This study focuses on the debates about holding perpetrators of massive human rights violations accountable. It also focuses on the experience, in Latin America and elsewhere, of attempts to restore truth and justice to the legacy of abuse remaining from the recent past. That experience is necessarily diverse and rich in variations, but it offers some principles of universal applicability.
In only a few years the international community has made considerable progress toward the recognition that a legacy of grave and systematic violations generates obligations that the state owes to the victims and to society. Considerable disagreement remains, however, as to the content of those obligations and as to how they should be fulfilled. This article attempts to show that those obligations (a) are multifaceted and can be fulfilled separately, but (b) should not be seen as alternatives to one another. The different obligations are not a menu from which a government can pick a solution; they are in fact distinct duties, each one of which must be complied with to the best of the government's abilities. In this context, prosecutions and trials, as long as they are held under strict fair trial guarantees, are a necessary and even desirable ingredient in any serious effort at accountability.
This article challenges the view that because democratic leaders know best what their societies need at any given time, the international community should not attempt to impose any rules about what should be done [End Page 255] about the recent past. This article also disputes the view that democratic leaders should strive to restore truth to the analysis of the recent past and, in general, forego attempts to restore justice, at least by way of criminal prosecutions (even while accepting that universal principles govern the problem).
II. Multiple Dimensions of the Accountability Problem
The experience accumulated since the early 1980s on this topic continues to be enriched. For example, the new South Africa is embarking on the most ambitious program to combine truth telling, clemency and prosecution, and eventual reconciliation. Undoubtedly, those who have designed the South African program have benefitted from the Latin American and East European experiences, but it is hard to find a place where so much thoughtfulness and creativity has gone into this grave matter of public policy as in South Africa. The ongoing experiment with justice that the United Nations has begun with the creation of the war crimes tribunals for the former Yugoslavia and for Rwanda has also served to broaden horizons and challenge assumptions. Two or three years from now, analysts will have to reexamine everything said today about truth and justice in light of what these experiments produce.
The accountability problem has legal, ethical, and political dimensions, and it is imperative to recognize and tackle all three. It is a mistake for the human rights movement to allow itself to be painted into the corner of either a "legalistic" or a "moralistic" position. Inevitably, many have seen the movement as grossly uncompromising, intransigent, terribly naive about political realities, vindictive, and opposed to reconciliation. One must, therefore, be ready to take a sober and realistic view of political constraints in proposing accountability measures. But such a view does not necessarily result in realpolitik and surrender of principle. In fact, it is possible to argue that a program of truth and justice is not only the right thing to do, but also politically desirable because it goes a long way toward realizing our idea of democracy.
III. Framing the Questions
The multiplicity of dimensions mentioned above has changed the way human rights organizations conceive their work and how they work to promote and defend fundamental freedoms. They no longer look strictly for the facts that constitute a violation of a universal standard, but trace how governmental institutions respond to each episode. They apply this approach [End Page 256] not only to a recent epidemic of abuses directed against a political enemy, but also to the "endemic violations" present in our democracies: police brutality, rural violence, poor prison conditions, the plight of minorities, and domestic violence. This notion of an institutional response recognizes that abuses will happen even in the most advanced societies, but correctly places the burden on the state to mobilize its resources to restore the imbalance and provide redress. The measuring stick of true commitment to democracy is the degree to which governments are willing to "organize the governmental apparatus and, in general, all the structures through which public power is exercised, so that they are capable of juridically ensuring the free and full enjoyment of human rights." 1
Beyond human rights work, accountability experiences inform the way we think about the related but distinct areas of promotion of democracy, peacemaking, and peacekeeping. Both the need to consolidate a shaky democracy and the need to stop the fighting in a conflict situation undoubtedly condition the possibilities of redressing past wrongs, placing limits on what a policy of accountability can achieve. Those urgent demands, however, by no means diminish the objectives of truth and justice. On the contrary, it is increasingly recognized that making state criminals accountable says something about the democracy that we are trying to establish, and that preserving memory and settling human rights accounts can be part of the formula for a lasting peace, as opposed to a lull in the fighting.
Some commentators have suggested that considerations of accountability for past human rights violations arise only in the context of transitions to democracy. José Zalaquett, in an influential article entitled Confronting Human Rights Violations: Principles Applicable and Political Constraints, 2 framed the discussion in this way, as did the US Institute of Peace in its high quality collection of materials titled Transitional Justice. 3 However, the broad range of contexts in which accountability problems occur suggests that accountability for past abuses must be considered not only in transitions to democracy, but in seeking solutions to armed conflicts as well. Ending conflict situations presents political challenges to accountability issues that are not present in transitions to democracy. For instance, in places like El Salvador and Guatemala, the human rights movement is not [End Page 257] necessarily confronting a regime that has changed, but governments coming to grips with violations committed in great part under their own watch. Different political considerations also arise in attempts to impose accountability through demands of the international community, as in the creation of war crimes tribunals for the former Yugoslavia and Rwanda. Moreover, restricting the analysis to transitions leaves out the approaches taken by organizations of civil society and even by governments to overcome impunity in situations of ongoing violations like in Colombia and Peru.
Another reason to review the analytical framework is that issues of accountability have proven to have lives of their own. They last beyond the short term of what can reasonably be called the transition. Witness the renewal of public debate about what the state owes the families of the disappeared in Argentina after the revelations of naval officer Adolfo Scilingo in 1995, 4 and again in March 1996, on the occasion of the twentieth anniversary of the coup d'êtat. Coming after more than a dozen years of democracy and after the measures taken in the 1980s both to reckon with the past and to attempt to bury it, the issue of the victims' rights in Argentina has far exceeded the limits of the transition.
