, unlike rehabilitation, are predicated on notions of individual autonomy. Both theories "presume that offenders are rational individuals able to make free moral choices for which they may be held liable." hi at 344. Offenders can legitimately be called to account, whether by making good or suffering proportionate punishment. See id. 2002 / A Frameworkfor UnderstandingTransitionalJustice derstood within a conflict resolution paradigm. In both cases, as in the domestic context, victims claim a central role.23 4This Section will first look at reparations as a form of transitional justice, and then assess the capacity of prosecutions to reconcile victims, offenders, and society as a whole. Reparations in the wake of massive human rights violations are designed to provide at least partial restitution to victims. While compensation typically involves monetary payments to individuals, other types of restitution, like building memorials or naming streets for victims, are less tangible and less directly focused on individuals. The truth itself can also be understood as a form of reparation. When the silence is broken, and families learn where the bodies are buried or victims discover the identities of their torturers, the injury caused by past abuse may begin to be repaired. Reparations should not only acknowledge the survivor's loss, but also repair the harm caused. While this is a valuable goal, it is often difficult to achieve in practice. Where former regimes have harmed both individuals and entire communities, should scarce resources be used to pay compensation to individual victims, or to rebuild a society victimized by poverty, appalling health care and lack of education? Moreover, monetary measures cannot remedy non-monetary harms, like the loss of a child or the agony of remembered torture.235 "[NIo market measures exist," writes Minow, "for the value of living an ordinary life, without nightmares or survivor guilt."2 36 Reparations are likely to be grossly disproportionate to the damage caused, and many thus trivialize suffering.237 In fact, victims frequently express only modest demands for reparations, such as a tombstone or death certificate for their loved ones or the removal of bullets from their own bodies338 Often, their most significant demand is for the truth. Thus while restitution can never fully compensate victims, it can serve an important symbolic function. Proponents of prosecution sometimes argue that trials are useful as a basis for reparations.2 39 This justification seems dubious, at least insofar as it is used to support the primacy of prosecutions. Since prosecution focuses on establishing the guilt of a few individual perpetrators, it is not an ideal tool for identifying victims who deserve compensation. Moreover, because prosecutions are adversarial, defendants may never reveal those facts about the past which victims most want to know. Arguably, alternative mechanisms are better able to identify victims, as well as to provide them with a more 234. Se4 ag., M Now, supranote 8, at 60 (arguing that the most distinctive element oftruth commissions, in comparison with prosecution, is the focus on victims). 235. Seeid at 93. 236. Id at 104. 237. Seeid at 93. 238. Id at 105-06 (describing the modest requests of South African victims before the Truth and Reconciliation Commission and worrying whether this modesty represents "dignifird assertions made bindividuals who have no illusions about the possibility of external repair for their Iossts," or the 'lowered expectations of the persistently oppressed"). 239. See, eg., Cassel,supranote 11 (citing reparation ofvictims as one of the goals of the lntemaranal Criminal Court). HarvardHuman RightsJournal / Vol. 15 comprehensive, and more personal, account of the past.240 Truth commissions and civil suits offer at least the possibility that victims will be com- pensated.241 Reparations may even take the form ofgovernmental or international aid programs providing medical treatment, scholarships for victims' children, or preferential access to government services such as public housing or transportation. Any of these approaches is more likely to provide real restitution to victims than prosecutions. Thus the reparative paradigm of restorative justice offers little justification for prosecution. How do prosecutions hold up under the conflict resolution paradigm of restorative justice? First, we must ask if reconciliation or forgiveness are even possible after massive human rights violations. In the domestic context, restorative justice programs are usually restricted to minor offenses, Yet conflict resolution in the context of transitional justice requires interaction between those who carried out and those who suffered from horrific atrocities. "Healing," worries Minow, may be "an absurd or even obscene notion for those who have died," as well as for the survivors, who often feel as if they "have died or live among the dead."2 42 Susan Dwyer expresses a similar concern that "[rleconciliation is being urged upon people who have been bitter and murderous enemies, upon victims and perpetrators of terrible human rights abuses, upon groups of individuals whose very self-conceptions have been structured in terms of historical and often state-sanctioned relations of dominance and submission."243 Absent a clear explanation of reconciliation and what it requires, she continues, "proposing reconciliation will seem like a political sop aimed at masking moral defeat."2 t So, what is reconciliation? Scholars of transitional justice distinguish between the repair of relationships that will suffice for a society to move forward and unrealistic expectations of transformative interactions between victims and perpetrators. Dwyer, for example, believes apologies and forgiveness are not absolutely required for future interaction. She contrasts personal reconciliation between victims and perpetrators, which may be too much to ask, with national reconciliation, which she argues is more possi- ble.245 What is necessary if perpetrators and victims are to live together in 240. See, e.g., Hayner, supranote 20, at 655 (noting that truth commissions can go hand in hand with reparations); John Dugard, RetrospeciveJusticeInternationalLaw and the South Afrian Mod, in TRANSITIONAL JUSTICE AND THE RULE OF LAW IN NEW DEMOciAcIEs supra note 7, at 269, 277 (noting that reparation was one of the goals of the South African Truth and Reconciliation Commission); Van Zyl, supra note 4, at 661 (reparations granted by the South African Truth and Reconciliation Commission were a "fairer, more nuanced and holistic means of redress than that which could be obtained through the courts"); Landsman, supra note 5, at 88 (arguing that truth commissions areat least as good as trials at establishing the predicate for victim compensation). 241. Murphy, supranote 17, at 48. 242. MiNow, supra note 8, at 62. 243. Dwyer, supranote 9, at 82. 244. Id. at 89. 