Expert Witnesses and International War Crimes Trials: Making Sense of Large-Scale Violence in Rwanda

Abstract

This chapter considers international criminal courts as institutions that produce their own narratives about extreme violence. It also explores war crimes courts as changing social spaces, with their own cultures and dynamics. The chapter argues that the judicial decisions made by the Yugoslav and Rwanda Tribunals in the early stages of their work were shaped by a meta-narrative about the nature of large-scale conflict that predominated in the 1990s. Expert witness evidence on the social and political contexts of the conflicts provided the tribunals with a particular form of knowledge that allowed them to make sense of the conflicts in Yugoslavia and Rwanda in terms consonant with this meta-narrative. This chapter then explores the role of expert knowledge at the Rwanda Tribunal, and in particular, the expert testimony of Alison Des Forges. Comparing two important cases before the Rwanda Tribunal, Akayesu and Bagasora, the chapter explores the ways in which Des Forges' expert evidence was influential in shaping the work of the Tribunal and its understanding of the causes and consequences of the 1994 genocide.

The international prosecution of war crimes—by the Yugoslav and Rwanda Tribunals or the permanent International Criminal Court—generates lengthy, detailed judgments about what happened during times of extreme political and military conflict. These international courts are increasingly expected to provide a careful accounting of the patterns of the conflicts, the nature of the different harms suffered, the circumstances that lead to the violence and the roles of different individuals, armies and militias in creating the conditions in which crimes were committed. In this painstaking work, the courts amass volumes of evidence, witness testimony, documents, videos and photos. The resulting written judgments, many running into hundreds of pages, condense this massive archive into a detailed analysis and accounting of the conflict and violence. The judgments both author a narrative about the causes and contexts of large-scale violence and are part of the record about what happened.

In this chapter, I am interested in international criminal courts as institutions that produce their own narratives about extreme violence. International courts are complex institutions, comprised of different bodies—the registry, trial chambers, the Office of the Prosecutor (OtP)—which act in varied, and sometimes contradictory, ways. Judicial accounts of ‘what happened’ emerge from this complex and changing space. Within the limits of this chapter, I want to begin the process of examining some of the practices, dynamics and actors that underpin and shape how trial chambers come to understand and produce an account of extreme violence. My focus here is on the Rwanda and, less-so, Yugoslav war crimes tribunals which have been in existence for more than 15 years and have generated their own substantial record through judicial decisions and judgments of the events in these two regions. As institutions of comparative long duration, the Rwanda and Yugoslav Tribunals also have had their own life cycle and knowledge trajectories.

Both institutions have relied upon expert witnesses to provide some of the evidence about the historical, social, political and economic contexts of the conflicts in the two regions. In this chapter, I consider the role of ‘context expert witnesses’, as they are sometimes called, in shaping the tribunals’ account of what happened. My focus is specifically on the expert testimony of Alison Des Forges, a renowned international human rights activist and Africanist scholar, who testified in 11 cases at the Rwanda Tribunal. I explore the transcripts of her testimony in two Rwanda Tribunal cases, Prosecutor v. Akayesu ¹ and Prosecutor v. Bagasora ², to explore three themes. First, and as noted above, I am interested in international criminal courts as dynamic, multivocal and multifaceted institutions. Within transitional justice literature, there is a growing interest in international courts as structured by social relations (Kelly and Dembour 2007). Scholars of the sociology of knowledge have similarly emphasized the importance of exploring the ‘day-to-day actions and processes through which’ knowledge is made in order to understand the ‘specific historical contexts’ and ‘multidimensionality’ of social action (Camic et al. 2011). This chapter explores the microdynamics within the Rwanda Tribunal to highlight some of the people, circumstances and practices that shaped, at least in part, the tribunal’s account of ‘what happened’ in Rwanda in the 1990s.

Second, Alison Des Forges’ testimony provides a useful lens for exploring the relationship between expert evidence as a type of knowledge and the Rwanda Tribunal’s use of it to author an account of the causes and contexts of large-scale violence. Mary Poovey (1998; see also Hacking 2002; Valverde 2003) has highlighted the relationship between the format of knowledge and the ‘available ways of organizing and making sense of the world’ (Poovey 1998, p. xv). That is, the ways in which knowledge is presented and represented can tell us something about the world views that are possible at a particular time and place.

Finally, I argue in this chapter that some expert witnesses, Des Forges in particular, were more influential than is sometimes credited by tribunal insiders. Des Forges’ testimony, I argue, had a significant impact in individual cases, particularly in the early stages of the Rwanda Tribunal. Her evidence provided a compelling framework within which the genocide was understood and provided the basis for the legal determination that the crime of genocide was even applicable. Her evidence was additionally influential in subsequent cases by shaping, in part, the way defence challenges to the dominant narrative of the genocide unfolded. I explore some aspects of this dynamic through a discussion of Des Forges’ evidence in Bagasora. The transcripts of her evidence in this case also reveal a changing Rwanda Tribunal that was evolving as an institution with its own personalities and cultures. The resulting judgment in Bagasora, I argue, reveals a court that was both more knowledgeable about Rwanda while less confident in its ability to know the causes and contexts of the 1994 genocide.

This chapter is based on a series of interviews and documentary analysis of both the Yugoslav and Rwanda Tribunals, but with a much more focused look at the International Criminal Tribunal for Rwanda (ICTR) .³ My arguments here are sometimes comparative, particularly in the first two sections of the chapter, where I trace the meta-narrative about the conflicts that shaped, in part, the establishment of the two tribunals in the 1990s. In other respects, this chapter provides a close reading only of the Rwanda Tribunal and only of certain cases heard at that institution. Legal arenas are curious knowledge-producing sites . Each case provides its own context in which decisions are made about what knowledge will be considered authoritative (Valverde 2003, Chap. 1). At the same time, certain institutional norms, practices and cultures are constituted and travel from one case (or ‘trial chamber’ in the ICTR/Y context) to another through legal precedents and the movement of people (i.e. judges at the ICTR often heard more than one case at the same time) (Eltringham 2011). In a further complication, some of the trials at both tribunals, but the ICTR in particular, took years to complete and generated their own cultures, contexts and courtroom dynamics. My discussion here is thus rooted, as much as possible, in the particular circumstances of individual cases. But, as I discuss below, the dynamics in one trial chamber can and do have impacts on other cases.

