Amnesty on trial: impunity, accountability, and the norms of international law1 This paper is a much revised and expanded version of the first of three Oliver Smithies Lectures delivered at Balliol College, University of Oxford, in May 2007.

Amnesty on trial: impunity, accountability, and the norms of international law1

Max Pensky*

Department of Philosophy, Binghamton University, USA


An emerging consensus regards domestic amnesties for international crimes as generally inconsistent with international law. This legal consensus rests on a norm against impunity: the chief role of international criminal law, and of the fledgling International Criminal Court (ICC), is to end impunity for violators of the worst of criminal acts. But the anti-impunity norm, and the anti-amnesty consensus that has arisen from it, now face serious difficulties. The ICC's role in the ongoing conflict in Northern Uganda illustrates the deadlock that has now emerged between countries wishing to retain the power to use domestic politics and criminal law as tools for negotiation with current or outgoing perpetrators, on the one side, and the ICC's determination to apply a consistent international anti-impunity norm, on the other. The paper argues that the anti-impunity norm itself is based on a narrowly retributivist conception of criminal justice. A broader norm for democratic accountability, by contrast, would continue to prefer prosecutions over amnesties in international law, less for the opportunity for deserved retribution for perpetrators than for the public enactment of the deliberative procedures associated with the rule of law.

Keywords: amnesty; impunity; international law; international criminal court

*Correspondence to: Max Pensky, Department of Philosophy, Binghamton University, USA. Email:

Citation: Ethics & Global Politics, Vol. 1, No. 1-2, 2008. DOI: 10.3402/egp.v1i1.1816
© 2008 M Pensky. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License (, permiting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

International criminal law stands at a crossroads. Its chief institutional embodiment, the International Criminal Court (ICC), can no longer postpone the determination of a basic dilemma. Will the norms of international criminal law continue to resemble those of domestic law, but transferred beyond domestic jurisdiction to transnational courts where international relations will compete with criminal justice for influence and determinacy? Or, conversely, will international criminal law embody and enforce justice and accountability norms more familiar from the global struggle for human rights and dignity? The former option may risk irrelevance of institutional innovations that have come about at tremendous cost and continue to hold tremendous potential. The latter option may demand a fundamental realignment of our prevailing conceptions of the nature and uses of the criminal law, perhaps even requiring that one of the most basic norms of criminal justice—sanctioning criminal behavior and meting out punishment to deserving perpetrators—lose much of its primacy. The attitude of the ICC to legal amnesties granted to perpetrators of international crimes is the concrete context in which this more basic choice is now presented. In this paper I will attempt to describe the range of issues involved, and indicate why the latter of these two options will emerge as the more attractive.

The argument will proceed in six steps. In the paper's first section, I describe the current situation of the ICC's involvement in the ongoing civil war in Northern Uganda, where the ICC Prosecutor's refusal to withdraw an indictment of rebel leader Joseph Kony and honor the Ugandan government's offer of amnesty now stands for better or worse as the chief obstacle to a comprehensive peace settlement for the region. In the paper's second section, I discuss how the ICC has come to incorporate into its core mission not just the establishment of a new class of international crimes but a legal and political norm against impunity: a duty to prosecute and punish violators of such crimes, whether by taking its own action or by influencing domestic courts to do so. At the heart of that anti-impunity norm, I argue, is the claim that domestic amnesties for international crimes are to be understood as incompatible with international criminal law; in the strongest reading, even as signals by domestic states of their unwillingness to discharge their duty to prosecute and hence their violation of the anti-impunity norm itself. The paper's third section analyses the function of the domestic amnesty and the international response to it beyond their status as legal instruments. Whatever the ICC and its Office of the Prosecutor (OTP) may wish, I argue that policy decisions over enforcement of an anti-amnesty norm transcend the narrower confines of criminal law in their implications and their effects. Domestic amnesties for international crimes, and the effort to prohibit or discourage them, are at heart part of a political debate about the status and extent of traditional national-state sovereignty in a rapidly globalizing political order. In the fourth section I discuss how the more general political–legal anti-amnesty norm ramifies specifically for the International Criminal Court; the paper's fifth section then attempts an evaluation of the arguments, pro and con, regarding the role of domestic amnesties given the results of the foregoing analysis. In the sixth section I conclude that the anti-amnesty norm cannot plausibly depend on merely deontological retributivist arguments for its defense—and yet these appear to be all that the norm against impunity provides. I propose that the anti-amnesty norm requires a broader consequentialist defense, and this defense is to be found only by a broad reconceptualization of the ends that the norm itself seeks. With a proper understanding of what amnesty is, a very broad conception of democratic accountability, rather than a narrow and retributivist conception of legal punishment, ought to be taken as the goal that implementation and enforcement of the anti-amnesty norm pursues.


By the Spring of 2008, it appeared that an end to Africa's longest-running conflict, a hideous and brutal insurgency in Northern Uganda pitting the Lord's Resistance Army (LRA) against the Ugandan Government and the Ugandan People's Defense Forces (UPDF) might finally be within grasp. Painfully slow and complex negotiations spanning over 19 months in Juba, Sudan, between representatives of the LRA and the Ugandan Government had yielded a tentative agreement on one of the last remaining sticking points blocking a comprehensive peace agreement. The February agreement specified that severe crimes committed by LRA rebels during the war, and not covered by amnesty legislation passed by the Ugandan government eight years earlier, would be tried by a newly created ‘special division’ of the Ugandan High Court.2

Those severe crimes certainly are numerous. While the LRA has never been an especially large rebel group, what it lacks in size it has more than made up for in viciousness, destructiveness, and persistence. Its leader, Joseph Kony, emerged in the military turmoil in Northern Uganda in the early 1990s, a follower of an indigenous spiritual movement pledging to protect the Acholi people from a legacy of abuse and discrimination by the Ugandan government.3 Kony proclaimed himself a messianic liberator, an indestructible spirit medium and a flaming sword of God's justice on earth, immune from government bullets and entitled to an unlimited number of Acholi children to run his revolution.4

How Kony proposed to protect the Acholi by murdering and tormenting them has never been entirely clear. Attempts to examine his political convictions according to familiar Western standards have proved frustrating, since those convictions appear to be an unsorted amalgam of indigenous spirituality and highly idiosyncratic appropriations of Christian doctrine.5 Kony seems impervious to familiar incentives that have led to successful peace negotiations in other national contexts. He has no interest in wealth or indeed any of the kind of inducements that have coaxed other intransigent democratic spoilers into more or less pampered exile. Barring the ultimate victory of the LRA in Uganda, the overthrow of the Museveni government, and the realization of God's commandments on earth via the person of Mr. Kony himself, the LRA up until recently has apparently been satisfied to maintain its bases in South Sudan and the north of the Democratic Republic of Congo and make life for the Acholi a living hell.6

In this at least the LRA has been successful. Its members (virtually all Acholi themselves, as is Kony) have managed to kill tens of thousands of Acholi civilians—by shooting, stabbing, clubbing and beating them to death—and have mutilated uncounted thousands more in a reign of terror on the Acholi villages they claim to protect. The hacking off of lips has become a signature. The LRA has made a specialization in the kidnapping, terrorization and cooptation of children, whom it forces to fight, or serve as pack animals or sexual slaves, often forcing children to commit brutal crimes themselves. Beyond its dedication to physical violence, the LRA's terror tactics have also displaced nearly two million people, forcing virtually the entire Acholi population of Northern Uganda to abandon their villages. In this massive internal refugee crisis the LRA has had the help of the Museveni government, which has used the insurgency as a rationale to forcibly resettle the Acholi into a chain of overcrowded internal displacement camps, where miserable conditions and lack of basic services ensure a wretched existence, rampant disease, and the complete dependence of the Acholi internees on donated food from foreign states and international relief organizations.

For the government of Ugandan president Yoweri Museveni—a former rebel leader himself who came to power in a successful military coup in 1986 that overthrew the Acholi Tito Okello—the crisis in Northern Uganda has resolutely refused to go away. A string of military operations has resulted in little more than massive civilian casualties, and atrocities and human rights violations perpetrated by Ugandan soldiers. Intermittent attempts at peace negotiations repeatedly broke down largely due to a lack of political will to end the violence on both sides.

In 1998, a newly established Acholi civil society organization, the Acholi Religious Leader's Peace Initiative, began an ultimately successful campaign to lobby the Museveni government to offer an amnesty to LRA soldiers as part of an incentive package to enter more serious peace negotiations. The Uganda Amnesty Act, passed in 2000, guaranteed relief from all domestic prosecution for anyone—LRA member or UPDF soldier—who had committed any criminal act in association with LRA activities, on the condition that the perpetrator make a formal (and largely pro forma) admission to an amnesty commission and renounce any further violence.7

By 2006, with the agricultural base of Northern Uganda in shambles, and the LRA extending their attacks beyond Acholiland, a new round of peace negotiations opened in Juba, Sudan between representatives of the LRA and the Museveni government.8 Opening those talks, President Museveni reaffirmed his generally tepid commitment to the terms of the 2000 Amnesty Act and offered various guarantees of Kony's safety, including an explicit promise personally to shield Kony from any possible international prosecution—to offer him a full and unconditional amnesty for acts that constituted international crimes. In August of 2006, a bilateral ceasefire agreement signed at Juba appeared to many Ugandans and international observers to be the definitive moment where Africa's longest conflict would finally end. In Acholiland, white flags began to appear on rooftops and car aerials, celebrating the ceasefire and its promise of a permanent end to violence. The number and rate of applications by former LRA members for amnesty spiked. People began to speak tentatively about reconciliation mechanisms, and traditional healing ceremonies; about quitting the hated internal displacement camps, returning to village life, planting crops and sending their children to school.

There was just one problem.

In December of 2003, nearly four years prior to the 2006 ceasefire agreement, President Museveni—operating with no doubt complex motives—had made a formal referral of the situation in Northern Uganda to the ICC. The ICC had officially entered into existence only in the previous year, in July of 2002, when the 60th sovereign nation had ratified the Rome Statute establishing the ICC as a court specifically dedicated to investigate, indict and prosecute individuals suspected of international crimes as specified in the Statute's opening articles—genocide, crimes against humanity, and war crimes9 —in cases where the state or states that would otherwise maintain jurisdiction were unwilling or genuinely unable to investigate and prosecute on their own.10

Museveni's referral set in motion a chain of events that quickly exceeded his or any state representative's control, a situation that the President himself came very quickly to regret. Once a referral is received by the Court, the state party itself no longer has any legal authority to rescind or modify it. (Museveni himself, who has repeatedly demanded that the Office of the Prosecutor withdraw its warrants and has been repeatedly refused, may not have entirely understood this condition when he submitted the referral.) The Statute also includes a provision (Article 53 Rome Statute) for the Prosecutor to decline to act on a referral if in the opinion of the OTP doing so is not ‘in the interest of justice,’ a deliberately indeterminate clause granting the OTP a degree of prosecutorial discretion.11

A little over a year after Museveni's referral, in January of 2004, the OTP announced the initiation of a formal investigation of the situation in Northern Uganda. Appearing at a press conference at Kampala, the ICC's Chief Prosecutor Luis Moreno de Ocampo announced the launch of its first case.12 I said earlier that the difficulty negotiating with Joseph Kony arises in large part from his apparent imperviousness to the familiar panoply of incentives, whether positive or negative, to go away. But the ICC's investigation and the threat of indictment and prosecution proved that there was one crucial exception to this. Kony was and remains extremely unwilling to submit himself to prosecution, whether by a Ugandan or any other court. Well prior to the 2006 ceasefire, Kony had made it clear that any lasting peace settlement would require guarantees of his own personal immunity from prosecution in any form—a guarantee that Museveni, once his referral had been accepted by the ICC, was no longer able to make. It quickly became clear that the threat of indictment and prosecution by the ICC was the chief remaining stumbling block to a negotiated end to a horrific conflict.13

In October 2005 the ICC unsealed a criminal indictment of Joseph Kony and four of his top military commanders, charging them with a range of crimes against humanity and war crimes, and authorizing their arrest and extradition to the Hague to stand trial. President Museveni immediately made it clear that he had no intention of acting to enforce the international arrest warrant, and in any case Kony was by then in residence either in southern Sudan (which is not a signatory of the Rome Statute) or the Democratic Republic of the Congo (which is, but has no evident interest or capacity in acting to enforce the warrant either.) Despite the current ceasefire, Kony remains at large, and regularly breaks its terms with kidnappings, organized assaults, and murders of civilians. Chief Prosecutor Ocampo has been clear that he has no intention to withdraw Kony's indictment, and no interest (or authority) in negotiating with Kony or any other LRA representative.14 And caught between Kony and the Court, the Acholi remain in their camps, gnawing on donated high-protein biscuits, waving away the flies, and waiting to go home.


The ICC's predicament in Northern Uganda is the most visible, public face of a deeper crisis over the Court's normative function and prospects. For its framers and advocates, the ICC was to have played a key moment in the emergence of a new set of legal, political, and even moral norms regarding the worst of crimes. Part of this emergent set of norms is the crystallizing legal norm expressing the duty of sovereign nation-states to prosecute serious crimes that are under their jurisdiction, and the authority of the international community of states to take limited measures to enforce criminal law in cases where state parties themselves were unwilling or unable to do so. This anti-impunity norm15 contradicts a long tradition that tacitly accepted, as part of Realpolitik, that control of the domestic system of criminal law is an integral feature of national sovereignty, and therefore that states using their own discretion on whether to investigate, prosecute and punish offenders of the worst of crimes are merely using the reason of state that such sovereignty entails. Despite its near-total lack of enforcement power, the ICC's Uganda predicament demonstrates powerfully that its normative force in pushing this norm has is considerable—and not entirely under its own control.16

The anti-impunity norm generated a remarkable reversal in the attitude toward national amnesties. By the second half of the twentieth century, domestic amnesties were firmly entrenched as standard political tools.17 In many cases amnesties for political prisoners were bargaining chips that oppressive regimes could offer the international community; Amnesty International named itself on this basis. In other cases, such as the notorious ‘self-amnesties’ of political and military leaders and their henchmen in countries such as Chile and Argentina, amnesty operated blatantly as a means to shield perpetrators from the legal consequences of their own crimes. In still others, blanket amnesties—that is, amnesties covering entire groups of persons unconditionally—were efforts to circumvent domestic criminal law, whether as parts of broader peace negotiations or international treaties, or as executive decisions simply to cut short any legal consequences of an ended or ongoing conflict. With the advent of the anti-impunity norm, the overall attitude toward amnesties in the international legal community underwent a remarkable 180 degree shift. Rather than a yearned-for release of innocents from captivity, domestic amnesties for international crimes became the poster child for the most egregious forms of impunity. ‘Amnesty, the symbol of freedom,’ as Louis Joinet reconstructed this transformation to the United Nations, ‘was more and more seen as a kind of “insurance on impunity” with the emergence, then proliferation, of “self-amnesty” laws proclaimed by declining military dictatorships anxious to arrange their own impunity while there was still time.’18 Now saddled with a less-than-descriptive name, Amnesty International became the first influential human rights NGO to articulate an anti-amnesty position favoring domestic and/or international prosecution as the preferred mechanism for addressing past atrocities in post-conflict states, and has been among the most vocal critics of national amnesty policies—including the innovative amnesty committee of South Africa's Truth and Reconciliation Commission—ever since.