245. Id.at 95. Dwyer cites a remark by anti-apartheid activist Marius Schoon about the Truth and Reconciliation Commission, which heard an amnesty petition from the bomber who killed Schoon's family: "I think [the Commission] is going to bring about national reconciliation. In my case, it's not 2002 / A Frameworkfor UnderstandingTransitionalJustice the future, she claims, is the development of a common national historical narrative, based on agreed-upon facts and a shared interpretation of them.21 ' Shriver, in contrast, prefers the concept of "forgiveness."247 Believing victims must have more courage than Dwyer requires, Shriver asks victims "to face still-rankling past evils with first regard for the truth of what actually happened; with resistance to the lures of revenge; with empathy-and no excusing-for all the agents and sufferers of the evil; and with real intent on the part of the sufferers to resume life alongside the evildoers or their political successors. ' 24 8 Whatever words we use to describe it, the concept of restorative justice is certainly relevant to transitional justice. The horrors of the past can be seen as a form of conflict, and the goal of transitional justice as conflict resolution. Transitional justice can be organized so as to give victims a central role and repair relationships between them, perpetrators, and society at large. In this way, transitional justice can strive for at least enough forgiveness, reconciliation, or healing to make coexistence possible. A reconciliation-based argument for prosecution is premised on the notion that retribution is a precondition for societal healing. There has been considerable debate about the tension between "justice," the requirement that those who violate human rights be punished, and "peace," the desire for both social cohesion and an end to human rights violations.249 Historically, the concern has been that pushing for justice would jeopardize peace. Increasingly, however, diplomats involved in settling violent disputes and politicians seeking to move their countries forward recognize that a focus on past atrocities is not an obstacle to stability and conflict resolution, but is, in Kritz's words, "an integral and unavoidable element of the peace process." 250 Proponents of prosecution have equated this new framework, in which justice is a precondition for peace, to an argument that retributive punishment must precede social healing.25 1 Accordingly, before there can be reconciliation in places like Bosnia and Rwanda, there must be retribution, at least as far as the most serious offenders are concerned. 2"2To achieve closure, "individuals emerging from massive abuse and trauma [must] develop apgoing to bring about personal reconciliation." Dwyer concludes that Schoon was "able to se how the narrative ofhis country can be revised in ways that his own personal soty cannot bm Id. 246. Idat 89-90. 247. SHRlVER, supranote 102, at 67. 248. Id. 249. Sometimes the tension is framed in terms of truth versus justice, similar to the peace versus justice framework. Truth is arguably a component of peace, for long-term cohesion cannot In-achieved without an understanding and acknowledgment ofthe past. 250. Kritz, supra note 21, at 128. 251. Of course, as a matter of theory, reconciliation is nor the justification for retribution. Ac'most it can be a beneficial side effect. Offenders must receive their -justdeserts. whether or not this promates reconciliation. 252. See David Little, A Different KirdqfJustice: Dealing uith Hor-.n Reis Ut'Atat:rnTraneut:=! Sodeties, 13 Em'ics AND INT'L AFF. 65, 66 (1999). HarvardHumanRightsJournal / Vol. 15 propriate mechanisms to confront and process that past experience," 253 But the need to face one's past does not adequately explain why such a confrontation must be retributive. Forgiving, after all, is not actually the same as for- getting. 254 Perhaps, as is frequently asserted, one cannot forgive what one cannot punish. 255 While this argument has strong intuitive appeal, it is problematic. First, it is a conclusion, not a verifiable fact. It seems equally plausible that those most likely to demand punishment are those least likely to forgive. Second, the extent to which forgiveness is contingent on punishment likely depends on context. Some offenses are easier to forgive in the absence of punishment than others, and some cultures and religions promote forgiveness more than others. Third, if one cannot punish all, how can one possibly forgive all? If only the ringleaders are punished, can one really expect victims to forgive when they must pass their tormentors in the street on a daily basis? If punishment is a prerequisite, reconciliation between the perpetrators and their victims is impossible. Fourth, and most importantly, the argument that one cannot forgive what one cannot punish suggests that where one cannot adequately punish extraordinary evil, one cannot forgive. If reconciliation depends on forgiveness, and forgiveness depends on punishment, does the impossibility of proportionate retribution in the context of extraordinary evil render reconciliation impossible? Truth commissions might be better suited to a restorative model of transitional justice than prosecutions. Truth commissions can focus on the victims, craft a shared narrative about the past as the basis for a shared future, and facilitate the active involvement of victims, perpetrators, and the larger community. In essence, truth commissions return the conflict to those who participated in it. The more open, discursive nature of truth commissions better addresses the problem of widespread complicity in massive human rights violations. Moreover, the informality of the process brings other benefits such as the ability of officials to grieve publicly with victims. "[Truth and Reconciliation Commission Chairperson Archbishop Desmond) Tutu cries," Justice Albie Sachs of the South African Constitutional Court has said. "A judge does not cry."256 Furthermore, restorative justice is multi- 253. Kritz, supranote 21, at 127. 254. Donald Shriver eloquently makes this point: Forgiveness begins with memory suffused with moral judgment. Popular use of the wordifrgiveness sometimes implies that to forgive is to forger, to abandon primary concern for the crimes of an enemy. Quite the reverse: "Remember and forgive" would be a more accurate slogan. Forgiveness begins with a remembering and a moral judgment of wrong, injustice, and injury. For this very reason, alleged wrongdoers are wary ofbeing told that someone "forgives" them. Immediately they sense that they are being subjected to some moral assessment, and they may not consent to it. Absent a preliminary agreement between two or more parties that there is something from the past to beforgiven, forgiveness stalls at that starting gate. SHRIVER, supra note 102, at 7. 255. See,e.g., Landsman, supranote 5, at 84 (arguing that prosecution may be essential to healing a society's wounds after massive human rights abuses because one cannot forgive what one cannot punish). 256. Quotedin MINOW, supranote 8, at 73. 2002 / A Frameworkfor UnderstandingTramitionalJustice directional. Prosecutions may be appropriate where individuals can be clearly classified as either victims or perpetrators. Truth commissions, on the other hand, recognize that this distinction is not always clear. During periods of mass atrocity or repression, individuals usually assume different roles over the course of the conflict. As a result, apologies must be both given and received. Ofcourse, truth commissions will not always work as a form of restorative justice. Repairing relationships through discussion and confrontation requires not only that all parties be engaged in the process, but also that perpetrators recognize their blameworthiness and accept responsibility for past actions. Amnesty can be understood as an annulment of the appropriate retributive penalty in return for truthtelling.2 "7 But offenders may be unwilling to tell the truth or to express contrition.258 Moreover, to be meaningful, apologies must be linked to restitution. Reconciliation in the wake ofatrocity requires "the credibility that can be established only by implementation of social and economic programs that concretely address the substantive injustices" of the past.259 Even those who acknowledge their own culpability or complicity through a national truth-telling process may recoil at the loss ofprivileges and power that true accountability demands. How much reconciliation can be achieved if in post-apartheid South Africa, for example, whites admit that their economic, social, and political status was based on a morally bankrupt system, but then refuse to accept sharply redistributive taxation? The fundamental difficulty with restorative justice is that it cannot adequately address demands for retribution.260This is not really a problem with restorative justice itself, but rather with the difficulty of reconciling two competing goals of transitional justice. This issue nevertheless deserves comment here, as it lies at the heart ofmany critiques of truth commissions. Retribution requires the isolation of perpetrators and the definition of their behavior as outside the realm ofhuman interaction. Restorative justice, in contrast, requires that we recognize the offenders' humanity and the pos- 257. SeeLittle, supranote 252, at 73. 