This chapter is not about the Rwanda genocide. It is much more narrowly a consideration of the knowledge practices of the international tribunal tasked with prosecuting crimes committed during the genocide. Underlying this study is a concern with the ways in which the causes and contexts of large-scale violence, such as the Rwanda genocide, are understood in ‘official’ knowledge-producing sites. The complexities of the Rwandan genocide and the politics around how the genocide is talked about and understood have been the subject of extensive, sophisticated analysis, (See e.g. Eltringham 2004; Hintjens 2008; Pottier 2002) a review of which is beyond the limits of this chapter. This scholarship, and the concerns expressed about the reduction of complex violence to simple causal explanations, influenced my thinking in this chapter. But my concerns with the simplification of complex forms of violence should not be interpreted as in any way denying the scale and the devastation of the violence in Rwanda in the 1990s including the 1994 genocide.

This chapter begins with a brief explanation of expert evidence and its role in international crimes prosecutions.

The Place of Experts in International Criminal Trials

Expert evidence is the oral and written testimony of ‘a person whom by virtue of some specialised knowledge, skill or training can assist the trier of fact to understand or determine an issue in dispute’.⁴ In criminal trials, expert evidence is generally thought of as coming from the forensic sciences: DNA testing, fingerprints, lie detectors and the like. The expert in DNA, for example, might testify before a criminal court that the DNA found on the weapon matched the DNA sample taken from the defendant.

International criminal courts also rely on expert evidence of this type and might include, for example, forensic anthropologists who testify about the state of bodies recovered from a mass grave or an expert on typewriters who can testify that a document likely was produced by a certain brand of typewriter (and hence linked to a particular political or military office). But, expert evidence can also come from the social sciences and humanities, though this form of expertise tends to receive much less scholarly attention.

At the Yugoslav and Rwanda Tribunals, experts with backgrounds in anthropology, history, political science, law and sociology testified in multiple cases about the social, political, historical contexts that lead to, and then shaped, the outbreak of armed conflict and violence in these two regions. This expert evidence—often referred to as ‘context’ or ‘linkage’ evidence—was primarily driven by two factors relating to the structure and objectives of international war crimes trials established in the 1990s.

The first factor is the definition of the crimes prosecuted by the courts, particularly genocide and crimes against humanity, which require evidence of ethnic, racial or religious social groups and the historical context of social relations between groups. A crime against humanity is a criminal act—torture, killing, rape—committed as part of, and in furtherance of, a larger attack against a population. The Rwanda Tribunal statute goes further than the Yugoslav Tribunal to specify that the attack must be against a population identified by ‘national, political, ethnic, racial or religious grounds’. Genocide similarly is the commission of certain acts with the intent of destroying a ‘national, ethnical, racial or religious’ group. Evidence is thus needed to establish both the existence of distinct groups or populations and their relations overtime (to explain, for example, why a particular act should be understood as part of a larger attack against a group). For example, in the Rwanda Tribunal decision in Akayesu (discussed in more detail below), the Court heard lengthy evidence about how the conception of ‘Tutsi’ and ‘Hutu’ as distinct groups within Rwandan society was historically and socially constructed and then hierarchically ordered in colonial and post-colonial Rwanda. In Akayesu, the Rwanda Tribunal had the difficult task of determining if Hutu and Tutsi were distinct groups within the meaning of the crime of genocide which is limited to ‘national, ethnical, racial or religious groups’.⁵ Further, the court heard evidence about the history of post-colonial Rwanda, where acts of violence were committed against Rwandan Tutsi and Hutu at different historical junctures. This historical context was offered to demonstrate that in different periods of political uncertainty, ethnically inflected violence was orchestrated by Rwandan elites to consolidate their hold on power.

The second major driver for context expert evidence is the tribunals’ focus on prosecuting the leaders said to be responsible for large-scale violence. Other legal processes within the affected regions, such as local criminal trials or ‘traditional’ justice measures, are meant to address the ‘foot soldiers’ or ordinary citizens who engaged in violence and atrocity. While both tribunals have prosecuted some defendants who might be seen as comparatively minor figures, both institutions have dedicated most of their resources in the pursuit of complex, often lengthy ‘leadership trials’, against individuals like Slobodan Milošević, Radovan Karadžić, Ferdinand Nahimana and Théoneste Bagasora, who were seen as the intellectual, political and military leaders in the regions.

The leadership focus of the tribunals necessitates expert evidence first to demonstrate that the accused were indeed leaders, something that might not be apparent from their formal title, and second, to establish a link between the ‘leader’, who may be distant from the physical locations of violence, and the atrocities committed ‘on the ground’.

Context expert witnesses have attracted some controversy within the tribunals. Some lawyers and judges I interviewed felt these experts complicated and distracted the proceedings from the more central focus on prosecuting and defending individual accused. This view resonates within some transitional justice scholarship that has portrayed the production of a historical accounting of ‘what happened’ as incompatible with the narrower criminal trial structure and focus on individual guilt. Some scholars have suggested that war crimes trials are too prone to political interference (over which version of history will be authored), (Arendt 1963) or too limited as legal venues, with the rigid rules of evidence narrowly focused on guilt or innocence, to produce a thick accounting of the causes and circumstances of large-scale conflict and atrocity (Eltringham 2009; Petrovic 2009; Simpson 2007).

More recently, Richard Wilson (2011), in an extensive study of the use and production of history in contemporary international criminal courts, has concluded that historical evidence, and the production of a record, is now a regular feature of war crimes prosecutions, required by legal elements of the crimes prosecuted and made possible by the large amount of documentary, witness and expert evidence compiled by the courts. Not only do contemporary international courts routinely engage in ‘historical forays’, he argues, the resulting accounts of history could even be seen as ‘reputable’ (Wilson 2005).