Black's Law Dictionary defines amnesty as ‘a pardon extended by the government to a group or class of persons, usually for a political offense; the act of a sovereign power officially forgiving certain classes of persons who are subject to trial but have not yet been convicted […] Unlike an ordinary pardon, amnesty is usually addressed to crimes against state sovereignty—that is, to political offenses with respect to which forgiveness is deemed more expedient for the public welfare than prosecution and punishment.’19 This is in fact a highly controversial definition, insofar as it conflates amnesty (the granting of immunity from prosecution) with pardon (the relief from sanctioned retribution associated with a conviction). In this essay's last section I hope to show why this distinction makes a potentially enormous difference. For the purposes of this study, I will consistently regard amnesty as an official act of government broadly construed, that is, an executive order and/or legislative act, or a treaty or peace negotiation conducted with duly appointed representatives of the state authority, with or without any popular legitimacy via referendum or plebiscite.20 The emphasis on the official nature of amnesty is intended to make a rough distinction (which inevitably will get messier in individual cases) between official acts of government as de jure amnesties, on one side, and on the other the large range of de facto amnesties covered, say, under prosecutorial discretion, or an inefficient or corrupt criminal justice system, or legislation intended to shield nationals from international indictments and prosecutions.21

This emphasis on amnesties as official state acts means to focus attention on a basic open question about the legal status of amnesties, which after all can range from the relatively mild and uncontroversial—say, tax amnesties with straightforwardly pragmatic justifications—through much more contested forms of official amnesties for illegal aliens, all the way to highly morally, legally and politically charged amnesties for international crimes. Amnesty is a sovereign act in these cases, demonstrating the (usually executive) power to effect significant alterations in the normal application and range of the domestic criminal law in order to bring about a politically desirable effect.22

There is of course a very broad range of official acts that can fall under this definition, and historically amnesties have varied considerably, both in their terms and instruments and in their compatibility with basic expectations of criminal justice. While self-amnesties of the most egregious political players and blanket amnesties for their henchmen and enforcers are the most visible at one end of a scale of acceptability, surely there are other amnesties that are far more complex. Amnesties can be made conditional on all sorts of desirable conditions and outcomes; they can be used as negotiating carrots for the peaceful exit of dictators and their cronies; they can be offered as incentives for entering into peace negotiations or securing temporary cease-fires. Most famously, of course, they can, in the case of the South African Truth and Reconciliation Commission, be largely removed from the hands of the executive and moved to the work of an extraordinary and temporary body that is neither entirely political nor legal, where individual petitioners receive amnesty for a narrowly defined set of criminal acts only conditionally on their satisfaction of numerous and stringent conditions, most significantly their willingness to provide testimony concerning the nature of their acts, their place in a chain of command, their subjective motivations, the disposition and whereabouts of their accomplices and victims, and so on.23

As integrated parts of a broader effort at transitional justice involving various justice mechanisms meant to offset a lacking prosecution, such conditional, individualized amnesties certainly challenge an inflexible anti-amnesty norm that equates amnesty with impunity at every turn. On the other hand, there are some compelling reasons why the South African experiment with commission-based conditional amnesties has not really been repeated elsewhere. Even the best-intentioned and most practically useful amnesty policy compels a legally sanctioned act of forgetting or omission, and in this case the amnesty's presumptive injustice must be counterbalanced by its potential to avoid foreseeable harms that convincingly outweigh the injustice of the amnesty itself.24 An open question remains whether a serious and good-faith effort to replicate the South African TRC model, offering amnesty not just for small fish but for some rather large fish as well, in return for admission of wrongdoing and testimony before a specially empanelled commission, would or would not run afoul of the anti-impunity norm and mission of the ICC (whose jurisdiction extends temporally only as far as the coming-into-effect of the Rome Statute in 2002). What remains true, this complex question notwithstanding, is that South Africa's complicated experiment in balancing the needs of democratic peace with criminal justice, whatever else we may make of it, was certainly also a traditional application of the sovereign power of a nation-state, with a new ruling political party eager to establish its sovereignty and legitimacy, to authorize quite dramatic deviations from the ordinary course of the domestic rule of law, in pursuit of some other desired political outcome.25

Does this imply that domestic amnesties for international crimes are in principle contrary to the rule of law, whether domestic or international? No clear definitive answer exists for this as a matter of settled law, but in an important and basic sense, such a deviation or suspension of the rule of law, with specific political outcomes in mind, is just what amnesties are intended to be. Indeed for our purposes we can stipulate that domestic amnesties for international crimes are presumptive departures from the rule of law, which is precisely what the emergent anti-amnesty norm in question expresses. The question is whether such an anti-amnesty norm ought to gain the status of settled international law, and, if that were to happen, whether further Uganda-like impasses are to be expected.26

The presumptive moral harm of the domestic amnesty arises in the first instance in the context of retributive justice. As relief from prosecution, amnesties necessarily also relieve from legal sanctions those who (may) deserve them. A positive retributivist account of punishment regards the guilt of the perpetrator as sufficient reason to punish, such that failure to punish is itself an injustice. The injustice of the criminal act is doubled by the injustice of the state authority in failing to fulfill its retributive duties.27

As we'll see in a moment, positive retributivism, interpreted as the state's duty to prosecute—and hence justifying contractual arrangements for international prosecution in the event of the state's failure of this duty—is the most direct link between the narrower category of criminal justice and the broader and deeper moral commitments that underlie the anti-amnesty norm, in the form of a normative demand to end impunity.28 A less direct but more powerful argument, as I will again explore in more detail in a moment, points out that the domestic amnesty violates a norm of legal, political and moral equality. To the set of all those granted amnesties, who are effectively beyond law's reach for acts that otherwise would warrant legal attention, corresponds another set of persons—their victims—whom the amnesty has deprived of effective legal remedy, a basic right. And the deprivation of such a right in the interests of public safety or security would appear to require consequentialist arguments of extraordinary weight. Indeed this very point—the impermissibility of the deprivation of a class of citizens of their rights to legal remedies for harms done by them—was the basis for the most serious challenge to the amnesty clause of the 1994 South African Interim Constitution, in the famous Azanian People's Organization (AZAPO) case.29


Domestic amnesties are thus presumptive injustices in the name of a domestic good—and in this sense the good in question, the provision of domestic security, cannot be analyzed meaningfully in the absence of the underlying function of such amnesties to reassert sovereign state control over the system of domestic criminal law. Under current international conditions, at the end of a 15-year-long experiment with border-crossing courts and jurisdictions, domestic amnesties have lost whatever shred of innocence they may have retained. Whether blatant efforts at impunity or as parts of broader peace negotiations or reconciliation approaches, domestic amnesties for international crimes are always such assertions of traditional sovereign power. In the ‘Schmittian’ sense, one could say, the executive amnesty is an expression of the state's sovereign power to declare an exception to the normal parameters of the rule of law, to break the rule of law requiring like cases to be treated alike, as a way of furthering some other political end. For this reason, the question of the legal status of the domestic amnesty in international law is always tangled with the larger policy context in which the amnesty happens—that's trivially true. But it's also true that if we accept even in principle a degree of validity to such amnesties, we must also grant some grudging acceptance to the principle that deviations from the rule of law are indeed the prerogative of the sovereign state power. Inevitably, judgments concerning the legitimacy of domestic amnesties for international crimes—even the question of whether international courts will see themselves in any way affected by them—are also judgments about the nature and extent of national sovereignty.30 As a visible symbol of intact sovereign power, amnesties are for better or worse super-charged in this way, making the collision between the international community and recalcitrant, beleaguered and fragile states all the more fraught, since the latter will always have much more at stake in the resolution of such questions of sovereignty than the international community does.

With this in mind it becomes clear that the narrow legal question at hand—the legitimacy and standing of the domestic amnesty for international crimes, and the proper attitude of international courts toward them—is really not so narrow. In fact it is ultimately a deeply political question about the nature and extent of a very traditional conception of the Westphalian sovereign state under pressure from an emergent global criminal law regime. This is the political background for the emergence of a norm within the international criminal law community that declares domestic amnesties to be contrary to international law. International courts and other transnational legal bodies must oppose such amnesties as part of a larger battle to end state-sponsored impunity for atrocities. Indeed it was just this commitment to end state-sponsored impunity that produced the ICC. The ICC's refusal to see itself bound by any domestic amnesty acts issued to Joseph Kony by the government of Uganda is an effort to enforce an emergent legal norm that regards amnesties as contrary to international law because of the function of amnesties to generate situations of impunity, but also because of the status of such amnesties as domestic, as roadblocks to the success of a global regime of international criminal law.

As I describe the emergence of this anti-amnesty norm,31 it's important to be clear on the category of emergence. The status of the domestic amnesty for international crimes is not at present a matter of settled international law.32 It is emergent in the sense that it is arguably in the process of, or on the way toward, such settled law status. The lack of extra-territorial applicability of the domestic amnesty is reflected in a growing body of legal opinions and analysis by scholars and ‘publicists’ in law textbooks and academic journals, and thus the anti-amnesty norm has begun to guide the procedures and conclusions of a number of national, regional, and transnational courts and tribunals.33 The norm does not, however, appear in the Rome Statute of the ICC.

Beginning in the late 1980s and early 1990s, a number of influential legal scholars and jurists grew increasingly suspicious of the claims that national processes of post-conflict reconciliation could justify state practices that permitted egregious breaches of states’ legal (and moral) obligations to bring minimal standards of justice to perpetrators of atrocities. Over the past 15 years in particular a small but rapidly growing body of rulings by domestic, regional, and transnational courts has come increasingly to assert, in ever bolder terms, the status of the domestic amnesty for international crimes as contrary to international law.34 As Diane Orentlicher has recently written, the effects of self-amnesty in Latin and South America, above all in Chile and Argentina, were foundational for a generation of jurists and legal scholars who saw international limits on amnesties as one means among many others to pressure states to combat this sort of officially sanctioned impunity.35 And in that Latin and South American context, where former dictators and their militaries remained as standing credible threats to fragile democracies, impunity was not just a legal insult but a continuous and ultimately unbearable and untenable political dilemma. On the one hand, acceptance of blanket self-amnesties was virtually the only fragile political means to keep the old powers in check; on the other, ‘for the newly elected successor governments to honor nakedly self-serving claims of untouchability would betray the very principles they had pledged to restore and safeguard.’36

In this sense, the anti-impunity norm, and the rejection of amnesties it demanded, remained a constant even as the context in which it was deployed and defended changed yet again. Over the course of the 1990s, state amnesties were no longer always a matter of blatant and cynical misuse by entrenched oppressive regimes in Latin and South America, even though such amnesties remained the rule in Africa and elsewhere. In many other post-conflict countries, however, amnesty was employed not just as a means for shielding perpetrators from their legal nemesis, but as part of broader programs of post-conflict national reconciliation, most notably with the rapid and remarkable popularity of national truth and reconciliation commissions.37 The normative context for evaluating the justice of amnesties became deeply complex. And in this context, the underlying justification of the anti-impunity norm itself underwent a subtle shift as well. Now no longer just a battle cry against the dictators, the anti-impunity norm was in addition a more measured, principled objection to the very wide latitude that some nations (most notably South Africa) granted themselves in interpreting their own legal obligations under international law as they ‘settled accounts,’ and balanced the need for legal justice for perpetrators of atrocities against the need for broad measures of national reconciliation and peacemaking. Hence the anti-amnesty norm, not surprisingly, was also an emergent political norm, expressing the growing confidence in an international human rights consensus, and the hope that institutional experiments across the 1990s and into the new millennium would effectively enact that normative consensus, challenging the orthodoxy of the Westphalian conception of national legal sovereignty and contributing to a ‘good’ globalization dynamic.