258. See id at 74 (arguing that contrition on the part ofSouth African offenders appeanng bfore the Truth and Reconciliation Commission seemed in short supply). 259. Dwyer,supranote 9, at 95. 260. Note that the problems here are not the same as in a domestic criminal justice system. First, in domestic systems the fact that some victims are vengeful and others forgiving can result in disparues in the treatment oftwo equally culpable offenders. Since a restorative vision of transitional justice dots not typically involve negotiation between individual victims and offenders, but rather a whole society in discussion with itself, such inequalities present less ofa problem. Instead, the difficulty lies in the f1, that those in the society who favor a more retributive model will be shortchanged. Second, d scopc of restorative justice in the domestic context differs from that in the transitional justice context. In domestic systems, efforts at conflict resolution between victims and offenders are usually an -ddmdum to a much larger and relatively comprehensive system of sanctions. In transitional justice, preacutions can reach only a fraction of offenders. Thus truth commissions or other non-prosecution alternatives. even when used in conjunction with prosecution, can be a mechanism for ensuring some accountability from the vast majority ofoffenders. HarvardHumanRightsJournal / Vol. 15 sibility of a continuing relationship with them. Forgiveness, argues Shriver, is about repairing fractured human relationships, and therefore "insists on the humanity of enemies even in their commission of dehumanizing deeds, and ...values the justice that restores political community above the justice that destroys it.''261 In the context of domestic crime, criminality is sometimes understood as in part the product of social and economic circumstances. In the context of massive atrocities, however, it is far easier to be appalled and to demand vengeance than it is to ask ourselves if, in similar circumstances, we might have done the same thing. Among the most important lessons to be learned from confronting radical evil is just how banal and widespread it really is. "Affirming common humanity," as Minow notes, "does not mean turning the other cheek or forgetting what happened. 262 But it does require that accountability mechanisms not dehumanize perpetrators. In essence, restorative justice seeks to affirm the humanity of those who have behaved inhumanely. It offers accountability, not vengeance. 263 Both retributive and restorative justice envision reconciliation as a product of full accountability for wrongdoing. 264 But while retributive accountability involves proportionate punishment, restorative accountability demands an acknowledgment by offenders of their culpability and a willingness to make good. While retributive justice allows society to punish an offender as a means of achieving reconciliation, restorative justice requires society to include the offender in the process of reconciliation. Neither approach is perfect; society cannot create adequate punishments for all offenders, nor can it force all offenders to be contrite. Whether either strategy will actually bring about reconciliation will depend on the extent to which punishment, apology, and reparation can form the basis for a common future. "Simple justice is elusive," concludes Shriver. In fact, "forgiveness thrives upon the tension between justice-as-punishment and justice-as-restora- tion."265 Retribution can play a restorative role, while "forgiveness can make ...room for punishment while making wider room yet for the repair of damages and the renewal of relations among enemies." 266 261. SHRIVER,$ulpra note 102, at 9. 262. MINOW, supranote 8, at 146. 263. Perhaps there are forms of retributive justice that also recognize the offender's humanity. This question, however, leads back to the debate about whether retribution is different from vengeance, What is clear is that in practice retributive justice often functions as an outlet for vengeance, and there is usually a demand for vengeance after mass atrocities. Kevin Minor andJ,T.Morrison argue that the central problem with restorative justice, and by extension with truth commissions, is that they are "not a vengeful response to crime, and vengeful responses are perhaps the most blatant means of displaying emotional outrage." Kevin . Minor and J. T. Morrison, A Theoretical Study and Critique of Restorative Justice, in REsToRATIvE JUSTICE ON TRIAL: PITFALLS AND POTENTIALS OF VICTIM-OFFENDER MEDIATION-INTERNATIONAL RESEARCH PERSPECTIVES supra note 220, at 117, 119. 264. See Little, supra note 252, at 79. 265. SHRIVER, supranote 102, at 32. 266. Id. 2002 / A Frameworkfor UniderstandingTransitionalJustice E. Comnmnication/CondemnzationlSocialSolidari0 1. Theory For tmile Durkheim, punishment is a form of moral communication used to express condemnation and strengthen social solidarity." 6' Durkheim argues that the "true function [of punishment) is to maintain social cohesion intact."268 Crimes, according to Durkheim, are acts that violate a society's fundamental moral code of sacred norms, thereby weakening those norms. Punishment plays a critical role in preventing the collapse of the moral order by limiting the "demoralizing" effects of crime.' 9 Punishment also functions as a collective response that demonstrates and reaffirms the real force of the common moral order.270 By punishing, a society expresses its shared moral outrage, strengthening and reinforcing the norms of social life. 7 Punitive rituals, by articulating shared sentiments, help to reflect and sustain a society's moral values, thereby strengthening the bonds of community.272 For Durkheim, sanctioning offenders is a way to communicate the continuing validity of the law, a "language" which "expresses the feeling inspired by the disapproved behaviour. ' 275 Thus the two central premises of the "Durkheimian school" of thought are that punishment communicates social condemnation, and that it therefore plays a role in reaffirming, or even creating, social identity and/or social solidarity.274 267. This discussion of Durkheim draws heavily on DAVID GARLAND, PuVsitItNT AND MDLLN SOCIETY: A STUDY IN SOCIAL THEORY (1990). 268. Durkheim further argues that punishment -does not serve, or else only stnes quite srondany; in correcting the culpable or in intimidating possible followers. From this point of view its eikelhS is justly doubtful and, in any case, mediocre." &.MrLE DuRKHrir s,Til DivisioN OmLon w.SojLT 103 (1933). 269. See GARLAND, supra note 267, at 42. 270. Seea at 28-35. 271. See id. 272. See id at 33-34. Of course, punishment is not the sole force creating socal solidantF Durkhrim believed that punishment plays a more important solidarity-producing function in .ome societies. such a5 "simple societies," than in others, such as -modem societies," where te division of labor becomes the predominant source of solidarity. 