<<previous  |  index  |  next>>

V. Justice for the Genocide

The challenge of delivering justice for the genocide in Rwanda has been huge. It started with the basic realities of life in a society in which both lives and institutions were shattered by genocide. It involved coping with criminal acts on an immense scale. It required establishing the legal basis of prosecution—and establishing institutions and systems that could bring justice.

Practical Problems: Turning on the Lights

Like other parts of the Rwandan government, the judicial system suffered serious losses in staff, facilities, and equipment during the war. Of some 600 judges in service before April 1994, for example, only 237 were available to resume work in August 1994 and only 53 of these sat in courts with jurisdiction over serious crimes. Similar losses had thinned the ranks of prosecutors, judicial officers, police officers, clerks, and lawyers. The ministry of justice recruited hundreds of new employees but was able to provide them with only minimal training before putting them to work.¹

Other practical problems were similarly daunting. War-time damage to the judicial ministry building was so serious that the new minister of justice worked from his hotel room, filing documents in boxes under his bed. Other court buildings had been stripped of furniture and electrical fixtures. At the national prosecutor’s office, judicial officers had trouble finding paper and pens to record the interrogations that they wrote out by hand.²

International donors, both multilateral and bilateral, provided substantial assistance in rebuilding the staff and infrastructure of the system, but it took time for their support to take effect.

Identifying and Prosecuting Perpetrators: the Issue of Scale

In August 1994 the new government had to decide how many of the hundreds of thousands of participants in the genocide to prosecute, how to evaluate their levels of guilt, and how severely to punish them. The minister of justice and other officials began with the premise that officials and political leaders of the previous government had deliberately misled ordinary people into seeing Tutsi civilians as enemies, the equivalent of combatants to be attacked and killed. They were clear that such leaders should be prosecuted, as should those who had killed most often and most brutally, a number that the minister estimated would amount to some ten thousand persons. ³ He was firm on the need for such persons to be tried in conventional courts and rejected any use of the gacaca process. Himself a Hutu, he said that gacaca proceedings would “trivialize the genocide” and diminish the credibility of convictions. He feared that failure to clearly establish the guilt of some Hutu would lead to the generalization of guilt to all Hutu.⁴

The minister held that the massive number of others who had participated by manning barriers, doing patrols, and damaging or stealing property should not be imprisoned but should instead be educated to see the harm they had done and should be obliged to make restitution for all property damaged or stolen.⁵

Despite the intention of prosecuting a narrowly defined group of suspects, however, the government permitted soldiers and others without legal authority to arrest persons accused of genocide, sometimes on the basis of a single unverified accusation. Numbers of detainees mushroomed. By October 1994, an estimated 58,000 persons were detained in space meant for 12,000,⁶ and by 1998, the number jailed had grown to 135,000.⁷ Overcrowding and inadequate sanitation, food, and medical care created conditions that were universally acknowledged to be inhumane. Many persons were held for years with no investigations done and no charges specified, a situation that violated Rwandan law until legislators passed several exceptional measures permitting such detentions.⁸

The Legal Basis for Prosecution

The 1996 Law and Categorization of Crimes

Rwanda had signed and ratified the Convention on the Prevention and Punishment of the Crime of Genocide, the Geneva Conventions and other conventions of international humanitarian law in 1975, but as of 1994 it had no provision for prosecuting and punishing these crimes in its domestic penal code.⁹ Although justice officials recognized the need to immediately create a legal basis for prosecuting genocide, the necessary law—Organic law 8/96—was not adopted until two years after the new government had taken power.¹⁰ 

To avoid offending the principle of non-retroactivity (the prohibition on punishing persons for crimes that were not defined as crimes when they were committed),¹¹ the lawmakers grounded the authority to punish these offences in the prior Rwandan ratification of the relevant international conventions. The law punished crimes specified in the Rwandan penal code, such as murder, which were also listed as crimes in the international convention against genocide.¹² In addition, it punished as constituent acts of genocide violations of the Rwandan penal code that were not listed in the international convention but that were committed “in relation with events” surrounding the genocide and crimes against humanity.¹³

With this provision, the law neglected an essential part of the definition of genocide according to the international convention, that is, the intent of the actor to eliminate all or part of a listed group. Thus persons convicted of crimes like theft committed between April and June 1994, could be—and were—convicted of genocide with no consideration of whether they were merely seeking to profit opportunistically from the situation or whether they actually sought to eliminate persons of the Tutsi ethnic group.¹⁴

