What truth means to a commission has sometimes been articulated, but most often not. Like the commissions that preceded them, neither the Haitian nor Guatemalan commissions had an explicit exposition of truth because they assumed truth would automatically result from a rigorous application of their mandate, the relevant legal standards, and a consistent methodology. 26 The South African commission articulated a complex and postmodern understanding of truth, but the definitions seem post-hoc and have little connection to the analysis in the volumes of the report in which historical narratives are presented. 27
In the CNVJ report, the methodology section describes the legal framework that established the commission, the use of primary (interview) and secondary sources, and how three levels of proof were used to grade the quality of each decision. 28 Similarly, the CEH report argues that in each of the cases investigated, there was at least one human rights violation by the government or an act of violence by the guerrilla insurgents: "in order to achieve this first objective, the Commission has used principally juridical categories from International Human Rights Law and International Humanitarian Law." 29
But inside the CNVJ and the CEH, debates about truth raged. In the Haitian commission, a conflict between legal and social science approaches to investigation and classification of violations resulted in three fundamental changes to the basic data classification schemes, requiring the analysts to recode the entire set of testimonies twice. The CNVJ's chair was a sociologist, but most of the other commissioners were lawyers, some of whom openly disdained social science methods. In Guatemala, although [End Page 9] there were significant differences in approach and tensions between legal and social scientific concepts, 30 Commission researchers incorporated history, anthropology, sociology, economics, and military science into the analysis. 31
Of all truth commissions, the TRC was the most self-conscious and intentional about its conception of truth. Its report distinguishes between four notions of truth: factual or forensic truth, personal or narrative truth, social or "dialogue" truth, and healing and restorative truth. 32 Of these four approaches, only factual or forensic truth refers to the impartial and objective evidence that most truth commissions have understood as their mandate. The other three concepts of truth articulated in the report sometimes competed with and sometimes complemented the objective and analytical approach to truth-finding. The TRC's "forms of truth" encode as "truths" ideas that other commissions may have had but have expressed as goals to be achieved, not as alternative--and competing--forms of truth.
The TRC's first kind of truth is impartial evidence that tells truth at two levels: (1) truth about individual events, cases, and people, and (2) about "nature, causes, and extent of gross violations of human rights, including the antecedents, circumstances, factors, context, motives, and perspectives that led to such violations." 33 This is the way that most truth commissions have understood truth. The TRC called this "scientific and forensic" or "microscope" truth, both of which imply that the TRC understood this kind of truth to play a very limited role in its work. The TRC report somewhat grudgingly links the use of a social science methodology enabling it to analyze and interpret information contained in its database to its mandate to make findings both at the micro-level (on particular incidents and in respect to specific people) and at the macro-level (to identify the broader patterns underlying gross violations of human rights). 34 Yet, it is obvious that the TRC understood this form of truth to play a limited role. The same passage of the report quotes Michael Ignatieff's remark--"[a]ll that a truth commission can achieve is to reduce the number of lies that can be circulated unchallenged in public discourse." 35 The narrowness with which this kind of truth was understood by the TRC is perhaps surprising since this is explicitly what other truth commissions claimed to seek through rigorous methods, though [End Page 10] this "micro-truth" was for most commissions a starting point from which generalizations could be drawn.
The second truth the TRC defines is narrative truth, explicitly evoking the cathartic benefits of storytelling. Victims make meaning and sense out of their experiences through narration, and under certain circumstances, storytelling contributes to psychological healing after trauma. 36 Most commissions noted the importance of this idea, including the Chilean and Guatemalan commissions, though it was not understood as a kind of "truth." 37 Narrative truth was central to the work of the TRC, especially to the hearings of the Human Rights Violations Committee to whom victims told their stories in public hearings.
The emphasis on "social" or "dialogue" truth is the one in which the TRC was truly unique. By this term, they referred to the process and dialogue which surrounded their work: as the TRC report said, "[t]he public was engaged through open hearings and the media. The Commission itself was also subjected to constant public scrutiny and critique." 38 The report traces the idea of "social truth" to a distinction made by Albie Sachs, who played an important role in the establishment of the Commission and now serves as a Constitutional Court judge. 39 He contrasted "microscope truth" with "dialogue truth," arguing that the former is factual, verifiable, and can be documented. 40 In contrast, dialogue truth "is social truth, the truth of experience that is established through interaction, discussion and debate." 41 Because all other commissions functioned largely out of public scrutiny, "social truth" was less important in other national contexts.
The fourth kind of truth--restorative truth--is the truth that comes from putting facts in their contexts: in political context of power between social actors, in historical context of the sequence of contingent events, and in the ideological context in which contending visions of the social world compete. As the commission acknowledges and recognizes what happened, it validates the experience of people. The TRC wrote that "[a]cknowledgement is an affirmation that a person's pain is real and worthy of attention. It is thus central to the restoration of the dignity of victims." 42 This [End Page 11] elegantly restates the overarching political objective of almost all truth commissions. The CEH, for example, has the following quotation from a testimony on the back of every published volume.
A witness showed us the remains of one of the victim's bones. He had these remains wrapped in plastic in a string bag: "It hurts so much to carry them . . . it's like carrying death. . . . I'm not going to bury them yet . . . I do so want him to rest, and to rest myself, too. But I can't, not yet . . . this is the evidence for my testimony . . . I'm not going to bury them yet, I want a piece of paper that will say to me 'they killed him . . . he had committed no crime, he was innocent . . .' and then we will rest." 43
The idea that public acknowledgment of suffering--the truth about injustice--will begin to restore victims' dignity is perhaps the central premise on which truth commissions are founded. Yet framing acknowledgment as a distinct kind of truth suggests that restorative truth (and the resulting reconciliation) might be somehow different from (or even in contradiction to) the forensic, legal truth demanded by the CEH witness in the quotation above. This may be one of the roots of a contradiction that beset the TRC's report, and indeed, its entire public life: was the TRC the "truth and reconciliation commission," or the "truth or reconciliation commission?" In the Methods and Report under Section D, parts 2 and 3, we will argue that when the TRC faced decisions between pursuing truth or pursuing reconciliation, they chose reconciliation.