273. ?ur.E DURKHEI , MoRAL EDUCATION 176 (1973). 274. What here is called the "Durkheimian school" actually encompasses a range of quite varied approaches. David Garland, for example, criticizes Durkheim's understanding of what mMage the law communicates but agrees that punishment is a form of communication which at least in part is a product of deeply felt popular beliefs. See GARLAND, supra note 267, at 53. Similarly, Antony Duff has suggested that punishment is "a mode of rational, transparent communication and persuasion," which is aimed at convincing the offender of his wrongfulness, but also has the effect of assuring the vttnm anJ the public at large that "the law means what it says." Antony Duff, Daiertacd Perit:ze,tn PitNCIrLsi SLNT .%cI.%GREADINGS ON THEORY & POLICY supra note 23, at 161, 161. According toJoel Feinberg. it is precisely this "expressive function" which distinguishes punishments from what he terms "pnaltites "Joel Feinberg, The Exprssive Functionof Punishrent,in A READER ON PuNtIsisma%"r at ', 71 (Antony Dull & David Garland eds., 1994). Punishments convey the community's sentiments of condemnation. rprobation, and resentment, while penalties, such as flunking a class, losing a job, or paying a parking ticket. do not. Through punishing, the scare goes on the record about its values. For example. Feinberg sa)-. it paramour killings go unpunished, the law expresses the judgment that the vindication of the cuckoldtd spouse is more important than the very life of the murdered lover. Sr iat. at 75-78. lake Feinbcrg. Herbert Morris sees the communicative component of punishment as a defining characteristic, An amton HarvardHumanRightsJournal / Vol. 15 In examining the Durkheimian school, three points are worth noting. First, to argue that punishment is a form of communication sheds no light on the message being conveyed. Different forms of punishment communicate different messages and produce different forms of social solidarity. Moreover, because other approaches to criminal justice rely on the communication of particular messages, like those of deterrence or moral reformation, the communicative paradigm is intertwined with, though conceptually distinct from, the theories discussed above. Second, in describing the impact of punishment on the construction of social identity, Durkheimian theorists focus on why a society punishes, not why a society says it punishes. After all, in adopting purpose provisions for penal codes, legislatures do not express a desire to shore up the social order, but rather use the language of retribution, deterrence, rehabilitation, or restorative justice. Moreover, descriptions of the consequences of punishment for social control or identity do not explain why one should punish. David Garland may well be right that punishment is sometimes used to express and construct the dominant moral order.275 But the idea that ruling elites employ punishment as a way to establish the legitimacy of their social vision is not an affirmative reason to punish. Third, if punishment is imposed in order to express societal reprobation, then condemnation, not punitive treatment, is the central goal. Under this view, while a prisoner-of-war camp may be harsher than a prison, such incarceration does not constitute "punishment," since it is not an expression of social condemnation.27 6 If one punishes in order to communicate moral valtaken inrevenge or retaliation may have no purpose but to harm the other. By contrast, punishment isan educational process whereby offenders are taught that their conduct was wrong inthe hope that they will choose to act differently in the future. See Herbert Morris, A PaternalisticTheory of Punisbment, in A READER ON PUNISHMENT at 95, 97, 100 (Antony Duff & David Garland eds., 1994). Thus the law "plays an indispensable role in our knowing what for society is good and evil. Failure to punish serious wrongdoing, punishment for wrongdoing in circumstances where fault is absent, would serve only to baffle our moral understanding and threaten what is so often already precarious." Id.at 104. 275. Garland, for example, sees punishment as a "cultural agent" which shapes public values and social identity. See GARLAND, supra note 267, at 249-76. Punishment does not reflect Durkheim's "conscience collective," but rather a "dominant moral order." "Establishing society is not just a problem of socializing deviant individuals," Garland writes. "liltis also, and crucially, a matter of subduing competing social movements and social groups." Id. at 52. Unlike Durkheim, Garland thus understands the penal law not as an expression ofa pre-existing moral consensus, but rather as a tool used in the political struggle to transform and shape public sentiment in accordance with a particular vision ofsociety. Se id. at 54. In other words, "punishment does not just restrain or discipline 'society'-punishment helps create it." Id. at 276. 276. Feinberg, supra note 274, at 75, citing Henry Hart, The Aims of the Criminal Law, 23 LAV/ & CONTEMP. PROBs.401 (1958). As Feinberg suggests, in an expressive theory ofpunishment, [w]hat justice demands is that the condemnatory aspect of the punishment suit the crime, that the crime be of a kind that is truly worthy of reprobation. Further, the degree of disapproval expressed by the punishment should 'fit' the crime only in the unproblematic sense that the more serious crimes should receive stronger disapproval than the less serious ones.... That is quite another thing than requiring that the 'hard treatment' component, considered apart from its symbolic function, should 'fit' the moral quality ofa specific criminal act .... Id. at 89. Feinberg notes that it is a convention ofmodern society that "condemnation is expressed by hard treatment, and the degree of harshness of the latter expresses the degree of reprobation of the for- 2002 / A Frameworkfor UnderstandingTransitionalJustice ues, it matters less that offenders pay for their misdeeds, than it does that society finds a way of adequately expressing its disapproval. Prosecuting and sanctioning offenders is an important way of communicating reprobation, but it is not the only one. If one justifies punishment as a means of communication, one must explain why punitive sanctions are the most appropriate form of censure.2 77 2. Communication/Condemnation/SocialSolidarityin TransitionalJustice In transitional justice, prosecutions are frequently portrayed as a way to communicate, to condemn, and to rebuild social solidarity in the wake of fratricide. Diane Orentlicher, in arguing for a duty to prosecute, insists that "[iif the international community cannot prevent, at least it must not condone. Its words of censure or approval eventually filter through."278 Human rights observers cheer each new prosecution as potent evidence that everyone, including heads of state, is bound by international human rights law.279 Indeed, trials have even been described, in Bruce Ackerman's famous phrase, as "constitutional moments" when a society is reborn.231 Implicit in such statements are two strands of the Durkheimian argument: that punishment communicates social condemnation, and that it reaffirms social solidarity or shapes social identity. First, as regards the communicative force of punishment, the prosecution ofthose who commit genocide, war crimes, and other atrocities indisputably conveys a powerful message of condemnation. Punishment of those who violate human rights conveys the message that violating human rights is profoundly wrong, and that wrongdoers must accept the consequences of their actions. The failure to respond to mass atrocities would send a very different message. As the moral philosopher Antony Duff has argued, "we are not sincere in our condemnation of an act if we are silent in the face of its met." IaStill, this is a convention, not a logical requirement. 