Even for clear transitional situations, this limited approach to accountability issues begins with the assumption that newly democratic governments are constrained in what they can do to correct past wrongs. 5 While this assumption is correct, unfortunately it has led many to expect too little of what governments can effectively do under the circumstances, and not enough of what they ought to do. In regards to Latin America especially, it has been all too common to think that any attempt to break the cycle of impunity would threaten democratic stability, as if a lesser form of democracy without equality before the law was all to which Latin Americans could aspire.
The framework of transitions has been extremely helpful in shaping debates until now, but a new approach is in order. Latin American democracies now seem more secure (whether they seem so because impunity prevailed or because some accountability has been accomplished is another matter), and experience shows that demands for accountability arise in a variety of historical contexts. It may be time to look at these problems in a broader scope. [End Page 258]
IV. Emerging Principles
A strong legal argument can be made for an emerging principle in international law that states have affirmative obligations in response to massive and systematic violations of fundamental rights. 6 Although existing international instruments do not specify the content of those obligations, the International Covenant on Civil and Political Rights (ICCPR) 7 establishes that each state party to the instrument "undertakes to respect and to ensure to all individuals" 8 the rights it recognizes. The duty to ensure means that states are obliged to take specific steps to redress the wrong committed by each violation of a right. In addition, most instruments establish the right of the victim of a violation to an effective remedy and to equal protection of the laws without discrimination. 9 The UN Human Rights Committee, which is the authoritative interpreter of the ICCPR, 10 has said that blanket amnesty laws and pardons are inconsistent with the Covenant because they create "a climate of impunity" and deny the victims this "right to a remedy." 11 International law also specifies that certain rights are so fundamental that [End Page 259] they cannot be suspended even in the event of an emergency that threatens the life of the nation or its national security. Those "core rights" are the ones that are violated by extrajudicial execution, torture, disappearances, and prolonged arbitrary arrests. 12 Immunity for these crimes constitutes an impermissible ex post facto derogation of rights which could not have been suspended at the time the acts were committed.
Many binding norms of international law point in the direction of an obligation to overcome impunity for crimes of this kind. The Genocide Convention 13 establishes the obligation to punish. 14 The more recent Torture Convention 15 obliges its signatories to make torture punishable within their domestic jurisdictions, to arrest suspected torturers, to extradite them to other jurisdictions or to prosecute them, and to cooperate fully with the prosecuting jurisdiction in the gathering and preservation of evidence. 16 Other conventions and customary norms rule on the inapplicability of statutes of limitations to crimes against humanity, on the inapplicability of the "political offense" defense against extradition for such crimes, and on universal jurisdiction to prosecute them. 17
Taken as a whole, these scattered norms point unequivocally to a trend in international law to punish the perpetrators of these crimes. In fact, it is hard to find disagreement on the point that the crimes' occurrence gives rise to [End Page 260] certain obligations; if anything, the disagreement (or skepticism) is on the content of the obligation or on its justiciability--though the latter is a problem for all international law obligations. Observation of this trend and other trends to expand universally applicable norms suggests that new principles are emerging. These principles hold that a state is obliged to carry out a number of tasks in response to crimes against humanity. These tasks are:
1. to investigate, prosecute, and punish the perpetrators;
2. to disclose to the victims, their families, and society all that can be reliably established about those events;
3. to offer the victims adequate reparations; and
4. to separate known perpetrators from law enforcement bodies and other positions of authority.
Some of these obligations are present when the state violates any right set forth in the universal instruments. The whole complex of obligations, however, applies only to situations of massive and systematic violations of the most basic rights to life, liberty, and physical integrity. In other words, a single case of torture gives rise to these obligations only if it is part and parcel of a systematic pattern of similar violations. The reason for this heightened protection is that human rights violations of this magnitude, when committed massively and systematically, are crimes against humanity.
From the point of view of individual and collective persons entitled to a specific duty from the state, the state's obligations correspond to a set of rights:
1. a right of the victim to see justice done;
2. a right to know the truth;
3. an entitlement to compensation and also to nonmonetary forms of restitution; and
4. a right to new, reorganized, and accountable institutions.
Society at large and not the victim is the titular head of this last right; for the first three, the right lies first and foremost with the victims and their families, and then with society.
The reference to these correlative rights and obligations as "emerging principles" and not as binding international law obligations signifies their present status: only in part do they find justification in existing norms of universal applicability. These rights and obligations result primarily from the recent expansion of existing norms through the creation of law, particularly international law in the form of nonbinding resolutions, judicial and quasi-judicial precedent, the practice of nations, and opinio juris. For example, [End Page 261] the Geneva Conventions of 1949 18 clearly establish an obligation to punish "grave breaches" or war crimes that happen in international conflict. In the last year alone, that notion has been extended to similar crimes committed in the context of conflicts not of an international character. The Security Council resolution establishing the International Criminal Tribunal for Rwanda and the landmark jurisdictional decision of the International Criminal Tribunal for the former Yugoslavia in the Tadic case 19 both led to this extension. With respect to the right to know the truth, even though the international community has only recently begun discussing this right, a recent meeting of experts convened by the United Nations has argued that the right to know the truth has achieved the status of a customary international law norm. 20
These examples indicate that the law on this issue is developing rapidly. The fact that these principles are not "hard law" in all their aspects, furthermore, does not mean that they do not constitute obligations. In most cases it may not be possible to obtain a judgment ordering performance of these duties, and such a judgment would be hard to enforce in any case. Nevertheless, these principles can be invoked to advocate certain measures by states that like to see themselves as contributing to an international lawful order. More importantly, a government that confronts a situation of massive state crimes can and should be judged by how much it attempts to do in order to comply with these principles. [End Page 262]
B. The Obligations Are Separate and Distinct
A second observation with regard to the four state obligations is that each of them is both integral to a fair policy of accountability and yet separate and distinct from the other three. Every government should strive to comply with each one of these obligations, and a high measure of compliance in one area does not excuse noncompliance in another. For example, the Menem administration in Argentina has enacted a comprehensive and generous policy of monetary compensation available to victims of the "dirty war" of the 1970s, but does little to tell each family of the disappeared what can be known about the fate and whereabouts of their loved ones, and then only upon request and in the most quiet of ways. 21 The government has done even less to purge the armed and security forces of the many perpetrators that continue to serve and advance through the ranks.