Reflecting early thinking on differences in levels of responsibility among perpetrators, the 1996 law divided accused persons into four categories on the basis of the gravity of the crimes committed: category one included leaders, organizers, and the most notorious killers, category two included killers and rapists¹⁵, category three included those who killed or inflicted bodily harm without the intention to kill, and category four included those who stole or damaged property. Penalties ranged from death or life imprisonment for persons convicted of category one crimes to restitution of property but no imprisonment for persons convicted of category four crimes. The 1996 law specified that the national prosecutor was to draw up and publish a list of all persons accused of category one crimes. Consistent with the notion that many Rwandans had been misled and would come to recognize the wrongness of their actions, the law introduced a form of “plea-bargaining,” permitting sentences to be reduced for those who confessed their crimes and gave full information on their accomplices.¹⁶

In 2001 a new law shifted most genocide prosecutions to gacaca jurisdictions (see below), but retained the division of accused into categories according to the gravity of the crime allegedly committed.¹⁷ The 2001 law, and its 2004 amended version, generally followed the definition of genocide found in the 1996 law, but required that violations of the penal code be committed with genocidal intent—rather than merely in relation with the events of the genocide—in order to be qualified as genocide.¹⁸

Prosecution in Conventional Courts

Judicial officials began genocide trials in conventional courts in December 1996 but two years later were already preparing to embark on another strategy to dispose of the growing backlog of cases. By 1998 only 1,292 persons had been judged and relatively few accused persons had confessed, disappointing hopes that plea-bargains would reduce the enormous number of persons to be tried. If the same rate of prosecutions were to continue, it appeared sure to take decades to prosecute the estimated 135,000 detainees.¹⁹ Rather than invest additional resources in speeding the delivery of justice in conventional courts, the government turned to another plan, the gacaca jurisdictions.

The new direction was announced in January 1998 by Vice-President Paul Kagame, already the dominant political figure in Rwanda. After remarking that Rwanda could not afford the U.S. $ 20 million a year necessary to support the huge population then in prison, he proposed that the most guilty be executed, and that others be dealt with through customary judicial mechanisms, with those convicted being sentenced to terms of enforced labor on public works.²⁰ The first of the measures was carried out in part three months later when 22 persons convicted of genocide were executed, the first and only formal executions carried out as a consequence of the genocide.²¹ The second proposal resulted three years later in the establishment of gacaca jurisdictions to try all but the most serious cases of genocide.

By mid-2002, conventional courts had judged 7,181 persons accused of genocide. After 2002, the rate of prosecutions slowed as prosecutors shifted their efforts to preparing cases for transfer to gacaca jurisdictions. Then the courts halted work for months as they took account of organizational changes and other aspects of the extensive judicial reforms of 2004.

From the resumption of court activity in 2005 to March 2008, conventional courts tried only 222 genocide cases. ²² In 2008 judicial officials proposed and the legislature adopted a law sending virtually all remaining cases of genocide to gacaca, with the exception of accused persons who served as national or provincial leaders during the genocide and those sent back to Rwanda for trial from other national or international jurisdictions. As a result of the new law, the prosecution of genocide cases by conventional courts will soon be effectively ended after what amounts to less than a decade of full judicial activity.

Gacaca: Popular or Political Justice?

When gacaca jurisdictions were established in 2001, they were meant to judge all but the most serious crimes of genocide (those of category one), which were left to the conventional courts. It was hoped that the gacaca process would speed the resolution of the huge backlog of cases, reduce the prison population, and contribute to reconciliation.

Gacaca jurisdictions brought together modified elements of customary practices for resolving conflicts and aspects of a conventional state-run punitive justice system.²³ In essence it involved popularly elected judges deciding cases related to the genocide before a gathering of the local population. The judges, who were to guide the hearings and then finally to deliver the verdict of the community, were chosen on the basis of their integrity rather than their formal learning. Some did not read or write, although all received several days of training on the relevant laws and procedures. The transparency of the process and participation of the entire community was supposed to assure the legitimacy of the proceedings and to protect the rights of all participants, making unnecessary the kinds of fair trial guarantees provided by Rwandan law and international conventions. The accused had no access to counsel in gacaca jurisdictions, for example, although that right is guaranteed by the Rwandan constitution and by the International Convention on Civil and Political Rights to which Rwanda is party.²⁴

After years of preparation—election of judges, explanatory meetings for citizens, collecting information and accusations from the local community—actual trials began on a pilot basis in about 10 percent of the country in 2005. Even before trials had begun throughout the rest of Rwanda in July 2006, judicial authorities had announced a projected end to the process in 2007, a deadline later extended into late 2008.²⁵

From its inception through its relatively brief period of operation, the gacaca jurisdictions were altered four times by law (2004, 2006, 2007 and 2008), ²⁶ as well as by administrative order or informal but official directions on several other occasions. Some of the changes, such as in the size and number of jurisdictions, seem of relatively little importance, but others altered the fundamental nature of the process, reducing the independence of the jurisdictions and diminishing the transparency and participation that were supposed to protect the rights of the participants.