277. See Duff,supranote 274, at 164: A central task fbr a communicative theory of punishment is to explain why punishment should take the fbrm of hard treatment. Penal hard treatment can communicate censure: but since censure can also be communicated by the offender's conviction itself, or by purely symbolic measures which are burdensome or unwelcome sold) by virtue ofthe censure which they communicare, we must ask what could justify the state in using hard treatment as the method of com- munication. 278. Orentlicher,supra note 1, at 2615. 279. See Warren Hoge, PinochetIs Ruled Unfitfr TrialardAa),Be Fred,N.Y.TissEs, Jan. 12, 2000, at Al (citing Kenneth Roth ofHuman Rights Watch, who argued that, despite the release of Pinochet. the case set a precedent that "Eh]eads ofstate no longer enjoy impunity for crimes against humanity-); Onishi, supra note 202, at A3 (citing human rights officials who claim that the case against Habni is significant for Africa because it shows that even the "ultimate unmuchable[s]" may not violate the law). 280. BRUCE AcKER.AN,WE THE PEOPLE (1991). Nino argues that rals, because they are -great occasions fbr social deliberation and far collective e.amination of the moral values underlying public institutions," are often more likely to be "constitutional moments" than the enactment of a new constitution. NiNo, supranote 16, at 131. HarvardHuman RightsJournal / Vol. 15 commission: to ignore manifest breaches ofwhat we have declared to be important norms is in effect to deny what we claim to have declared." 281 Our choices, however, are not limited to prosecution and inaction. Truth commissions, reports by investigatory bodies, lustration, and even works of art, literature, and history can also send a powerful message. Are trials more effective than these other methods of communication? Douglas Cassel claims that "[tlhere is no more powerful social condemnation of evil than to label it as a serious crime, for which serious punishment may be imposed. ' " 82 Criminalizing mass atrocity and prosecuting perpetrators are indeed among the strongest ways to express our horror. However, even if we accept that prosecution is the most potent vehicle for expressing the wrongfulness of human rights abuses, it does not necessarily follow that trials are always the best method for communicating moral lessons. While we will always want to condemn, just what and whom we want to condemn will vary. Fundamentally, prosecutions communicate a lesson of individual responsibility; they rebuke particular criminals in the strongest possible terms. Prosecution, especially trials of the most senior officials, may also be used to air a country's past more generally. Eichmann, notes one Israeli scholar, "rather swiftly became peripheral to his own trial. 28 3 Nevertheless, if we respect due process, prosecutions cannot be allowed to slip into the realm of show trials or scapegoating.284 "It is one thing," writes Mark Osiel, "to acknowledge that prosecutors have a legitimate range of dramaturgical discretion; it is quite another for them to attempt a staging of Hamlet without the prince."125 Even when the "prince" can be found and brought to trial, however, the binary oppositions in criminal law-guilty and innocent, blamers and blamed-may distort the lessons prosecutions teach.286 Evidence relevant to the message may not be relevant to the trial. Prosecutors, whose goal is to convict particular individuals, are limited in their ability to expose or condemn wider patterns of conduct underlying mass atrocities, such as economic inequality or the opportunistic exploitation of ethnic rivalries. The criminal law often exculpates most of the morally culpable parties because, in Osiel's words, "they are 'culpable' in ways that liberal jurisprudence does not recognize. '28 7 281. Duff,supranote 274, at 162. 282. Cassel, supra note 11, at 533. 283. Moshe Halbertal, The Seventh Million. The Israelis andthe Holocaust, NEw REPUBLIC, Oct. 18, 1993, at 40, 43 (book review). 284. For a comprehensive discussion of the benefits and drawbacks of using trials to shape public memory in the wake of mass atrocity, see OSIEL, supra note 34. See alsoIan Buruma, Tf-in WAGrS Or GUILT: MEMORIES OF WAR IN GiuANY & JAPAN 142 (1995) ("[W]hen the court of law is used for history lessons, then the risk ofshow trials cannot be faroff."). 285. OSIEL, supranote 34, at 139. 286. Id.at 159. 287. Id.at 163. 2002 / A Frameworkfor UnderstandingTransitionalJustice The message of prosecution is shaped not only by its focus on individual culpability, but also by its inevitable selectivity. Defactoamnesty for the vast majority of perpetrators, coupled with an emphasis on the personal guilt of the prosecuted few, communicates that it is Hamlet, not the supporting cast, who is to blame. Often an expression of condemnation that effectively ignores the guilt of the vast majority of actors is inadequate and misleading. Ifone wants to express outrage at a few wicked individuals, prosecutions are an excellent tool. But if one wants to condemn the offenses of the many, as well as voice disapproval of the social, economic, and political conditions underlying a conflict, non-prosecution alternatives like truth commissions may do a better job. The second strand of Durkheimian analysis-the idea that punishment contributes to social solidarity by reaffirming shared moral norms-also has implications for the choice between prosecution and non-prosecution alternatives. Clearly, prosecution can have a powerful effect on a nation's collective conscience. Chilean psychologists report that the arrest of Pinochet produced "a great catharsis that has begun to break the silence," with hundreds if not thousands of torture victims seeking therapy for the first time.2 Commentators claim that trials of Nazi officials effected a "symbolic severing of ties to the past," allowing Germans to forge a new identity for the future.28 9 It is important to recognize, however, that while trials of human rights abusers, as expressions of collective outrage, can certainly reinforce moral norms and provide opportunities for the creation of a new social identity, they will not do so in every instance. A few years after the Argentinian prosecutions, writes Osiel, "state-sponsored torture and murder-now by police, of common criminals-was applauded by many Argentines, including those in prominent public office."2 90 While prosecutions can have tremendous communicative force, several factors limit their usefulness as tools for social transformation. First, because prosecutions are selective and focus on individual responsibility, the truth they produce is not always suited to fostering social consensus. As a matter ofpersonal moral responsibility and criminal law, the fact that A's ancestors killed B's ancestors cannot justify A's killing of B. After all, the penal law, unlike the law of torts, almost entirely excludes "comparative fault. '-' In transitional justice, however, acknowledging the broader context is often critical if the goal is to establish a shared condemnation of past atrocities as the basis for a new common identity. Truth commissions confront similar problems rooted in the inevitably limited scope of their inquiry.' - Nevertheless, such mechanisms generally articulate many sides of the same story, 288. Clifford Krauss, PinodetCaseRetiring oea cftk:Tcrturd,N.Y. T.srs.Jan. 3. 20U1t. atA I 289. OSIEL, sypra note 34, at 193. 290. Id at 196. 291. Id at 123. 