Sheltering the Architects

In July 1994 the government that had organised the genocide, its army, its militia, its media, its treasury and roughly a million and a half civilians crossed the border into Zaire. They did not scatter. They moved as an organised state, and they were received, housed and fed by the United Nations and by international non-governmental organisations at a cost of something in excess of a billion dollars in the first year.

What happened next set the terms of the next thirty years in the Great Lakes region (including two enormous wars in Congo), and produced the peculiar situation in which the political project of 1994 still exists, still has spokesmen, and now operates in the language of human rights.

The corridor out

Chapter 25 described Opération Turquoise: the French intervention of June–August 1994, authorised by the Security Council, which established a "safe humanitarian zone" in the south-west of Rwanda in the last month of the genocide.

Whatever its stated purpose, its practical effect is not disputed. The zone, roughly a fifth of the country, covered the area still held by the government and its forces. Under its protection, the interim government's ministers, the army's command, the militia leadership and the personnel of RTLM moved across the border into Zaire. Nobody was disarmed. Nobody was arrested. The government that had ordered a genocide walked out with its structures intact.

The pattern had been set in the genocide's first week. On 9 April 1994, three days into the killing, the French aircraft of Opération Amaryllis, sent to evacuate foreign nationals, also carried out Agathe Habyarimana, the president's widow, with members of her family and entourage: the centre of the akazu circle described in chapter 25. She was flown to safety and settled in France, where she has lived ever since. In 2007 the French refugee office refused her asylum on the grounds of her proximity to the machinery of the genocide, a refusal upheld by the Conseil d'État; a Rwandan extradition request was refused in 2011; a French judicial inquiry opened in 2008 has never come to trial. She denies any role. For thirty years she has been held too implicated to be a refugee and yet has never been made a defendant, and that suspended condition, neither cleared nor charged, is the French file on Rwanda in miniature.

The state's property crossed too. The interim government left with what remained of the national bank's currency and with control of the accounts the state held abroad, resources that kept the government-in-exile solvent, salaried and, soon, armed. And then came the population. In the four days around 14 July 1994 roughly 850,000 people crossed at Gisenyi into the Zairean border town of Goma, one of the largest and fastest movements of people ever recorded; within weeks well over a million were encamped in North Kivu, with hundreds of thousands more around Bukavu and Uvira. Most were civilians who had been told by their own administration and by RTLM's broadcasts that the RPF would exterminate them. The fear was manufactured, but it was real, and it delivered to the leadership in exile the one asset a state-in-exile needs: people to govern.

The camps

The camps around Goma, Bukavu and Uvira became the most notorious case study in the modern history of humanitarian aid, and they generated a literature of institutional self-examination that is still assigned in the field.

The camps were governed by the perpetrators. The former administrative hierarchy (prefects, burgomasters, councillors) reconstituted itself in the camps and was, in many cases, the structure through which aid agencies distributed food, because it was the structure that existed. Agencies that had to feed hundreds of thousands of people needed lists, and the men who had the lists were the men who had used lists in April. Ration cards passed through the hands of men who had organised killings.

Aid was taxed and diverted. Food was resold; the militia levied charges on rations and on the salaries of camp residents employed by the agencies; a war economy formed on top of the relief operation. Estimates of the proportion diverted vary, but the mechanism is uncontested.

The camps functioned as a rear base. Military training continued in and around the camps; recruitment among the young men was open; weapons were bought. Human Rights Watch documented, in Rearming with Impunity (1995), deliveries of arms to the defeated government forces through Goma after the Security Council's embargo, across the same airfield the relief flights were using, with the connivance of the Zairean authorities. From 1995 armed groups launched cross-border raids into Rwanda, killing survivors and witnesses, a campaign specifically directed at those who could testify.¹

The cholera epidemic in the first weeks at Goma killed tens of thousands of people, at its worst thousands in a single day, mostly civilians who had been told by the same leadership to flee. The camps stood on volcanic rock too hard to dig graves in; the dead were rolled in straw mats, laid along the roads and collected by truck.

And then the inversion, which deserves a paragraph of its own. The world's cameras, which had largely declined to come for the genocide, came for the cholera. The images from Goma produced an extraordinary mobilisation: airlifts, celebrity appeals, American military engineers flying in water-purification equipment, a public that gave generously to "the Rwandan refugees" with the machinery that had driven them there cropped out of the frame. By most accountings the camps received more international assistance in the genocide's immediate aftermath than the survivors inside Rwanda did, where a new government had inherited a looted treasury, a destroyed judiciary and a population of widows and orphans. Within weeks of the last massacres, the perpetrators' community had become the object of the world's compassion, and the survivors watched the airlift pass overhead. No single fact from 1994 did more lasting damage to Rwanda's opinion of international moral judgement.

The arrangement also suited its host. Mobutu Sese Seko, an ailing dictator whose Western patrons had lost interest in him when the Cold War ended, discovered that a million refugees and an exiled army made him indispensable again: the camps had to be negotiated over, and every negotiation ran through him. He protected the former army's leadership, tolerated its rearmament, and used the crisis to re-enter the diplomacy of the region. The camps were, among everything else, the instrument of his last relevance, which is one reason they stood for two years.

