How commissions arrive at their findings is deeply if subtly influenced by how they gather their information. The methods, focus, and availability of [End Page 36] data all shape the ultimate conclusions. This process is beset by complexity and countervailing pressures, and the responses of the commission to these pressures occasionally create biases and inconsistencies. All commissions suffer from biases of this kind, as the examples below make clear.
A. Misspecifying Historical Patterns at the CEH
One of the core concerns in an assessment of "macro-truth" is to identify when violence occurred. Fewer than 2 percent of the 24,910 killings documented by the CEH occurred in the period 1960-1977. 118 Yet in the CEH projection of the total number of dead, about 13.6 percent of the estimated killings occurred in this period; in effect, the CEH undercounted killings before 1977 by more than a factor of six. 119 "Undercounted," in this context, means more than simply not having statistics on these people. Their stories were not heard, which means that the social conflicts in which these people were killed were not documented by primary sources, and consequently played a less prominent role in the analysis than would have been true with unbiased data collection. What happened?
The killings omitted from the CEH's testimonies and statistics occurred during the late 1960s and early 1970s and were focused on rural agricultural workers in the eastern departments of Zacapa and Izabal, as well as in the flat areas of the southern coast. The primarily Ladino (i.e., not indigenous) workers were organizing trade unions, and the developing authoritarian state practiced its first counterinsurgency tactics against them. Approximately 22,000 people were killed between 1960 and 1977. 120
There are three interrelated reasons that these killings were not well covered by the CEH. First, the political structures in which the victims had been organized were, in the Guatemalan term, "disarticulated." That is, the repression had succeeded in killing enough people that those few who survived left the areas and stopped participating in political activism. The CEH depended strongly on political groups to organize their social bases to give testimonies (by arranging transportation from remote areas, food, social support). Social sectors that were no longer organized were less able to [End Page 37] bring their people to the commission. Relative to the victims of the 1968 and 1971 killings, the victims of the early 1980s were well-organized.
Second, during the killings of the late 1960s and early 1970s, there were very few human rights NGOs in Guatemala around to monitor and document, nor were there many international groups that could focus international attention. There were a few groups documenting state violence in the late 1960s, 121 but ten to fifteen years later, the relatively much greater number of domestic and international NGOs assured that most human rights violations would eventually be documented.
Partly as a result of the extensive documentation by civil society of the atrocities of the early 1980s, partly by the sheer enormity of the killings, and partly due to the 1992 Nobel Peace Prize awarded to Rigoberta Menchú and the educational campaigns in indigenous communities about 500 years of European colonialization, the 1990s version of the story about human rights violations in Guatemala is about violence in the Western highlands in the early 1980s (i.e., violence against communities populated primarily by indigenous people). The CEH was affected by this, and with its conclusions, the CEH helped to ratify this finding. The historical analysis in the report attempted to rectify the statistical misspecification, and the late 1960s violence is covered in the narrative of events. But the earlier period does not strongly influence the analysis of specific kinds of human rights violations nor how specific perpetrating units evolved.
This bias does not fundamentally change the story told by the CEH. The number of people killed in the 1968 and 1971 period was certainly many fewer than those killed 1979 and 1982. However, the example suggests how, even after elaborate and sophisticated investigations and carefully inductive report drafting, a truth commission may ultimately confirm the implicit hypotheses with which they begin. The TRC provides another example of this problem.
The primary method by which the TRC elicited social and dialogue truth in their effort to promote restorative truth was by means of public hearings at which victims or survivors gave testimony. But if more than 21,000 people gave statements, and only 1,818 appeared in hearings, how did the TRC choose the people who appeared in hearings? The TRC report claims that hearings "should reflect accounts from all sides of the political conflicts of [End Page 38] the past," and that women, men, and youth should all be heard. 122 TRC staff often commented on the commissioners' intense interest in finding white, Asian, and colored people to appear. Since hearing time was limited, emphasizing representation from some groups necessarily meant less time and fewer appearances for other groups.