The separate and distinct nature of these obligations also dictates, however, that if one of these duties is rendered legally or factually impossible, for example by a blanket amnesty law which prevents criminal prosecutions, the other duties remain in full force. The UN Human Rights Committee, in its latest periodic review of Argentina, 22 rightly rejected the argument put forth by Minister of Justice Rodolfo Barra that the pseudo-amnesty laws of the 1980s and the presumption of innocence prevented the government from forcing known criminals into retirement by administrative or disciplinary procedures. 23 Similarly, those laws may be insurmountable barriers to criminal prosecution, but they do not relieve the Argentine government from its duty to use all the means at its disposal to tell each family what can be known about the fate and whereabouts of the disappeared. 24[End Page 263]
This analysis of the four obligations and their correlative rights has the advantage of allowing the international community to insist on certain measures beyond the artificial attempt by governments to draw a line on the matter and "move on." The approach should also serve as a way for the human rights movement to avoid self-defeatism and reject all solutions because one of them, punishment, becomes unavailable. While continuing to condemn amnesties and pardons as inconsistent with these obligations, the victims and society can still demand the complete truth, reparations, and law enforcement bodies that are effectively purged of criminals.
C. Obligations of Means, Not of Results
A third and final observation about these principles is that, in all four cases, they constitute, in the familiar distinction made by civil law, "obligations of means" and not of "results." This means that a state fully complies with its duty to punish even if the trial results in acquittal, as long as the prosecution has been conducted in good faith and not as a ritual "preordained to be ineffective." 25 In that regard, these obligations are subject to conditions of legitimacy in their performance. 26 These conditions constrain governments in two opposite directions. In the first place, governments are mandated to make good faith efforts to achieve the desired results. It follows that, if a government has leads, documentation, known actors whose testimony can be obtained, and so on, it has a duty to do all that is reasonably within its means to tell each family the truth. In this context, however, "reasonably within its means" signifies an affirmative organization of efforts and tasks that must be placed in the service of truth, not merely a pro forma act of bureaucratic compliance. 27
In the opposite direction, another condition of legitimacy is that these efforts must be conducted in full respect for due process standards mandated by international law. 28 What process is "due" depends, however, on the object of the exercise and its likely result. 29 Prosecution and punishment require the highest degree of deference for the rights of the accused because the outcome may be the loss of liberty of a person. 30 On [End Page 264] the other hand, disciplinary or administrative separation from the armed or security force requires, arguably, a lesser standard of due process. 31
Considerable dispute exists in this regard on the issue of whether truth commissions should "name names" of officials accused of the violations they describe. Zalaquett objects to naming names under all circumstances, considering it a violation of the due process rights of the persons so named and also an invasion of judicial functions that truth commissions, by definition, do not exercise. 32 An absolute position against naming names, however, can in certain circumstances result in an unacceptable limitation of the "full truth" that governments are bound to disclose and that truth commissions are charged with rendering.
The legitimacy of a decision--either to name names or to withhold them--depends on whether prosecutions and trials (and with them a more exhaustive exploration of the truth) are available after the truth commission issues its report. If the possibility of trials is open, it is perhaps a good idea to allow the courts to rule on matters of personal criminal liability after a fair trial takes place--a fair trial necessitating the withholding of names in the truth commission report. If, on the other hand, the report is likely to be the last opportunity to air these matters, an honest deference to truth suggests the need to disclose reliable information about the behavior of certain individuals because they can hide behind the impunity given to them by amnesties or pardons. Even in that case, however, some measure of due process is necessary: at the very least, the truth commission is required to give them a chance to rebut the incriminating information. If names are going to be named, it is also important that the truth commission deal honestly and impartially with the information and be perceived by the public as having done so. Because the Truth Commission for El Salvador was widely seen as having received many more names than it published, the duty should have been incumbent upon it to be more clear and forthright as to the criteria by which some names were published while other names were suppressed. Possibly improving on that experience, the South African Truth and Reconciliation Commission is charged not only with investigating and disclosing the circumstances of each case, but also with identifying the perpetrators. 33 At the same time, however, the statute that created the Commission requires that persons accused of human rights [End Page 265] violations be given a chance to respond before their names find their way to the final report. 34
V. Avoiding False Dilemmas
A. Misconception One: There Are No Rules for States
The first misconception about the accountability problem is that there are no rules governing what states must do in response to massive violations that occurred in the recent past. Early in the 1980s, this view prevailed among many democratic and even human rights minded observers of Latin America. This mind-set resulted in a lack of support for efforts to inaugurate an era of accountability in the new wave of democratization. The views of former President Jimmy Carter most prominently reflect this position in its present incarnation. The position not only defers to the judgment of democratic leaders but expresses a preference for settlement of disputes based on forgiving and forgetting by fiat. In fact, President Carter most recently applied this idea in his offer of amnesty to General Raoul Cedras as a condition for allowing democracy to return to Haiti, an offer made over the express objections of Haiti's democratically elected leader, Jean-Bertrand Aristide.