The events of the 1990s also required the anti-amnesty norm to undergo a number of important refinements and qualifications, all reflecting the emergence of new and highly experimental mechanisms for addressing the needs of post-conflict justice in newly democratizing societies. The anti-amnesty norm came to include a sub-principle of selectivity of prosecution. That is, the norm's applicability in the context of post-conflict criminal justice has to be integrated into larger programs of disarmament, demobilization, and reintegration (DDR) which recognize that lower-level combatants, often conscripts, occasionally kidnapped, very often minors, have mitigated culpability. In such cases, prosecutions of the broad bottom of a pyramid of criminal culpability would be so burdensome, with such little potential rewards, that amnesties for small fish, whether entirely domestic or negotiated in peace agreements with international representation, pose no significant obstacle in international law, and indeed are a welcome part of a larger plan for transitional justice in a post-conflict state.38

The principle of selectivity derives from the Nuremberg-era idea of command responsibility that, in the context of the kinds of low-level chronic conflicts that international courts are likely to address in the decades to come, poses some serious difficulties. Still, for reasons both normative and pragmatic, the principle of selectivity has become an important rider to the anti-amnesty norm: always prosecute the big fish, and only prosecute the big fish. But drawing a legal bright line between the top perpetrators, whose amnesty would constitute an unacceptable state-sponsored impunity, and the small fish, whose amnesty may be not just tolerable but welcome as part of a broader program of national reconciliation, makes visible a problem internal to the anti-amnesty norm—the bigger the fish, the more potential leverage the domestic amnesty may have for the purposes of establishing and securing conditions of peace and security. Whether individual little fish are amnestied, prosecuted, or merely let alone is ultimately a matter of very limited practical importance in the effort to establish minimal conditions of security. But the greater the amount of power a perpetrator maintains, the larger his potential to act as a democratic spoiler, and the greater interest he will have in making cooperation contingent on an amnesty for himself and his closest allies. Therefore the stronger the perpetrator, the more the domestic government has to gain by the (low-cost) amnesty, and the more it has to lose by insisting on a (high-cost) prosecution.39

This implies that the selectivity condition placed on the anti-amnesty norm does little if anything to resolve the chief problem arising as international organizations pressure states to abandon amnesty processes. Indeed it may sharpen it. Like other experiments with transnational criminal law institutions over the course of the 1990s and 2000s, the ICC was intended to operate under a principle of complementarity to resolve just this problem.40 But before we move to the ICC itself as an institutional embodiment of the anti-amnesty norm, it's important to reconstruct those institutional experiments, and to be clear on how the Court, as the only one of them likely to survive into the twenty-first century, depends for its legitimacy on its principle of complementarity. Put into political practice, complementarity allows powerful and wealthy Western European states, which have pushed for the ICC's existence and continue to define its institutional rationale, to demand legal limits to state members’ sovereign control over their own domestic policy (as part of the states’ treaty obligations if they voluntarily ratify the Rome Statute), but at the same time to direct (poor, post-colonial, Southern) states to remain the presumptive first recourse for high-risk and high-cost prosecutions.

These transnational legal experiments, crucially, developed parallel with the experimental use of alternative mechanisms for post-conflict justice, most often in poor states emerging painfully from a long period of post-colonial or post-communist misrule. Truth commissions, memory politics, vetting of corrupt public officials, reconciliation programs, educational reform all came to supplement, and often compete with, the ‘standard model’ of criminal prosecution as transitional justice mechanisms. Complementarity—the international preference for domestic prosecution as the first among equals in these mechanisms—significantly shifts costs and risks from the normatively self-authorized international legal community to those states that can putatively afford prosecution least. As we review those international law institutions that arose during this period to address this situation, then, we need to understand that their ambivalent status—partly aid-like institutions — and partly rule-like institutions, is structural, not accidental.

The institutional experiments over the last 15 years directed at supplementing this presumption of domestic prosecution break down essentially into three.41 First, the 1990s witnessed a number of experiments with international or hybrid international–domestic courts: UN-sponsored criminal tribunals were set up in the former Yugoslavia and Rwanda, and hybrid courts (with components and personnel from both domestic and international law systems) in Sierra Leone, East Timor, and Cambodia.42 Second, the same period saw numerous efforts by sovereign states to extend their claims of legal jurisdiction dramatically by investigating and prosecuting non-nationals outside of their own territory. While the most famous of these experiments in ‘universal jurisdiction’ were the legal proceedings against former Chilean dictator Augusto Pinochet, arrested in London on a Spanish warrant, other (West European) countries, most notably Belgium, also made concerted attempts to transform international criminal law by asserting their duty and jurisdiction to prosecute international crimes according to an interpretation of customary international law in which violations of jus cogens, that is, acts that must be regarded as criminal by any civilized nation and hence are contrary to law regardless of any treaty obligation, implied that all states were under non-derogable obligations to prosecute them.43


Neither of these experiments seems likely to survive far into the twenty-first century.44 The ICC—the third great contemporary experiment in the institutionalization of the anti-impunity norm—was conceived largely as a means to correct the obvious failings of other juridical experiments and to solidify their gains in a permanent world criminal court. The anti-impunity norm takes center stage as the core ‘mission statement’ of the Court. The preamble of the Rome Statute reads in part that ‘the most serious crimes of concern to the international community must not go unpunished,’ and that ‘their effective prosecution must be ensured by taking measures at the national level and by enhancing international cooperation.’45 The ICC's very purpose is therefore ‘to put an end to impunity for the perpetrators of these crimes and thus to contribute to the prevention of such crimes,’ even as the very next sentence of the Statute insists ‘that it is the duty of every State to exercise its criminal jurisdiction over those responsible for international crimes.’46

One way to interpret this preamble language—the preferred way, certainly for many if not most of its over 100 treaty member states—is that the Court serves merely as a ‘backstop,’ a kind of insurance policy that states may wish to use in cases where their domestic duty to prosecute international crimes seems to pose insurmountable problems: where protracted violence has left the domestic criminal justice system disabled or corrupt, for instance, or where the size and complexity of investigations and prosecutions would overburden domestic capacities and resources.47 There is certainly another possible interpretation of the commitment to ‘enhance international cooperation,’ however, and that is that the ICC is an institutional embodiment of an international justice directive that works toward prohibiting states from permitting impunity for international crimes by erecting various negative incentives for doing so. The ICC therefore enforces an anti-impunity norm that state members might otherwise be inclined to ignore, whether in the name of simple expediency, or defensive reactions to perceived threats to national sovereignty, or some admixture of both.48 In this rather Trojan Horse reading, the ICC is less a resource that embattled states can draw on when necessity requires than a legal force meant to influence or even compel states to fulfill their international legal duties for prosecution; duties which state parties to the Rome Statute have presumptively voluntarily contracted to assume in the act of signing and ratifying the Statute, but also duties that exist, in customary international law, independent of the Statute and thus that apply also to non-signatory states.49

This second, more active sense of the ICC's commitment to the anti-impunity norm has the effect of communicating, in treaty language, the emergent principle that domestic amnesties for international crimes are in principle contrary to international law. And again, this anti-amnesty interpretation of the anti-impunity norm has both a weak, negative reading and a far stronger positive one. Negatively and weakly, the implications on domestic amnesties of the anti-impunity norm are simply that such domestic amnesties cannot be presumed to have extra-territorial effect. This seems uncontroversial. There is no special reason to think that a perpetrator of mass crimes who has received a state amnesty from his country of nationality should, for instance, expect that any other state should, absent any treaty, regard such an amnesty as having any legally binding effect on its own criminal justice system.50

But even this fairly modest, negative reading of the anti-amnesty norm would constitute a remarkable departure from a tacit code of international relations that has survived virtually intact into the end of the twentieth century. The effects of this break are visible, for instance, in the growing reluctance of ‘‘receiving’’ states to permit exiled dictators and their henchmen to take up residence in their territories free of any reasonable expectation of legal action against them. The anti-amnesty norm read as lack of extra-territorial effect has the indirect effect of lowering the negotiation value of exile-for-peace offers by raising the risk that exiled perpetrators will face legal proceedings in exile countries. This is perhaps an unintended effect, but a real and measurable one just the same, with ambiguous consequences for any state that wants to retain amnesty offers as a useful negotiating incentive.51

A positive and far stronger reading of the anti-amnesty norm becomes practically possible only with the advent of international legal bodies; initially, international and hybrid criminal tribunals but now specifically the ICC. This version contains the weaker claim that domestic amnesties for international crimes are in principle contrary to international law and thus have no extra-territorial validity. But in addition it draws the implication that international courts not only may regard such amnesties as without effect on them, but further that international courts must regard such amnesties as invalid and thus must recognize themselves under legal obligations, at least in principle, regarding those acts that the amnesties cover. In this positive sense, the amnesty itself serves as a trigger of international legal attention. If the amnesty is contrary to international law, this does not mean that the amnesty itself is a violation of the law: neither treaty-based nor customary international law contains any such provision.

If amnesties provide impunity, and impunity is equated to criminal injustice, this doesn't mean that amnesties are illegal any more than the failure to punish a crime is itself criminal. What it does mean, however, is that domestic amnesties for international crimes cannot in principle remain a matter of indifference to the international legal community. The amnestying state, in effect, has by its act signaled to the international criminal law community its intention to fail in its international obligations. This signal may in some instances be interpreted as a cry for help; in others as an act of defiance. But in either case, the anti-amnesty norm has the effect of transforming the status and function of the domestic amnesty for international crimes as a kind of international signal, drawing the attention of the international community, which, whether via states, or transnational courts, or non-governmental organizations, or cooperation amongst all three, asserts its own rights to make the acts amnestied into matters of international interest and possible legal action.52

These facts serve to sharpen the ICC's standing, like it or not, as a political body, and this emphasizes the conclusion reached in the preceding section: under the Court, the anti-amnesty norm is not just a legal but also a political norm, asserting the rights of a more confident international normative consensus regarding the minimum standards of criminal justice that members in good standing of the international community are expected to adhere to.53 But like all such political norms of global justice, the political effects of the pursuit and implementation of the anti-amnesty norm are complex, subtle, and beyond unilateral control. If it's to succeed in imposing conformity to the norm, the international legal community will require increased power and influence. And to the extent that it does have such power and influence sufficient to affect the internal decisions and policies of states, those states will be obliged to acknowledge the institutions imposing the anti-amnesty norm as new players in the field of international politics. Such acknowledgment may well take the form of normative integration, certainly—accepting the legitimacy and due influence of new players like the ICC as part and parcel of accepting the validity and binding nature of the new norm.

But as a mode of states’ reactions to changing international conditions, normative integration is only one option among many. It would be surprising if states, whatever their position on amnesty, didn't respond to a new influential player rationally and intelligently, by adapting various policy and law positions, domestically and externally, to accommodate the new normative and institutional reality in ways that harmonize with, or at least are not harmful to, the national interest. One way of thinking of the anti-amnesty norm, for example, is that, together with the ICC's stated principle of complementarity54 —the court only prosecutes when and if domestic prosecution is impossible or refused—it implies something like a negative externality, the imposition of an unfunded mandate on poor states by the international law community. The ICC, like the UN and other international bodies, respects national sovereignty insofar as it has a stated preference for domestic over international prosecution. The Rome Statute's Article 17 on admissibility establishes the world Court as a court of last resort in the event of the failure or refusal of domestic law.55 The norm against amnesty is a norm mandating investigation and prosecution, actions whose potentially very high political and financial costs are thus effectively shifted from the international community to the states themselves, which are often in poor positions to absorb such costs, especially when the potential benefits tend to be comparatively abstract and deferred. No wonder that many states, even signatories to the Rome Statute, are wary indeed of the ICC's attention, well aware that as the presumptively responsible parties for acts falling under the Statute's specifications of international crimes, they will themselves be asked to foot the bill and absorb the risks. In the Ugandan case, this externalization of cost and risk appears, potentially at least, to be dramatic. Withdrawing its warrants for Kony risks the Court loss of face and international legitimacy. Bad enough—but presumably stronger arguments than the threat of losses such as these would be necessary to justify the continuation of conflict, with the guarantee of loss of life and the misery of nearly two million people.

Conversely, it's equally possible that this same rational, intelligent response to a new international player might provoke what William Burke-White has described as an international version of moral hazard: the presence of an effective world court may act as an incentive for states to engage in riskier behavior regarding negotiations (or refusals to negotiate) with powerful perpetrators than they might otherwise permit themselves, precisely because they know that they can appeal to the backstop court to prosecute if they demonstrate their own unwillingness or inability to do so. (Burke-White speculates that this kind of moral hazard effect may be the best explanation for President Museveni's otherwise baffling use of the ICC to increase pressure on the LRA in a purely domestic strategic game.)56

The ICC's insistence that it is a legal body, intent only on enforcing the law and not interfering in domestic politics, is ultimately not credible. For better and for worse, its capacity to influence and in many cases even alter state parties’ behavior makes it an international agent. Its decisions, specifically the measures it is willing to take to enforce conformity with a putatively purely legal norm such as the opposition to domestic amnesties, are also political ones. What is less clear is what the ICC will do, and what in ought to do, if (as has happened it its very first referral) the political and legal strands of a situation are so entangled as to force the ICC to depart from its own self-understanding as a law body, and to make decisions that manifestly straddle the border between law and politics.


In the situation in Northern Uganda, it seems very strongly that the Court's hands are tied. But the legacy of that predicament must be a more serious and sustained examination of the full range of normative arguments regarding the status of the domestic amnesty, for the circumstances in Uganda are not especially strange, amnesties remain popular, and the Court will find itself in this or a similar situation again. It's worthwhile attempting to construct a kind of balance sheet of arguments pro and con. In what follows, I appeal to a standard distinction in normative ethics between consequentialist and deontological arguments regarding the justification of a contested norm. I assume that the stronger arguments in favor of flexibility regarding the domestic amnesty for international crimes—that the Court ought in certain cases to honor such an amnesty—are consequentialist. My position is that the Court's anti-amnesty norm does not fare especially well against such consequentialist arguments if it remains defended by a deontological argument alone: in this case, by the retributivist claim that deserved punishment of perpetrators is such a powerful intrinsic good that it is the source of a duty urgent enough to trump other considerations. That deontological position, I'll attempt to show, is only entailed insofar as we continue to regard the anti-amnesty norm as an inseparable part of an anti-impunity norm. But amnesty rightly understood is not ultimately a violation of a norm forbidding impunity; or at least not just or even primarily such a violation. I will conclude by arguing that it is far better to understand amnesty as a violation of a norm that demands accountability in law and politics. Accountability, unlike impunity, is far more open to support by both deontological and consequentialist arguments. But this generally stronger position comes at a price: unlike impunity, accountability is a good that is ambiguous between law and politics. It is thus not clear whether the Court is in a position to articulate it.

As a way of beginning an evaluation of the range of arguments justifying the anti-amnesty norm, it's helpful to look first at its opponents. These opponents are just as inappropriately thought of as ‘pro-amnesty’ as arguments for the limited toleration of abortions are thought of as ‘pro-abortion.’ These are consequentialist arguments showing that the permissibility of some form of amnesty for international crimes in some well-defined contexts is justified by the extreme costs imposed by a preference for prosecution.

