Rwandan Genocide

An exile army invaded from Uganda after thirty years in which Rwanda had refused the return of its refugees. France backed the government and Uganda the insurgents, so a domestic dispute acquired external patrons and the capacity for a long war.

The accords divided power and integrated the armies on terms the ruling elite understood as its own elimination, and that elite then organised to prevent implementation. Settlements that require a faction to accept its own destruction generate spoilers by design.

The Council authorised a small force with a mandate to assist implementation and rules of engagement that prohibited it from using force except in self-defence. This was peacekeeping designed for a peace, deployed into a country where organised preparations for mass killing were under way.

Whoever fired the missiles, the killings began within an hour with lists, roadblocks and prepared militia. Preparation on that scale is not a reaction, and the assassination functioned as the trigger for a plan that already existed.

The killing was organised through the state's own administrative structure, from prefects down to local officials, using identity cards issued for other purposes. Genocide of this speed requires a bureaucracy, and Rwanda's was unusually dense and effective.

The Council cut the force from two and a half thousand to a few hundred while the killing was at its height, because contributing states would not risk casualties. This is the clearest available measure of what collective security amounted to when no interest was at stake.

Five and a half thousand troops were authorised in May and arrived largely after the killing had ended in July, because no state provided them promptly. Authorisation without forces is a resolution, and the gap between the vote and the deployment was fatal.

The genocide ended because one side won the civil war, not because anyone intervened. That is the fact against which every account of the international response has to be measured: the killing was stopped by an army with its own objectives.

The refugee population in Uganda had been a source of instability for a generation, providing recruits, grievance and a base. Unresolved displacement is a structural cause of war in the region, and it was left unaddressed until it produced an army.

The agreement was reached under pressure from donors and neighbours rather than because either party had exhausted itself. External pressure can produce a signature and cannot produce the willingness to implement, and the distinction was fatal here.

Contributing states had just lost soldiers in Somalia and were unwilling to risk more in Africa, so the mission was sized to what was politically available rather than to the task. Supply of peacekeepers, not demand, determined the deployment.

The Burundian president died in the same aircraft, and the region already contained a Burundian civil war, Ugandan involvement and Zairean instability. The genocide's aftermath drew all of them into the wars that followed, so a regional system was destabilised for a decade.

No state had an interest sufficient to intervene, and the one with the closest relationship had backed the government being overthrown. In the absence of a great power with a stake, the capacity of the international system to act was effectively zero.

The United States had lost soldiers in Mogadishu six months earlier and had issued a directive restricting participation in peacekeeping. One state's domestic reaction to an unrelated failure determined the Council's response to a genocide, which is how the system actually works.

The United Nations owns no troops and no equipment and must borrow both, and the states with airlift declined to provide it or charged for armoured vehicles. Dependence on member states' capabilities is the organisation's permanent structural weakness.

Two million people fled into Zaire including the defeated forces and many perpetrators, who reorganised in the camps. The displacement exported the conflict and led directly to two Congo wars that killed millions over the following decade.

Rwanda was a one-party state under external pressure to democratise, and the invasion allowed the regime to define its domestic opponents as accomplices of the enemy. Liberalisation under war conditions gave extremists a platform and a target simultaneously.

A broad-based transitional government, multiparty competition and refugee return: the accords were a model liberal settlement on paper. They allocated power among parties that had no intention of sharing it, and the design contained no answer to that.

The mission's commander cabled headquarters in January with detailed information from an informant about weapons caches and killing lists, and was told to inform the government the plotters belonged to. Reporting lines designed for consent-based missions routed the warning to the wrong place.

The first targets were moderate Hutu politicians, including the prime minister, who were killed before the mass killing began. Eliminating the domestic opposition first is the standard sequence, because it removes anyone who might have organised resistance from within the state.

Neighbours killed neighbours, using local knowledge to find people, which required a mobilisation of ordinary participation that officials organised through existing structures. The social mechanism, obligation, fear, and reward, has been documented in enormous detail.

Belgium withdrew after its soldiers were killed and lobbied for full withdrawal so as not to be alone in leaving. Democratic governments answering to grieving publics behaved exactly as expected, and the aggregate of those domestic decisions was the abandonment of the mission.

Public and press attention arrived only as the refugee crisis unfolded in Zaire, after the genocide, and then produced a large humanitarian response. Attention came for the aftermath and not the crime, and the resources mobilised for camps exceeded anything offered earlier.

The new government inherited a destroyed state with most of its professionals dead or in flight and later pursued accountability through a hundred thousand prosecutions and community courts. Rebuilding institutions after this was an undertaking without precedent.