Some agencies drew the conclusion and left. Médecins Sans Frontières–France withdrew from the camps in December 1994 on the grounds that it was sustaining an armed force that had committed genocide; the International Rescue Committee and others followed. Others stayed, arguing that abandoning several hundred thousand civilians was not a humanitarian option either. Both positions were serious. The episode produced the "do no harm" literature and Fiona Terry's Condemned to Repeat?, which remains the standard analysis.²

Justice, slowly and elsewhere

The International Criminal Tribunal for Rwanda, established by the Security Council in November 1994 and seated in Arusha, Tanzania, indicted 93 people over twenty years and convicted 62. It produced real jurisprudence: the Akayesu judgement of 1998 was the first conviction for genocide by an international tribunal and the first to recognise rape as an act of genocide; the "Media Trial" convicted the organisers of RTLM and Kangura for incitement; Théoneste Bagosora, the defence ministry official at the centre of the machinery in the crucial days, was convicted in 2008 and died in a Malian prison in 2021. The tribunal also showed what international justice costs. It spent roughly two billion dollars. The Butare case, six accused including Pauline Nyiramasuhuko, the former minister for family and women's affairs and the first woman convicted of genocide by an international court, ran fourteen years from opening to final appeal. The tribunal sat outside Rwanda, could not impose the death penalty while ordinary Rwandan courts could, and housed its convicted in conditions markedly better than those of the country where the crimes occurred, a discrepancy that Rwandans noticed.⁴

Meanwhile Rwanda itself had over a hundred thousand people in detention and a destroyed judiciary. Its answer was gacaca: a modified version of a customary community dispute process, scaled to hear roughly two million cases between 2002 and 2012 before locally elected lay judges. It was criticised, with justification, for weak due process, coerced testimony, and vulnerability to score-settling. It also processed a caseload that no conventional court system on earth could have handled, in the country where the crimes happened, in the language the witnesses spoke.⁵

Where the fugitives went

The pattern is the part that belongs in this chronicle.

• Father Athanase Seromba, who had his church bulldozed onto the people inside it, was serving as a priest in Italy under an assumed name until located by a journalist in 2001. He surrendered to the tribunal the following year and was sentenced on appeal to life imprisonment. The question of who in the church knew his identity, and arranged his parish, has never been formally answered.
• Father Wenceslas Munyeshyaka, accused in connection with events at the Sainte-Famille church in Kigali, reached France in 1994 and was given a parish in the diocese of Évreux. A complaint was filed against him in 1995. In 2004 the European Court of Human Rights condemned France for the length of the proceedings, which had by then produced nothing in nine years. They continued for another decade after the condemnation, and ended in dismissal in 2015. He was convicted in absentia in Rwanda; in France he remained in ministry.
• Félicien Kabuga, the businessman accused of financing the Interahamwe and RTLM, lived under aliases for twenty-six years, latterly in a flat in Asnières-sur-Seine outside Paris, where he was arrested in May 2020. In 2023 the residual Mechanism's judges found him unfit to stand trial by reason of dementia. He will now be neither convicted nor acquitted.
• Two other names at the top of the wanted list were resolved in the same period, in the same way: Augustin Bizimana, the interim government's defence minister, was confirmed in 2020 to have died years earlier; Protais Mpiranya, commander of the Presidential Guard, was confirmed in 2022 to have died in Zimbabwe in 2006 and to have been buried in Harare under a false name. Both died free.
• Belgium, alone among the implicated states, started early. In 2001 a Brussels assize jury tried the "Butare Four" (among them the two Sovu nuns of chapter 25) in the first jury trial of the genocide's crimes under universal jurisdiction, and Belgian juries have returned to the subject repeatedly since.
• France did not hold its first genocide trial until 2014, twenty years after the fact: Pascal Simbikangwa, convicted and sentenced to twenty-five years. Others followed: Ngenzi and Barahira in 2016, Bucyibaruta in 2022, and in 2023 Sosthène Munyemana, who had practised as a hospital doctor in the south-west of France for nearly three decades while the complaint against him, filed in 1995, worked its way through the system.

Two decades of delay in the country most implicated in supporting the regime is not an accident of court scheduling. Suspects lived openly, published, gave interviews, and organised. The reckoning, when it finally came, came in the wrong order: the state's own historians (the Duclert commission of chapter 25) delivered their verdict on France's responsibilities in 2021, seven years after the first French courtroom heard a Rwandan case, and twenty-seven years after the men in question arrived. Most of the delay was not investigative difficulty. It was the absence of any institution that wanted to know.⁶

The project continues

The most consequential legacy is not the individual fugitive. It is that the political project of 1994 survived, adapted, and found a new vocabulary.