The remaining discussion in this chapter builds in part on Wilson’s conclusions that the production of a historical record is now a recognized feature of contemporary war crimes trials. But while Wilson may be correct that the Yugoslav Tribunal, at least, is producing ‘reputable’ accounts of the conflicts, this begs the question: which accounts of what happened become reputable and at what junctures? In the next section, I argue that context expert evidence at the two tribunals was driven in part by a meta-narrative about the causes of the conflicts in Rwanda and Yugoslavia that prevailed in the 1990s. That narrative, I suggest, depicted the conflict and genocide in Rwanda as elite-orchestrated, ethnically directed violence, designed to create the conditions to secure or maintain power.

Making Sense of Violence

In the mid-late 1990s, when both tribunals were beginning their work, there was a growing consensus in some international policy and scholarly circles that the conflicts in these two regions were problematically portrayed as resulting from tribal, atavistic hatreds that were beyond rational explanation or effective intervention. Scholars and policy practitioners began emphasizing the conflicts as constructed, elite driven, modern and intentional.⁶ Robert Hayden, a US-based scholar of the Balkan region and two-time expert witness at the Yugoslav Tribunal, describes Western state and institutional responses to the war in Yugoslavia as ‘informed by a particular teleology—in which the demise of Yugoslavia was an aberration, a disaster caused by evil politicians, whose culpability needs to be shown so that normalcy can be obtained’ (Hayden 1999). Scott Straus (2006), writing on the Rwanda genocide, refers to a ‘new consensus’ about Rwanda that emerged in scholarly and activist work in the 1990s. ‘Rather than seeing the violence as chaotic frenzy, as state failure, or an explosion of atavistic animosities, scholars and human rights activists alike stress the violence was modern, systematic and intentional’.

The ‘new consensus’, as Strauss calls it, emerged partly in response to simplistic characterisations by Western media and some political leaders of the conflicts as ‘ethnic’ or ‘tribal’ violence. Warren Christopher, then secretary of state in the USA, described the violence in the Yugoslav regions in 1993 as a ‘problem from hell’: ‘The hatred between all three groups…is almost unbelievable. It’s almost terrifying, and it's centuries old’.⁷ Christopher’s phrase—‘a problem from hell’—became the title of Samantha Power’s Pulitzer prize-winning book which was a strongly argued tour de force about US government failure to officially recognize and respond to genocide and mass violence in various non-US locations, including Yugoslavia and Rwanda. For Samantha Power, Christopher’s comment reflected a US government strategy to absolve itself of responsibility by characterizing the violence in Yugoslavia as an ‘amoral mess’, centuries-old grievances that were beyond US intervention (Power 2002, p. 306).

Against the powerful trope of ‘tribal violence’, the new consensus was, on one level, an attempt to insert complexity and responsibility (not only of local leaders but also Western decision-makers) into the discussions about mass atrocity. As some scholars have noted, however, the responsibility called for in analyses like Power’s is narrowly focused on international/US responsibility to save the regions from violence (Orford 2003). Other scholars have noted that ‘elite responsibility’ explanations for some contemporary conflicts can generate their own distorting policy prescriptions that fail to take account of the complex, multi-level contexts within which violence has unfolded (Autesserre 2010; Kalyvas 2006).

For my purposes, an understanding of violence as elite orchestrated (rather than mindless group violence) suggests an epistemological claim that has important implications for international criminal prosecutions. If the conflicts are understood as caused by rational, calculating individuals, then what happened in Rwanda and Yugoslavia ‘was not tribalism run amok; it was genocide’ (Straus 2006, p. 33) or, other, related, international crimes. The dominant conception of the conflicts as stemming from criminal actions of individual political, military and social elites makes it possible to conceive of international criminal prosecutions as a justifiable international response to large-scale violence.

The understanding of large-scale violence as resulting from cynically contrived, local, elite-orchestrated ethnic conflict, I suggest, exerted a powerful conceptual pull in the early days of the tribunals. The lawyers, judges, clerks hired to work at the tribunals were drawn from various regions across the globe and almost none of them had any working knowledge of the conflict regions or even the local languages. Faced with the enormity of prosecuting the most serious of international crimes—genocide and crimes against humanity—the judges and lawyers sought a framework within which they could explain what had happened. ‘What they [International Criminal Tribunal for the former Yugoslavia (ICTY) judges] pretty much all feel is a need for a vehicle to frame the alien world they are being asked to render judgments on’, according to Robert Donia, a 14-time expert witness at the Yugoslav Tribunal.⁸ Navanethem Pillay, former judge and eventual president of the Rwanda Tribunal, described the approach of judges in the first Rwanda Tribunal case of Akayesu in similar terms: ‘We judges agreed that you can’t avoid this question of history of Rwanda, otherwise it’s just one ethnic group killing another ethnic group with no reason why. History is necessary for an understanding of why the conflict occurred. Our first judgment—Akayesu—did this’ (Quoted in Wilson 2011, p. 72).

The prosecutors Robert Donia encountered at the Yugoslav Tribunal were not looking for just any kind of historical narrative, he found, but ‘one that could be inserted in indictments, could impute motivations of actors as rational rather than crazed and wild. In general, make some sense’.⁹ Making ‘sense’, for both Donia and Judge Pillay, is rooted in a dichotomous way of conceiving the violence; as tribal or atavistic hatred on the one hand or intentional elite orchestration on the other. Within the logic of this dichotomy, the tribunals’ focus on elite responsibility is not seen as a ‘version’ of what happened but simply as common sense; a laudable rejection of the ‘ancient ethnic hatreds’ characterization.

The context experts who testified in the first 5 or so years at the tribunals, but the Rwanda Tribunal in particular, provided the judges and the prosecutors with the means to construct such an explanatory framework. The most often cited expert witnesses in the first few years of the tribunals were from Europe or North America, with doctoral degrees in topics related to the regions (in various disciplines: history, political science, sociology, anthropology, law), who had a grasp of the local languages, knew the history, culture and political contexts and had done (often considerable) field research in the regions.¹⁰ These experts were able to explain the regions to the lawyers and judges in ways that provided a framework within which the tribunal personnel could ‘make sense’ of what happened.

In the following section, I explore this ‘framing’ role for expert context evidence in the first case before the Rwanda Tribunal—Prosecutor v Akayesu, which resulted in the first conviction for genocide by that court in 1998. In this discussion, I focus on the expert evidence of Alison Des Forges, the woman who would become the most frequently appearing expert witness at the tribunal and an internationally celebrated expert on the Rwanda genocide. Her evidence, I suggest, had a significant impact in that first case in framing the judges and prosecutors’ understanding of what happened and shaping, in turn, how defendants in subsequent cases responded.