President Carter explains that he is less concerned about the past violations than about avoiding "the next ones." Even so, it is far from proven that a policy of forgiving and forgetting automatically deters future abuses. In fact, at least in Haiti one can more easily make the case that the opposite is true: each self-amnesty by the military has only led to further interruptions of democracy and to further atrocities. 35 This deference to democratically elected leaders, who supposedly know better than anyone what is best for their country and what the traffic will bear, is unwarranted. In the mid 1980s, such deference allowed Guatemala's President Vinicio Cerezo to get away with actively impeding any attempt to invalidate the shameless self-amnesty with which his military predecessors left office; at the end of Cerezo's term, members of the military committed more violations of human rights with further impunity. 36[End Page 266]
B. Misconception Two: Truth Is Always Preferable to Justice
A second pernicious position in this debate postulates that, even in the context of trying to settle accounts, truth is always preferable to justice. Columnist Charles Krauthammer, relying in part on Zalaquett's ideas and presumably extolling the virtues of the Chilean and South African experiences, has recently offered as a proposition that "truth reports" should be written but no trials should take place. 37 With respect to the crisis in Bosnia, some observers have proposed closing down the war crimes tribunal and replacing it with a "truth commission" modeled after those in Chile and El Salvador. 38
In arguing that truth telling always promotes reconciliation while trials are vindictive, Krauthammer simplifies the facts about the positions espoused by Chile's new democratic leaders and by Nelson Mandela. He fails to take into account that Chile successfully prosecuted and convicted Pinochet's right-hand henchman, General Manuel Contreras, and that Mandela has allowed justice to take its course in the case against former defense minister, General Magnus Malan. Contreras is serving time in jail and Malan spent several months behind bars before being acquitted, while Mandela steadfastly refused to intervene. Does that mean that Aylwin and Frei and Mandela have turned vindictive?
It is hard to see what could possibly be gained by another "truth report" in the Balkans, after UN thematic and country-specific rapporteurs have documented the crimes so painstakingly while the international community prefers not to hear. The United Nations has had a "truth commission" for the former Yugoslavia already, in the form of the commission headed by Professor M. Cherif Bassiouni, which preceded the creation of the special tribunal. 39 Incidentally, every eminent jurist appointed by the United Nations has proposed that the next inevitable step should be to do justice; Robert Pastor and David Forsythe suggest that the human rights movement should ignore these recommendations until there is a commission that proposes immunity. 40[End Page 267]
It is of course true that a process of accountability that neglects or downplays the truth would be unacceptable. Zalaquett even includes full knowledge of the truth as one of his "conditions of legitimacy." 41 A policy that might result in convictions without a full airing of the facts, say by plea bargaining, would not pass muster. That is not the same as saying that truth is preferable or superior to justice, or even that truth telling should come first. 42 Other than in the unlikely plea bargaining scenario, a trial without full discussion of the facts would violate due process in any event. More importantly, posing the question as a contest between truth and justice fails to give credit to prosecutions for their specific contribution to the public's knowledge of the facts.
In the position exemplified by Krauthammer, Pastor, and Forsythe, truth is actually proposed as an alternative to justice. In fact, the best exercises in truth telling so far have not been predicated on the prospect of immunity from prosecutions. Both the Sabato commission in Argentina and the Rettig commission in Chile withheld names of accused perpetrators in their final reports, but submitted them with the relevant evidence to the courts as a way of contributing to justice. 43 If the commissions had been charged with telling the facts as the last step in accountability, their reports without names probably would have been received less favorably by the public; the "truth" they were telling would have been perceived as a preordained and incomplete one and not as a truth that lets all the facts speak for themselves, and leading to wherever it should.
The third Latin American exercise in truth telling was, in fact, the last word to be heard on accountability in El Salvador. The three international jurists who formed the UN-sponsored Truth Commission decided to be silent in their report on the matter of an eventual amnesty, despite the fact that the Salvadoran government had talked openly about one. 44 In press statements in the days following publication of their report, each of the commissioners specifically argued against a blanket amnesty, but to no avail. President Alfredo Cristiani and the Congress dominated by his party immediately passed an unconditional amnesty law. Whatever the intentions of the Truth Commission members, Cristiani's amnesty actually took away much of the weight that the report otherwise would have had in Salvadoran [End Page 268] society. Despite the subsequent impunity, if the report did contribute to peace in that country, it was precisely because the investigation that preceded it was not premised on a preordained amnesty. In contrast, a truth report designed as an alternative to justice would have been an exercise in tokenism, and as such it would have been doomed to be forgotten rather quickly by the society it purported to serve. Justifiably, the public expects the truth telling to be a step in the direction of accountability, not a poor alternative to it.
This is not to say that we should reject truth commissions unless prosecutions and punishment are also on the horizon. As stated earlier, if punishment is rendered legally or factually impossible within legitimate conditions, the state is still obligated to investigate and disclose the facts, to acknowledge the wrongs committed in its name. However, that situation is different from the one in which governments sanction impunity and, as a gesture to the victims, give them a report instead, hoping that everything will soon be forgotten.