The mandate assigned to a specific truth commission plays a decisive role in shaping its truth-finding. Its terms of reference determine its priorities and the nature of the truth it will attempt to investigate. Many truth commissions have restricted terms of reference that reflect the political compromises that led to their creation. Typically, truth commissions emerge out of negotiated settlements in which there are not clear victors and vanquished. Often the architects of the violence and abuses or their supporters retain political influence and power. Those ceding their positions of power frequently attempt to impose conditions that will restrict the powers and findings of the bodies that will be able to investigate the past. Or the mandate may reflect the priorities and concerns of those drafting the mechanisms under which they were established, many of whom believe it is in the country's interest to have a short transition period. [End Page 12]
Their assigned mandates circumscribe the truth finding of truth commissions in a variety of ways. Despite the brutality of both Duvalier dictatorships that governed Haiti for some forty years beginning in 1957, the CNVJ was mandated to deal only with the human rights abuses of the Cedras de facto regime. General Raoul Cedras overthrew President Bertrand Aristide in September 1991 and governed until Aristide returned to Port-au-Prince on 14 October 1994. 44
The CEH did have responsibility for a longer period, approximately coinciding with the internal armed confrontation beginning in 1962 through the signing of the 1994 peace accords. But this periodization led to the understanding that the guerrilla insurgents began the war; if the period had begun in 1954 with the CIA-organized military coup d'etat against a democratically elected government, the implication would have been that the coup was the root cause of the conflict. 45 The CEH was prohibited by its mandate from naming individual perpetrators, and many critics argued that this terribly weakened their ability to shed light on the past. 46 The apparent handicap of "no names," however, became a strength, in part because it widened the CEH's scope, encouraging it to examine the roles of institutions and the social structure that produced the violence, instead of identifying individual perpetrators. That is, the prohibition forced the CEH to focus on the macro issues which are a truth commission's most important goal.
The mandate of the TRC focused it exclusively on "gross violations of human rights." In the South African case, there was a further qualification; to fall under its purview the gross human rights abuses had to be committed with a political motive. The period the TRC was assigned to investigate was 1 March 1960 through the elections in May 1994, but certainly not the full sweep of apartheid history. Gross human rights abuses were defined as "the killing, abduction, torture or severe ill-treatment of any person; or any attempt, conspiracy, incitement, instigation, command or procurement to commit [such an] act . . . with a political motive." 47 Because the mandate of the TRC was restricted to gross human rights abuses, it did not assess the impact of the institutionalized racism of the apartheid system that arguably had a far more profound and abusive impact on the population. Forced removals alone accounted for many more deaths than direct state violence.
The TRC commissioners could have interpreted their mandate more [End Page 13] broadly than they did. Among its assignments, the TRC was to establish "as complete a picture as possible of the causes, nature and extent of the gross violations of human rights . . . including the antecedents, circumstances, factors and context of such violations." 48 As it tried to do in some of the sectoral hearings, the TRC could have used its mandate to determine the causes, antecedents, circumstances and factors to undertake a more systematic analysis of apartheid. Somewhat ironically, despite its limited mandate and decades of international attention to apartheid, the CEH undertook a far more penetrating analysis of the official policy of racism and social exclusion than did the TRC: among the three historical causes of the armed conflict, the second theme is "racism, the subordination and exclusion of the indigenous community." 49
The approach of the TRC resulted in the individualization of the gross abuses of the past. 50 Gross violations of human rights were treated more as the product of individuals' decisions and actions than the outcome of the dynamics of the apartheid system. This enabled white South Africans who were the supporters and beneficiaries of apartheid to escape from acknowledging their complicity in the system's abuses. As one analyst points out, "The increasingly familiar refrain within white South African communities, that apartheid was merely a 'mistake' for which no-one was responsible, that somehow the system propelled itself impersonally, may be one of the more ironic, unintended consequences of the TRC's rendition of the past." 51
Truth commissions have significant time constraints built into their mandates, and indeed, a limited period of existence is part of what distinguishes a commission from a permanent governmental human rights body. Most begin with an explicit time limit, often just six to nine months to complete all investigations and write a report. 52 Sometimes the initial period is extended. The Haitian truth commission worked for nine months (April 1995-January 1996), the full TRC two years and eight months (February 1996 through October 1998), and the Guatemalan commission eighteen months plus three months pre-staff preparation (August 1997 through February 1999). While the South African and Guatemalan commissions functioned for longer than prior commissions, they also had longer historical periods to cover. The short official duration of truth commissions, quite inconsistent with an assignment to discover the truth of a disputed [End Page 14] period of history, necessarily limits the scope of the investigations and the depth of their analysis.
The enabling legislation setting forth the powers of specific truth commissions also has a significant impact on its functioning and findings. Of the recent truth commissions, only the TRC has been given extensive powers of subpoena, search, and seizure--although the TRC utilized this advantage very sparingly. None of the Latin American commissions was granted the power to compel witnesses or perpetrators to come forward with evidence, and they generally had considerable difficulty in obtaining official written records from the government and armed forces. 53 The quasi-judicial power to grant amnesty to individual perpetrators assigned to the TRC was also unprecedented although in the end it did not elicit the detailed accounts from perpetrators and institutions that had been anticipated.