Framing Genocide: Prosecutor v. Akayesu

When the Rwanda Tribunal was established in 1994, the OtP turned for assistance to three Rwanda experts: Alison Des Forges, a US-based, long-time scholar of Rwanda who, at the time of the genocide, was working closely as a volunteer (later as an employee) with the organization that would become Human Rights Watch, André Guichaoua, Professor of Sociology at Université des Sciences et Technologie de Lille, France and Filip Reyntjens, Professor of African Law and Politics at the University of Antwerp, Belgium. These three acted as advisors and provided training to the OtP. In the early days of the tribunal, the three would meet with the OtP in various locations, even at Reyntjens’ home in Antwerp, where the prosecutors would, as Reyntjens’ described it, ‘pick their brains’ about Rwanda and the genocide.¹¹

In 1994, when the genocide in Rwanda unfolded, few scholars or diplomats were familiar with the country and Des Forges, Guichaoua and Reyntjens constituted a sizeable part of the scholarly community undertaking research on Rwanda.¹²Filip Reyntjens¹³ describes the researchers working on Rwanda as a ‘very small community’, who, in the days immediately following the genocide, saw their role primarily as activists: ‘trying to save people, lobby the international community, et cetera’. This small research community, Reyntjens notes, ‘implicitly’ divided the research labour of studying the genocide among themselves. Reyntjens settled on an intensive study of the 3 days following the downing of President Habyarimana’s plane on 6 April 1994, seen as the start of the genocide (Reyntjens 1995), while Des Forges, along with a team of researchers from Human Rights Watch, undertook a massive study of the genocide itself (Des Forges 1998). All three—Reyntjens, Des Forges and Guichaoua—became an important resource for the OtP with each testifying in numerous cases (with Reyntjens also testifying once for the Defence).¹⁴

Of the three, Alison Des Forges testified the most often, giving evidence in 11 ICTR cases, as well as appearing in numerous legal proceedings in national courts (Switzerland, Canada and Belgium, for example) involving the Rwanda genocide and several immigration cases (in Canada and Belgium).¹⁵ At the time of her death in 2009 in a plane crash in New York State, Des Forges was a leading figure in human rights circles, in part because of her commitment to testifying about the Rwanda genocide.¹⁶ Des Forges’ study of the genocide, published as Leave None to Tell the Story, became seen by lawyers and experts at the tribunal as the ICTR’s ‘bible’.¹⁷ It won the Raphael Lemkin award by the Association of Genocide Scholars and was listed as one of best books of the year by the Los Angeles Times. Des Forges herself was named a MacArthur Fellow and given a ‘Genius Grant’ in 2000.¹⁸

Alison Des Forges was the prosecutor’s main expert witness in the first ICTR case of Prosecutor v Akayesu and she was on the witness stand for 8 days in 1997. The transcripts of her evidence¹⁹ suggest a very positive interaction between her and the three judges who comprised the trial chamber: the presiding judge, Laïty Kama, from Senegal, Lennart Aspegren, from Sweden and Navanethem Pillay, from South Africa.

In the blended common and civil law procedures at the tribunal, judges can and do ask follow-up questions. In Akayesu, the judges asked wide-ranging, open-ended questions about a myriad of things, from the meaning of ethnic identity as a sociological term²⁰ to very specific questions about Rwandan history and the events leading to the genocide.²¹ At one point, the presiding judge even commented to Des Forges about the large number of questions asked of her: ‘Madam, I hope you will understand that we are taking a lot of your time because your testimony is particularly of great importance. You know very well Rwanda, as well as its history. And also the judges will be benefiting from your knowledge…I hope that we will not be accused of asking too many questions’.²²

Des Forges’ role at this point in the tribunal’s life cycle was largely educative. The transcripts of her testimony read like a series of erudite mini lectures delivered to the trial chamber. While she was cross-examined by defence counsel for many days, her mini lectures unfolded with little challenge or interruption.²³ The prosecution’s questions, like the judges’, were also wide-ranging covering topics not limited to the matters raised by the indictment. At one point, the prosecutor asked Des Forges about the conduct of the Muslim community during the genocide, a subject not related to the charges against the defendant, Jean-Paul Akayesu. ‘As you know’, the prosecutor explained, ‘this process, in as much as it is a trial, is also a process for keeping accurate records for posterity’.²⁴

The written judgment in Akayesu relies significantly on Des Forges’ evidence, particularly in establishing the history and context leading up to the 1994 genocide. The second section of the judgment, entitled ‘Historical Context of the Events in Rwanda in 1994’, contains 33 paragraphs summarizing Rwandan history from the start of German colonial rule in 1897 and concluding with a brief overview of the events in April 1994, which the trial chamber then explores in more detail later in the judgment (ruling that a genocide did occur). Alison Des Forges is the only witness specifically referenced as an authority for the chambers’ summary of Rwandan history.

The ‘Historical Context’ section of the judgment is an outline of the judges’ interpretation of the aspects of Rwandan history they saw as relevant to the events in 1994. As an account of Rwanda history, 33 paragraphs in a legal judgment are obviously too short and there are inevitable gaps. But for a legal judgment, devoting 33 paragraphs to a review of Rwandan history is significant.

The 33 paragraphs tell the story of Rwandan history in terms of the political forces that shaped and manipulated Hutu and Tutsi as identity categories over the course of Rwanda’s colonial and post-colonial history. The first part of the ‘Historical Context’ section of the judgment, for example, details the construction of Tutsi and Hutu as identity categories, by German colonial authorities and their hierarchical ordering under Belgian rule. Hutu and Tutsi elites responded, in turn, to consolidate their hold on power or vie for additional power through the manipulation of those categories (Paragraphs 78–85).

The end of colonialism and the rise of the first Rwandan Republic in the late 1950s to the early 1970s is described as producing an often violent consolidation of Hutu and Tutsi identities (Paragraphs 86–91); elections in 1957, for example, saw ethnic rather than ideological voting patterns, and cross-border attacks by exiled Tutsi in the late 1950s and 1960s led to reprisal killings against Tutsi within Rwanda, generating in turn, more departures from Rwanda.