All consequentialist arguments57 must contend with familiar internal problems, and these will all be present in arguments permitting amnesties for international crimes. Consequentialist arguments can often harbor a tendency to justify unpalatable rights violations once the potential benefits of a given policy rise to a sufficient level. The introduction of side-constraints, or rights consequentialism, if not a form of question-begging, in this context asks the concrete question of whose rights, can be violated and to what extent, in pursuit of the goals of establishing security and consolidating the democratic rule of law. Any rights-consequentialist argument for the tolerability of amnesties in certain specified circumstances would, once again, necessarily have to argue that the rights-violations that such amnesties demand—the rights of legal remedy of victims—would not qualify as ‘off limits’ side-constraints, and such an argument would, it seems, additionally require some nontrivial rank-ordering of political and legal rights that would describe why the transitional situation, in which a rights-based legal regime is under special strains and scrutiny, might justify such a rank-ordering when ‘normal’ politics might not.

It's doubtful whether any successful democratic transition has ever managed to refrain entirely from deviating from the rule of law in the pursuit of democratically legitimate outcomes: you play the cards you've been dealt in such transitions, and flexibility—including a high tolerance for not-quite-clean policy—may be a sine qua non for the efficient domestic politician.58 But can such tolerance extend to the systematic deprivation of rights to legal remedy by those most in need of them, the victims of amnestied perpetrators? Clearly, if a side-constrained consequentialism were to regard those rights as off limits, then no such amnesties would be possible.

Further, insofar as consequentialist arguments argue inductively, the demand for amounts and quality of relevant information can rarely if ever be satisfied to the degree necessary for a confident application of a high-risk policy.59 Is it empirically documented that amnesties do in fact, overall and on the whole, maintain significant causal correlations to positive political outcomes? Do they work? We should not be surprised that no consensus at all emerges from the relevant quantitative studies. Those larger longitudinal empirical studies that look at the short-term and medium-term effects of amnesty policies in democratic transitions contain so many independent variables that even claiming a causal correlation is can seem hubristic. Amnesties have certainly played a role in numerous negotiated peace settlements and democratic transitions worldwide over the past 20 years.60 But no studies can document negatives: they cannot show that but for the amnesty, the resultant peace would not have been achieved, nor can they show that failed transitions would not have failed but for the missing incentive of amnesty.

Finally, consequentialism notoriously struggles with the inherent incommensurability of intersubjective assessments of worthwhile policy goals, and the absence of any consequence-independent standard to settle such conflicts over what counts as the best outcome.61 Stability apart from all other considerations cannot in and of itself count as a context-independent summum bonum absent some convincing argument to that effect, for instance, one that would establish stability's status as a necessary but insufficient condition for the possibility of other political goods,62 and (a far higher hurdle) that would document convincingly that the means by which initial stability is established need not affect the longer-term prospects for stability ‘for the right reasons.’

The international relations theorists Jack Snyder and Leslie Vinjamuri have offered a direct-consequentialist objection to the anti-amnesty norm with remarkable force.63 For them, it's unquestionably the case that domestic amnesties have in numerous cases served as useful negotiating tools in brokering peace agreements and easing democratic spoilers out of power. International efforts to deprive post-conflict states of such powerful and low-cost incentives run the risk of needlessly prolonging and in many cases exacerbating conflicts or even creating new ones.64 Indeed, they argue, once decision-making on the part of conflict-plagued states is seen less in terms of successful adoption of preferred international norms, and more in terms of efficiency for preferred end results, arguments for the anti-amnesty norm evaporate:

Preventing atrocities and enhancing respect for the law will frequently depend on striking politically expedient bargains that create effective political coalitions to contain the power of potential perpetrators of abuses (or so-called spoilers). Amnesty—or simply ignoring past abuses—may be a necessary tool in this bargaining. Once such deals are struck, institutions based on the rule of law become more feasible.65

Snyder and Vinjamuri are international relations experts, and not lawyers. They attach no special intrinsic value to the delivery of criminal justice. This is a case where disciplinary affiliation is a distinction that makes a difference: they are generally unimpressed by the anti-impunity norm, seeing it as an unpersuasive, lower-ranking political good and a legal artifact without much relevance in the field of international politics apart from its capacity to compel state actors to change behaviors. And insofar as the establishment of peace and security is a clear trump over the goal of ending impunity, difficult decisions about the status of domestic amnesties should in any case not be made by international courts; such decisions are properly political ones and should be reserved for national legislatures. For Snyder and Vinjamuri, the overall poor record of internationalized criminal tribunals either to generate convictions or to contribute to political stability is documentation of what Judith Shklar described as legalism.66 Decisions regarding prosecutorial discretion in transnational courts begin to translate, often poorly, what are in fact complex political and policy decisions that require both policy expertise and democratic legitimacy. In the absence of any such legitimacy and without meaningful enforcement powers, these courts have little incentive to yield to pragmatic bargaining, what Snyder and Vinjamuri call a logic of consequences rather than a logic of (normative) appropriateness.67

Arguments such as these don't simply recalculate the social costs of criminal justice. They ask whether criminal justice, as the determination of individual criminal guilt or innocence, is at all appropriate as a guiding norm in the field of international politics. If international criminal law and its new institutions are to defend themselves from this argument, then the norm of impunity, which has guided the dramatic expansion of international law so powerfully, needs to be reinterpreted to mean something more, and inevitably something more complex, than simply the demand to deliver deserved punishment to the guilty. For such an interpretation seems to invite only strongly retributivist defenses.68

In short, the anti-amnesty norm has been justified predominantly as a component of a battle against impunity, and impunity has been understood virtually entirely in non-consequentialist terms, as a retributive principle of narrowly defined criminal justice according to which the desert of punishment alone grounds a positive duty to prosecute and punish, on deontic grounds. If this positive duty—either alone or in combination with the rights of victims to legal remedy—is taken as a serious side constraint to any consequentialist evaluation of policy options for overcoming endemic violence, then the anti-amnesty norm seems to me to be fatally weak, intuitively implausible, and very defeasible by the kinds of consequentialist considerations all too vividly on display in the Northern Uganda situation.69 The positive duty undertaken by the ICC to bring Joseph Kony to his legal nemesis (coupled with the negative duty to take no actions that would deprive his victims of their legal rights) is a strong one, certainly. But we cannot expect even strong duties of this kind to withstand any and all arguments concerning the foreseeable and avoidable consequences of acts taken to fulfill such duties.

But even the well-grounded suspicion that Northern Uganda constitutes the grounds for one such argument does not in itself let us draw the conclusion that the ICC's position is without defense. Another and better route open to us is to question whether the purely retributivist, deontological interpretation of the anti-amnesty norm is the only or best one available. And one beginning step here is to re-examine the apparently conceptually necessary relationship between amnesty and impunity, as a way of seeing whether ‘impunity’ is even the appropriate subject matter of the anti-amnesty norm at all.

Impunity obviously refers specifically to punishment: it declares the absence or deprivation of deserved punishment of guilty perpetrators to be an injustice. That is certainly true. But a moment's reflection on the very idea of amnesty indicates that this cannot be an exhaustive account. Amnesties provide pre-prosecution relief. They indemnify persons from the legal consequences for acts that otherwise, under the normal operation of domestic criminal law, would have been proper objects for legal action. Amnesties thus differ sharply from pardons insofar as the latter are post-conviction relief from legal sanction. This may seem like a fine distinction. In fact it's crucial, since any argument for the justice of prosecution must be careful to distinguish between prosecutions and the sanctions imposed only subsequent to conviction. Prosecutions, with all due process observed, can also lead to acquittal. It is easy to forget that even villains like Kony enjoy the presumption of innocence together with the panoply of legal, procedural rights and protections that establish their status as a defendant, not a perpetrator. Battling impunity cannot be taken to imply a waiver of due process. In fact in the international arena, there are powerful arguments for placing more emphasis on the public adherence to due process norms than on retribution.

It would be a strange kind of legal justice—though one not so unfamiliar to prosecutors—that saw the benefits of the criminal law to reside entirely in sanctioned punishment. The rule of law ought to be understood to deliver justice independent of its sanctioning power. For the law to fulfill its crucial expressivist function—to communicate to a polity of citizens the force of the law and the law's status as a legitimate set of norms that govern one another's behavior—then the law itself must contain reasons for general compliance apart from the mere form of lawfulness; that is, the law must be capable of being rationally approved of by those whose acts it governs. Duly enacted law expresses the reasonable consensus of citizens that they approve of the mode of social life it makes minimally possible. Criminal law doesn't just offer negative incentives that make criminal acts sufficiently costly. It expresses the willingness of citizens to interact with one another via impersonal procedures establishing the equality, the reciprocity, and the publicity of freedom. It is this expressivist, communicative function of criminal law that is lost, or at least badly foreshortened, as we take the narrower retributivist view that the chief function of the criminal law is to punish the wicked.70

In this sense, the injustice of amnesties is less the withheld criminal sanction than the suspended public process of the giving and taking of justificatory reasons before the law. In fact, what amnesties threaten is not impunity at all, since prior to a due determination of criminal guilt or innocence there simply can be no talk of what sanctions are deserved. Instead, what is lost in the amnesty is a broader normative good, and one that is importantly indeterminate between law and democratic politics: accountability.

My argument here—and it's not only mine71 —is that fighting impunity ought to be taken merely as a part, and perhaps not such a significant part, of a broader effort to use the tools of international criminal law to introduce the institutional means for a regime of accountability of persons and states, to other persons, states, and the international community, as a part of the larger mission of a normative transformation of the international community itself.72 Unlike other commentators, however, who see the core or paradigm instance of accountability in the criminal trial where the defendant must ‘answer for his crimes,’ or get his day in court whether he wishes it or not, I want to appeal to a theory of deliberative democracy in order to interpret ‘accountability’ in a sense deliberately indeterminate between a legal and a political, procedural norm.73 Accountability mediates between a purely legal conception of the rule of law, on one side, and on the other the political norms of democracy, the equal freedoms each agent, together with the fullest possible inclusion of all in a political community.

Criminal trials are institutional embodiments of a discursive relationship that is reciprocal. Trials are institutionalized contests between antagonists, of course, and not collective searches for consensus. The goal of prosecution is the determination of individual legal guilt or innocence, rather than the formation of some deliberative agreement. Nor does due process expect a defendant to behave in any other way than strategically. Particularly in international trials, we should expect recalcitrant and obstinate defendants who will do their best to discredit and subvert the criminal trial. Even here, however, the small but growing body of prosecutorial expertise in managing high-level trials of this kind can ultimately turn even such obstinacy into a broader public depiction of the justice norms that inhabit the rule of law. Obstinacy itself, or the effort to discredit a court, may well backfire in the perception of the public, provided that the rule of law is seen to be the final authority.74

The procedure of reason-giving and reason-taking in the context of a group of legally equal consociates is the core norm in question.75 Accountability in this broader sense reduces to the paradigm case of all being equally answerable to one another by no other means than the giving and taking of reasons that are then publicly justifiable. Law is both a system of coercive norms and a set of communicative acts demanding acknowledgment of their legitimacy, and legitimacy arises only in the constantly circulating performance of public deliberation, both within and between institutions and an unfettered political public sphere.76 This broader accountability norm, as a predominantly procedural norm, spans political (democratic-parliamentary), legal, and un-institutionalized public forms of discourse. In this sense, accountability, unlike culpability, is a quintessentially intersubjective category. A person must be accountable to others; reciprocally, to be accountable to others means that others are accountable to me, by the same giving and taking of reasons in an institutionalized form of public discourse. Accountability denotes an expected deliberative performance from a reason-giver to a set of appropriately designated reason-takers. The determination of individual criminal guilt is merely the most visible outcome of a larger accountability process that includes all those involved—indeed that imbricates the entire political community that seeks to regulate its common political existence by means of positive law.77

Accountability on the part of elected officials is the core meaning of the public dimension of deliberative democracy, and implies that democratic practices are legitimate if and only if such practices are to a suitable degree and under appropriate conditions capable of being justified by those who initiate and execute them to the appropriate audience: those members of a political community who count as reason-takers. But accountable politicians need an active and involved citizenry who are prepared to demand explanations from political figures for policies, under the condition that the citizenry is accountable to its democratic leadership, whether (weakly) via periodic elections or (strongly and significantly) via active inclusion in a political public sphere and institutions of civil society in which political opinions and positions are transmitted from the citizenry to parliamentary bodies.

National policies are to various extents liable to justification via reasons to those who are members of the national polity, and in this sense political officials are personally answerable and accountable, according to duly enacted deliberative procedures. When such national policies clearly and significantly affect persons who are not members of the national polity, then the demand for a transnational or global democracy is intelligible as the demand for inclusion of affected persons into accountability mechanisms beyond or across national discourses.78 Such is the appeal of arguments for cosmopolitan democracy generally, an institutionalization of the principle that people ought to be able to participate in the political processes whose outcomes affect them. This institutionalization, as Jürgen Habermas and others have argued, conforms poorly to the Westphalian national-state structure; this is especially, viscerally apparent in the internationalization of criminal law.79


International criminal law is one such forum where this kind of demand for expanded inclusion has generated new institutional experiments with public reason-giving and reason-taking on behalf of responsible officials and affected persons. Criminal law contains procedural norms—due process, publicity, personal and collective responsibility and culpability, the designation of harm and victimhood—that are, in important respects, purely legal, in the sense that the kinds of reasons given and taken in criminal legal procedures need not borrow from any extra-legal discourse in order to apply apt criteria for what will count as legitimate or illegitimate justificatory reasons. The law says what is prohibited; prosecution and trial intend to determine individual guilt or innocence. But criminal law is a part, occasionally the most visible part, of the rule of law, and the rule of law, if it is to be more than a merely formal specification of conformity to rules, bears structural and institutional connections to the broader norms of democratic life. In an influential work the legal theorist Martha Minow has put the matter succinctly:

To respond to mass atrocity with legal prosecutions is to embrace the rule of law. This common phrase combines several elements. First, there is a commitment to redress harms with the application of general, preexisting norms. Second, the rule of law calls for administration by a formal system itself committed to fairness and opportunities for individuals to be heard both in accusation and defense. Further, a government proceeding under the rule of law aims to treat each individual person in light of particular, demonstrated evidence. In the Western legal tradition the rule of law also entails the presumption of innocence, litigation under the adversary system, and the ideal of government by laws, rather than by persons. No one is above or outside the law, and no one should be legally condemned or sanctioned outside legal procedures.80

Not just equality but the warranted expectation of giving account for norms and one's attitudes toward them is also a part of the rule of law.81

A century ago, Emile Durkheim had already argued that the equality norm expressed by criminal law is older than democracy, and a motor for a pre-modern, ‘mechanical’ social solidarity, the affectively laden reaction of resentment of a collective provoked by the criminal's violation of a norm of communal equality.82 Desert-based defenses of retribution often still appeal to an unfair-advantage argument to justify sanctioning the criminal. But international criminal law for a variety of reasons cannot rest contented with this essentially pre-modern range of justificatory reasons.