B. Balancing Truth-Finding with Other Objectives
Further, truth commissions have been vested with a wide range of responsibilities ostensibly related to truth-finding, some explicit and others implicit. The dilemma is that these tasks are quite different in nature and entail very different approaches, some of which are likely to be in conflict with others. Typically the mandate includes several of the following responsibilities:
establish an authoritative record of the past;
overcome communal and official denial of the atrocities, violence, or abuses and gain official and public acknowledgment of them;
restore dignity to victims and promote psychological healing;
end and prevent violence and future human rights abuses;
create a "collective memory" or common history for a new future not determined by the past;
forge the basis for a democratic political order that respects and protects human rights;
identify the architects of the past violence and exclude, shame, and diminish perpetrators for their offences;
legitimate and promote the stability of the new regime;
promote reconciliation across social divisions; [End Page 15]
educate the population about the past; and
recommend ways to deter future violations and atrocities. 54
Those who write commissions' mandates seem to assume that a single body vested with the responsibility to discover the truth about the past can manage all or many of these responsibilities. However, these tasks are quite different in nature, and some reflect divergent conceptions of truth and approaches to truth-finding. Establishing an authoritative record of the past, for example, requires an exhaustive research effort using scientific methods. Documenting criminal action will depend on the relevant legal definitions and specific juridical investigation techniques. In contrast, restoring dignity to victims and promoting psychological healing calls for attention to the subjective rather than objective dimensions of the work of a truth commission. Creating a collective memory or common history, nation building, and educating the population emphasize the public dimensions and accessibility of the findings; but if the quality, representativeness or rigor of the content are sacrificed, the gains of the commission's public dimensions may be short-lived. Forging the basis for a democratic political order, legitimating and promoting the stability of a new regime, promoting reconciliation across social divisions, or recommending ways to deter future violations all involve the formulation of long-term public policy. Very little attention has been paid to potential incompatibilities in these assignments and how they affect the work of truth commissions.
Truth commissions typically have severe resource constraints. Nevertheless, the magnitude of this problem varies quite considerably. At one end of the spectrum, for example, the Haitian Commission probably operated on a little more than $1 million, and the report complains that the lack of sufficient funds greatly reduced their ability to complete their work. 55 The Guatemalan commission had a total budget of approximately $10 million, very generous by truth commission standards. While the South African commission also operated under strained financial conditions, its budget was considerably greater than any other truth commission. The TRC received something in the range of $28 million from the Treasury 56 and [End Page 16] another $5 million from foreign donors. 57 Consistent with its greater financial resources, the TRC also had a much larger staff, over 400 persons at its peak, several times the size of other commissions, and a separate research department. In comparison, the staff size of the Guatemalan commission was approximately 200 people; the Haitian commission was composed of seventy-five persons.
D. Role of Commissioners and Staff
Those vested with the responsibility of establishing the truth come to the task influenced by their life experiences, particularly in relationship to the conflict, training, and current roles and status. No matter how qualified and well-trained, commissioners and key staff members cannot be expected to be completely neutral. At the least, their perspectives and training will influence their priorities and approach. To highlight just a few considerations, commissions have differed in whether their commissioners have been selected from within (e.g., Chile, South Africa) or outside of the country (e.g., El Salvador) or a combination thereof (e.g., Guatemala and Haiti). Some commissioners are chosen in order to represent particular ethnic or political constituencies, while others are chosen as nonpartisan experts on key subjects. Whether commissioners choose a legal or social science approach to truth-finding is also important.
Because human rights violations have been largely understood as violations of law, the composition of most truth commissions favors commissioners with legal training. Rather than providing balance, the mix of social scientists (2) and lawyers (5) among commissioners at the CNVJ led to repeated conflicts among them and contributed to a poor outcome. The commissioners were equally divided between three foreigners who were well-respected international jurists, and four Haitians who represented parts of the government and important nongovernmental organizations (NGOs).
There was also differences of approach in Guatemala between the lawyers and social scientists among the commissioners (two lawyers and a social scientist) and the staff. The foreign commissioner had previously been a UN Special Rapporteur to Guatemala, and was clearly an expert. The two Guatemalan commissioners were chosen to represent the country's Ladino and indigenous communities. Perhaps because Guatemala's intellectual life has been dominated by anthropologists committed to intellectual pluralism, or because human rights NGOs in Guatemala had extensive experience [End Page 17] mixing legal and social scientific approaches to investigation, or for some other reason, the clash of perspectives proved fruitful at the CEH.
The TRC was unusual in the large number of commissioners, the diversity of their professional backgrounds, and the central role of activists and religious thinkers and clergy. Of the seventeen members, four came from the religious community; five from medicine, psychology, and nursing; seven from the law; three from politics; and three from NGOs. Nearly all of them represented an identifiable political, sectoral, or ethnic constituency. The chair, vice chair, director of research, and several other commissioners all had religious backgrounds. And given the presence of Archbishop Desmond Tutu as the Chair, some of its public hearings had a decidedly religious character. Commentators pointed out that the hearings frequently resembled a church service more than a judicial proceeding, with a definite "liturgical character." 58 The Christian atmosphere and discourse of the TRC, and particularly Archbishop Tutu's frequent framing of issues in terms of repentance and forgiveness, was applauded by many South Africans for whom Christian ideals served as the basis for an ethical critique of apartheid, but it was distasteful to others, including some of the commissioners themselves. 59 The latter category included both secular academics and some victims who complained about "the imposition of a Christian morality of forgiveness." 60
In addition to the question of their backgrounds influencing their approach to truth-finding, the simple number of commissioners appointed to the commission, and whether they work full or part-time, also influences the kind of role that they are likely to play. In Haiti, three of the seven (later six) commissioners lived abroad and visited only for brief meetings. The Guatemalan commission had only three commissioners, all of whom served on a part-time basis. The role of a small number of part-time commissioners is necessarily strategic, to set policy and make major decisions, while senior staff have the responsibility for implementing policy.
In contrast, the seventeen South African commissioners, 61 all of whom worked on a full-time basis, had tactical roles, with each of them directly administering some office or aspect of the work of the Commission. 62 It was [End Page 18] virtually impossible for such a large and deliberately diverse group to reach consensus on specific issues or collectively coordinate the TRC. For example, one commissioner resigned for substantive reasons, and another issued a minority report. 63 Furthermore, each commissioner administering a committee or regional office was accountable only to the whole group of commissioners; essentially, the commissioners were accountable to themselves. Staff were directly managed by commissioners who could change direction or refocus attention very rapidly. This put senior staff in structurally weak positions in that it was very difficult to plan work. In the words of the TRC itself, "[t]his explains the structural overlaps and operational duplication that occurred." 64
Commissions also differ regarding whether their personnel configurations are static or reorganized for different phases of the process. The Haitian commission deployed eighty interviewers into the field for less than two months; for the other months of its existence, only a small team of analysts, support personnel, and the Haitian commissioners worked in the organization. The TRC maintained the same organizational structure from April 1996 until May 1998, with minor adjustments in August-September 1996. 65 In contrast, the CEH went through continuous reorganization, with five distinct structures from July 1997 through February 1999, although the most technical areas (i.e., the database) were kept together. 66 The chaos that resulted from the CEH's constant reorganization prevented turf battles simply because staff never had time to become accustomed to any given role.