The success of national truth commissions has prompted the international community to use the model in situations in which a state or its citizens calls on the community to help put an end to a conflict. The United Nations proposed and funded the Truth Commission for El Salvador and, more recently, the one in Haiti. The former was composed only of non-Salvadorans; the latter included a minority of three non-Haitians. Of these, the human rights movement generally considers the Salvadoran effort successful because its report is a credible, persuasive, and honest appraisal of the atrocities committed during the twelve year war in that country. Its silence over the amnesty question and the controversy over names named and names not named are negative factors. More importantly, the fact that the commission was formed and staffed exclusively by non-Salvadorans also limits its value: the different sectors of Salvadoran society see the final document as yet another external "report" on their country and have not taken stock of its findings. 45 The Haiti Truth Commission is much more decidedly a failure. After months of deliberation it resolved not to publish its report and to release only a few pages of inane "recommendations." Six months later the Commission did release its full report, but printed and distributed only a small number of copies. A truth commission that withholds publication of its own findings is at the very least a contradiction in terms. 46[End Page 269]
The UN involvement in the efforts in Haiti and El Salvador does not explain their partial failures. In addition to being simplistic, such an assessment would be terribly unfair to international commissioners and staff who tried their best to make the most of very difficult situations. It is fair to say that truth commissions succeed when they are seen by the public as an effort by a national community to come to grips with its own reality of atrocious behavior by its members toward one another. In this regard, President Nelson Mandela, heeding the advice of a community panel that proposed candidates for the Commission on Truth and Reconciliation, correctly decided not to use the slots for foreign members contemplated in the statute and filled the Commission exclusively with South Africans. This conviction that truth telling efforts are primarily a national enterprise leads one to be highly skeptical of a recent proposal to create a UN-sponsored "permanent truth commission." 47 As mentioned elsewhere, the special rapporteurs, working groups, and other UN mechanisms do an uneven but generally adequate job of describing the facts and analyzing them in light of international human rights law; there is no need for a permanent body which would only duplicate them. What matters most to the international community is seeing justice done. The proposal for a permanent truth commission does not intend to belittle the importance of the success of the tribunals for the former Yugoslavia and for Rwanda, but it does miss the point that the major challenge ahead is to uphold the decisions of those tribunals and not to let world leaders find excuses to ignore them.
To the extent that the "truth as an alternative" position cites Zalaquett, it misrepresents his views. Zalaquett has frequently agreed that crimes against humanity give rise to an obligation to punish, even while arguing that this concept needed further elaboration. 48 He has also never suggested that states should not punish when they can, but rather that the rest of the world should not hold them to a high standard of achievement in that regard. In addition, he has praised President Aylwin for not broadening the scope of immunity in Chile, as opposed to the pseudo-amnesty laws passed by democratic governments in Argentina and Uruguay. 49[End Page 270]
It would be a mistake, however, to assume that Krauthammer and others have simply not read Zalaquett carefully. Zalaquett's articles do favor truth telling over justice and reflect skepticism about prosecutions. His insistence that leaders who must reckon with a legacy of human rights violations must be guided by an "ethics of responsibility" reinforces that impression. 50 Zalaquett quotes Max Weber for the need to follow an "ethics of responsibility" as opposed to an "ethics of convictions" 51 (also translated as an "ethics of ultimate ends"). 52 What this distinction actually adds to the debate is unclear because, according to Zalaquett, Weber readily acknowledges that, just as an ethic of responsibility does not imply a lack of conviction, neither does an ethic of convictions imply a lack of responsibility.
Weber has contributed greatly to modern sociology, but the incursion into moral philosophy that Zalaquett cites is less convincing. That every person should act responsibly (in the sense of measuring the likely though unwanted results of one's own actions) does not say much about the intrinsic morality or immorality of those actions. In fact, it leads to judging those actions exclusively by those unwanted results and not by the purposes or means of human conduct. In this sense, Weber's dichotomy aligns Zalaquett with a consequentialist philosophy that is increasingly called into question. More precisely, applying the "ethics of responsibility" to issues of accountability means that going too far in the direction of accountability creates the serious risk that the enemies of justice, who still retain residual power, will interrupt democracy again and return to a policy of human rights violations. That may well be so, but an "ethics of responsibility" analysis unfairly burdens the well-intentioned democratic forces with all the consequences of a bad result--a result which is owed primarily to the behavior of other actors, and only secondarily, if at all, to that of the democrats.
It is easy to agree that it makes no sense to urge leaders to act recklessly or irresponsibly. Insisting on punishment without due process and in disregard for the principles of res judicata, for example, would be irresponsible not because of the consequences (the reactions of the enemies of democracy), but because the quest for justice would thereby surrender the moral high ground of principle. In the absence of insurmountable legal obstacles to prosecutions, however, the problem lies with establishing the limits of what can be accomplished within the cards that the particularities of each transition have dealt each society. Urging leaders first and foremost to "be responsible" seems to open a large door to excuses for inaction and [End Page 271] for accepting the status quo of impunity for egregious abuses. Worse, it conveys the message that it is highly ethical to govern by yielding to the blatant blackmail of powerful undemocratic forces.
In particular, setting a universal standard of "responsible leadership" without simultaneously stressing the debts that are owed to the victims assumes that prosecutions are inherently destabilizing, while a truth commission report will presumably be swallowed more easily by the enemies of democracy. Experience shows, however, that those enemies are probably just as terrified at the prospect of truth being revealed as they are that some of them will have to face trials. It would be best if the international community strongly set forth the duties owed to the victims and to society, and then carefully and in a particularized way analyzed what can be done and what would be irresponsible to attempt.
The emphasis on the limits as opposed to the possibilities also assumes that whatever is done is the most that could have been done under the circumstances. Although each transition is certainly different, this cannot possibly account for the wide gap between how much accountability was accomplished in Argentina compared to Uruguay or Brazil, for example, unless one incorporates the factor of the relative adherence of political leaders to the values of human rights and the rule of law, as well as the ability of the human rights movement (domestic as well as international) to push its agenda onto the national debate. Highly respected democratic leaders like Julio Sanguinetti and Wilson Ferreira of Uruguay could claim to have been acting responsibly when they lent their influence to a policy of immunity that yielded to the blackmail of the Uruguayan military. 53 In retrospect, it would have been helpful if Sanguinetti and Ferreira would have heard a stronger voice of support from the international community for the efforts of the Uruguayan victims and human rights community to set the record straight.