Table 1, below, confirms the anecdotal observations and the TRC policy noted above. Column A shows the relative population proportions of each of South Africa's racial categories, indicating for example, that Africans constitute 76.1 percent of all South Africans. Columns B and C show the number and proportion of statements given to the TRC by members of each racial category: Africans gave 89.9 percent of all statements, while whites gave 1.1 percent of statements. Columns D and E show the number and proportion of hearing appearances by the racial category of the person appearing. So, for example, 87.5 percent of all people who appeared in hearings were African. Column F combines columns B and D. It shows the proportion of all statement-givers who appeared in hearings. Thus while 8.3 percent of African statement-givers appeared in hearings, 36.4 percent of white statement-givers appeared in hearings.
The dramatic outcome in Column F, that white statement-givers were roughly four times as likely to be selected to appear in hearings as African [End Page 39] statement-givers, is the product of a complex process. Note that Africans are over represented among statement-givers, relative to their proportion of the South African population (Column A), and other groups are correspondingly under represented. However, as the TRC report argues:
If the conflicts of the past had affected the population groups equally, one would expect that the numbers of deponents in each category would be proportionate to the national population. However, the table shows that the number of deponents [i.e., statement-givers] who described themselves as African is much higher than would be expected from the population statistics. . . . [And t]he reality is that the conflicts of the past affected very few whites in comparison to the rest of the population, so very few came forward to make statements. 124
Despite this recognition in the report, the TRC commissioners emphasized the appearance in hearings of non-Africans, selecting them preferentially from among statement-givers. This emphasis was likely one of the effects of Archbishop Tutu's belief (in effect, treated as a working hypothesis) that South Africa "is soaked in the blood of her children of all races." 125 The effort to show victims of all races may make sense if the objective is to balance hearing appearances as much as possible to represent the population of South Africa, and this effort itself may be the logical outcome of a strategy to organize hearings in order to promote reconciliation.
However unintentionally, the emphasis on selecting non-Africans created an unequal opportunity for Africans to appear in hearings, exacerbating the resentment many statement-givers felt at not being tapped for hearings. This resentment has been problematic for the long-term achievement of restorative truth. 126 More directly, the racial imbalances between statement-givers and hearing appearances damaged empirical truth. By giving white victims space in hearings far beyond their proportion of statement-givers and out of proportion to their level of victimization, the TRC created the truth that the "children of all races" suffered violations more or less equally. Balancing victim groups also balanced perpetrator groups, and consequently created the impression that all the political actors committed violations in more or less equal ways. The moral equivalence among perpetrating parties and among victim groups implied by the choices [End Page 40] of victims to appear in the hearings and in the Archbishop's reconciliation-focused foreword to the report were so powerful that the detailed histories in volumes two and three, and the unequivocal conclusions in volume five are insufficient to overcome them. The lack of clarity, equivocal positions, and the lack of connection between the TRC's impressive body of evidence with the findings in the report have made it vulnerable to politically motivated critiques 127 --which is perhaps the worst outcome for a commission charged with articulating the truth.
V. Conclusions and Recommendations
This article has explored the nature of the "truth" that truth commissions are mandated to find and the factors affecting the process of truth finding. Clearly, the documentation and interpretation of truth is considerably more complex and ambiguous than many analysts of truth commissions assume--proponents and critics alike. Far from being generic bodies, truth commissions have very different approaches to the kind of "truth" they are seeking. Their official mandates, the perceptions and priorities of their commissioners and key staff, the methodological orientations utilized, and the level of resources available shape the nature of their findings and the type of report they produce. Decisions made by those working within a specific truth commission regarding how they organize their work and perform their tasks may also have important consequences for truth-finding.
What then are the implications of the analysis in this article? What kind of "truth" can and should truth commissions seek? First, it is our view that truth commissions are far better suited to pursue what we have termed "macro-truth," the assessment of contexts, causes, and patterns of human rights violations, than "micro-truth" dealing with the specifics of particular events, cases, and people. To be able to make macro-determinations requires utilizing more of a social science than a legal approach to truth finding. Second, we believe that truth commissions should focus on the objective rather than the subjective dimensions of truth. Doing otherwise amounts to confusing process and outcome.
Patterns, trends, tendencies, and the big picture are often the pieces most missing from the history of transitional societies. Victims and their communities know quite a bit about the daily micro-processes of persecution and abuse: they lived it, after all. Although many victims and their survivors may want acknowledgments from perpetrators or the details of [End Page 41] particular stories, they also want a sense that they were not alone, and that the perpetrators were not a few "bad apples," but rather that an entire legal, ideological, political, and military system was responsible for years, or decades, of human rights violations. This broad explanation should not be left to scholarly debate, but should be the core of a commission's work.