The frequent--or constant--reorganization served the most important benefit to the CEH by enabling it to cope with the most difficult challenge which faces a commission. Commissions are never able to define the most important thematic concerns until very late in the life of the organization. How can a team of people investigate massacres if there is no common definition of a massacre until the fieldwork is long complete? How can several dozen analysts begin working on a report before the report has a table of contents and identified core issues? It is difficult to manage collective research without a written plan by which each subgroup knows precisely what parts of the final product are its responsibility. The CEH management team moved people rapidly from role to role, and created new writing assignments for each team in waves. This management process [End Page 19] delayed decisions until facts had been gathered and due deliberation given, but it provided short-term work assignments even before final decisions had been made. 67 This process, called the insumo model for drafting the report (described in the Report section) was the central product of this management strategy.
1. Legal versus Scientific Methods
In their first moments of existence, commissions must decide what they intend to argue, or stated differently, what their working hypotheses will be, and how data will be gathered to support the hypotheses. But what commissions set out to do--argue a case or test a hypothesis--reflects different epistemological and methodological perspectives. Commissions often experience tensions between these two perspectives, legal arguments versus social science evaluation. Commission lawyers will establish a set of legal definitions about what constitutes a violation of a given norm from domestic law, international human rights law, or international humanitarian law. They may then seek examples, in the form of particular cases, which demonstrate the violation of the norm in question. This kind of work seeks to argue that a norm was violated, making the claim in the strongest possible terms. By contrast, social scientists, who tend to be rare in truth commissions, would ask "how often was the norm violated, in absolute and proportional terms? Was the norm violated more frequently in some circumstances or during some periods than in others? Why might the norm have been respected on occasion? And most importantly, how can we find evidence to address these questions by methods which are not self-fulfilling?"
Truth commissions are charged with broad objectives that require answers to the social science questions posed above. By its mandate, the TRC was directed to examine the "nature . . . and extent" of gross human rights, looking at the "context, motives, and perspectives which led to such violation," and then identifying "systematic pattern[s] of abuse." 68 Similarly, the CEH interpreted their work to require an "examination of the causes and origins of the internal armed confrontation, the strategies and mechanisms of the violence and its consequences and effects." 69 None of the broad [End Page 20] structural generalizations demanded by the mandate language of either commission can be satisfied by a simple aggregation of cases, no matter how many cases the aggregation includes, because the simple aggregation of cases might exclude cases that would tend to contradict the generalization. Instead, methods must be devised that would find evidence to confirm or, if it exists, evidence that would contradict the assumptions with which the commission began work. However, truth commissions do not usually employ methods that would allow them to frame, test, and potentially reject a series of hypotheses.
One way to make generalizations is to seek out all the cases and other evidence which support the claim, and if there are enough such cases, and if they are sufficiently dramatic, conclude that the generalization is valid. A second method is to appoint an arbitrator, permit advocates and opponents of a claim to each make the strongest possible arguments before the arbitrator, and then allow the arbitrator to rule on the generalization. Both of these methods are derived from legal and political procedure, and in fact, these methods dominate decision making at truth commissions. As will be discussed in the methods section below, the TRC's conclusion that all of South Africa's people (not only people of color, or Africans) were victims of apartheid is apparently based on evidence marshaled by the first technique. The CEH's conclusion that genocide was committed against some Mayan communities by Guatemalan state forces was made (in part) after a series of quasi-juridical internal debates that took place in 1998. The first method is entirely vulnerable to the exclusion of information that does not support the hypothesis. The second method is better but is nonetheless vulnerable to a weak presentation of one of the two contending positions.
Scientific methods proceed quite differently. Beginning with an hypothesis, the scientific method seeks data with which to test the hypothesis. 70 For example, to test the hypothesis that South Africans of all races suffered under apartheid, a random selection of people of all races could be taken and the proportion of each who suffered a gross human rights abuse measured. If the proportions were roughly equal, the hypothesis would be confirmed; 71 otherwise it would be rejected. The randomness of the selection of respondents protects the conclusion against the possibility that [End Page 21] the evidentiary basis was constructed deliberately to support the conclusion. The analysis of genocide would require a comparative measure of rates of killing among groups hypothesized to be subjected to genocide and similar rates among otherwise similar groups which are not thought to have suffered genocide. If the killing rates were essentially the same among all groups, then this would be evidence of a very violent moment, but not genocide. However, if killing rates were much higher among the groups in which genocide had been hypothesized, this finding would support the genocide hypothesis. 72 Not all scientific methods are statistical, obviously; forensic methods, in particular, have been enormously useful to truth commissions. But what scientific methods all share is the possibility of finding the unexpected and refuting underlying assumptions. Legal and other methods based on unsystematically collected evidence do not have this characteristic.
Another difference between legal and social science generalization methods is the relative importance each assigns to trends and patterns versus individual cases. The TRC used the "balance of probabilities" method by which something was judged true if there was more evidence in favor than against. 73 The balance of probabilities method, derived from civil litigation, served the TRC well in judgments about individual cases. But the TRC clearly lacked a method for making generalizations from the thousands of findings made by this method. Although charged with finding "systematic patterns," the TRC's report almost exclusively makes findings on cases: there is little broader macro analysis. Commissions prior to the TRC were not much more successful elucidating or applying methods for systematic generalizations. However, in Guatemala the CEH explicitly invoked a range of social science disciplines, and a range of social scientists were employed there. 74 This may partly explain the CEH's greater relative success at establishing macro findings.