Undoubtedly, to insist on prosecutions in the presence of an important legal obstacle like a preexisting amnesty law that has had firm legal effects would be irresponsible, because it would subvert the very rule of law that [End Page 272] the human rights movement proclaims and because it would violate the cardinal principle of nullum crimen nulla poena sine lege (defendants are always entitled to be judged by the criminal law most benevolent to them that exists at or after the time of the commission of the crime). Advocating amnesties and pardons to be enacted by democratic authorities is quite a different matter. In the case of Chile, one can accept the objective limitations that Pinochet's orderly retreat has put on the new democratic government, without prejudice to criticizing the self-amnesty of 1978 as a shameful act. The international community should praise Chilean society for its ability to produce a great deal of accountability within those objective limitations, but there is no need to turn necessity into virtue.
C. Misconception Three: Prosecutions Are Inherently Inimical to Peace and Reconciliation
A frequent misconception in this debate is that prosecutions are inherently inimical to peace and to reconciliation. Protocol II to the Geneva Conventions 54 (applicable to internal conflicts of a particularly intense nature) actually promotes broad and generous mutual amnesties to put an end to the conflict. 55 That amnesty, however, refers to the offenses of rebellion or sedition and to comparably minor infractions of the laws of war on the governmental side. It is not meant to encourage immunity for attacks on civilians or for serious crimes against life and the integrity of the person of the adversary. For "grave breaches" of the laws of war, on the contrary, there is a clear obligation to punish. 56 Beyond the question of a binding legal obligation, however, the threat of prosecution can be a clear disincentive for actors in an armed conflict to give up their resort to violence. At the same time, it does not necessarily follow that the interest of peace can only be served in these cases by a blanket amnesty for both sides. In fact, one can make a strong argument that a lasting peace is only possible if the process by which it is attained carefully and honestly addresses human rights and laws of war violations by all sides.
That the object of the whole exercise is to obtain reconciliation in societies torn by conflict should be undisputed. Unfortunately, the human [End Page 273] rights movement has been reluctant to embrace reconciliation as a goal, perhaps because as used by the proponents of immunity, the word has achieved a bad name. The net result, in any case, is that human rights organizations have been unjustly labeled as enemies of reconciliation and obstacles to leaving the past behind. Human rights organizations should, instead, adopt reconciliation as their own agenda, but insist on its true meaning. In the first place, true reconciliation cannot be imposed by decree; it has to be built in the hearts and minds of all members of society through a process that recognizes every human being's worth and dignity. Second, reconciliation requires knowledge of the facts. Forgiveness cannot be demanded (or even expected) unless the person who is asked to forgive knows exactly what it is that he or she is forgiving. Third, reconciliation can only come after atonement. It seems to add a new unfairness to the crimes of the past to demand forgiveness from the victims without any gesture of contrition or any acknowledgement of wrongdoing from those who will benefit from that forgiveness.
A final objection to prosecutions is that they stand no chance of addressing all the possible violations or of trying every single perpetrator. This inevitable selectivity discredits the effort because it instantly smacks of discrimination and favoritism. 57 The risk of selectivity is present in truth telling exercises as well, although this fact does not seem to bother proponents of truth reports as alternatives to justice. Selectivity is certainly inevitable, but it is also part of the rules of the game of trials. No system of justice in the world even pretends that it punishes each and every case of antinormative behavior. In cases of this sort, an initial self-selecting factor exists in the fact that the evidence needed to initiate prosecutions will plainly be lacking in many cases. If all perpetrators were bound to be convicted and punished without regard to the evidence, the trials would not be consistent with the rule of law. Cases can be lost for lack of evidence and should be lost if fair trial guarantees are violated; this chance element and especially the fact that all defendants have the opportunity to prevail is what distinguishes fair trials from mockeries of justice. 58 Beyond that, a certain selection on the basis of degrees of culpability is not only necessary but quite legitimate. There is nothing wrong with selectivity as long as the rules are clear and do not discriminate on the basis of a proscribed category. 59 Those rules should also be reasonable and clear to the public, and not [End Page 274] subject to change to suit the political needs of the moment or in response to undue pressures on the democratic leadership. 60 Human rights activists largely agree that prosecutions should start at the top where possible, without, however, allowing "due obedience" to orders as a defense where a clear opportunity to resist an immoral order was available.
VI. Arguments Favoring Prosecutions
A. Moral Justifications for Prosecutions
These ideas are not meant to establish an absolute preference for prosecutions over truth telling or the other two state obligations for dealing with accountability. 61 As stated earlier, the state must meet all four obligations to the best of its abilities. However, the obligations definitely conflict with a rationale that makes prosecutions less preferable. One cannot question that prosecutions are the hardest choice of the four paths and should therefore be thoroughly justified in moral terms.
In this sense, a position that justifies prosecution exclusively on the basis of its deterrent effect does not provide a sufficient foundation on which to advocate trials. As Aryeh Neier has said repeatedly, deterrence of future violations is unreliable as the foundation of punishment because one cannot predict the future behavior of the relevant actors. 62 Societies can only hope that punishment will deter the transgressor as well as other potential offenders, but they can never assume deterrence. As the human rights [End Page 275] movement justifiably criticizes the dogma that accepts the death penalty as an effective deterrent to crime when statistical information proves the contrary, the movement cannot just as easily affirm that prosecution and punishment of state crimes will prevent future violations by the same or by other state actors. In some country-specific situations one can certainly show that a policy of impunity--by way of repeated amnesty laws or simply de facto refusal to investigate crimes of the security forces--results in encouragement of further human rights violations. 63 But the converse is not necessarily true. It may well be true that in a given situation a policy favoring forgiveness is better suited to avoiding the recurrence of egregious violations, but that proposition cannot be proven categorically.