International criminal law foregrounds the broader shift from substantive to procedural equality. Its normative commitment is to a form of equal accountability—that all those under the law can regard themselves and one another as equally entitled and burdened with the responsibility to be answerable to others for those acts and omissions that become visible as possible violations of the law.83 Legal procedure is not just a legitimate medium to extract retribution against bad men. It's also a public performance of the deliberative and procedural norms of democracy, and thus a kind of public catechism of democratic virtues.84 Even criminal trials, where deliberation is adversarial and strategic, and not consensus-driven, nevertheless are an institutionalization of the norm of the deliberative exchange of justificatory reasons between equally situated agents. Hence even trials—in the transitional situation, indeed perhaps especially trials—are public, procedural demonstrations of the rule of law, and hence also of democratic norms. This insight invites a broader, longer-range brand of consequentialism for trials and against domestic amnesties, as Carlos Nino, drawing on both the Nuremberg and Argentine experiences, had already argued against Diane Orentlicher's impunity norm at the very beginning of the 1990s:

The trials promote public deliberation in a unique manner. Public deliberation counteracts the authoritarian tendencies which had led, and continue to lead, to a weakening of the democratic system and massive human rights violations. All public deliberation has this effect, but even more so when the subject of the public discussion is those very authoritarian tendencies. The disclosure of the truth through the trials feeds public discussion and generates a collective consciousness and process of self-examination. Questions like, ‘Where were you, Dad, when these things were going on?’ become part of daily discourse. The contrast between the legality of the trials and the way the defendants acted is prominently noticed in public discussion and further contributes to the collective appreciation of the rule of law. Public discussion also serves as an escape valve for the victims’ emotions and promotes public solidarity which, in turn, contributes to the victims recovering their self-respect.85

This consequentialist argument for criminal trials is also an argument about the more subtle long-term harms that amnesties can inflict on transitional societies where the rule of law must be re-established, where a vacuum of civic trust86 exists between citizens and their institutions of government, and where a broad cultural acceptance of the norms of the rule of law and of due process, part of what John Rawls termed stability ‘for the right reasons,’ may be lacking.87 Domestic trials after serious political conflict certainly can generate legal and political accountability for those societies most in need of it. In this vein, the ‘justice cascade’ that Katherine Sikkink has empirically documented in South and Central America—the catalyzing effects of domestic criminal trials not just for other national justice processes but for regional democratization and stabilization—can be regarded as a powerful empirical documentation for a form of normative integration guided by procedural norms and not retributive outcomes.88 And this emphasis on procedural norms is crucial for the argument to de-emphasize impunity and adopt a broader, meta-legal norm of democratic accountability to house the anti-amnesty norm. For among its other virtues, such a norm, by de-emphasizing the significance of criminal sanction, can also more readily accommodate the plea for flexibility and pragmatism that Snyder and Vinjamuri present. Moreover, this harmonizes with the well-documented empirical claim that domestic criminal trials need to be pursued together with other transitional justice mechanisms—public, transparent and accountable truth commissions charged with the drafting of an official report on incidents, the vetting of judges, public officials and security forces, memorials and reparations—if they are to be perceived by those affected, above all by victims, as successful efforts by the state authority to deliver justice.89

The standard description of the basic transitional dilemma—peace versus justice—is a request to evaluate consequentialist and deontological arguments for radically different outcomes. But if we take accountability rather than impunity as the consequence of trials, this dilemma must be recast. It is better seen as a spectrum of tough choices over potentially incommensurable goods, whose rank-ordering is both highly contentious on moral and pragmatic grounds, and also foundational to the needs of a fragile democracy. Moreover, each argument implies policy decisions that carry risks that in context are difficult if not impossible to evaluate, recommending a form of rational prudence in which the worst-case scenario is the repetition of the very same mass political violence and widespread collapse of respect for human rights that members of a society have already experienced. Both sides of the ‘peace versus justice’ dilemma are misnamed. There is the basic Hobbesian point that the promise of short-term political stability or the temporary cessation of hostilities is not the same as peace.90 At the same time, the initiation of criminal investigation and prosecution of suspects of international crimes is not justice, or only a very modest part of it. Both options fail to deliver accountability. The former buys short-term calm at the price of a massive and public abandonment of the norms of the rule of law, with unknown negative effects for the middle-term or long-term prospects for a healthy democratic political culture; the latter mistakes the exaction of legal retribution on a small number of wicked men for the provision of the institutional and even spiritual resources necessary (if never sufficient) for a post-conflict society to recover itself.91

Yet however tough such prosecutorial decisions are, they are still distinct from the new set of choices that the new international legal regime has produced: evaluating the alternatives of domestic or international prosecution. The ICC's principle of complementarity, if it's not to operate as an unfunded mandate on states, must take seriously the possibility that an international trial under its auspices may permissibly have a far more diffuse and indirect relation to the sanctioning force of criminal law than domestic trials. It may for instance be worth considering the possibility of international trials without criminal sanctions, or with an altered conception of proportionality in sentencing.92 Such a possibility is intuitively unattractive on many levels. But without concrete experience to draw from, it's difficult to say how much such intuitive claims on the centrality of the sanctioning force and function of criminal law are justified in the international sphere.

As adversarial deliberative encounters aimed at the establishment of facts, the disclosure and public justification and criticism of reasons, and the expression of public censure upon conviction, trials bear intermediate, expressivist sanctions independent of punishment, surely. It is not the case that only punishment of offenders keeps their victims from ‘lumping it.’93 And it may well be that international trials, as trials over the extraordinary class of international crimes that ‘shock the conscience of the world,’ should be rare, extraordinary, highly publicized efforts to express just such opprobrium—in just those cases where the opportunities to express opprobrium by means of retributive punishment are reduced or lacking entirely. The broader goals of democratic accountability—both for transitional societies and for the relations between states—are better served in any event by the public expression of a normative consensus of this kind, than meting out justice to individual wrongdoers by imprisoning them in the Netherlands.

The intuitive repugnance at the lost opportunity for legal retribution may be softened by considering the broader potentials of international law to contribute to the larger project of border-crossing, cosmopolitan forms of democratic inclusion. By transferring domestic criminal harms into an international arena, international criminal law also broadens and complicates the range of polities, of included groups, with potential deliberative access to legal proceedings, and such inclusion may serve to significantly enhance the influence of transnational NGOs, organized victims’ advocacy groups, horizontal linkages amongst different domestic courts, and so on. Indeed one welcome effect of the dramatic demonstration of the ICC's political difficulties in Africa may be (and should be) a continuation and expansion of that spirit of institutional experimentation that emerged in the wake of democratization in the 1980s and 1990s, and which led to efforts such as truth commissions and UN-backed tribunals which, while certainly flawed, were also important catalysts for the growth of transnational civil society. Thus simply on the procedural, institutional level, the transfer of legal competence to a transnational level offers modest but measurable steps toward a cosmopolitan law. The ICC cannot be expected to do more than its share in this effort.

Beyond institutional innovations, normative dimensions in cosmopolitan democracy now compete with the law norm of just retribution. International criminal law establishes crimes such as crimes against humanity or genocide that cannot be adequately expressed or captured in domestic law systems, because of the intrinsic heinousness of the acts themselves—they are acts that constitute intolerable moral harms to the very idea of an international community, and not simply acts that threaten the security of the international system.94 The groups of all those potentially affected expand to include all legal persons; crimes against humanity know no nationality.

International trials therefore must operate according to a different kind of normative framework. Beyond the important strategic considerations of how best to bring the guilty to justice, how to impose and enforce the rule of law, and how best to meet the needs of victims, international trials also express a fragile transnational consensus concerning the exceptionless and universal nature of the dignity of the human person, a norm that is newer than the norm of just retribution, and one which bears internal links to the procedural norms of due process and equality before the law. This accountability norm's realization through the new and largely untested form of the international criminal trial is, then, at least a potentially very great good which domestic amnesties check.95

But strengthening the anti-amnesty norm in pursuit of this expanded, meta-legal norm of accountability implies a heightened degree of pragmatism and flexibility. Since its inception, the ICC has been faced with the question of what kinds of non-prosecutorial justice mechanisms (possibly including amnesties) at the domestic level a State party could offer that would supplement or even replace criminal justice, to the point that the Court would be satisfied that the State had adhered to the ‘genuine effort’ requirement to investigate and address criminal acts articulated by Article 17 of the Rome Statute.96 The test case has been whether a State party could effectively reproduce the South African experience—amnesties for big fish, conditional upon a number of stringently enforced conditions, and conducted in a court-like setting with empanelled judges and legal representation—and yet still rise to meet the requirements that the Rome Statute places on state parties for the provision of domestic criminal justice.

This question is not settled, and the Ugandan situation will not do much to settle it. The February agreements in Juba, calling for war crimes by LRA members to be tried in a special new division of the Ugandan High Court, can be interpreted both as honest efforts to deliver justice and as a tactical effort to pressure the ICC to withdraw its indictments. ‘Traditional’ Acholi justice practices, such as matu oput, have been offered as reconciliation mechanisms again as a potential replacement for international trials.97 More formalized truth commissions, together with serious commitments to DDR beyond the weak provisions in the original 2000 Amnesty Act, have been proposed but remain disputed. What remains clear for the moment is that the ICC will not rescind its indictments, nor will it be able to arrest and extradite Joseph Kony and his henchmen to the Hague for trial. The fragile peace settlement agreed to in the Spring of 2008 collapsed over this refusal, with Kony retreating once again into the bush. The Court has no attractive options. It has itself assumed the role of spoiler in Uganda, but cannot rescind its indictments without losing its credibility. Its Prosecutor's single-minded dedication to a narrow conception of prosecutorial zeal in pursuit of legal retribution has effectively painted the entire court into a corner. To escape from that corner, and more effectively join a broader effort at globalizing democracy and democratic accountability, the court may need to rethink its role as a provider of justice.


My thanks to the Centre for the Study of Social Justice, Department of Politics and International Relations at the University of Oxford, for supporting the research on which this paper is based, and to Dapo Akande, Bill Burke-White, Phil Clark, Vaughan Lowe, Pablo de Greiff, Sir Adam Roberts, Adam Swift, Henry Shue and Leslie Vinjamuri for their helpful advice and conversation. I owe special thanks to Mark Freeman, Director of the Brussels Office at the International Center for Transitional Justice, for his enormously helpful conversations. Thanks finally to two anonymous reviewers for Ethics and Global Politics for very helpful criticism.


1This paper is a much revised and expanded version of the first of three Oliver Smithies Lectures delivered at Balliol College, University of Oxford, in May 2007.

2 Rebels and Ugandan Government Agree to Terms of Prosecution of War Crimes, New York Times, 20 February 2008.

3 For a comprehensive analysis of the root causes and development of the LRA and the conflict in Northern Uganda see Tim Allen War and justice in Northern Uganda: an assessment of the International Criminal Court's intervention, Working Paper of the Crisis States Research Centre, Development Studies Institute, London School of Economics; see also Tim Allen (2005) Trial justice: the international court and the Lord's Resistance Army London, Zed Books.

4 Ruddy Doom & Koen Vlassenroot, Kony's message: a new koine? The Lord's Resistance Army in Northern Uganda, African Affairs, no. 98.

5 Human Rights Watch (2003) Abducted and abused: renewed conflict in Northern Uganda, Human Rights Watch Report, 15(12a). Available online at:

6 See Erin K. Baines (2007) The haunting of Alice: local approaches to justice and reconciliation in Northern Uganda. International Journal of Transitional Justice, 1(1), 91–114.

7 See Kasaija Phillip Apuuli, Amnesty and international law: the case of the Lord's Resistance Army insurgents in Northern Uganda, Refugee Law Centre Publication; Lucy Hovil & Zachary Lomo (2005) Whose justice? Perceptions of Uganda's Amnesty Act 2000: the potential for conflict resolution and long-term reconciliation, Refugee Law Project Working Paper 15, February 2005.

8 Jeffrey Gettleman (2006) Uganda peace hinges on amnesty for brutality, New York Times, 15 September 2006.

9 Rome Statute of the International Criminal Court, Article 5, section 1, items (a) through (c), specifying war crimes, crimes against humanity, and genocide as the crimes for which the Court asserts jurisdiction. (A fourth category, crimes of aggression, has yet to find a settled definition and thus awaits a definitive ruling on jurisdiction.) For the purposes of this paper, ‘international crimes’ will be equated with these three crimes as specified in the relevant article, which refers to them as ‘the most serious crimes of concern to the international community as a whole.’

10 The language here ‘unwilling or unable genuinely to investigate on their own’ is consciously quoting article 17 of the Rome Statute, which specifies admissibility of crimes before the Court. See the further discussion of the ‘genuine’ test of article 17 below.

11 The powers and duties conferred on the Prosecutor by Article 53 are highly controversial. It's not clear whether ‘the interests of justice’ refer specifically and exclusively to criminal justice, or whether the Article directs the Prosecutor to consider broader justice criteria and mechanisms such as testimonial evidence submitted to truth commissions in exchange for individual amnesty (the South African model), traditional or local justice and reconciliation practices, or reparations. Indeed the OTP itself was uncertain enough about the meaning and extent of the phrase in the context of the Northern Uganda situation that it requested position papers from a range of global NGOs for study. Of these the most incisive is that by Human Rights Watch, which urges the OTP to cleave to the narrower definition of justice as criminal justice, as a way to tie itself to the mast and resist the siren call of non-criminal justice mechanisms that would allow high-level perpetrators to escape prosecution. See Human Rights Watch Policy Paper: The Meaning of ‘the Interests of Justice’ in Article 53 of the Rome Statute, June 2005.