The dominant focus on ethnicity in this section of the ‘Historical Context’ is not entirely surprising given the judges’ need to resolve the legal question of whether ‘Hutu’ and ‘Tutsi’ are distinct ethnic groups and hence, if the crime of genocide was even applicable. This legal question is addressed in another section of the judgment where the judges rule that Hutu and Tutsi emerged over the course of Rwandan history as stable and permanent groups within the legal definition of genocide.²⁵ In reaching this decision, the judgment contains a lengthy quotation from Des Forges’ testimony which explains identity groups as being subjectively determined, shaped by a mix of ‘the actual conditions and peoples’ subjective perception of those conditions’.²⁶ The ‘Historical Context’ section, set out above, could be read as providing the raw materials for an explanation of how the ‘actual conditions’ of ethnic differentiation were established and then consolidated through the contrivance of various elites.²⁷

These ‘raw materials’ of ethnic differentiation are relevant when the ‘Historical Context’ moves to a consideration of the events of the 1990s and the factors that shaped the deteriorating political and military situation in Rwanda. At the very end of a long paragraph describing the end of the first Republic in 1975 and the conduct of the second president, Juvenal Habyarimana, in the 1970s and 1980s, the tribunal notes that on 1 October 1990, an army of Tutsi exiles, known as the Rwandan Patriotic Front (RPF), based in Uganda attacked northern Rwanda. ‘The attack’, the tribunal noted, ‘provided a pretext for the arrest of thousands of opposition members in Rwanda considered as supporters of the RPF’.

In the next 12 paragraphs, the ‘Historical Context’ explains the 4-year period leading to the genocide in terms of ‘the worsening internal situation’ in Rwanda caused by a number of developments including: increasing pressure on President Habyarimana’s government from growing opposition movements, as well as international donors, to share power and democratize; the war with the RPF and RPF demands to return to Rwanda (as well as their plans to overthrow the government, Para. 95) and the Arusha peace accords between the RPF and the Habyarimana government that provided for power sharing and the integration of the RPF into the Rwandan army.

This section explains how a deteriorating political situation was used by different political elites to vie for power, often through the manipulation of ethnic affiliation. For example, the trial chamber describes how the power-sharing arrangements in the Arusha peace negotiations, as well as the 1993 assassination of Burundian president Melchior Ndadye, a Hutu, by Burundian Tutsi army officers, were exploited by Hutu extremists within Rwanda to call for ‘solidarity among all the Hutu’ (Para. 103). President Habyarimana’s inner circle used a similar tactic, according to the judgment, intentionally heightening ethnic tensions to mask its own efforts to stall political change and hold onto/reclaim power. ‘To make the economic, social and political conflict look more like an ethnic conflict, the president’s entourage, in particular, the army, persistently launched propaganda campaigns which often consisted of fabricating events’, the trial chamber concluded.²⁸

The ‘Historical Context’ section ends with the start of the genocide on April 6, 1994, when President Habyarimana’s plane was shot down, and a brief outline of some of the main events in April 1994. From there, the judgment moves to a new section and a consideration of whether a genocide happened in 1994 in Rwanda or not. Not only was there a genocide, the trial chamber concluded, but it was also centrally organized. Here too, Des Forges’ evidence is cited prominently. ‘[T]he Chamber concludes from all the foregoing that genocide was, indeed, committed in Rwanda in 1994 against the Tutsi as a group. Furthermore, in the opinion of the Chamber, this genocide appears to have been meticulously organized. In fact, Dr. Alison Desforges testifying before the Chamber on 24 May 1997, talked of “centrally organized and supervised massacres”. Indeed, some evidence supports this view that the genocide had been planned’.²⁹

The trial chamber eventually concluded that the defendant in this first case, Jean Paul Akayesu, major of Taba commune, was guilty of genocide, incitement to genocide and crimes against humanity. Although Akayesu was not a ‘leader’ in orchestrating the genocide, the decision was a watershed in establishing some of the parameters of the judges’ understanding of the causes and dynamics of the genocide. It also reflects the judges’ confidence in summarizing Rwanda’s history and political developments, a confidence that is strikingly absent in later decisions (see below). In a 2003 speech, Judge Pillay, one of the three judges in Akayesu and who later served as president of the tribunal, observed that ‘it has been said that “those who cannot remember the past are condemned to repeat it.” The ICTR , through its jurisprudence and trial proceedings, is establishing a historical record of what happened in Rwanda between April and July 1994—a record which will help keep alive the world’s collective memory’.³⁰

Not all ICTR Trial Chambers in every case, even in the early years, provided a summary of ‘historical context’, and some even expressed misgivings about doing so.³¹ Yet, Des Forges’ evidence, I would argue, had a substantial impact on the Rwanda Tribunal even in cases where her testimony, or context evidence in general, was not specifically referenced. First and foremost, the Akayesu Trial Chamber’s conclusion, from Des Forges’ evidence, that Tutsi and Hutu were stable and permanent groups within the legal meaning of genocide, was a foundational move. As other scholars have noted, the Trial Chamber’s conclusions here appear premised on a problematic understanding of both ethnicity and the legal elements of the crime of genocide (Eltringham 2004, Chap. 1; Wilson 2011, Chap. 7). Nonetheless, from that determination, the tribunal was able to consider genocide charges against subsequent defendants. In 2006, the Appeals Chamber³² definitively concluded that the 1994 genocide and the existence of Hutu and Tutsi as stable and permanent groups were ‘facts of such notoriety, so well known and acknowledged that no reasonable individual with relevant concern can possibly dispute them’.³³ Trial chambers at the Rwanda Tribunal could, thus, take ‘judicial notice’ of these facts without having them proved. Importantly, the Appeals Chamber decision refused to categorize Hutu and Tutsi as stable and permanent ethnic groups, but relied on the Akayesu determination that they were stable and permanent groups within the meaning of the crime of genocide,³⁴ which was itself based on Des Forges’ testimony and an understanding of these groups as ethnic.