Regional and French mediation produced a succession of ceasefires and eventually a negotiation, none of which addressed the question of who would actually hold power in Kigali. Processes that manage violence without resolving the underlying distribution postpone the contest rather than settling it, and the postponement here lasted three years.

Implementation depended on a sequence of steps each of which could be blocked, and blocking was cost-free. Agreements need mechanisms that penalise obstruction, and Arusha's had none beyond the presence of a peacekeeping force yet to be deployed.

The restriction on raiding arms caches followed from a mandate written for a consensual operation. Institutional rules encode assumptions about the situation, and when the situation differs the rules prevent adaptation rather than guiding it.

Ten Belgian peacekeepers guarding the prime minister were disarmed and murdered, an act calculated to force the withdrawal of the contingent. The vulnerability of a consent-based force was identified and exploited deliberately, and it worked within a fortnight.

Radio Mille Collines coordinated the killing, naming individuals and directing militia to locations, and jamming it was considered and rejected as expensive and legally doubtful. A cheap intervention against the coordinating mechanism was available and was declined.

The residual force under Dallaire protected several thousand people at a few sites with no mandate and almost no resources. What a fraction of a force achieved without authority indicates what the full force might have done with it.

The French intervention in June under Council authority created a protected zone and also allowed many perpetrators to escape into Zaire. An intervention by a party with prior ties to one side produces results that no impartial force would have produced.

The Rwanda tribunal, established in November, produced the first conviction for genocide by an international court and the first finding that rape constituted genocide. The jurisprudence developed there is now the foundation of international criminal law in this area.

An invasion across an international border by forces based in a neighbouring state engaged the clearest rules available, and nothing followed. Where no great power's interest is engaged, the society's rules operate as description rather than constraint.

Regional states, the OAU and the United Nations all participated in the mediation, which gave the accords the endorsement of international society. That endorsement did nothing to protect the agreement, and the same institutions were absent when it collapsed.

International society deployed a force whose presence implied protection while withholding the authority to protect. The appearance of commitment without its substance is a recurring pattern, and it induces reliance that makes the eventual failure worse.

Responsibility for shooting down the aircraft has never been established, and separate inquiries in France and in Rwanda reached opposing conclusions about who fired. Where no impartial mechanism exists to determine such a question, a foundational fact remains permanently contested, and both sides continue to find the ambiguity politically useful.

Officials avoided the word genocide for weeks because using it would have engaged obligations under the 1948 Convention. That the terminology was managed to avoid a legal duty is documented in the record, and it is the clearest example of a norm being evaded rather than breached.

The society of states reduced its presence during a genocide, which its own convention obliged it to prevent. The contradiction between the obligation and the conduct was visible at the time and has been formally acknowledged since by every institution involved.

Every institution involved has since issued a report acknowledging its own failure, including an independent inquiry commissioned by the Secretary-General and reports from Belgium, France and the OAU. Institutional self-examination on that scale is genuinely rare, and it produced the doctrinal changes of the decade that followed.

The scale of the failure forced international society to revisit sovereignty and intervention, producing the responsibility to protect within seven years. Norms change after catastrophes, and this catastrophe changed the doctrine more than any other since 1945.

The colonial classification of Hutu and Tutsi, hardened into identity cards by Belgian administration, supplied the categories the war was fought in. Those categories were an administrative invention that had become a lived reality, which is what made them lethal.

Extremist radio began describing the accords as a Tutsi takeover months before signature, preparing a population to reject them. Building the interpretation in advance is how a peace agreement is defeated, and this campaign was documented as it happened.

Rwanda was understood in New York as a peacekeeping success in preparation, a country moving from war toward a settlement. That framing made warnings of a very different scenario difficult to hear, because they did not fit the file the mission belonged to.

Radio broadcasts within the hour named the RPF as responsible and called on the population to defend itself, converting an assassination into a licence. The interpretation was supplied faster than any fact could be, which is how the mobilisation was achieved.

Years of propaganda had established Tutsi as inyenzi, cockroaches, and as an alien population that had oppressed the majority. Dehumanising language prepared over years is what made mass participation possible, and it was a deliberate campaign with identifiable authors.

Officials in several capitals used tribal violence and chaos to describe an organised campaign, a framing that made intervention look futile. How the killing was categorised determined what response seemed available, and the categorisation was in part deliberate.

Rwanda became the standing reference for the cost of inaction, invoked in every subsequent argument for intervention. The lesson drawn was that failing to act is itself a decision, and it produced the responsibility to protect doctrine within a decade.

Around eight hundred thousand people were killed in a hundred days, largely with hand tools, by neighbours. That this was possible in a country that had been at peace within living memory is the fact that has most disturbed the understanding of what makes such events possible.