12 He did so with President Museveni by his side—a publicity event that unfortunately created the very false impression that Museveni was in control of the OTP's investigation, and that he could limit that investigation to LRA members only. This misconception seems to have persisted even in the mind of the President, even though at the press conference itself Ocampo made it clear that UPDF soldiers or indeed anyone suspected of criminal wrongdoing was equally liable to investigation and indictment.

13 For a review of the arguments in favor of the withdrawal of the warrants, see Zachary Lomo, Why the International Criminal Court must withdraw indictments against the top LRA leaders: a legal perspective, Refugee Law Project Working Paper, August 2006. For a rejoinder, see Phil Clark (2007) Dilemmas of Justice, Prospect, magazine issue 134, May 2007. See also Southwick, Katherine (2005) Investigating War in Northern Uganda: dilemmas for the International Criminal Court, Yale Journal of International Affairs (Summer/Fall 2005).

14 Statement of the Office of the Prosecutor on Uganda, Press Release, March 4 2008, accessed at:

15 The anti-impunity norm, as I am calling it here, was most clearly and powerfully articulated in the early and mid-1990s in a pair of commissioned UN reports by special rapporteurs, Diane Orentlicher and Louis Joinet. See especially the history of the development of the anti-impunity norm in Louis Joinet, ‘The Administration of Justice and the Human Rights of Detainees: Question of the Impunity of Perpetrators of Human Rights Violations (Civil and Political),’ E/CN.4/Sub.2/1997/20/Rev.1, p 2. See also Diane Orentlicher, Promotion and Protection of Human Rights: Impunity, UN Document E/CN/4/2004/88.

16 For an overall assessment of the implications of the ICC for national sovereignty, see Atul Bharadwaj (2003) The International Criminal Court and the question of sovereignty, Strategic Analysis, 27(1), Jan–March.

17 A comprehensive analytical survey of the number and range of national amnesty policies in post-conflict situations over the past 50 years, the Amnesty Law Database, has now been compiled by Louise Mallinder at Queens University, Belfast. It documents 421 different amnesty acts and laws enacted in 127 countries. Mallinder's discussion of her findings can be found in Louse Mallinder (2007) Exploring the practices of states in introducing amnesties, Working Paper for Workshop 4, Legal Framework, International conference: building a future on peace and justice, Nuremberg, 25–27 June 2007.

18 Joinet, 2–3.

19 Bryan A. Garner (Ed.) (2004) Black's law dictionary (8th edn). New York, Thompson-West, 93.

20 Amnesties had their legal origin in treaty language ending international conflicts and were regarded as mechanisms to establish the basis for lasting peace among former enemies. They therefore emerged from the language of international treaty law as tools for the legal enforcement of reciprocal ‘oblivion’ between formerly hostile states for offenses at the conclusion of international wars, as an expedient for the immediate work of the restoration of peaceful international relations. This international status of amnesties as components of treaties is still registered in the provisions for amnesties in the Geneva Conventions governing international armed conflict. It is not identical to situations where states consider amnesties for their own citizens for atrocities, where the calculation of expediency must be morally weighted quite differently. The Geneva Convention provision for post-conflict amnesties in international conflicts has frequently been used as a justification for blanket amnesties for internal disputes, even in the case of South Africa, where the Constitutional Court cited the Convention as precedent to justify its decision in the AZAPO case, where plaintiffs had charged that the amnesties were contrary to international law. See Michael P. Scharf Amnesty, in Encyclopedia of genocide and crimes against humanity. On Amnesty's Greek history, see Jon Elster (2004) Closing the books: transitional justice in historical perspective. Cambridge, Cambridge University Press, chapter 1.

21 For a lucid description of the various conditions and terms that have been attached to national amnesties, and their relative legal and moral valences, see Kent Greenawalt (2000) Amnesty's Justice, in: Robert I. Rotberg & Dennis Thompson (Eds), Truth v. justice: the morality of truth commissions. Princeton, Princeton University Press.

22 The decision of whether to regard the official granting of amnesties for offences and delicts that otherwise (in ‘normal’ criminal justice) would justify prosecutorial attention is in fact at the core of the evaluation of the political (democratic) status of amnesty policies. In some measure a deciding factor is whether such policies are democratically legitimate, and such factors as the distinction between executive and parliamentary decrees is important. (The main examples of blanket amnesties in Latin America and Africa have been executive, have had little or no democratic legitimation or transparency, and have frequently, as in the cases of both Chile and Argentina, been self-amnesties. On the other hand, even in the rare cases of amnesties with popular and parliamentary backing, such as Uruguay, the question of whether popular plebiscite justifies deviations from normal course of criminal justice remains.) To a great extent, this is a conceptual question regarding the seating of domestic criminal justice systems in the institutions of democratic governance, and involves the evaluation of prosecutorial discretion, executive pardon, and other ‘exceptions’ as parts of normal law. On the status of amnesty in domestic criminal law, see Ronald Slye (2002) ‘The legitimacy of amnesties under international law and general principles of Anglo-American law: is a legitimate amnesty possible? 43 Virginia Journal of International Law, 43, 173.

23 Among the many analyses of the Amnesty Committee of the South African Truth and Reconciliation Commission, see especially James L. Gibson (2004) Overcoming apartheid: can truth reconcile a divided nation? New York, Russell Sage Foundation, and the overall critical political analysis in Richard Wilson (2001) The politics of truth and reconciliation in South Africa: legitimizing the post-apartheid state. Cambridge, Cambridge University Press.

24 The etymological connection between amnesty and forgetting should not be over-played; on the other hand the frequent connection of amnesty and oblivion from the Greek example to the language of modern international treaty law makes the functional connection between willed forgetting and acts of immunity compelling. See Leslie Sebba Amnesty and pardon, in: Encyclopedia of crime and justice, 1, 59: ‘Amnesty…derives from the Greek amnestia (“forgetting”), and has come to be used to describe measures of a more general nature, directed to offenses whose criminality is considered better forgotten.’ See Andreas O'Shea (2001) Amnesty for crime in international law and practice. The Hague, Kluwer, 22–30; Jon Elster (2004) Closing the books: transitional justice in historical perspective. Cambridge, 77ff; see generally Bigger, Nigel (Ed.) (2001) Burying the past. Making peace and doing justice after civil conflict Georgetown University Press; Crocker, David (1999) Reckoning with past wrongs: a normative framework, Ethics and International Affairs, 13.

25 Unlike any other similar transitional state, the South African amnesties were authorized by means of constitutional law, rather than executive order, parliamentary legislation, or popular referendum. But this may actually serve to strengthen the underlying objection that sovereign deviations from the rule of law are in principle unjust irrespective of whatever other effects they may produce, opening the door to a legitimate question concerning the legitimacy of effective international legal and political constraints on such exercises of sovereign power.

26 The putative harms of amnesty as deviation from presumptively appropriate legal procedures are magnified in the international law context, for the simple reason that those acts denominated as international crimes are acts whose status as malum en se, as wrongs prior to legal determination, transcend what the domestic criminal law is capable of recognizing due to its own limits. Genocide, war crimes or crimes against humanity may be composed of many separate acts, each severally recognizable in the domestic criminal code, but which are larger than the sum of their parts and thus have come to require special status. Therefore, the legal injustice of domestic amnesties can also extend to the failure of states to adhere to the terms of valid international treaties of which they are signatories, and which commit states to a duty to be the courts of first resort for the investigation and prosecution of international crimes.

27 For a classic presentation of positive retributivism, see Kant's famous claim that even a self-dissolving civil society would be commanded by duty to hang the last of its murderers before the lights were turned out, ‘so that each has done to him what his deeds deserve and blood guilt does not cling to the people for not having insisted on this punishment; for otherwise the people can be regarded as collaborators in this public violation of justice.’ Immanuel Kant (1991) The metaphysics of morals. Cambridge, Cambridge University Press, 142. Kant straightforwardly associated positive retribution with lex talionis; modern retributivism cannot so easily do so but expresses the intuitive justice claim of desert in terms of the implicit claims of equality and the correcting of unfair advantage. See among others George Sher (1987) Desert. Princeton, Princeton University Press.

28 The classic recent statement of the duty to prosecute is Diane Orentlicher (1991) Settling accounts: the duty to prosecute human rights violations of a prior regime, Yale Law Review, 100. See also Jaime Malamud-Goti (1990) Transitional governments in the breach: why punish state criminals? Human Rights Quarterly, 12. For an overall analysis of the international law duty to prosecute and its effects on national law processes, see Naomi Roht-Arriaza (1995) Impunity and human rights in international law and practice. Oxford, Oxford University Press. For a strong defense of international obligations to prosecute, see Geoffrey Robertson (2004) Crimes against humanity: the struggle for global justice (3rd edn). New York, Penguin.

29 The suit filed by among others representing the AZAPO against the interim government of South Africa, claiming that the provisions for amnesty in the interim South African Constitution of 1994 were unconstitutional insofar as their discharge would require South Africa to renege on its international treaty obligations and would deprive a class of citizens (victims of apartheid violence) from their constitutional rights to legal remedy for harms. AZAPO and Others v President of the Republic of South Africa (1996) 8 B.C.L.R. 1015(CC). See also Hippler Bello, J. & Wilhelm, D.F. (1997) International decisions, American Journal of International Law, 91, 63.

30 On this point see especially Diane Orentlicher (2004) Whose justice? Reconciling universal jurisdiction with democratic principles, Georgetown Law Journal, 92.

31 In using this term I am borrowing, with some changes, from the work of Mark Freeman, who refers to an ‘amnesty norm’ to refer to just what I mean here by ‘anti-amnesty norm.’ I make this change in the interest of clarity but acknowledge that the term is originally his. See Mark Freeman, ‘Intended and unintended consequences of the new intolerance for amnesties,’ lecture, University of Oxford, 19 June 2007. See also Mark Freeman (forthcoming) Amnesty and accountability: the leniency dilemma. Cambridge, Cambridge University Press for a full account of Freeman's position, which I cannot hope to do justice to in this context.

32 See Robert Cryer (2007) An introduction to criminal law and procedure, 33, for a description of some of the more recent legal precedents for the ‘crystallization’ of the anti-amnesty norm. Compare this law textbook from 2007 with a parallel definitive statement from Cassese's International Law in 2003, where the author writes that ‘[t]here is not yet any general obligation for States to refrain from amnesty laws on [international] crimes. Consequently, if a State passes any such law, it does not breach a customary rule. Nevertheless if a court of another State having in custody persons accused of international crimes decides to prosecute them although in their national state they would benefit from an amnesty law, such court would thereby not act contrary to general international law, in particular in respect to the principle of respect for the sovereignty of other States.’ Antonio Cassese (2003) International law. Oxford, Oxford University Press, 609. One way to make clear the process of ‘crystallization’ or emergence of a new international law norm, then, is to take this five-year old statement of a publicist and note that it subsequently is expanded from a negative form—merely that domestic amnesties do not prima facie have any extraterritorial effect—to a strongly positive form that such domestic amnesties in fact signal to other courts, whether other states or international tribunals, that a breach of international law has taken place insofar as an act has been amnestied, and not simply for the act itself. Again, domestic amnesties are not violations of international law according to the emergent anti-amnesty norm. It is not a crime not to punish a crime. But the non-performance of punishment now at least in principle triggers the legal scrutiny of the international community—and in this case it is not entirely true (nor entirely false) to say that such amnesties are conformable with customary international law. For a more complete discussion, see Michael P. Scharf (2004) The amnesty exception to the jurisdiction of the International Criminal Court, The International Criminal Court; Hoepfel, Frank & Claudia Angermaier (2005) ‘Adjudicating international crimes’, in: Handbook of transnational crime and justice. New York, Sage Publications; Alexander Orakhelashvili (2006) State immunity and international public order revisited, German yearbook of international law, 49, 327–365; Theodor Meron (2006) Reflections on the prosecution of war crimes by international tribunals, American Journal of International Law, 100(3), 551–579.

33 Many academics recognize the path breaking work of Diane Orentlicher, which over the course of the 1990s combined expertise in international law and international relations to push for a duty to prosecute on the part of states with jurisdiction. See the canonical essay by Diane Orentlicher (1991) Settling accounts: the duty to prosecute human rights violations of a prior regime, Yale Law Journal, 100, 2537–2615. Orentlicher herself has frequently recalled a watershed conference in 1988, in which a group of prominent international legal scholars resolved on a small set of core principles necessary to battle impunity. One of those principles, Orentlicher has written, was the core belief that the international community had to find institutional means to assist states in their efforts to bring to justice those who were guilty of past atrocities—even if at the time those institutional means were largely conjectural. What remained (and remains largely) unsettled is how much discretion states may take to ‘deal’ with such past abuses in a manner compatible both with their own demands for reconciliation and democratic stability, and their international legal obligations.