In this example and others, Des Forges’ evidence in Akayesu had an additional impact by travelling within the tribunal. The transcripts of her evidence, for example, were required to be read by OtP staff who arrived at the institution without any background on Rwanda.³⁵ More significantly, aspects of her evidence in Akayesu were used in other cases. For example, the ‘Historical Context’ segment of Akayesu was relied upon extensively in the ‘Media’ case,³⁶ in which three ‘ideologues’ of the genocide were on trial. The three judges who sat on that case included Navanethem Pillay as the presiding judge, Erik Mose³⁷ and Asoka de Zoysa Gunawardana, from Sri Lanka.³⁸ The judgment in that case ‘accepts the importance of this history [of Rwanda], particularly in this case, and for this reason sets forth largely in extenso the comprehensive review of the historical context as described in the Akayesu judgement’.³⁹ The judgment then reproduces verbatim 26 of the 33 paragraphs from the ‘Historical Context’ section of Akayesu.

The impact of Des Forges’ evidence is arguably felt in a third way; helping the trial chamber establish the framework within which the events leading up to, and during the genocide, were understood. In the words of one tribunal defence lawyer, the judges ‘rely on Des Forges even when they don’t say in their judgments how much they rely on her. Her evidence puts them in a state of mind for analyzing the rest of the evidence’.⁴⁰ The ‘Historical Context’ section in Akayesu based on Des Forge’s evidence can be seen as providing the very sort of framework that Robert Donia suggested the judges and prosecutors needed in order to ‘make sense’ of mass violence . It offers a narrative structure within which the events in Rwanda are explained as resulting from very modern, systemic factors (internal stress caused by the transition to democracy) combined with the nefarious actions of desperate elites who manipulated ethnic identity in order to secure power.

Aspects of this narrative, as it is reproduced in the ‘Historical Context’ section, provide an account of the events in the 1990s that are challenged in subsequent cases. Defence lawyers have argued that the war between the Rwandan army and the RPF had more of an impact on the events in the 1990s and the genocide than was portrayed in Akayesu and in Des Forges’ testimony. And indeed, the Akayesu judgment, while certainly taking note of the war, mostly discusses the role of the RPF in terms of the actions by Hutu hardliners to hold onto power and resist the Arusha peace accords between the Rwanda government and the RPF. The ‘Historical Context’ section only first mentions the 1990 RPF invasion of Rwanda, which started the war, at the end of a very long paragraph covering the period 1975–1990. When the invasion is mentioned again three paragraphs later, it is wrongly identified as occurring in 1991.

The judgment in Akayesu explains that while ‘the genocide against the Tutsi occurred concomitantly’ with the armed conflict between the RPF and the Rwandan army, that war can ‘in no way be considered as an extenuating circumstance’ for the genocide (Para. 128). But the arguments advanced about the role of the RPF in subsequent cases are used to make a different claim: to challenge the conclusion reached by Des Forges and the Akayesu judgment that the genocide was planned prior to April 1994. The trial chamber in Akayesu concluded that this planning did occur, the evidence of which can be found in the circulation of ‘lists of Tutsis to be eliminated’, ‘arms caches in Kigali’, the ‘training of militiamen by the Rwandan Armed Forces and of course; and the psychological preparation of the population to attack the Tutsi…by some news media’ (Para. 126).

Des Forges’ testimony in Akayesu was relatively uncontested over the course of the trial. This all changed in subsequent cases. Four years after Akayesu, Des Forges began testifying in the case against Colonel Théoneste Bagasora, the man alleged to have been a central actor in preparing for the genocide in the 1990s. In this case, some of the conclusions reached by the Trial Chamber in Akayesu were more centrally in issue and in particular Bagasora’s role in planning for the genocide through: compiling the lists of names of Tutsi and Hutu who would be exterminated,⁴¹ establishing, training and then arming militia groups⁴² and finally, presiding over the military commission established to determine how to ‘defeat the enemy militarily, in the media and politically’. That commission, it was alleged, resulted in various efforts to depict all Tutsi as the enemy and eventually led to the incitement to genocide.⁴³

In the following section, I examine the transcripts of Des Forges’ testimony in Bagasora which suggest a very different experience from her testimony in Akayesu. I explore a change in the courtroom climate from the relatively warm embrace of Des Forges’ evidence in Akayesu to a more combative and sometimes volatile environment in Bagasora. This change, I suggest, speaks to a shift within the tribunal itself, which was growing and developing as an institution with its own cultures, personalities and dynamics. More importantly, however, Des Forges’ experience in Bagasora and the resulting trial judgment suggest a change in both knowledge practices—how expert witnesses were viewed and relied upon, for example—and on the tribunal’s knowledge about and accounting of the events in Rwanda in the 1990s.

Contesting the Frame: Prosecutor v. Bagasora

By September 2002, when hearings began in Prosecutor v. Bagasora, a series of ‘scandals’ had significantly tarnished the tribunal’s reputation and placed it on a public relations collision course with an increasingly hostile RPF-led Rwandan government (Peskin 2008; Cruvellier 2006; Del Ponte and Sudetic 2008). Members of the defence counsel were also vocal about their concerns that no charges had been brought against the RPF for conducting its own massacres and killings in the 1990s (Del Ponte and Sudetic 2008, Chap. 7). And indeed, no charges would ever be laid by the tribunal, a fact that caused Filip Reyntjens in 2005, to write a strongly worded letter to then Chief Prosecutor Jallow, withdrawing his support from the OtP.⁴⁴

In response to concerns about the slow pace of the Rwanda Tribunal, the chief prosecutor in the early 2000s, Carla Del Ponte, was pushing forward with several large trials against multiple accused seen as leaders of the genocide. These were grouped into five, large, multiple defendant trials known in tribunal parlance as: Media, Government I, Government II, Military I and Military II. ‘Military I’ was the trial involving Bagasora and three other military officers. Bagasora was alleged to have been one of the main conspirators planning the genocide prior to 6 April and who then enabled the installation of the interim government following Habyarimana’s death, the same government in power throughout the genocide.⁴⁵

The lead prosecutor at the commencement of Bagasora, Chile Eboe-Osuji, called Alison Des Forges as his first witness. He explained to the trial chamber that Des Forges’ role as the first witness was to set the context for the remainder of the trial. She would, among other things, ‘give the story of the history of Rwanda’ and would help the judges ‘make better sense’ of the factual evidence that would follow.⁴⁶ In other words, the prosecutor was beginning with Des Forges so that her testimony would frame the rest of the evidence.