The micro-level truths, the forensic or microscope truth, are important: the facts must be gotten right. However, commissions are not particularly good at determining the details of hundreds or thousands of cases. They lack the time, staff, and resources to undertake such a massive investigatory task. Moreover, when a commission does make micro-truth findings, inevitably there are likely to be errors, which in the aggregate have little or no effect on the macro-truth. But these errors provide critics with a basis for claiming that the entire process is discredited. Beyond a small number of illustrative cases and a limited number of hearings on events or institutions, micro-level truth should be better left to judicial tribunals which specialize in weighing evidence, on a case-by-case basis.
Three of the forms of truth identified by the TRC (narrative truth, social truth, and restorative truth) are each important goals of a truth commission, and they deserve substantial attention. But these are subjective, not objective. They are process goals, not forms of truth. The distinction between the subjective and objective components of truth commissions does not denigrate the importance of subjective processes to validating victims' experiences and ultimately contributing to reconciliation. However, the conflation of the subjective with objective truth-finding weakens the political and moral importance of truth by making truth a matter of personal opinion, and not the product of verifiable scientific best practices.
Our analysis in this paper leads to the following recommendations:
1. Mandate: The mandate should specify that the function of the Commission is to make broad findings about the antecedents, causes, patterns, trends, perpetrators' motives, and impact on victims of the period of violence being studied. The mandate should explicitly specify that the findings of the Commission must be based on legal and scientific best practices to ensure that working assumptions are tested and that the findings will be robust to sustained and informed criticism.
2. Commissioners: Commissioners should play a strictly advisory role. That is, they should not participate in the day-to-day operations of the commission, but instead should function as a governing board in [End Page 42] setting policy, making broad determinations, and guiding the Commission's strategy. To promote their focus on strategic issues (and not on tactical concerns), there should be relatively few commissioners, and they should work part-time.
3. Staff: Staff should be professional and entrusted with the quotidian operations and all technical questions of the commission. They should be drawn from a variety of intellectual disciplines with an emphasis on qualitative and quantitative rigor.
4. Attribution of responsibility for violations: The goal of the Commission should be primarily to identify institutions, parties, structures, and ideologies that permitted or committed gross human rights violations. Only secondarily should a Commission identify particular individuals who played roles in the abuses. The attribution of responsibility should be centered on the broad proportions or patterns of specific violations attributed to the various parties to the conflict, rather than excessive attention to one or a few illustrative cases.
5. Function of the findings: The findings of a commission, made via event or sectoral hearings or other public processes as well as in the final report, must be unequivocal, massive, objective, and undeniable, and made according to scientific best practices. The findings should be victim-centered, telling the story from their point of view and validating their experiences. A truth commission cannot by itself change the future of a democratizing nation. But at the very least, a commission must present a narrative that becomes the central component in the debates about the past which shape the future.
Audrey R. Chapman is the director of two programs at the American Association for the Advancement of Science, Science and Human Rights and the Dialogue on Science, Ethics, and Religion. She has published widely on a variety of topics related to economic, social, and cultural rights; bioethics; and truth commissions. Currently she is the director of a project with several South African collaborators assessing the legacy and impact of the South African Truth and Reconciliation Commission, specifically its ability to balance truth finding with promoting forgiveness and reconciliation. She is the author, coauthor, or editor of fourteen books including Human Rights and Health: The Legacy of Apartheid.
Patrick Ball is Deputy Director of the American Assocation for the Advancement of Science (AAAS) Science and Human Rights Program. Since 1991, he has designed information management systems and conducted quantitative analysis for large-scale human rights data projects for truth commissions, nongovernmental organizations, and UN missions in El Salvador, Ethiopia, Guatemala, Haiti, South Africa, and Kosovo. In his most recent report from AAAS, "Policy or Panic? The Flight of Ethnic Albanians from Kosovo, March-May 1999," he reconstructs the flow of refugees by villages across time to show that only a small fraction of Kosovar Albanians fled as a result of NATO bombing raids. Instead, the migration patterns were so regular that they must have been the result of a coordinated policy.