34 We can summarize the most influential of these more recent rulings: first, certainly, we can cite what Naomi Roht-Arriaza has dubbed the ‘Pinochet effect,’ in which the former Chilean president was not successful in his claim that both his self-amnesty and his status as a former head of state (and sitting senator in the Chilean parliament) granted him immunity from prosecution in foreign courts; in 1998 the Law Lords of the British House of Lords ruled that there can be no such immunity for international crimes, and refused his appeal to vacate an indictment for arrest issued by the government of Spain. For a thorough account of this emperor-has-no clothes moment—including the powerfully catalyzing effect that the British decision had on the Chilean government, which eventually reversed its own amnesty law and initiated (but never completed) a domestic prosecution of Pinochet, see Naomi Roht-Arriaza (2005) The pinochet effect: transitional justice in the age of human rights. Philadelphia: University of Pennsylvania Press. Perhaps earlier still, though less influential, the Inter-American Court of Human Rights produced a string of rulings that laid out a legal justification for the serious enforcement of states’ legal duties to prosecute international crimes occurring within their own jurisdiction. Most notably in Velasquez v Rodriguez (Inter. Am. Ct. Hum. Rights (Ser. C), number 4, 1988), the IACHR insisted that the American Convention on Human Rights must be interpreted to ‘require states to investigate seriously, identify, and punish offenders as well as compensate victims’ of international crimes. States, in other words, are not unlimited in their discretion to respond to such crimes as they see fit: the default obligation is prosecution. On the IACHR ruling, see Steven R. Ratner & Jason S. Abrams (2003) Accountability for human rights atrocities in international law (2nd edn). Oxford, Oxford University Press. In Prosecutor v Furundzija, the International Criminal Tribunal for the Former Yugoslavia ruled that incidents of torture are unambiguously violations of jus cogens norms; that is, are violations of international law even apart from any treaty obligations of state parties, and therefore by inference any state was capable of claiming universal jurisdiction in cases where the state with relevant jurisdiction was unwilling or unable to prosecute. In the now-infamous case of the ‘oral reservation’ to the 1999 Lomé peace accord in Sierra Leone, the UN special representative to the negotiations to bring about a peace settlement in that country, when presented with a final document providing blanket amnesty to leaders who had committed numerous war crimes and crimes against humanity (including the rare bird, a peace treaty that granted blanket amnesty by name to the rebel leader Foday Sankoh) quickly penciled in a ‘reservation’ next to his signature—namely, that the UN did not regard as valid and would not honor any domestic amnesty for international crimes. The reservation is generally regarded as evidence of the slow but definite evolution of the principled position of the UN on amnesties over the last 10 years: from pragmatic accommodation of amnesties provided they demonstrably contributed to negotiated peace settlements, to a position that regards such amnesties as principled violations of international law that are incompatible with the UN's core mission. On the UN Security Council's and the UN Human Rights Commission as (not always vigorous) agents of the anti-amnesty consensus see the UN HRC's Report of the Human Rights Committee on the Situation in Nigeria, Vol. 1, HN GAOR 51st Session, supplement 40, para. 284 [UN DOC A/51/40 (1996)]. On the Lomé Accord and the reservation see UN DOC S/1999/836; for a detailed interpretation, see Patricia Hayner (2007) Negotiating peace in Sierra Leone: confronting the justice challenge, Expert Report, International Center for Transitional Justice and Centre for Humanitarian Dialogue, December 2007, accessed 2/2/2008. Available online at: The ‘Princeton Principles,’ while of course having no source in the rulings of any court, did serve to state explicitly that ‘blanket amnesties generally are inconsistent with a nation's obligation to hold individuals accountable for these crimes.’ Certainly the hedges in this definition—namely, the fuzzy meaning of a ‘blanket’ amnesty and the ‘general’ nature of the incompatibility of amnesties with international law—are directives for future courts to interpret cases, rather than adhere to international law principles. In practice, though, these hedges seem to have worked all too well: as has often been pointed out, the tough cases of amnesty, on which domestic and international courts could actually use some clear guidelines, are not at all touched by provisions for the mere generality of the legal status of amnesty. See Princeton Principles on Universal Jurisdiction. Available online at: http://lapa.princeton/hostedocs.unive_jur.pdf.

35 Diane Orentlicher (2007) Settling accounts’ revisited: reconciling global norms with local agency, International Journal of Transitional Justice, 1(1), 11.

36 Orentlicher, 11.

37 A helpful summary of truth commissions that depended on negotiations in which blanket amnesties were offered to outgoing political leadership is in Chandra Sriram (2004) Confronting past human rights violations. London, Frank Cass, especially Chapter 2. Prominent examples of commissions in exchange for granting or acknowledgment of prior amnesties are Argentina, Honduras, Uruguay, Guatemala, Chile, Romania, South Africa (unique in many aspects), Sri Lanka, and with some modifications Sierra Leone (where amnesties offered under the Lomé Accord and affirmed by the Sierra Leone truth commission were later declared illegal by the hybrid Special Court for Sierra Leone).

38 The principle of selectivity certainly has problems of its own. For a thorough analysis of the legal and political problems with the principle of selectivity, see Robert Cryer (2005) Prosecuting international crimes: selectivity and the international criminal law regime. Cambridge, Cambridge University Press, Chapters 2 and 3.

39 Critics of internationalized or hybrid tribunals have made much of their extraordinary costs in relation to their relatively low conviction rate. Helena Cobban for instance has pointed out that with costs running in excess of one billion dollars and a conviction total of only 25, the International Criminal Tribunal for Rwanda averages roughly 40 million dollars per conviction. Helena Cobban (2006) Think again: international courts, Foreign Policy, March/April. These claims are misleading in many respects, however. Even at such a high rate, such costs are not out of line with prosecution costs for serious, high-level perpetrators in domestic cases. See David Wippman (2006) The costs of international justice, American Journal of International Law, 100(4), , 861–881.

40 Complementarity was most certainly a concession that the drafters of the Rome Statute adopted on realist grounds: few potential signatory states would have been prepared to accept both that international criminal law would be supreme over parallel domestic codes and that an International Criminal Court would be in a position to arrest, extradite and prosecute their nationals without state consent, in the first instance. On the legal and political implications of the Court's principle of complementarity, see Muhamad M. El Zeidy (2002) The principle of complementarity: a new machinery to implement international law, Michigan Journal of International Law, 23; Jann Kleffner (2003) The impact of complementarity on national implementation of substantive international criminal law, Journal of International Criminal Justice, 1.

41 For an excellent general account, see Naomi Roht-Arriaza (2001) The role of international actors in national accountability processes, in: Alexandra Barahona de Brito & Carmen Gonzalez-Enriquez (Eds), The politics of memory: transitional justice in democratizing societies. Oxford, Oxford University Press.

42 For a balanced and not uncritical account of the ICTR in Rwanda, see Payam Akhavan (1997) Justice and reconciliation in the Great Lakes Region of Africa: the contribution of the international criminal tribunal for Rwanda, Duke Journal of Comparative and International Law, 7.

43 See the essays collected in Stephen Macedo (Ed.) (2003) Universal jurisdiction: national courts and the prosecution of serious crimes under international law. Philadelphia, University of Pennsylvania Press. For a critical account of the practice of universal jurisdiction and the international state system, see Larry May (2005) Crimes against humanity: a normative account. Cambridge, Cambridge University Press.

44 See Ralph Zacklin (2004) The failings of ad hoc international tribunals, Journal of International Criminal Justice, 2; Theodor Meron (2006) Reflections on the prosecution of war crimes by international tribunals, American Journal of International Law, 3, 551–579.

45 Rome Statute of the International Criminal Court, Preamble.

46 Rome Statute of the International Criminal Court, Preamble.

47 For a critical view on this backstop role, see David Wippman (2006) Exaggerating the ICC, in: Joanna Harrington, Michael Milde, & Richard Vernon (Eds), Bringing power to justice: the prospects of the International Criminal Court. Montreal, McGill University Press. On the formation of the preamble to the Rome Statute, see Antonio Cassese et al. (Eds) (2002) The Rome statute of the International Criminal Court: a commentary, vol. 1. Oxford, Oxford University Press.

48 For a fuller defense of this interpretation, see Yasmin Naqvi (2003) Amnesty for war crimes: defining the limits of international recognition, International Review of the Red Cross, 85(851), 583–626.

49 Claudia Angermaier (2004) The ICC and amnesty: can the court accommodate a model of restorative justice? Eyes on the ICC, 131, 144; Dinah Shelton (Ed.) (2000) International crime, peace, and human rights: the role of the International Criminal Court. Ardsley NY, Transnational Publishers; Diba Majzub (2002) Peace or justice? Amnesties and the International Criminal Court, Melbourne Journal of International Law, 3.

50 Not that weakly, however. This claim, together with the decision on the limits of sovereign immunity, grounded the international legitimacy of universal jurisdiction of Spain in the arrest of Augusto Pinochet. It reached its limit in the more recent ruling of the International Court of Justice in Belgium v Congo, which held that the government of Belgium was not justified in attempting to arrest the sitting foreign minister of the Democratic Republic of Congo; the ICJ held that a domestic state court did not have the appropriate jurisdiction to arrest a sitting state authority—but that an international court may well have such authority. Hence it was the crucial distinction between domestic and international criminal law on the sovereign immunity of sitting heads of state that the ICJ appealed to in its decision.

51 See Simon Romero (2007) Living in exile isn't what it used to be, New York Times, October 7. The most dramatic recent example of the reduced hard currency value of the domestic amnesty is the successful indictment, arrest and extradition of former Liberian president Charles Taylor on an international warrant issued by the hybrid Special Court for Sierra Leone. The government of Nigeria—with its own domestic and international agenda quite apart from acceptance of international justice norms—reneged on its own negotiated exile and sanctuary agreement with Taylor and gave him up to international law enforcement once it became convinced that he would not hold to his own commitment to refrain from meddling in Liberian or Nigerian internal affairs. For former political leaders and potential spoilers accustomed to freedom of travel, even the potential threat of arrest abroad may have a definite deterrent effect. See Marc Weller (1999) On the hazards of foreign travel for dictators and other international criminals, International Affairs, July, 599–617.

52 See Rhys David Evans (2007) Amnesties, pardons, and complementarity: does the International Criminal Court have the tools to end impunity? Unpublished paper, accessed on 13 February 2007 at

53 For a discussion of the impact of the ICC on traditional state sovereignty, see Bruce Broomhall (2003) International justice and the International Criminal Court: between sovereignty and the rule of law. Oxford, Oxford University Press.

54 John Holmes (1999) The principle of complementarity, in: Roy Lee (Ed.) The International Criminal Court: the making of the Rome Statute—issues, negotiations, results. Dordrecht, Martinus Nijhof.

55 On the overall position of the United Nations, Amnesty International, Human Rights Watch, and other larger transnational organizations that domestic criminal prosecutions for perpetrators are the ‘gold standard’ for justice in transitional contexts see the policy paper by the International Center for Transitional Justice; United Nations General Assembly RES/60/147, ‘Basic Principles and Guidelines on the Right to a Remedy and Reparation for Victims of Gross Violations of International Human Rights Law and Serious Violations of International Humanitarian Law.’ On the changing and occasionally contradictory position of the United Nations on the default preference for prosecutions of perpetrators of state-backed atrocities, see in general Carsten Stahn (2002) United Nations peace-building, amnesties and alternative forms of justice: a change in practice? International Record of the Red Cross, 84(845), 191–205.

56 William W. Burke-White (2006) Double edged tribunals: the political effects of international criminal tribunals, Guest Lecture Series of the Office of the Prosecutor, ICC, 28 July. For a liberal-IR analysis of how states (taken in a disaggregated form) respond to international pressures for prosecution via internationalized courts, see in general the work of Anne-Marie Slaughter (1993) in particular International law and international relations theory: a dual agenda, American Journal of International Law, 87; Anne-Marie Slaughter (2004) A new world order. Princeton, Princeton University Press, describing the liberal-cosmopolitan project of transforming international criminal law system into ‘a global legal hierarchy, with a world supreme court such as the International Court of Justice resolving disputes between states and pronouncing on rules of international law that would then be applied by national courts around the world. What is in fact emerging is messier and much more complex. It is a system composed of both horizontal and vertical networks of national and international judges, usually arising from jurisdiction over a common area of the law or a particular region of the world. The judges who are participating in these networks are motivated not out of respect for international law per se, or even out of any conscious desire to build a global system. They are instead driven by a host of more prosaic concerns, such as judicial politics, the demands of a heavy caseload, and the new impact of international rules on national litigants’ (p. 67).

57 There is clearly a great deal more to say about the nature, variety and aspects of consequentialism here that I have neither the competence nor the space to provide. I can at best offer a promissory note to the effect that the particular kind of consequentialism that best opposes an inflexible anti-amnesty norm is highly contested.

58 One way, for instance, to remove a stubborn potential spoiler from the transitional scene is an amnesty as an effective inducement to exile or to join the new political process; another may be extralegal assassination. Move too far along this scale and the ‘democratic’ transitional regime begins to look suspiciously like the oppressive regime it attempts to replace. Too little—so the argument would run—and it doesn't replace it at all. But where does amnesty lie on this scale of tolerable deviations?

59 See David Brink (2006) Some forms and limits of consequentialism, in: D. Copp (Ed.) Oxford handbook in ethical theory. Oxford, Clarendon Press.

60 For an argument (though troublingly thin on empirical specifics) that domestic amnesties have not demonstrably contributed significantly to long-term political stability, see Leila Sadat (2005) Exile, amnesty, and international law, Washington University School of Law Working Paper No. 05-04-03 (April 2005).

61 See Ruth Chang (1997) Incommensurability, incomparability, and practical reason. Cambridge, Harvard University Press; Peter Railton (2003) Facts, values, and norms: essays toward a morality of consequence. Cambridge, Cambridge University Press, 249–291.

62 Although see the classic argument that security counts as a basic right beyond utilitarian calculation in Henry Shue (1996) Basic rights (2nd edn). Princeton, Princeton University Press, 21–22.

63 Jack Snyder & Leslie Vinjamuri (2003–2004) Trials and errors: principle and pragmatism in strategies of international justice, International Security, 28(3), Winter, 5–44.

64 Snyder and Vinjamuri describe what they take as definitive cases where domestic amnesties have enabled peace agreements and/or the work of other non-legal justice mechanisms like truth commissions to operate in Afghanistan, Macedonia, Sierra Leone and the Ivory Coast.

65 65Snyder and Vinjamuri, 6.

66 See Judith N. Shklar (1998) American citizenship: the quest for inclusion. Cambridge, MA, Harvard University Press. For a positive interpretation of Shklar's conception of legalism and international criminal tribunals, see Thom Ringer ‘Legalism and its virtue: tribunals and the concept of war crime legalism in Judith Shklar's ‘legalism and Gary Jonathan Bass's ‘Stay the hand of vengeance.’

67 More problematically, Vinjamuri and Snyder make the pragmatic assumption that the only plausible goal for the range of transitional justice measures is deterrence—a questionable assumption at best. Thus deterrence and its pursuit ‘must take precedence over the objective of retroactive punishment when those goals are in conflict. Where human rights violators are too weak to derail the strengthening of the rule of law, they can be put on trial. But where they have the ability to lash out in renewed violations to try to reinforce their power, the international community faces a hard choice: either commit the resources to contain the backlash or offer the potential spoilers a deal that will leave them weak but secure. Efforts to prosecute individuals for crimes must also be sensitive to the impact of these efforts on relations between dominant groups in a future governing coalition. Where trials threaten to create or perpetuate intra-coalition antagonisms in a new government, they should be avoided.’ 12.