Because truth commissions cannot investigate all potential cases or events, they frequently choose representative or significant ones, what the TRC referred to as "window cases" and the CEH called "illustrative cases." There are hundreds of window cases in volumes two and three of the TRC report, spanning everything from deaths in detention to cases of arson. The presentations range from one paragraph to four pages. The CEH chose its eighty-five illustrative cases using the following criteria: that the case showed an important tactical or strategic change of one or the other of the [End Page 22] parties; 75 that by its seriousness, had a special impact on the national conscience; or that it illustrated a particular pattern of violence for a region or period. 76 These cases, from ten to thirty pages in length, were chosen by the legal method described above: they were chosen to make specific points, not to test a generalization.
As has been mentioned, the public hearings held by the TRC were a unique feature of the South African experience. Hundreds of hearings were convened by different components of the TRC for different purposes. The Amnesty Committee held hearings at which they questioned applicants and made determinations about whether amnesty should be granted; the Human Rights Violations (HRV) Committee heard victims chosen from among those who presented statements to the TRC, and to event hearings at which particular historical episodes were considered in depth by calling witnesses, victims, and alleged (or confessed) perpetrators; and ad hoc sectoral hearings were called at which the conduct of particular social areas were considered, such as the legal or medical sectors. 77 The hearings attracted massive public attention, both within South Africa (where they were the subject of nearly daily press reports and a weekly television show) and internationally. It is clear that the hearings were tremendously successful at generating attention and debate about the violence of the past. It is less clear that they provided objective data so that the debates about the broad truths of the past could be resolved in ways that would withstand subsequent criticism.
Data are central to the process of truth-finding. A major task of truth commissions is to collect and analyze critical data sources on which to base their findings. Most truth commissions seek and create truth through interviews with victims, witnesses, and survivors as a major source of their data. An interview is an intimate interaction between a representative of the commission and one or more deponents who tell their story. The presentation of information by respondents to the commission is guided by an interview protocol designed to elicit empirical and narrative truth. [End Page 23]
The National Commission on Truth and Justice in Haiti collected some 6,000 semi-structured interviews utilizing both closed and open-ended questions. Perhaps surprisingly given its other problems, the Haitian commission's fieldwork was excellent and culminated in a dozen area reports, each of which was hundreds of pages long, combining statistics, secondary materials, analysis of interviews, and direct investigations. 78 The Haitian commission also sponsored a series of forensic investigations of mass graves undertaken by a multinational team of experts assembled by AAAS. Unfortunately the deteriorating condition of many of the sites did not render conclusive evidence. A third data source came from morgue records from the University Hospital in Port-au-Prince. Comparing trends in morgue admissions with the trends from interviews found that the two matched almost identically, thereby validating the interviews as a good measure of political violence. The Haitian commission did not hold any public hearings.
The CEH conducted 7,200 interviews. In rural areas populated primarily by indigenous people interviews were sometimes conducted with groups of people simultaneously; sometimes as many as fifteen people met with the interviewer at the same time. This was the closest analogy to the South African hearing model, but no public or members of the press were present. There was also extensive secondary research by a team of Guatemalan historians and public meetings and workshops to discuss policy.
Because truth commissions are constrained by limited time and resources, the availability of completed research on which they can build can play an important role. In Guatemala, for example, there were teams of forensic anthropologists, including an initial team founded and trained by AAAS, that had been conducting exhumations of mass graves for seven years prior to the commission. By specifying the cause and manner of death of the victims, the reports of the forensic anthropologists clearly refuted military claims about massacres. 79 AAAS also worked with a coalition of human rights groups to construct a database of violations. By the time the truth commission began its work, this database had documented over 37,000 killings. 80 A project sponsored by the Archdiocese of Guatemala provided a third source. 81 All of these data were made available to the truth [End Page 24] commission. Many of the cases in one of the three databases were not replicated in the others. 82 The pool of data sources that the CEH could draw upon was far richer than it could have collected on its own, and the combination of databases made possible a scientific estimate of the total number of people killed in Guatemala. 83 In South Africa there was also an NGO project to present data to the TRC; the project was undertaken by a coalition of NGOs and was known as the Human Rights Data Project. However, two factors combined to reduce the impact of the Human Rights Data Project on the TRC's analysis. First, the way the NGO data were represented as a database weakened their statistical value because insufficient detail was recorded. Second, the TRC's lack of interest in macro-truth meant that the kinds of statistical and comparative uses to which NGO data were put in other commissions were less important to the TRC than to the commissions in El Salvador and Guatemala.
The TRC conducted the most massive data collection effort in the history of truth commissions, yet they collected only a fraction of what was potentially available to it. One rationale for the amnesty provision was that it would provide the TRC with data about the architects of apartheid and the internal structures which implemented policy, but the results were disappointing. Very few of the former apartheid leaders came forward to take advantage of the offer of amnesty for full disclosure. The flood of applications just before the deadline came primarily from middle and low-level functionaries responding to the successful prosecution of Eugene de Kock. The TRC's own assessment was that:
the spirit of generosity and reconciliation enshrined in the founding Act was not matched by those at whom it was mainly directed. . . . With rare individual exceptions, the response of the former state, its leaders, institutions and the predominant organs of civil society of that era, was to hedge and obfuscate. 84
But the TRC itself made little use of the materials that came out of the amnesty process. Not until 1998 did the amnesty application materials become part of the TRC database, and as a result, the data from the Amnesty process played almost no role in the substantive evaluations in the TRC report.