Together with deterrence, retribution is the object traditionally assigned to criminal punishment. For the present purposes, it seems unnecessary to "sign up" to one theory of punishment or the other. Provisionally, however, deterrence alone does not seem to explain why societies punish, even though deterrence may well be the goal that is hoped for each time a criminal sanction is meted out. Societies punish offenders in some situations in which recurrence is unlikely and in which the deterrent effect over the conduct of others is not demonstrable.
That modern penology strives for rehabilitation of the offender does not contradict a theory of punishment that recognizes some place for retribution because rehabilitation is the object of the penalty, not the reason why the behavior is penalized. In any event, one need not see retribution as a policy of vindictiveness. In its true form, it simply says that society does not brook behavior that breaks the rules. This policy is all the more important when those rules protect the innocent and defenseless. An enlightened theory of punishment, therefore, puts the victim at the center of the need to redress wrongs: societies punish because they wish to signify to the victim that his or her plight will not go unheeded. 64 This is not tantamount to resurrecting a theory of "victims' rights" which, in the United States, is often used to tie the hands of judges in imposing penalties, including capital punishment, and on occasion even invoked to justify removing a prosecutor from a case because the prosecutor's "toughness" does not match that of the politician. 65 Victims do not have a right to a certain form or amount of penalty, but they [End Page 276] do have a right to see justice done by means of a process. Zalaquett's concern, therefore, that victims should not have a veto over how society decides to punish is misplaced. Majorities in society do not have a right to tell the victims that their cases will be forgotten for the sake of a higher "good." Victims have a right to a process that fully restores them in the enjoyment of their rights and in the dignity and worth that society owes to each of its members. Clemency, if advisable, can only come after that process is fulfilled.
A moral justification of punishment offered in Argentina is probably a variation of a "just deserts" theory, but it is especially attractive in this setting. The theory holds that societies must punish acts of torture, murder, and disappearance out of respect for the norm that prohibits such conduct. The purpose of punishment reaches beyond merely restoring the rule of law and of doing so simply because the rule of law protects the individual against other powers in society. The heightened value of those particular norms (i.e., prohibiting torture, state-sponsored murder, and disappearance) gives rise to the duty to punish them. 66 In countries striving to move out of dictatorship and authoritarianism, this argument tells much about the new society that the people are trying to establish. In those cases, therefore, societies punish torture, murder, and disappearances out of a desire to draw a "thick line" with the past: from now on there will be no privileged defendants, the plight of victims will not go unrecognized, and the abuse of power will be checked. 67
A final argument for prosecutions is that they are the most effective means of separating collective guilt from individual guilt, and thus to remove the stigma of historic misdeeds from the innocent members of communities that are collectively blamed for the atrocities committed on other communities. Judge Richard Goldstone, the General Prosecutor for the international criminal tribunals for the former Yugoslavia and Rwanda recently made an eloquent plea to that effect. 68 This argument is especially applicable to situations where the human rights movement seeks to break the cycle of ethnic violence because trials would allow the victimized communities to distinguish between ordinary members of rival ethnic groups and those who manipulate their fears for political ends. It also applies, however, mutatis mutandi, to countries in which the abuses were not linked to ethnic strife. In Argentina, for example, the civilian population [End Page 277] might today be more reconciled with democratic armed institutions if the pseudo-amnesties had not allowed the relative minority of very culpable officers to seek refuge in a misguided esprit de corps. As it happened, it is taking a long time for Argentines to recognize that it was a gang of murderers in uniform and not necessarily the armed forces as such that committed the massacres of the 1970s.
B. Trials and Memory
Even from the perspective of the all-important goal of telling the truth where the killers cling to silence and denial, it does not seem apparent that the report of a truth commission is automatically more effective than trials. There are two clear advantages to truth commissions in this regard: one is the concentration of effort in a limited time and the capacity to assemble information from varied sources; the other is the process that is usually established by which the victims and their families are "listened to" and respected, being treated with dignity unlike they were treated before. In contrast, courts necessarily need to take on cases in a relatively piecemeal fashion, must treat victims more or less neutrally as witnesses, and are constrained in their assembly of evidence--and even in their analysis--by stricter rules of admissibility. 69 These comparative advantages in favor of truth commissions, however, assume that the exercise is conducted thoughtfully, in good faith, and with adequate time and resources, which is not always the case.
For their part, trials offer their own advantage in promoting a measure of truth and acknowledgement. The adversarial format, with the ability to compete with equal arms in the establishment of the truth and to confront and cross-examine the opponent's evidence, results in a verdict that is harder to contest. The judicial approach to evidence is certainly not infallible, but the truth thus established has a "tested" quality that makes it all the more persuasive. This notion also presupposes fair trial guarantees, but as stated earlier, the international community should reject any effort that falls below that standard.