Militarily, the organisation never dissolved; it renamed itself. When the camps were dismantled in 1996 the ex-FAR and Interahamwe remnants regrouped as ALiR, which in 1997 and 1998 fought an insurgency in Rwanda's north-west that specifically targeted survivors and witnesses, and then, from 2000, as the FDLR, based in eastern Congo, which has operated there for three decades and has been sanctioned by the UN Security Council. For years it was run from Europe: its president, Ignace Murwanashyaka, directed the organisation from Mannheim by phone and email until German prosecutors tried him in Stuttgart, where he was convicted in 2015 of leading a terrorist organisation; he died in custody in 2019. The FDLR's existence (a genocidal force encamped on Rwanda's border, never disarmed by anyone) is the fact around which the entire subsequent Great Lakes conflict turns, and it is the argument Rwanda invokes for every intervention it has made in Congo since. Whether that argument justifies what has been done in its name is chapter 28's problem; that the force is real is not in doubt.

Politically, a diaspora network in Belgium, France, the Netherlands, Canada and the United States sustains the ideology in a form adapted for European audiences. Its core is the "double genocide" thesis: the claim that what occurred in 1994 was a reciprocal civil war rather than a genocide, or that a second, equal genocide was committed against Hutu. The thesis takes real events (the RPF killings described in chapter 26, the Congo campaigns described above), and inflates them into an equivalence that the evidentiary record does not support.⁷

Rhetorically, and this is the sophisticated part, the project now speaks the language of human rights, press freedom, democratic accountability and rule of law. Organisations with human-rights names campaign in European capitals; op-eds appear; parliamentary questions are tabled. Some of the criticisms carried by these channels are true (chapter 28 will grant several of them), and that is exactly what makes the technique effective. A denial movement that only denied would be ignored. One that campaigns on genuine grievances while quietly relitigating 1994 gets an audience.

This is why the ordinary reader has such difficulty. Two things are simultaneously the case: Rwanda's government does things that deserve criticism, and there exists a well-organised movement whose interest in that criticism is that it launders a genocidal politics. Distinguishing them requires knowing who is speaking and what they said in 1994, which is the entire purpose of a chronicle like this one.

The pattern this chronicle has seen before

Look back over the preceding chapters and this is not a Rwandan anomaly. It is the standard ending.

The men who organised the Congo Free State's rubber terror were not tried. Nobody was punished for the Herero and Nama extermination; Germany recognised it as genocide in 2021, one hundred and seventeen years later. The German industrialists of chapter 15 were released early and became rich; the scientists were recruited. The architects of the Guatemalan and Indonesian killings of chapter 16 died in their beds. Belgium's role in Rwanda in 1933 and 1959 has never been the subject of a prosecution and never will be.

The rule is consistent enough to state as a finding: the international system prosecutes the losers of wars and the leaders of poor countries. Rwanda's génocidaires were prosecuted, eventually, in significant numbers, because they lost and because their country was too weak to protect them. The people who supplied the theory, the cards, the arms and the diplomatic cover have not been prosecuted at all.

Chapter 28 asks what happens when the country they wrecked declines to stay wrecked.

Footnotes

1. Human Rights Watch Arms Project, Rwanda/Zaire: Rearming with Impunity: International Support for the Perpetrators of the Rwandan Genocide (May 1995); Gérard Prunier, The Rwanda Crisis: History of a Genocide (New York: Columbia University Press, 1995), on the exodus and the government-in-exile. ↩
2. Fiona Terry, Condemned to Repeat? The Paradox of Humanitarian Action (Ithaca: Cornell University Press, 2002); Mary B. Anderson, Do No Harm: How Aid Can Support Peace — or War (Boulder: Lynne Rienner, 1999); Gérard Prunier, Africa's World War (Oxford: Oxford University Press, 2009), chs. 1–2. ↩
3. Office of the UN High Commissioner for Human Rights, Report of the Mapping Exercise documenting the most serious violations of human rights and international humanitarian law committed within the territory of the Democratic Republic of the Congo between March 1993 and June 2003 (August 2010), and the Government of Rwanda's official response of September 2010. ↩
4. Prosecutor v. Akayesu, ICTR-96-4-T (2 September 1998); Prosecutor v. Nahimana et al. ("the Media case"), ICTR-99-52; Prosecutor v. Bagosora et al., ICTR-98-41; Prosecutor v. Nyiramasuhuko et al. ("the Butare case"), ICTR-98-42. The tribunal's caseload statistics are published by the UN International Residual Mechanism for Criminal Tribunals, which inherited its functions. ↩
5. Phil Clark, The Gacaca Courts, Post-Genocide Justice and Reconciliation in Rwanda (Cambridge: Cambridge University Press, 2010), for the sympathetic reading; Human Rights Watch, Justice Compromised: The Legacy of Rwanda's Community-Based Gacaca Courts (2011), for the critical one. ↩
6. On Kabuga's unfitness, the decisions of the Residual Mechanism's chambers in 2023; on France's delays, European Court of Human Rights, Mutimura v. France (2004); the chronology of the French cases is maintained by the Collectif des Parties Civiles pour le Rwanda, the survivors' association that brought most of the original complaints. ↩
7. Jean-Pierre Chrétien and Marcel Kabanda, Rwanda: racisme et génocide. L'idéologie hamitique (Paris: Belin, 2013); Linda Melvern, Intent to Deceive: Denying the Genocide of the Tutsi (London: Verso, 2020). ↩