From the first meetings of the TRC, with the Amnesty Committee in January 1996 and then with the Database Development Group in February 1996, 85 the judges who composed the majority on the Amnesty Committee [End Page 25] made clear that they would not permit the collection of the kind of data on which a systematic evaluation of the structures of repression might be made. These data, including perpetrators' career histories, unit command structures across time, and other personnel information, had been precisely the key to identifying the military units and particular officers most deeply involved in human rights violations in El Salvador. 86 Although the circumstances in South Africa seemed optimal for undertaking this kind of data collection and analysis, the Amnesty Committee's role as construed by the judges was narrowly defined in a legal framework that admitted only information precisely pertaining to the events for which applicants requested amnesty. The focus on the actual cases and acts, but not on the antecedents, causes, organizations, ideologies, and perspectives that gave rise to the acts, was one of the most significant opportunities lost due to the excessively legalistic approach that sometimes bedevils commissions. Because these data were not gathered, the analysis of the structure of organizations that perpetrated state violence is brief and superficial. 87 There is no assessment of how particular units came to be as violent as they did, nor how units evolved over time, nor how racist beliefs were indoctrinated among new recruits. The concluding section of the report on the state only details that communications links between "contra-mobilization" units in the mid-1980s were regarded as as especially destructive. 88 The TRC's analysis section, like other parts of the report, is a listing of cases with little discussion of why the cases occurred, how the structures changed over time, and how individual perpetrators were socialized differently in different parts of the state apparatus.
Even within the Human Rights Violation Committee, the quantity, format, and quality of data were subject to conflicting goals. The original interview protocol included narrative and structured components, similar to the format used in Haiti and to the format that would later be used in Guatemala. This model was scrapped at the initiative of one commissioner in August 1996, five months after the original forms had been deployed, in an effort to increase the rate at which statements could be taken--that is, to reduce the time each statement giver spent with a representative of the [End Page 26] commission. 89 The new protocol was dominated by closed-ended questions 90 and greatly reduced the space available for interviewers to record qualitative and narrative information. In an internal analysis, it was shown that the number of words written for each interview declined, as was intended by the format change. Data processors reported that the quality of the interviews consequently declined, including increases in the rate of missing data. 91
Although narrowing the scope of the conversation with statement givers would seem directly to contradict the TRC's goal to promote narrative truth, the proponents of the reduced interview format argued that it would allow more statements to be taken. 92 In order for South Africans to become eligible to receive reparations for gross human rights violations they suffered under apartheid, the TRC had to "find" that this person was a victim under one or several of the definitions set out in the TRC Act. 93 The need to accumulate the data required to make this administrative judgment about the maximum possible number of people therefore competed directly against the "truth" goals the TRC articulated in the final report. Narrowing the range of information sought in the interview led to lower quality data in each interview, thus diminishing the quality of the TRC's empirical data. The shortened and tightly structured interview process may have seemed to the statement giver more like a police interrogation than the original format (which was organized more like a social work quasi-therapeutic interview), and this change must have had negative implications for both narrative and [End Page 27] restorative truth. It should be kept in mind that notwithstanding the truth-damaging outcomes of the interview format change, the objective of the change was to open the doors to more South Africans becoming eligible for reparation, and so two laudable goals conflicted with each other. 94 After a few months, this format was in turn discarded and a third semi-structured interview protocol was substituted.
The volume and complexity of the data sources truth commissions collect require the application of rigorous and systematic methods. The information analyzed in the report must fairly represent the narratives presented by statement givers. Attempts to document the abuses and write official histories need objective and compelling means to describe patterns among tens of thousands of victims and violations, to evaluate the relative proportions of responsibility for violations, and to estimate the total number of victims killed or missing. The methodologies used play an important role in shaping the nature of the data available to the commission, the conclusions that can be drawn, and their validity and credibility. Information management systems and social science research methods can help make sense of data from thousands or tens of thousands of interviews--literally hundreds of thousands of pages of questionnaires and transcripts. 95
Quantitative tools can help investigators understand history and the story of human suffering in ways that would not have been possible using testimonies alone. For example, the TRC used statistics to show that the police were responsible for the overwhelming majority of killings before 1990. 96 As an important step in the analysis of genocide in Guatemala, the truth commission employed quantitative analysis to show that in several crucial regions rates of indigenous people killed by the state were five to eight times greater than rates among nonindigenous people. 97
When both commissioners and senior staff lack background in quantitative methods, they may not make full use of the potential analytic capability of the commission's data and information management system. [End Page 28] There was a considerable difference between the South African and Guatemalan commissions in this regard. At the TRC, few of the commissioners had any experience using social science methods, and since the publication of the report, some commissioners have criticized information-intensive approaches to truth-finding as lacking utility. In a public speaking engagement, for example, one of the commissioners attributed the creation of the victim database to legal requirements and claimed that virtually no use was made of these data. 98
Although CEH commissioners and staff were not statistical experts, they engaged the data constantly, so much so that a full-time statistical analyst had to be hired to keep up with the demand. 99 CEH researchers queried the database for lists of case summaries that fit elaborately precise criteria. Some CEH report chapters used precise statistical analyses, while others compared CEH data to NGO databases given to the commission. It is noteworthy that the first task upon every visit of the methodological advisor to the CEH was to brief some or all of the commissioners on the statistical patterns emerging from the testimonies.
Much of this difference can be explained by the dominance of hearings in the TRC process. The public hearings captured the South African public imagination, "defin[ing] the parameters of truth-seeking," 100 for over two years. The hearings appeared on television, on radio, and were covered nearly every day by newspapers. 101 But the hearings included only about 8.5 percent of the more than 21,000 people who gave statements to the TRC. By law, commissioners were required to make findings on all of the statements given to the TRC, a mundane and tedious process. The intensity of the hearings was such that the findings on statements were put off until the closing moments of the TRC's life. Participating in the hearings simultaneously occupied every available moment of the commissioners' lives and prevented them from engaging with the mass of statements. 102 Given the commissioners' quotidian participation in most aspects of the TRC's work, this meant that almost no staff worked with the statements until the final [End Page 29] months, except for those few staff actually engaged in taking statements or data processing. In the last three or four months before the report was published, with the hearings reduced to video and typed transcripts, TRC researchers discovered the power of structured data. Only then were statistics used on a broad scale, and even then, the statistics were used only by the researchers. 103
In Guatemala, the CEH had no primary data except from the interviews. With no hearings to distract them, nearly the entire commission's attention was focused on the interviews. The database was, by design, the easiest way to access the interviews, and as a result, it was used intensively. This is not to say that CEH staffers were happy with the database, but that their complaints had a fundamentally different character than the complaints in South Africa. CEH staffers' complaints tended to be very specific demands that the database should do a particular task that at that moment it did not do. 104 Complaints of this kind are indications that the tool is being used and that it needs to be further refined. By contrast, at the TRC it was said (constantly if vaguely) that "the database is not working," but few people had ideas or tasks they wanted to have done that any database would have been capable of doing.