But things did not go according to plan for the prosecution. The lawyers for the four defendants objected to Des Forges as the first prosecution witness arguing that experts normally testify only after the factual basis has been laid by material witnesses. While the judges eventually agreed to Des Forges as the first witness, the prosecutor’s strategy proved to be a procedural distraction.⁴⁷

Des Forges testified for what must have been a gruelling 18 days spread over the months of September and November 2002, overlapping with some of the time she was testifying in the Media case. Most of her days on the witness stand were taken up by cross-examination by defence counsel. Even before she began testifying to the substance of her report in Bagasora, the three defendants objected to Des Forges’ qualifications as an expert, even though she had, at that point, been qualified as such in two other ICTR cases (Akayesu and Media) . Each of the three defendants took a different position on her authority as an expert, but in principle they contested Des Forges as an expert in post-1990 events, suggesting that if she was an expert at all (and not everyone agreed she was) her expertise was pre-1990 Rwanda history and second, that she was not a human rights expert because she did not have formal training in that field and/or had not carried out her own research on human rights abuses.⁴⁸ The OtP responded strongly to the challenge to Des Forges’ expertise, with the chief prosecutor herself appearing in the trial chamber noting that it was ‘improper that Alison Des Forges is subject to this humiliation’.⁴⁹

While the trial chamber ruled that Des Forges was an expert, the judges themselves asked very few follow-up questions, unlike the judges in Akayesu. One reason for this change might have been that two of the three judges (Judge Lloyd Williams and Judge Pavel Dolenc) who began hearings in Bagasora,⁵⁰ were simultaneously sitting on two other cases (Prosecutor v. Ntagerura et al. and Prosecutor v. Semanza).⁵¹ As one judge explained to me, judges ‘learn as we work along. …when you’ve been in a case or two or three, you recognize what you are hearing’.⁵² It may have been that the judges in Bagasora felt they did not need the same tuition that Des Forges provided the trial chamber in Akayesu.

A second reason why the judges may have held back from asking questions was that Des Forges’ testimony was already subject to multiple interruptions and took much longer than was budgeted for in the trial schedule. As in Akayesu, her evidence was wide ranging, covering topics from early Rwandan history in the nineteenth century to events in the 1990s. But compared to Akayesu, her answers to the prosecutor’s questions were briefer and often cut off by challenges by defence counsel on various procedural grounds, ranging from the admissibility of certain documents to her use of the term ‘genocide’.⁵³

Defence lawyers cross-examined Des Forges on everything, including her methodology and professional history. Some of the questions sought to cast doubt on her reliability, suggesting that, for example, as a young undergraduate student doing volunteer work in refugee camps, she became sympathetic to (and biased in favour of) Rwandan.⁵⁴ In this and other lines of questioning, the defence lawyers drew on Des Forges’ testimony in previous Rwanda Tribunal cases as well as other judicial and quasi-judicial hearings, which they had read closely, exploring any possible inconsistencies in her evidence.

Mr. Degli (for Defendant Kabiligi): You told my colleague, Ogetto, today that this publication Leave None to Tell the Story is your work. Is that correct?

Answer: I said that I was the author of that book which was based upon research done by myself and colleagues.

Mr. Degli: Madam Des Forges, when I take this book, contrary to books that are published by individuals, I realized that it was published by our two organizations, International Federation of Human Rights and Human Rights Watch and when I open the French version to the second page of the book…I do not seem to see your name as author….

Answer: Indeed, Maître, you point to a complicated situation which has caused enormous distress to librarians all over the world… Since it is clear that organizations do not write books, it seems apparent that there is a person as author….

Mr. Degli: Madam Des Forges, is it correct that 16th June 1998 you appeared before the French parliamentary committee that listened to you?

Answer: That is correct.

Mr. Degli: Please look at the document which I have just passed to you, the first page of that document…I read the words that you said and which were noted by the committee. Talking about you, the committee said the following: ‘She stated or she indicated that the data she goes to present were the result of research conducted for 3 years on the field by a research team of FIDH and the Human Rights Watch association, and that it was a common project’…I emphasis the world ‘common’ or ‘collective’…. Is this the statement exactly what you made to the French parliamentary committee?

Answer: No, that statement is not mine.⁵⁵

While it is not unusual for a lawyer to try and discredit a witness, this and other lines of questioning point to a changed, more combative, climate in the courtroom. Most of the combativeness took place between the lawyers for the defendants and the lead prosecutor on the case, Eboe-Osuji, but Des Forges herself was also targeted. On different occasions, for example, the presiding judge intervened to correct questions by defence lawyers that suggested something improper about Des Forges.⁵⁶ And, on another occasion, Des Forges herself intervened to address what she saw as a challenge to her reputation.

The Witness (Des Forges): So the document which I used as a sample was one, which was in no way bound by attorney–client privilege, had been handed over with Colonel Bagosora’s approval to Professor Reyntjens. I want to make that clear in order to remove any suggestion that I might have inappropriately benefited from this document and also in order to remove any obstacle that might exist to the submission of this diary of Colonel Bagosora to the Court so that the Bench might have the opportunity to use to the full this piece of evidence.

(this is followed by a short procedural exchange about the document number)

Mr. Skolnik (counsel for Bagasora): My Lord. I see that my colleague (referring to the Prosecutor Eboe-Osuji) is trying to get the agenda into the evidence after there was ruling yesterday.

The Witness: Excuse me. The Prosecution has nothing to do with this. This is a point I raised because my honour has been called into question.⁵⁷

Substantively, the cross-examination of Des Forges was very detailed focusing primarily on the events in the 1990s. Here too, the defence counsel seems to be challenging the framing of the events offered by Des Forges. It is not clear from the transcripts that the cross-examination was particularly effective. As the defence lawyers themselves noted on several occasions, most were not trained in the common law tradition of cross-examination, and their questions, while wide ranging were also meandering and imprecise.