292. See id at 136 n.177 (noting that truth commissions, likecourts, must decide which airc'aics will be covered by their reports). HarvardHumanRightsJournal / Vol. 15 and therefore may be somewhat better at fostering social solidarity in societies where guilt runs in multiple directions. Second, the Durkheimian account of how punishment produces solidarity is predicated on the existence of a community organized around a shared moral code enshrined in the criminal law.293 However, as Osiel explains, "whether any such shared moral code exists and whether it is accurately reflected in the community's criminal law prove to be highly problematic propositions in the historical circumstances where large-scale administrative massacre occurs." 294 Post-transition societies are frequently characterized by what Nino terms an "extreme conceptual divergence," 295 whereby those who committed abuses are sincerely convinced that they acted in society's best interests. Moreover, while Durkheim "saw solidarity arising out of the shared indignation of the innocent many toward the guilty few,"296 transitional justice requires that such solidarity arise out of the shared indignation of the guilty many towards themselves. Of course, punishment and condemnation can be used to shape collective memory and social identity even in the absence of a broad moral consensus. As Garland has suggested, punishment sometimes reflects the "dominant moral order" rather than the "conscience collective."297 The winners in a particular conflict may craft prosecutions or non-prosecution alternatives to suit their political needs, reinforce their values, and enshrine their version of his- tory.298 The fact that such efforts are undertaken to establish the legitimacy of a particular social vision is not, however, a morally valid reason to condemn or punish. Moreover, whatever social solidarity is produced through such partisan mechanisms will likely be untenable; proceedings that are perceived as victors' justice will not convince the losers to adopt the moral values of the winners. Whether prosecutions are the most effective way to communicate condemnation or create social cohesion probably depends on the atrocity and society in question. It may be, as Nino argues, that "[tlhe formation of a social consciousness against human rights abuses depends more on the exposure of the atrocities and on the clear condemnation of them than on the number of people actually punished."299 If so, extensive coverage of past atrocities through a truth commission might be the most effective way to develop a new moral consensus. Truth commissions, like prosecutions, distinguish right from wrong, remind us that we live in a moral universe, and 293. See id.at 299. 294. Id. 295. NINO, supranote 16, atix. 296. OSIEL, supra note 34, at208. 297. GARLAND, supranote 267, at 52. 298. For example, critics claim that the outcomes of the prosecutions in Cambodia of Khmer Rouge leaders, if they are actually undertaken, "will be crafted to suit the political needs of Mr. Hun Sen (the Cambodian Prime Minister] rather than to examine the deaths of more than amillion people," Mydans, supranote 21. 299. Nino, supra note 22, at 2630. 2002 / A Frameworkfor UnderstandingTransitionalJustice communicate public disgust for the perpetrators' actions. As an alternative, some scholars have argued for civil liability as an important tool for establishing the full factual context in which the perpetrators committed their crimes ...thereby ... enhancing the prospects that the victims will have their suffering brought to the attention of the wider community and that a definitive, historically accurate account of the atrocities will be provided."43 Perhaps in ideologically divided societies, non-prosecution alternatives, which allow for public debate about moral standards, would be more effective in developing a public commitment to human rights values than prosecutions, which simply impose norms on society.3 1 On the other hand, perhaps prosecutions, by foreclosing debate about the content of norms and identifying certain atrocities as unequivocally wrong, would be more effective in reshaping public attitudes. In the end, the choice between prosecutions and non-prosecution alternatives is not simply about how best to express condemnation for mass atrocities. It also concerns the nature ofa society. The mechanism a society uses to communicate reprobation, and the nature of the message it sends, will shape moral norms and social cohesion. As Garland has said, "the ways in which we punish, and the ways in which we represent that action to ourselves, makes [sic] a difference to the way we are."30 2 III. A NEw FRAMEWORK FOR TRANSITIONAL JUSTICE A. PickingGoals Whether prosecutions are the optimal form of transitional justice depends on the goals we choose for this "justice system." A retribucivist will want to maximize the extent to which offenders are given their "just deserts." Prosecutions, which can produce stiff sentences, are the best method for meting out retribution. Although prosecutions cannot achieve true retributive justice in the wake of mass atrocity, only prosecutions approach the goal of punishing human rights abusers as much as they deserve. If our goal is deterrence, on the other hand, the yardstick for success is the prevention of future atrocities. It is plausible that prosecution has a marginal deterrence advantage over alternative post hoc mechanisms, such as truth commissions. But prosecutions are ineffective deterrents compared to other political and diplomatic tools, such as economic sanctions, diplomatic isolation, or militaty intervention. Under rehabilitative or communicative theories, the ob- 300. Murphy, supra note 17, at 48. 301. Some scholars suggest that trials themselves are a form of debate about norms. Ostel, forexampie, has suggested that although trials in bitterly divided societies cannot produce social solidanty in the Durkheimian sense of common moral values, they can foster public dialogue among da*ergent groups, which itself is a form of liberal social solidarity. Osius, suprw note 34. at 208, 283. While it may be true that trials have a discursive element, the point ofcriminalizing conduct is to establish that such behavior is primafade impermissible. 302. GARtS, supra note 267, at 276. HarvardHumanRightsJournal / Vol. 15 ject is to reshape moral norms. Prosecutions have no clear advantages over other mechanisms in this regard; what is most effective depends on cultural, economic, and political factors. Finally, if we aim for restorative justice, we will want to repair harm to victims and resolve societal conflicts. Alternative mechanisms like reparation schemes are preferable to prosecutions as a tool for compensating victims. In addition, non-prosecutorial mechanisms like truth commissions are generally better at fostering conflict resolution. Because prosecution is not always the most effective way to accomplish our desired goal, we should not blindly pursue prosecutions whenever they are possible. Nor should we assume that non-prosecution alternatives are necessarily morally inferior. We must first determine the goals of transitional justice, and then decide on the priority to be given to prosecution. Of course, we must also consider what options are realistic in the context of a particular transition. What is feasible will depend on economic and political conditions, social, cultural, and religious traditions, and the interests and involvement of other states.30 3 But the question of what is possible must follow the question of what is desirable. A full discussion of how to decide what goals to pursue and how best to pursue them is beyond the scope of this Article. Nevertheless, it may be useful to lay out a few factors to be considered in making this choice. First, economic, political, social, and cultural