The preparation and dissemination of a report play a central role in shaping the truth communicated and disseminated by a specific truth commission. The report should elaborate the truths required to fulfill the mandate, and it should be written with a rigor and quality to withstand sustained criticism. Yet the preparation of a report rarely receives the time and attention it deserves. Often the level of fatigue and conflict among the staff and commissioners is so considerable at this point that the commission begins to disintegrate. Key figures in the life of the commission may no longer be available or involved. Despite these limitations, and even if the report itself is not a commission's most important product, the report is what a commission leaves for history.
That the final reports of truth commissions vary considerably is clear from a cursory review, but only some of these differences reflect careful or conscious decisions. The Haitian commission provides the most extreme [End Page 30] example of terminal conflict and fatigue. Scheduled to complete the fieldwork by August 1995 and produce the report by mid-December 1995, the final report was actually transmitted to President Aristide in February 1996, a few months late. But the report submitted was extremely truncated. 105 It consisted of five pages of description of the modes of repression, seven of the structures and perpetrators, and twenty-one of recommendations. Most of the report was a listing of the cases of the victims who were identified, with several appendices based on supplementary research conducted by AAAS consultants. Although extensive statistical analysis of data from the interviews and from morgue records had been conducted, these findings were not included. Instead the report printed the headings for a few of the statistical tables without showing any of their data cells. This embarrassing omission reflects the condition of the commission rather than the intent of the drafters. In addition, there were serious problems with the translation and production of the appendices, particularly the forensic analysis of the exhumations of graves.
In part, the reduced report was the result of organizational collapse. But the hiring of an outside expert in December 1995 to write the CNVJ report also reflected deep disagreements among the commissioners about what the report should say. The regional reports with their rich analytical and contextual material (discussed above in subsection 2 on Data) were excluded in favor of a superficial chronology constituting sixty-one of the report's 105 pages. The analysis and synthesis--a social scientific model--was eliminated in favor of a legalistic review of the facts.
The TRC report consists of five volumes. The first volume includes an introduction by the Chair followed by philosophical and methodological discussions and administrative reports. The second and third volumes present chronologically organized overviews of perpetrating units and geographic areas and were written primarily by a small group of staff, with the Research Director playing the central role. The fourth volume reviews findings from institutional and sectoral hearings. The final report provides a preliminary list of victims, an interim report from the Amnesty Committee, additional details on how the Commission made findings, and in chapter six, the findings and conclusions of the TRC.
The TRC report presents substantially different conclusions in different sections. It opens--the first paragraph on the first page--with a statement by Archbishop Tutu that South Africa "is soaked in the blood of her children of all races." 106 His introduction turns on lists of incidents (e.g., paragraphs one and twenty-nine) which seem carefully calibrated to include atrocities [End Page 31] committed by a spectrum of perpetrators against a range of victims. Yet some 2,500 pages later, in volume five, the TRC concludes that "[t]he predominant portion of gross violations of human rights was committed by the former state through its security and law-enforcement agencies," and "[b]lack . . . youth were demonized as the 'enemy' by the security forces in particular." 107 The difference between these positions is stark. The first is focused on reconciliation, whereas the second is a broad empirical conclusion. While both could be true in a literal sense--there were some victims of every race, although black youth were by far the majority--the moral and empirical implications of the two sections are not consistent.
In direct contrast to the rhetorical foreword of the TRC report, the CEH introduces their work with unequivocal conclusions. On 25 February 1999, the report was published as a ninety-two page recommendations and conclusions volume, of which more than 40,000 copies were immediately distributed in Spanish and English. 108 The first paragraph includes broad statistical findings: that the CEH documented 42,275 victims, 83 percent of whom were Mayan, 29,830 of whom were killed or disappeared. In paragraph two, the CEH estimates that more than 200,000 people were killed, and in paragraph three, they conclude that "the violence was fundamentally directed by the State against the excluded, the poor and above all, the Mayan people, as well as against those who fought for justice and greater social equality." 109
The twelve volumes of the CEH report include most of the components in the South African report, but with very different emphases: the administrative report, legal mandate, and technical functioning get less treatment than in the South African report. The historical section is much more detailed, but there are no geographic or chronological reports. The bulk of CEH volumes two and three analyze specific kinds of violations such as sexual violence against women and genocide. None of these sections is strictly chronological. Instead, most sections begin with legal frameworks for what the crime means, offer some statistical overview of what the CEH found in this area (when and where it happened, against whom, and committed by what organizations), and proceed to theoretical arguments about how and why this kind of crime occurred. The expositions are based on frequent quotations from testimonies and occasional graphs or other statistics, as well as references to secondary materials. [End Page 32]
The CEH and TRC reports were drafted in fundamentally different ways. As described earlier, the TRC report is a compilation of essays written by individuals, beginning with the Archbishop's foreword, the discussion of kinds of truth by senior staff, reports from technical and administrative staff and consultants, researchers' reports on perpetrators and areas, and then reports by the commissioners who organized particular institutional hearings. In contrast, the CEH report was written in repeatedly restructured drafts, in which authorship of each section rotated among several people. When the CEH office heads returned from the field in February-March 1998, they were immediately required to write extensive area reports, many of which were several hundred pages long. These reports, called building blocks (insumos) became the inputs for the first of several rounds of evolving drafts of the report. Each geographic field report was then handed to new teams formed during June-August to work on thematic areas. These reports were then passed to new teams (in September-December) who wrote the actual material published in the report. The final stage was guided by a senior "Coherence Advisor" whose job was to assure that all the sections followed the broad outlines of the commissioners' analysis, and that the sections did not contradict each other.
