The most incisive analyses of the Arusha peace process and of its failures come from a key US policymaker of that time, Assistant Secretary Herman Cohen, and an academic and international consultant on peacekeeping, Bruce Jones. In a single chapter on Rwanda in his Intervening in Africa: Superpower Peacemaking in a Troubled Continent, Cohen, with pointed realism, faults US policy for its failure to understand the depth of hostility in Rwanda (an arena of low US interest), for being obsessed with the negotiating process, and for not realizing that “as the Arusha process unfolded . . . it inadvertently guaranteed the genocide.”8
During the late 1990s, Jones, who is now a senior fellow at Brookings, wrote his doctoral dissertation on Rwanda. He published his Peacemaking in Rwanda: The Dynamics of Failure in 2001 on the basis of his doctoral work. His chapter on the Arusha negotiations analyzes the negotiations through the optic of their peacemaking objectives. Although his narrative contains errors on chronology, procedural details, and negotiating dynamics, Jones tells the larger story with insight into why the Arusha negotiations looked so good and failed so miserably.
In his conclusion, Jones highlights what went wrong with third-party intervention in the Rwanda crisis: undue pressure to democratize; drawbacks in the facilitation of negotiations; lack of community-level buy-in with the peace process; and failure to plan for violent opposition or to account for spoilers and losers.9 I would agree with Cohen and Jones in their critiques of the Arusha process. But the social cataclysm that was the Rwandan genocide needs a more systematic explication of the international search for peace in Rwanda. That narrative begins with a look at what brought us to Arusha.
Antecedents
Like all Rwandan stories, this is a long one. Rwandan dynastic chronologies would push us back into the mystic past of the founding kings, Kigwa or Gihanga, depending on the legend. Some historians would credit stories about Ruganzu Bwimba, reputed to have built a kingdom in north-central Rwanda in the 1400s, only to have it fall apart in civil war five generations later. Others assert that it was rather Ruganzu Ndori who in the 1600s established the Nyiginya dynasty over the first Rwandan Kingdom of which we have reliable oral record.10
TRADITIONAL SOCIETY
Out of this ancient genesis, Rwandan social and political organization coalesced around descent groups. In Rwanda, heads of lineages formed alliances to build clans. “Being alliances rather than descent groups, clans were mutable . . . [and] the number of lineages composing a clan constantly varied over time according to the political adventures of the great families.”11
As with many other emerging polities of the Great Lakes area, Rwandan leaders knit ties with clans having reputed ritual power and built clientage through distribution of land and cattle. By the beginning of the 1700s, the Nyiginya Kingdom was a fragile coalition of lineages in south-central Rwanda. By the end of the century, “it had transformed into a unified, centralized and aggressive entity.”12 Three factors brought about this transformation: the centralization of power at the court; the extension of clientage to bring all land and cattle under notional royal authority; and continuous expansion under armies by which sons of the ruling elite proved their valor and into which all Rwandan citizens were incorporated.13
During the 1800s, the reach of the state (king and court) deepened. Cyrimina Rujugira (1770), himself a usurper, structured succession to a cycle of four regnal names, expanded court ritual, and increased the number of armies. Around the 1840s, as population and demands on land and pasture grew, the court reserved pasture domains for royal use and disposition. It then appointed chiefs of “tall grass” to oversee pastures and herds, as well as chiefs of land to manage farmers and their obligations. Add to these the commanders of now-ubiquitous armies, and a tripartite system of bureaucratic control emerges, especially in the districts close to the court.
This system did provide a sort of check against arbitrary power at the local level where a client, whether herder or farmer, could seek the protection of one lord against the predations of another. Nonetheless, under a clientage system now triply bureaucratized, exactions necessarily increased, impoverishing farmers and herders. Two “social categories” came to define subjects of the king: those who had lineage links to power and privilege—the Tutsi; and those who by dint of circumstance fell into servile status—the Hutu.14
That was the Rwanda that the Europeans “discovered” in the late 1800s: a country ruled by a king newly installed by the 1896 Rucunshu coup and hemmed in by courtiers, ritual, and intrigue; a territory densely populated at the center with farmers competing with herders for land; and a militarized society in which each person belonged to an army and was a client of a superior both for land and cattle. An overlapping system of chieftaincies administered central lands; at the periphery, sons of aristocrats with new armies pressed against surrounding polities, making Rwanda one of the more powerful kingdoms of the region.
COLONIAL RULE
Rwanda’s mountain fastness and warrior reputation had kept the outside world (including Swahili slavers and traders) at bay until German explorers crossed the land in the 1890s, with Count von Goetzen discovering Lake Kivu in 1894. Three years later, Hauptmann Ramsay offered the newly ascendant king, Mwami Musinga, a letter of protection. German forces and firepower helped Musinga consolidate his rule at the court and expand Rwandan control in Gisaka to the southeast and Bugoyi in the northwest.
In 1907, Germany’s military protection changed to civil administration by indirect rule from a residency at Kigali. But in 1916, Belgian forces from the Congo pushed the Germans south out of Rwanda and beyond the rail center of Tabora. Belgium held its conquered territory provisionally until the Milner-Orts Agreement of May 30, 1919, ceded to Belgium the two German provinces of Ruanda and Urundi as the mandate territory of Ruanda-Urundi. At the Paris peace talks, the Mandates Commission approved the Belgian mandate and the Great Power Council of Five accepted the Milner-Orts Agreement on August 7, 1919.15 In 1947, this mandate evolved into a trusteeship under the United Nations.
The Belgians, like the Germans, faced the dilemma of how to assert authority over a polity already endowed with a sovereign, an administration, and a society knit together by clientage relations. Under ideals of the League of Nations mandate, Belgium was to assure the uplift of the local population as well as the development of natural resources.16 Belgium intended to meet these goals while administering the territory efficiently, with minimum investment and maximum gain from the territory’s dense population and fertile lands. Tying into the Congo’s physical and administrative infrastructure seemed the appropriate approach.17
Whereas in Congo, Belgian administrators ruled directly over respective provinces, in Rwanda, Belgium would follow the German approach of indirect rule. The goal, according to Colonial Minister Louis Frank, was to “associate the native princes with our plans and bring the indigenous ruling class into our service.”18 In 1919, the administration established schools for chiefs’ sons to train the necessary cadre. Then in 1926, the administration began to rationalize and decentralize the tripartite chiefs system while setting contiguous boundaries for chieftaincies and subchieftaincies.
This streamlined structure was to uniformly administer the provinces throughout Rwanda, with chiefs trained first at Nyanza and other local schools and then at Butare. The new structure was also intended to relieve the “Hutu masses” of burdensome obligations to three different chiefs. However, the traditional system of clientage, both in land and cattle obligations, remained, now legally and efficiently enforced. Ignoring common social identities found in lineage and clan, as well as regional variances in social structure, the mandate administration saw Hutu and Tutsi as the overarching social constructs of personal identification and, in 1931, established identity cards classifying each person accordingly.19
In the name of progress, colonial policy discounted regional differences, reinforced clientage obligations, and crystallized identities along ethnic lines. Moreover, the colonial regime, through the local chiefs, assessed a head tax, required labor on public