68 The ‘traditional’ consequentialist defenses of the provision of punishment—deterrence, and the incapacitation and rehabilitation of the offender—seem weak at best in the context of international crimes. The peculiar status of high-level offenders both politically and psychologically makes any argument for the deterrent effect of prosecution problematic—and even in cases such as the Mugabe regime in Zimbabwe, where threat of international prosecution does indeed seem to be a motivating factor, the result of deterrence is far from clear. Incapacitation is indeed a concern, though again the relatively weak enforcement and sanctioning powers of international law are evidently relevant here, as is the potential that an imprisoned spoiler could generate attention and sympathy that would offset the benefits of his own incapacitation. Rehabilitation, in my view, is a fantasy in such cases.

69 For a presentation of consequentialist and retributivist considerations of the Court's role in terms of punishment—rather than in terms of process, which I am here developing—see Eric Blumenson, The challenge of a global standard of justice: peace, pluralism, and punishment at the International Criminal Court, Columbia Journal of Transnational Law, 44, 797. I am unable here to do full justice to Blumenson's rich and powerful presentation. I think on the whole he attaches far too much significance to the social communication that punishment can achieve; while he and I both would like to see the Court's retributive function in terms of the expressive communication of transnational justice norm, I am arguing that this communicative function is far more plausibly accomplished in the trial, not the punishment.

70 I cannot develop here my view that international criminal law's sanctioning power is far better justified through expressivist arguments than standard consequentialist or retributivist ones. For a comprehensive account of penal expressivism in its relation to contemporary political theory's accounts of the normative foundations of the democratic polity see above all the work of Jean Hampton (1981), specifically The moral education theory of punishment, Philosophy and Public Affairs, 209–238, and Forgiveness and mercy. Cambridge, Cambridge University Press, 1988. A complementary theory with more focus on the procedural aspect of the criminal trial itself can be found in R.A. Duff (2001) Punishment, communication, and community. Oxford, Oxford University Press. On expressivist penal theory in its particular relation to war crimes trials, see Bill Wringe (2006) Why punish war crimes? Victor's justice and expressive justifications of punishment, Law and Philosophy, 25, 159–191.

71 See especially William Burke-White (2001) Reframing impunity: applying liberal international law theory to an analysis of amnesty legislation, Harvard International Law Journal, 42. Burke-White and I share a number of important conclusions. His normative analysis of amnesty laws develops a norm of democratic legitimacy with important parallels to the norm of accountability I propose here; however, his use of liberal international relations theory obliges him to regard democratic legitimacy chiefly as the efficiency of domestic regimes to respond to the aggregated interests of its citizens via voting, while my use of deliberative democracy theory attempts to cash out a norm of discursive inclusion as the core normative dimension of democratic governance. Naturally I believe that the latter is the more appropriate normative orientation; while Burke-White concentrates on the legitimacy features of the process of legislative enactment of amnesty laws, I focus instead on the links between the procedures of criminal trials and the broader accountability mechanisms that they catalyze and promote.

72 One curiosity that emerges in re-reading the canonical academic texts and UN reports that first developed the anti-impunity norm in the 1990s is that authors such as Diane Orentlicher and Louis Joinet actually already saw impunity as merely the legal expression of accountability. And yet they rarely if ever subsequently reflected on the implications of the significant differences between these two norms, and allowed impunity to gradually serve as shorthand for accountability, even though the latter norm extended well beyond what any criminal law system could or was intended to produce. In a sense, this paper is simply an attempt to correct this slippage. A more recent key development in the emergence of accountability as a norm of transitional justice possibly independent of impunity is former UN Secretary General Kofi Annan's 2004 report, The rule of law and transitional justice in conflict and post-conflict societies, S/2004/616, 23 August 2004.

73 Of the many texts on a definition and defense of deliberative democracy see especially Amy Gutman & Dennis Thompson (2004) Why deliberative democracy? Princeton, Princeton University Press, and Jon Elster (1998) Deliberative democracy. Cambridge, Cambridge University Press.

74 See especially Michael Scharf (2006) Chaos in the courtroom: controlling disruptive defendants and contumacious counsel in war crimes trials, Guest lecture, Office of the Prosecutor of the International Criminal Court, 25 September 2006, accessed at:, arguing that international war crimes trials should waive defendants’ rights to self-representation common in domestic criminal trials. For a broader analysis, see Martti Koskenniemi (2002) Between impunity and show trials, Max Planck yearbook of United Nations law, 6.

75 On the deliberative-consensus-based norm that animates even antagonistic criminal trials see the classic analysis by Duff, R. A. (1986) Trials and punishments. Cambridge, Cambridge University Press. One interesting proposal is that international criminal trials have a greater claim to require an American-style jury system of trial by peers rather than a more conventional (European) trial before empanelled judges, the system that the ICC and all other international criminal tribunals employ. For an argument for a jury system for international trials with arguments that are very consonant with those in the present essay, see Amy Powell (2004) Three angry men: juries in international criminal adjudication, New York University Law Review, 79.

76 See Jürgen Habermas (1996) Between fact and norms: contributions to a discourse theory of law and democracy. Cambridge, MIT Press, 51ff.

77 I borrow here from a position developed by Pablo de Greiff (2002) Deliberative democracy and punishment, Buffalo Criminal Law Review, 5.

78 I am here drawing on a recent argument by Robert Keohane; see Robert O. Keohane (2006) Accountability in world politics, Nordic Political Science Association; Robert O. Keohane, Global governance and democratic accountability, unpublished manuscript.

79 See Jürgen Habermas (2001) The postnational constellation and the future of democracy, in: The postnational constellation. Cambridge, MIT Press.

80 80Martha Minow (1998) Between vengeance and forgiveness. Boston, Beacon Press, 25.

81 Habermas, Between facts and norms, 116ff.

82 Emile Durkheim (1997) The division of labor in society. New York, Free Press, 33ff.

83 Habermas's argument to establish the cooriginality of the normative foundation of the rule of law and of democratic popular sovereignty as originating in a ‘discourse principle’ would be central here. Persons in democratic constitutional states ought to be understood as equally entitled to maximal participation in those institutionalized discourses whose policy outcomes are likely to affect their interests to a significant degree. Political and legal rights codify and embody a range of protections to this more basic conception of individual agency. See Habermas, Between facts and norms, 121ff.

84 Here I borrow from the argument for the public, catalyzing function of criminal trials developed by Mark Osiel, who has argued that the high visibility of trials of high-level perpetrators is in fact their chief democratic virtue, and advocates minimizing both the deterrent and retributive justice goals of prosecution, suggesting that public outlays of funds for high visibility trials are monies well spent even when the trials end in acquittals, and even if, as is predictably the case, the former political leadership itself provides no dramatic confessions or apologies. For Osiel, high visibility criminal trials can catalyze both the procedural medium for extralegal public discussion—installing tolerance and mutual trust as default procedural norms—as well as providing a medium for public discussions of the meaning and relevance of the collective memory of moral catastrophe for a democratic polity spent even when the trials end in acquittals, and even if, as is predictably the case, the former political leadership itself provides no dramatic confessions or apologies. Mark Osiel (1997) Mass atrocity, collective memory, and the law. New Brunswick, Transaction. Osiel suggests that the missing justifications of deterrence and retribution in the transitional trial are compensated for in their potential discursive gains: ‘As an aim for criminal law, the cultivation of collective memory resembles deterrence in that it is directed toward the future, where enhanced solidarity is sought. But like retribution, it looks to the past, to provide the narrative content of what is to be shared in memory. Stated most modestly, its purpose is, as Thomas Scanlon puts it, ‘to achieve a general state of mind in the country in which the unacceptability of these acts is generally recognized, so that the perpetrators become pariahs…having done something that cannot be sustained and accepted’ (18).

85 85Carlos Nino (1996) Radical evil on trial. New Haven, Yale University Press, 147.

86 On the centrality of civic trust as a goal of transitional justice mechanisms generally see Pablo De Greiff (2006) Truth telling and the rule of law, in: Tristan Anne Borer (Ed.) Telling the truths: truth telling and peace building in post-conflict societies. Notre Dame, University of Notre Dame Press.

87 John Rawls (1999) A theory of justice (revised edn). Cambridge, Harvard University Press, 431ff; Political liberalism. New York, Columbia University Press, 1993, 459.

88 Katherine Sikkink & Ellen Lutz (2001) The justice cascade: the evolution and impact in Latin America of foreign human rights trials, Chicago Journal of International Law, 2(1), Spring 2001, 1–34. Snyder and Vinjamuri's work was intended as a direct challenge and refutation of this article; for Sikkink's own response to it, see Katherine Sikkink & Carrie Booth Walling (2005) Errors about trials: the political reality of the justice cascade, paper presented at the American Political Science Association, August 2005. [Reply to Snyder and Vinjamuri], accessed at:

89 See Harvey Weinstein & Eric Stover (2004) My neighbor, my enemy: justice and community in the aftermath of mass atrocity. New York, Cambridge University Press. On the theoretical premises supporting these studies—that transitional justice mechanisms must be mutually supporting in order to win public perception of successful justice efforts, see Pablo De Greiff (2008) Theorizing transitional justice, unpublished manuscript, March.

90 See the emergent literature speculating that the spectacularly high level of ‘ordinary’ (non-political) crime in South Africa has its origins (though causal arguments are elusive) with the perceptions about the relative value and legitimacy of the rule of law generated by the South African experience of amnesty in the mid-1990s. On this see the recent work of Diana Gordon (2006) Transformation and trouble: crime, justice, and participation in democratic South Africa. Ann Arbor, University of Michigan Press.

91 One perennial response to this kind of peace versus justice dilemma is a plea for sequencing: first secure the conditions of basic security compatible with the establishment of democratic stability, and then initiate prosecutions at such a time as conditions allow. Tempting though this sequencing solution may be, there are several reasons to be deeply suspicious of it. For a comprehensive analysis of the pitfalls of the sequencing solution, see Thomas Carruthers (2007) The sequencing fallacy, Journal of Democracy, 18(1).

92 This enormously complex issue cannot be done justice here. The notion of trial without legal sanction naturally brings to mind the quasi-judicial role of the Amnesty Committee of the South African Truth and Reconciliation Commission. While that extra-legal procedure certainly did have the potential for non-legal sanctions of shaming and public opprobrium. See James Gibson (2002) Truth, justice, and reconciliation: judging the fairness of amnesty in South Africa, American Journal of Political Science, 46(3); Dan Markel (1999) The justice of amnesty? Towards a theory of retributivism in recovering states, University of Toronto Law Journal, 49. My own somewhat contrarian view is that it was, unfortunately, the very parallels to a criminal trial that made the Amnesty Committee's longer-term legacy so unfortunate: it generated the public perception that a quasi-trial, without any legal sanction, was all the justice that the victims of Apartheid deserved. See Max Pensky (2008) Transition, amnesty, and social trust: lessons from South Africa, in: Joseph Lewandowski (Ed.), Trust and transitions: social capital in a changing world. Cambridge, Cambridge Scholars Press.

93 See Owen Fiss (1989) The awkwardness of the criminal law, Human Rights Quarterly, 11.

94 This is one possible interpretation of Larry May's ‘international harm principle’ justifying international trials not just insofar as acts threaten international security, but insofar as acts threaten ‘serious harm to the international community.’ See Larry May, Crimes against humanity, 83.

95 Not surprisingly, this social function of criminal trials has been the focus of bitter disagreement amongst analysts of transitional justice. A ‘first generation’ of pro-prosecution analysts, led by Diane Orentlicher, has over the past 10 years (and in large measure following the generally positive experiences of the South African TRC and the mixed-at-best experiences of the international criminal tribunals in the former Yugoslavia and Rwanda and the mixed tribunals in Sierra Leone, East Timor-Leste and Cambodia), been strongly challenged by a new wave of transitional justice experts who see alternatives to prosecution, specifically the social function of truth commissions, as offering a better overall approach. Both sides can muster considerable empirical support. But on the specific question of whether trials generate catalyzing social effects for deliberative procedures, both for the rule of law and for larger, diffuse forms of discourse within government and in the political public sphere, empirical findings are even more malleable. Take, for instance, the counter-argument by Laurel E. Fletcher & Harvey M. Weinstein (2002) Violence and social repair: rethinking the contribution of justice to reconciliation, Human Rights Quarterly, 24, 573–639, on the overall divisive effects of trials on civil society in the former Yugoslavia. Citing a large-scale study of Serbian responses to trials of former leaders, they argue that ‘[t]he study found that all participants sought to represent themselves as members of a national group that was the target of aggression in the conflict. None advanced the view that trials were necessary to learn what crimes ‘their’ forces had committed’ 589.

96 Article 17, which governs issues of admissibility of cases before the Court, is the relevant text of the Rome Statute for testing whether a national amnesty, as part of a larger domestic justice and reconciliation approach, could be interpreted by the Court making a ‘genuine’ effort at prosecution such that the principle of complementarity would be satisfied. While the Northern Uganda situation and the efforts to guarantee immunity to Joseph Kony clearly do not represent a genuine national effort to bring Kony to justice, it is an open question what the Ugandan government could do to end its confrontation with Kony in a way that would meet minimal standards of genuine justice. It might for instance make Kony's immunity contingent on an appearance before some judicial body, might compel him to make sworn promises to refrain from any further violence, might require him to provide sworn testimony, or pay reparations, or any number of steps separate from a criminal prosecution. From the position I am developing here, any meaningful and sustained attempt to include Kony in a deliberative procedure that compels him to give an account of himself and his actions in an institutionalized discourse in which the interests of all affected are included would go a long way to satisfying the admissibility of Article 17 and thus signal to the Court that justice (though not necessarily criminal justice) could be done domestically.

97 The scare quotes around ‘traditional’ here indicate both my belief that reconciliation rituals developed to resolve disputes over livestock are inappropriate for violations of international law, and my sense that such rituals are rarely, if ever, traditional in the sense of a homegrown alternative to ‘Western’ criminal justice measures. The story is always more complex and generally involves the invention of tradition. On the peculiar history of matu oput as the latest nominee for indigenous justice, see Tim Allen, note 3 above.

About The Author

Max Pensky
Binghamton University
United States

Professor, PhD
Department of Philosophy
Binghamton University
The State University of New York


Article Metrics

Metrics Loading ...

Metrics powered by PLOS ALM

Related Content