The CEH process could easily have collapsed: it required more than a year's time, dozens of analysts, and much, if not most, of the drafted material was wasted (in the sense that it did not appear in the final report). However, given sustained effort and sufficient time, money, and staff, the CEH model proved to produce a more complete, consistent, and coherent report than any other commission to date. The report's greatest achievement is not the balance of evidence, the use of data, or its legal sophistication. The report's tremendous strength comes from its simple unity of premise: that the origin of the conflict and the overwhelming majority of violent acts were the responsibility of the Guatemalan state. At the end of the day, it is far easier to clarify history than to find truth and promote reconciliation through public hearings.
The analysis of whether each commission was successful must be made in its own terms. The TRC report should not be considered the TRC's only product. Indeed, the TRC themselves describe the report as a report on their [End Page 33] activities, with the activities--hearings, primarily--being the central focus of their enterprise. If the impact of the TRC is to be evaluated, the hearings should be the focus. The success of the CEH rests much more uniquely on the reception and impact of the report because it was the Guatemalan commission's only product.
Nonetheless there are several problems with the TRC's approach to truth in their report. That the TRC commissioners believed that the process by which the truth was reached and whether it affirmed the dignity of the participants counted, perhaps as much as the actual outcome or findings, 110 had obvious implications for their truth-finding. The zero-sum balance between public hearings focused on reconciliation versus rigorous truth-finding was debated openly within the TRC. 111 However, the value of empirical truth--macro and micro--was debated less in terms of its long-term impact on the culture of human rights and more in terms of fulfilling particular articles in law that protected the TRC from legal attacks, although South African victims and survivors of gross human rights abuses were explicit that meaningful truth finding was a precondition to reconciliation. 112 In general, the TRC failed to recognize the extent to which its various approaches to truth conflict with one another and with reconciliation. Its characterization of factual or forensic truth is an inadequate and impoverished rendition of what objective research entails and what it can contribute to the analysis and findings of a truth commission. Nevertheless, it probably reflects the attitudes of many of the commissioners. Archbishop Tutu's personal commitment to forgiveness and reconciliation and the presence of others from a religious background likely shaped many other aspects of the work of the TRC. It may account at least in part for the moral rather than factual or analytic orientation of the report and its strenuous effort at fairness that had the unfortunate effect of seeming to balance the systematic abuses of the government with the occasional excesses of the ANC. 113 The concern with reconciliation may also be related to the reluctance of the TRC to exercise its considerable powers of subpoena, search, and seizure, which [End Page 34] were much stronger than those of most other truth commissions. It may have contributed toward the process of reconciliation by emphasizing the validation of the individual subjective experiences of those who had previously been silenced or voiceless. In many cases, the reconciliation achieved in the hearings is real. And whatever the criticisms of the process, it is certain that for over two years, the TRC successfully focused South Africa's attention on the evils of the apartheid period such that no South African can now claim that he or she does not know at least some of what happened.
The CEH offers a very different model that was in our judgment more successful than the TRC at establishing macro-truth. However, the November 1999 elections resulted in a landslide victory for the party of General Efraín Ríos Montt, who was the dictator responsible for many of the most egregious atrocities of the early 1980s. Guatemalans elected the General's party only a few months after the presentation of the report in which his government was indicted for acts of genocide. Given this outcome, it is not clear what positive political impact the CEH's report may have had, no matter how rigorous the report may have been. But the CEH's lessons should be noted, and a new book by the UN Office of Project Services for Guatemala examines the lessons in detail. 114
Acknowledgment, the translation of a private into a public truth, requires widespread dissemination of the findings of a truth commission. Here too truth commissions differ. Again the experience of the Haitian National Commission for Truth and Justice testifies to its breakdown. The final report transmitted to President Aristide during his last hours in office in February 1996 was not published until September 1996, and then only seventy-five copies were made. Conflicts between the supporters of Aristide and the administration of President René Preval made the report something of an embarrassment for President Preval. A second edition of 1500 copies was published in March 1997, but these were circulated only in Haiti. 115
Although the TRC placed a considerable emphasis on its processes and findings being publicly accessible, its strategy for disseminating its report fell far short of this standard. Its five-volume report was released coincident with its public presentation, but it was published by an outside printer rather [End Page 35] than the government press. As a consequence, the set was priced well beyond the means of most South Africans. The Truth Commission also placed the full text on an open web site, but only for a limited period of time. Its agreement with its publisher, CTP Book Printers, precluded a long-term posting on the Internet because it would have decreased sales for the print version. The original plan was to have a writer or journalist prepare a one volume summary for the general public, but eighteen months after the submission of the official report, this popular version has yet to surface.
In contrast, the Guatemalan Commission for Historical Clarification had a coordinated strategy for public dissemination of its findings. It printed 42,000 copies of the CEH's conclusions and recommendations (in English and Spanish) that were disseminated on the day of the report's presentation to the Parties of the Peace Accord. Simultaneously with the presentation in printed form, the full text of both the English and Spanish recommendations and conclusions was placed on the Internet. 116 Copies were distributed to the press, libraries, universities, and other sites at the same time. The Sunday following the public presentation, major newspapers in Guatemala carried supplements with much of the text of the summary. A week later the full text of the report was posted on a permanent dedicated web site, and notices of its availability and how to access it placed on other Internet sites focusing on Guatemala and newspapers. The first five volumes of the complete report were published in Guatemala in July 1999, and the seven supplemental annex volumes were published in October.
The difference in dissemination between the South African and Guatemalan commissions may be the difference between choosing a private sector publisher and a public sector publisher. The private sector publisher in South Africa has removed the report from the Internet, priced it at levels unreachable for most South Africans and for many internationals, and generally restricted the circulation of the information. 117 In Guatemala, the UN Office of Project Services (UNOPS) has given away tens of thousands of summary volumes, donated hundreds of complete sets to schools, libraries, and NGOs, and financed the publication of an indexed and hyper linked CD-ROM which was given away and sold for a nominal price.