Trials contribute to truth, however, only if they are used for what trials [End Page 278] are traditionally intended. Any attempt to turn them into the site of "historic" judgments, or the instrument to "settle" long-standing political, social, or ideological conflict runs the risk of a double failure. The trial can result in a mockery of justice and in a contrived and unsatisfactory "truth." In order to serve the purposes of truth, a court must strictly observe due process guarantees and restrict its analysis to the principles of criminal law and the law's insistence on individual responsibility for each person's conduct. Strict adherence to principles of criminal liability does not exclude convictions on theories of command responsibility, "control of the episode" (dominio del hecho), and even conspiracy, so long as each element of the crime is scrupulously established in the evidence. Adherence to such principles does exclude convictions for forms of collaboration, sycophancy, and cheerleading that may be morally contemptible but not criminally illegal at the time they took place. It follows that an attempt at sweeping historic "settlement of accounts" can result in a miscarriage of justice. 70 To quote another important author,
[j]ust as belief belongs in church, surely history education belongs in school. When the court of law is used for history lessons, then the risk of show trials cannot be far off. It may be that show trials can be good politics . . . [b]ut good politics don't necessarily serve the truth. 71
In fact, it is useless to try to settle a dispute over history by way of trials because history cannot ever be "officially" written with such an effect as to end all disputes about its interpretation. The most to which the human rights movement can aspire is to get the facts right, so that arguments can go on about their meaning for as long as necessary. Mark Osiel argues that trials can nonetheless be of great service over a continuing disagreement, by moving a conflict outside the realm of violence and constraining it within nonlethal bounds. 72 Osiel calls this approach liberal, and distinguishes it from Durkheim's belief that criminal trials can obtain settled consensus over moral fundamentals as well as from the postmodern view that any settlement is by definition impossible, and permanent disruption should therefore be celebrated. The merit of Osiel's argument is that it inspires respect among adversaries and promotes social solidarity as all sides accept the rules by which the facts are going to be established and responsibilities determined. The continuing disagreement can also proceed over a set of uncontested facts, and the approach "confront[s] those with something to hide with evidence they have tried to keep from coming to light." 73[End Page 279]
VII. Universal Principles and Different Solutions
One can clearly distinguish the problems of the transition in Eastern Europe and its own legacies of human rights violations of the communist era from those of Latin America. The most salient features of repression in Latin America were extrajudicial executions, disappearances, and torture over a relatively short but tragic period in the life of each country. For those actions, either the state or the international community can single out a manageable number of victims and, more importantly, identifiable perpetrators, instigators, and masterminds with relative ease. In contrast, repression in Eastern Europe lasted several generations. It imposed a pervasive social control through surveillance and informer networks, and asphyxiated dissent through professional and social ostracism; but in Eastern Europe, the repressive forces resorted to violations of the right to life and physical integrity much less than in Latin America. 74 In such a context, criminal responsibility is harder to establish when the actions were not illegal under the laws in force at the time, and it is perceived that moral responsibility is shared by large sectors of the population. For this reason, human rights activists in those countries have generally favored memory over punishment. 75 South Africa also embarked on an ambitious program of truth and reconciliation as part of its own attempt to leave apartheid behind. Arguably, repression in South Africa has contained the worst features of that of Latin America and of Eastern Europe; this makes the problems of accountability all the more complex there.
Nevertheless, the four obligations outlined in Part IV.A are universally applicable within the conditions of legitimacy also mentioned. In the first place, the obligation to punish that is postulated here applies to violations so serious that they qualify as crimes against humanity, not to acts of pervasive surveillance, denials of free speech and association, short term arbitrary detention, denial of due process, and suppression of religious freedom. Even for these, however, concern for the victims does require that the truth be known and acknowledged. The right to reparations and the obligation to purge the security forces of elements steeped in those abject traditions also apply.
Because the human rights movement must insist on scrupulous respect [End Page 280] for due process, the "lustration" laws that have been applied in different forms in Eastern Europe almost always constitute punishment without a fair trial, 76 and for that they deserve firm rejection. On the other hand, the human rights movement should encourage Eastern Europeans to find ways of having the full truth known and disseminated without adverse legal consequences to individuals who cannot defend themselves at trial. Governments must make the files of the intelligence services available to the public; the mistake is to take at face value the "truth" that they describe. A process that allows individuals to demonstrate the falsehood of the substantive information contained in those files can be incorporated into the decision to release those files. Democracy and the rule of law demand that state files do not remain secret; their disclosure will in fact serve the purposes of truth and reconciliation as long as it is balanced with the rights to privacy and reputation of the individuals who may see themselves stigmatized by the documents' publication.
Similarly, the human rights movement must criticize attempts at prosecution under contrived principles of what was legal and what was illegal under the laws applicable at the time. With equal firmness, the movement must condemn attempts to punish small cogs in the wheel of repression while allowing the head planners and perpetrators to go free. 77 The limitations and dangers of using courts to settle historic grievances can easily be applied to the legal contortions that seem to characterize some German attempts to punish human rights violations by East German officials. 78 Those "false starts" do not invalidate all attempts at accountability in Eastern Europe, nor do they justify the preference by many to sweep the recent past under the rug. In particular, truth telling, integral reparations, and purging security forces of known criminals should proceed whether or not there are trials; and for particularly egregious human rights violations, like the murder of Father Jerzy Popieluzko, prosecution and punishment in a fair trial is in order. [End Page 281]
Proponents of accountability have gained a lot of ground in the last twelve years. The theme is firmly established in the agenda of all major challenges of our time. A few ideological battles must still be waged, however, especially to overcome the lack of imagination and vision that often passes for prudence and realpolitik. As in the past, it is not enough to insist on moral principles. The international community must acknowledge the political constraints while insisting that everyone look at them objectively and without preconceptions. The most important point is not so much to impose a set of obligations upon democratic leaders, but to find the means by which the international community effectively supports the efforts made by some of those leaders--and by organizations of civil society--to achieve accountability. In a world marked by globalization, such encouragement is crucial to the legitimacy and acceptability of those efforts, and it has been sorely missing.
It is also time to review the theoretical framework under which the international community has discussed these issues. The presentation of the moral principles and the political constraints made early on by José Zalaquett and Aryeh Neier, among others, were immensely useful when the human rights movement confronted new situations with relatively little experience to guide it. These analyses will continue to be useful, to be sure, but after the rapid development of many new experiences, it is time to review all of their assumptions and to examine their continued applicability.
Juan E. Méndez is the Executive Director of the Inter-American Institute on Human Rights, though the arguments espoused here are his own. Previously, he was General Counsel at Human Rights Watch. Mr. Méndez wrote this article when he was a Visiting Scholar at the Helen Kellog Institute for International Relations of the University of Notre Dame.