Initially welcomed by victims and accused persons alike, gacaca jurisdictions have met popular expectations in some communities but failed to satisfy parties—whether victim or accused or both—in many others. A general assessment of the jurisdictions lies outside the parameters of this report. Here we deal only with official and other interference with gacaca and with other aspects of the gacaca jurisdictions that impinge on the operation of conventional courts.

“Justice is a Political Problem”

Even before the gacaca jurisdictions were actually trying accused persons, senior officials in the ministry of justice anticipated that the process would necessarily be political. In a November 2003, one such official told Human Rights Watch researchers repeatedly that “justice is a political problem that needs to be solved politically.” The minister of justice, present in the interview, did not question this assertion.²⁷

This recognition of the political nature of gacaca justice notwithstanding, the gacaca law of 2004 seemed to provide some measure of judicial independence for the jurisdictions by excluding political leaders, administrative officials, magistrates, and police officers and soldiers from serving as judges.²⁸ Administrative officials—often members of the dominant RPF party—were to provide logistical assistance, encourage the participation of the population and “monitor” (in ways unspecified) the jurisdictions.²⁹ They were not authorized to play any other role in the judicial process.

Limits on Judicial Independence in Gacaca

In November 2004, however, as authorities were preparing for accusations to be collected nationwide the National Service of Gacaca Jurisdictions (SNJG) granted local officials disproportionate power in the process of making accusations and assigning accused people to categories. Later in the process, in 2007, SNJG agents were permitted to wield extensive power over the categorization of accused persons. These changes in procedure opened the way for officials—and others working through them—to influence the judicial process, sometimes to the benefit of the accused, but probably more often to the benefit of the accusers.

Initially members of the community were to accuse alleged perpetrators in public gatherings, the local gacaca assembly, but in 2004 the SNJG mandated local administrative officials, and particularly the nyumbacumi or person in charge of ten households, to gather information from small groups or even by going door to door in the community. The nyumbacumi later presented the accusations to the assembly, but this was to permit local residents to check that the information had been correctly recorded rather than to test its truth. ³⁰ 

The increased role of the nyumbacumi lessened the importance of the popular assembly and of judges originally charged with responsibility for gathering information about accusations. By acquiring a special role behind the scenes, the nyumbacumi—and others who worked through him—had disproportionate power to influence the nature and amount of information that would form the basis of the judicial files of accused persons.

The relatively private way in which accusations were gathered under the changed policy clearly deprived the accused person of the guarantee that was supposed to be provided by the openness of the process. He or she had no opportunity to ensure that any information in his or her defense was recorded—in fact, the forms that the gacaca service provided administrative officials for recording information had space only for accusations, none for exculpatory information. ³¹  Nor had the accused any opportunity to contest charges when they were read out to the assembly.

With the establishment of the gacaca jurisdictions, gacaca judges at the cell level assumed responsibility for assigning accused persons to categories, a duty previously attributed to the national prosecutor and his office. The judges based their determination on information gathered during the accusation process and that the accused had not had the opportunity to contest. The placement in categories effectively determined the seriousness of punishment in case of conviction.³²

In addition, when persons placed in category one were brought to trial in conventional courts, judges in these courts sometimes relied on the untested information gathered during the gacaca accusation process in deciding the case. In one appeals trial in the High Court in 2006, a three judge panel used information from a gacaca jurisdiction to justify its confirmation of the conviction of the accused without independently assessing the credibility of the information. In another such case, the judges accepted only part of the information provided by the gacaca jurisdiction but again did not attempt to verify the part accepted.³³

Reclassification by Administrative Decision

As the accusation phase of gacaca ended, some 818,000 persons had been accused, 77,000 of them placed in category one and so designated for trial in conventional court. These numbers, particularly those in category one, far exceeded those originally foreseen by Rwandan officials. Recognizing that so many conventional trials would take decades—one of the eventualities that the government intended to avoid by creating gacaca—lawmakers redefined the categories in a March 2007 law and provided for some persons to be moved from category one to category two where they would be tried by gacaca jurisdictions.³⁴

During 2007 the National Service of Gacaca Jurisdictions sent its agents throughout the country to meet with local administrative officials and gacaca staff to select the persons who would benefit from this reduction in the assessed gravity of their crimes. The goal of the reclassification of the accused was to have no more than 10,000 and perhaps as few as 2,000 left in category one.³⁵

The qualifications of agents charged with the reclassification as well as the rules under which they operated were not publicly announced. Like the use of the nyumbacumi to prepare accusations, this process ran counter to the basic premise of gacaca—that is, open discussion with full community participation. Only those persons selected by the reclassification teams benefited from the change in the assessment of the gravity of their supposed crimes. Those not chosen were deprived of this benefit on what may have been an arbitrary basis, and without having had any opportunity to speak on their own behalf.