But the trial judgment does seem to suggest that the defence had an impact on challenging the certainty of the conclusions noted in Akayesu that the genocide was planned and that the various events in the 1990s prior to 6 April 1994 were part of that planning. There is no room in this chapter to discuss in detail the judgment or the defendants’ arguments. But in the following concluding comments, I highlight a few aspects of the Bagasora judgment that mark a shift in the trial chamber’s conception of the genocide and its own remit as a war crimes court.

Throughout its judgment, the Bagasora Trial Chamber notes that the specific events listed as evidence of planning for genocide, such as the establishment of local security forces, the compilation of lists of ‘enemies’ and the ethnicisation of security concerns, could equally be explained as steps taken to protect security during a time of military and political conflict.

After the death of President Habyarimana, these tools were clearly put to use to facilitate killings. When viewed against the backdrop of the targeted killings and massive slaughter perpetrated by civilian and military assailants between April and July 1994 as well as earlier cycles of violence, it is understandable why for many this evidence takes on new meaning and shows a prior conspiracy to commit genocide. Indeed, these preparations are completely consistent with a plan to commit genocide. However, they are also consistent with preparations for a political or military power struggle.⁵⁸

The judgment concludes that after 408 trial days, 242 witnesses and nearly 1,600 exhibits, the evidence of planning was simply inadequate. ‘It is possible that some military or civilian authorities did intend these preparations as part of a plan to commit genocide’, but this has not yet been proved.⁵⁹ While Bagasora himself is found guilty of superior responsibility for actions of Rwandan army soldiers committed in the first few days of the genocide, the tribunal rules there is insufficient evidence of a conspiracy among the defendants to plan and execute a genocide.

The implications of Bagasora are complicated. While this was one of the main leadership trials before the tribunal, there are others that explore the role of, for example, political leaders in orchestrating the genocide. But, Bagasora was one case where the OtP sought to prove that the genocide was planned prior to April 1994. The judgment in Bagasora begins to cast doubt on that conclusion, as does some scholarly study of the genocide (See e.g. Straus 2006, p. 33).

Filip Reyntjens suggests that he and the other scholars working on Rwanda have changed their views somewhat about the events in Rwanda in 1994. Substantively, his analysis is the same, he says, but some of his emphasis has shifted, for example, towards highlighting the actions of the RPF.⁶⁰ In an article on Des Forges’ legacy, he co-authored with David Newbury, Reyntjens also suggests that while Des Forges’ book, Leave None to Tell the Story, on which much of her testimony at the tribunal was based, will ‘remain the essential referential work’ about the 1994 genocide, ‘some of its material will need to be revisited’ (Newbury and Reyntjens 2010).

For both the experts and the tribunal then, knowledge of what happened in Rwanda in the 1990s is changing as more information becomes available. The results of this change for the tribunal are paradoxical. As the tribunal amasses more evidence and more experience, some trial chambers, such as in Bagasora, become less certain about what happened. ‘The process of a criminal trial cannot depict the entire picture of what happened in Rwanda’, it wrote, ‘even in a case of this magnitude’. Thus, writing years after the first decision in Akayesu, the Bagasora Trial Chamber sought to distance itself from authoring a record of what happened and opts instead, to depict its task in much narrower terms: ‘The Chamber’s task is narrowed by exacting standards of proof and procedure as well as its focus on the four Accused and the specific evidence placed before it in this case’.⁶¹

Conclusion

In this chapter, I suggested that the work of the Rwanda Tribunal was shaped from the beginning by a meta-narrative about the genocide as orchestrated by political, military and social elites for the purposes of securing power. The Akayesu judgment, the first issued by the Rwanda Tribunal, confirmed this view, authoring an account of what happened in Rwanda in the 1990s as a series of events and deliberate actions that resulted in a group of elites conspiring to incite ordinary Rwandan Hutus to kill Rwandan Tutsi and politically sympathetic Hutu, allowing the elites to consolidate their hold on power. This account of what happened in Rwanda in the 1990s, and the understanding of large-scale violence that it confirms, was substantially based at the outset on the research of several experts on Rwanda, including Alison Des Forges. Her testimony provided the basis on which the trial chamber was able to, first, depict the violence as inter-ethnic and constituting the international crime of genocide and second that the genocide was planned and then executed by a select group of elites.

I have suggested in my, far too brief, discussion of a later case at the tribunal, Prosecutor v. Bagasora, that aspects of this meta-narrative became less clear to the judges as the tribunal continues its work. My point here is not that the broad outlines of the account of genocide as elite orchestrated are necessarily wrong. Rather, my argument is that the meta-narrative, as I call it, is an epistemological framing that might not appear as such. As an account of violence, the meta-narrative sometimes appears as a common-sense truth. And while the evidence of elite orchestration of the genocide is substantial, the self evident ‘truth’ of the meta-narrative may have the unintended effect of obscuring the more uneven, changing and arguably less intentional aspects of the genocide. This is not to say that individual defendants are not guilty, but it does raise questions about how responsibility and guilt are understood in relation to large-scale, complex, multiple perpetrator violence.

Another argument pursued in this chapter is that international criminal courts acquire and produce knowledge about large-scale violence. What are the implications of an international criminal court that concludes after 10 years of amassing evidence that it cannot ‘depict the entire picture of what happened’, at the same time, that judges in these same institutions are becoming more assertive about their own acquired knowledge base (and their need to rely less on context experts)? The discomfort with producing a record expressed by the Bagasora Trial Chamber is somewhat understandable. Authoring a narrative about the causes and consequences of large-scale conflict is a difficult business and it is invariably enmeshed in political ramifications. The contexts of large-scale violence are extremely complex, with information and insight unfolding gradually, often years after the formal end of the violence. And yet international criminal courts do author an account of ‘what happened’ during times of large-scale violence, even when they depict their remit as narrowly about the guilt or innocence of individual defendants.

In this chapter, I have suggested that the narratives authored by these institutions about large-scale violence are contradictory and changing, and that some of that change reflects how, and through what means, trial chambers come to know the contexts they are tasked with judging. Recognizing international criminal courts as structured by social processes, including knowledge processes, is thus crucial for a better understanding of how, through what means and with which limitations international courts produce narratives about the causes and contexts of large-scale violence.