factors affect not only what form of transitional justice is possible in a particular society, but also what form of transitional justice is desirable. For example, a country's lack of financial resources may limit the number of prosecutions it can undertake or the form of reparations a truth commission can provide. The underlying economic structure may also affect what that society hopes transitional justice can achieve; a rehabilitative focus on ameliorating poverty may be especially appropriate where a conflict was the result of gross inequalities in wealth. Clearly, what works in one country may not work in another. As Richard Goldstone has noted, "ft]he correct approach to the past will depend upon a myriad of political, economic, and cultural factors which all operate and interact with each other.''3°4 The philosophical underpinnings of a culture will be particularly significant in influencing its choices about appropriate goals and mechanisms. For example, in some societies, morality and the law are conceptualized primarily in terms of Western liberal notions of individual rights and individual responsibilities. Because such a vision encourages an understanding of mass atrocity in terms of individual culpability, it also fosters a retributive approach to transitional justice and a focus on prosecution. By contrast, communitarian societies, because they place the needs ofsociety 303. For example, Landsman expresses the common belief that the viability of prosecutions depends on political context: "Nuremberg was ...an unusual case. The Nazis had been utterly defeated. They had unconditionally surrendered. Their society was in ruins and in need of reconstruction from the ground up." Landsman, supra note 5, at 87. 304. Goldstone, supranote 35, at 615-16. 2002 / A Frameworkfor UnderstandingTramsitionalJustice above those of the individual, may find restorative justice more attractive.- 11 While both individualists and communitarians will agree that atrocities should be prevented and condemned, societies that emphasize individual responsibility will want to structure transitional justice differently from those that more strongly value the community. Religion will also affect the choices a particular society makes about transitional justice.306 In many religions, forgiveness and reconciliation-as well as retribution and condemnation-have deep significance. But interpretations of these concepts will vary among different religions and between religious and non-religious societies.307 Some religions, as Stephan Landsman has noted, have concluded that "'an eye for an eye' is a sterile sort of justice that is far less satisfactory than breaking the cycle of violence or vengeance once and for all."308s Other religions envision a retributive God meting out retributive justice. The religious underpinnings of a society, like its individual or communitarian orientation, will shape both what a society will hope to achieve and what it can actually accomplish.30 9 Second, different patterns of human rights violations may require different approaches to transitional justice, just as goals for ordinary criminal justice vary depending on the particular crime and criminal. °10 While such distinctions are commonly recognized in the context of ordinary crime, in arguing about the best approach to transitional justice, scholars often lump together such diverse actions as genocide, war crimes, disappearances, torture, ethnic cleansing and communist repression. Awkward phrases like 305. The author is indebted to Catherine Mahnaz Amirfar for raising this point. Sir Catherine lahnaz Amirfar, Fighting the Liberal Myth of Prosecutions: Transitional Justice Redefined (unpublished manuscript, on file with author). 306. SeeDwyer, supra note 9, at 82-83 (arguing that reconciliation is grounded in religious nations. and that the concept that no one is without sin can provide inspiration for such efforts); SIuMvtn. ,LTa note 102, at 7 (arguing that while the term fbrgiveness 'has a religious ring in the ears of most modern westerners," forgiveness should "escape its religious captivity and enter the ranks of ordinary political virtues"). 307. A question worth further study is whether or not reconciliation is religiously contingent; is it an achievable goal fbr societies that do not share a religious commitment to forgiveness? 308. Landsman, supranote 5, at 87. 309. Religion and individualismlcommunitarianism might cut in conflicting directions. For example. while western liberalism's focus on individual culpability suggests a retributtve framework, an important concept in the Judeo-Christian tradition is forgiveness, which suggests a rstorauve justice pardhgm. One way of reconciling these conflicting approaches is by focusing on individual accountability and penance: restorative justice may be used when offenders are contrite, but retribution must be exacted when offenders are unwilling to acknowledge their culpability. The South African Truth and Reconalation Commission might be understood as embodying this mixed approach (though of course it was also heavily influenced by more communitarian African traditions). If offenders were willing publicly to acknowledge their crimes, they were eligible fbr amnesty If not, they could be, and often were, prmrcured. See Dugard, supra note 240, at 283. 310. While domestic penal goals may seem uniform as articulated through public polic; arguabl) criminal justice systems serve different purposes depending on the offenders and crimes Howemer. many domestic criminal justice systems have increasingly adopted a one-size-fits-all approach that tails ,Jrquately to distinguish among different types of crimes and criminals. In the United States, for example, the increasingly exclusive emphasis on retribution has been persuasively criticized as inappropriate for such groups as young offenders and such crimes as drug use. HarvardHumanRightsJournal / Vol. 15 "mass atrocities" and "widespread human right violations" encompass a wide range of different abuses. Much of the transitional justice literature, in focusing on when and how to "bring the perpetrators to justice," fails to ask if different patterns of abuse require different forms of "justice." Clearly, differences among crimes are significant in determining both what goals one should have for transitional justice, and what means one should use to achieve them. For example, in some cases of massive human rights violations, the line between perpetrators and victims is much clearer than in others. Administrative massacres are qualitatively different from fratricide. Retribution and prosecution may be more appropriate where there is unilinear responsibility, while restorative justice and truth commissions may be more helpful where there is long-standing social conflict and a measure ofguilt on all sides. Scholars must give more thought to how the nature of the atrocities themselves affects what a society will want to achieve through transitional justice. Finally, it is possible to pursue more than one goal simultaneously, and to use multiple mechanisms in pursuit of those goals. At the same time, we should honestly and openly confront whether and to what extent different goals and different mechanisms are contradictory or complementary. In some societies, retribution may be a precondition for social healing, while in others retributive justice will contribute to a cycle of vengeance and undermine efforts at reconciliation. Where goals are complementary, we need to structure the mechanisms of transitional justice to enhance the chances of achieving multiple aims. Where goals are contradictory, we must prioritize them and recognize that we may not be able to accomplish them all. In a world of limited resources, whatever choices we make will have opportunity costs. 3 1 1 B. PickingParadigms Given the difficulty of transposing the theoretical framework of domestic criminal justice to the transitional justice context, is ordinary crime really the best analogy for extraordinary evil? Hannah Arendt has written that it is impossible to punish radical evil adequately because "men are unable to forgive what they cannot punish... and are unable to punish what has turned out to be unforgivable .... Such offenses ... transcend the realm of human affairs and the potentialities of human power."312 Carlos Nino notes that, whether or not one accepts such powerlessness in the face of radical evil, Arendt's insight "points to the difficulty of responding to radical evil with the ordinary measures that are usually applied to common criminals." 1 ,3 Massive human rights violations, he argues, are "offenses against human 311. See Penrose, supra note 6, at 340 (arguing that "the millions" invested in the ICTY might have been better spent on refugee assistance). 312. HANNAH ARENDT, THE HubiAN CONDITION 241 (1958). 313. NINO, supranote16, atviii. 2002 / A Frameworkfor UnderstandingTransitionaIJustice dignity so widespread, persistent, and organized that normal moral assessment seems inappropriate. 314 Yet at the root of the preference for prosecution in transitional justice is the belief that normal moral assessment, expressed through the prosecution and punishment of individuals under the criminal law, is possible. The choice ofgoals will affect whether a paradigm focusing on individuals makes sense. If one desires retribution, collective guilt will be inappropriate, for it is individuals who murder, rape, and torture.31 As Karl Jaspers said at the end of World War II, "[i]t is nonsensical to charge a whole people with a crime. The criminal is always the individual."316 If, however, one adopts a non-retributive approach, a focus on individual wrongdoing will be less useful. In the context of transitional justice, deterrence, rehabilitation, restorative justice, and communication/condemnation/socia solidarity are primarily about societies, not individuals. Deterrence, while sometimes aimed at a few cynical leaders, is more often accomplished by creating credible threats that will keep a society from descending into mass violence. Rehabilitation is an attempt to develop democratic institutions that will reduce the risk of future violence, or change a society's moral values. Restorative justice seeks to create sufficient national reconciliation for the society to move forward; forgiveness or mutual understanding among specific victims and perpetrators is a helpful but not indispensable step towards this goal." And condemnation focuses on the reinforcement of a society's moral norms and the creation ofsocial solidarity. The criminal law paradigm fits neatly with the goal of retribution.318 But the analogy between extraordinary evil and ordinary crime is less compelling where one seeks to achieve one of the other goals, which appear not just in criminal justice, but in other contexts as well. Societies seek to prevent traffic fatalities by encouraging seatbelt use, and to contain infectious disease through vaccines. The ill are treated and rehabilitated; injuries are repaired through civil liability and insurance. Neighbors and loved ones reconcile by apologizing or by hashing things out. And societies reinforce moral norms and social solidarity through school curricula, national anthems, and the bully pulpit. The paradigms a society chooses in confronting particular social problems both reflect its goals and shape its understanding of how to solve these problems. For example, in domestic debates about gun violence, a focus on 314. atvii. 315. There can, nevertheless, be collective responsibility. Jaspers writes that "puhnical guilt" involves liability fir "the consequences ofthe deeds ofthe state whose power govems me and under wvile Urdt I live. Everybody is co-responsible fbr the wvay he is governed." JAsPERs, iqra note 110. at 159. Howevcr. political guilt does not mean "the criminal and the moral guilt of every single citizen for crimes committed in the name ofthe state." Id at 160. 316. Id at 163. 317. Note that the reparative prong of restorative jusuce is focused primarily on 1nJ1 ,duals 318. As the events of September 11, 2001 demonstrate, a '%-ar par-dAgm" tan also ucommixiate retributive impulses. HarvardHuman RightsJournal / Vol. 15 punishing individuals who use guns to commit murders, rapes, and robberies, will foster the understanding that gun violence is primarily a crime problem that should be addressed by prosecuting firearms offenders. If the goal is prevention, however, gun violence might be seen primarily as a regulatory problem, which should be handled by imposing controls on the manufacture and sale of firearms. Similarly, if the approach to mass atrocity is retribution, one will look to the criminal law and prosecutions. But if the approach is prevention, one might be more concerned about arms sales and oil embargoes, economic development and education programs. Adopting a framework that is non-retributive does not necessarily preclude individual accountability through the criminal law. Regulatory gun control is not incompatible with the punishment of those who use guns. Similarly, even when traffic fatalities are framed as a public safety issue, governments do not simply regulate car manufacturers and conduct public education campaigns on seatbelts. Governments also hold drunken, reckless, and speeding drivers individually responsible. In thinking about extraordinary evil, then, we should consider analogies other than ordinary crime. Such alternate paradigms might be particularly useful where the desired goal is non-retributive and the focus is on societies rather than individuals. For example, natural or humanitarian disasters, which like mass atrocities are wide-reaching catastrophes, could provide a model for how to inoculate societies against such horrors and to ameliorate the harm they cause. Because such alternative analogies shift the focus from specific perpetrators to society as a whole, they are arguably better suited to an understanding of the larger context within which widespread human rights abuses occur. Such analogies direct our attention to rebuilding societies by developing shared histories, establishing democratic institutions, or ensuring greater economic and political equality. And more importantly, such analogies encourage us to focus on preventing atrocities before they occur. Instead of simply concentrating on prosecuting past violations, we must, as Nino has noted, fulfill our "duty to safeguard human rights and to prevent future violations.' 319 A conceptual framework other than ordinary crime might help to highlight these priorities. In the end, there may be no better analogy to extraordinary evil than ordinary crime. Extraordinary evil, after all, is composed of ordinary crimesassaults and thefts, rapes and murders. If we adopt the ordinary crime analogy, however, we should recognize that this framework will inevitably shift our attention to retribution, and thus to prosecution. We must accept that even when the goals of domestic criminal and transitional justice are the same, the best method of achieving those goals may be different. Some of what prosecution can achieve in domestic criminal justice may be better accomplished in transitional justice through alternative mechanisms. After 319. NINO, supranote 16, at 188. 2002 / A Frameworkfor UndostandingTransitionalJustice 97 all, "ordinary crime" is only an analogy to "extraordinary evil." And analogies do not always hold, because analogies, by definition, compare different things.