WHO ARE the Hutu and who the Tutsi? Are they the same people, as many a militant in the Rwanda Patriotic Front (RPF) is prone to insist? Or are they distinct ethnic—even racial—groups, as proponents of Hutu Power claim? In my visits to postgenocide Rwanda, the question came up in literally every discussion. The answers were many. At one end were those who claimed the difference did not exist or that it was simply a “normal” socioeconomic difference: either a class difference between poor and rich, or a division of labor between agriculturalists and pastoralists. In either case, they said it was a type of difference that would exist “normally” within a single people, anywhere. At the other end were those who maintained the difference to be sociobiological: Hutu and Tutsi, they said, were two distinct peoples with separate histories, until Tutsi migrants conquered the settled Hutu communities and reduced them to the status of a servile population.
The genocide consolidated two opposed points of view: one Tutsi, the other Hutu. A Nigerian colleague made the point boldly at a conference organized by the Dakar-based Council for the Development of Social Research in Africa (CODESRIA) in Arusha early in 1995. If he went to a discussion on Rwanda and Burundi, he said, he could close his eyes and tell the identity of a speaker by the twist of his or her argument: if a person claimed there was no difference between Hutu and Tutsi, or that the difference was one of class, the speaker was most likely a Tutsi. A Hutu intellectual was more likely to argue otherwise, that the difference was one between distinct groups, ethnic or even racial. The “no difference” (or class difference) point of view has come to be identified with a pro-Tutsi orientation, the “distinct difference” point of view with partiality to the Hutu. In the inflamed atmosphere of postgenocide Rwanda studies, even the tiny coterie of Rwanda specialists among Western academics—mostly Belgian, French, and North American—has not escaped this litmus test. Depending on one’s point of view, each gets tagged as pro-Hutu or as pro-Tutsi.
Within the region of the African Great Lakes, the diversity of views on this question can be mapped around a dialectic defined by two opposed viewpoints: colonialism and nationalism. European explorers and missionaries who entered the region at the turn of the century began with what seemed a commonsense observation: Tutsi aristocrats looked different from Hutu commoners. As we shall see, colonial scholarship built on this observation and constructed Hutu and Tutsi as different. Faced with an accent on difference, the intellectuals of the anti-colonial movement took as their starting point a second common sense observation: no matter how different they looked, Hutu and Tutsi were part of a single economic and cultural community. Ergo, they were the same.
When I returned from my first visit to postgenocide Rwanda in 1995, I was perplexed by the same question: Who were the Hutu and who the Tutsi? My forays into Rwandan history brought me to two observations that form the point of departure of this book.First, I noticed that the colonial and the anticolonial standpoints, one highlighting difference and the other sameness, had given rise to distinct scholarly traditions, each with its own overall consensus and internal controversies. While the “distinct difference” point of view highlights separate origins of ancestors of Hutu and Tutsi, the “no difference” point of view emphasizes subsequent processes leading to cultural integration on the one hand, and occupational and wealth differentiation on the other—both within the framework of a single cultural and economic community. The relationship between scholarship and politics has been dialectical: grounded in scholarship, political perspective has in turn influenced the development of scholarship. This, it seemed to me, was reason to take the scholarship more, and not less, seriously.
I became one of a growing number of African academics who, though new to Rwandan scholarship, was beginning to think of the relationship between political power and political violence in the 1994 genocide. The more I delved into it, the more convinced I was of how effectively power had mapped the parameters within which scholars had pursued knowledge of Rwanda. This gruesome event not only stripped from scholarship the veil of objectivity with which it habitually claims a distance from the world of power and practice, it also threw specific light on how complicit history-writing on Rwanda had become in the imperialist project in twentieth-century Africa. If power classified the population of Rwanda into three “races”—Hutu, Tutsi, and Twa—then scholars accepted race as a transhistorical reality and wrote the history of Rwanda as a history of the coming together of three races. If power assumed that migration was central to the spread of civilization, particularly of statecraft, in Africa, then scholarship became preoccupied with the search for origins. If power read racial differences as cultural artefacts, and translated cultural into political differences, few indeed were the scholars who pointed out that what was at work was really a political project trying to naturalize political difference as a simple and unproblematic reflection of cultural and biological difference.
Precisely because the discussion on origins was from the outset framed in political terms by the colonial state, it took on a partisan character that goes beyond the intentions of any one author. The most esoteric debates unfolded in political terms framed by the colonial power, even the most “disinterested” scholars observed those parameters and got implicated, their intentions not withstanding. If postgenocide sobriety can teach us one thing, it is how colonial power has become etched on the pages of scholarly books no less than on the surface of public life in the region. As we shall see, competing interpretations got linked to changes in sociopolitical context in the region. When it came to major shifts in interpretation, they followed events that signaled a crisis of political power no less than that of intellectual perspective, such as the 1959 Revolution, and now, the 1994 genocide.
My second observation followed from this reflection. To the extent that it saw political difference as a simple consequence of cultural or biological difference, neither scholarly tradition had much to say about the question that interested me: the politicaldifference between Hutu and Tutsi. To make that point, and to begin my own exploration into the meaning of Hutu and Tutsi, I first need to sum up the main tendencies in both scholarly traditions.
SEARCH FOR ORIGINS: THE MIGRATION HYPOTHESIS
At least four different kinds of studies have tended to buttress the “distinct difference” school of thought. All four highlight particular facts of history. Chronologically, the first is a literature from physical anthropologists that takes as its starting point differences in phenotype, mainly in physical height. Akin to this, but more recent, is a second type of literature originating from a combination of physical anthropologists and natural scientists. Its focus is genotype: blood factors, the presence of the sickle cell trait, and the prevalent ability among adults to digest lactose, a milk sugar. The third type of literature comes from cultural anthropology, and takes as its source material the memory of the peoples of the region as its cultural archive. The final branch of this scholarship is constituted by the work of the more conventional historians. In the absence of written archival sources that go back more than a century, and a recognition that it is difficult to stretch the reliability of oral sources for more than a few generations, historians have looked for other source materials, mainly archaeological and linguistic, to piece together a narrative. Let us briefly look at each of these scholarly endeavors.
Colonial anthropologists began with a commonsense observation concerning a difference in physique between those who lived in Rwanda: the Twa, the Hutu, and the Tutsi. The Twa were short, like pygmies. The Hutu were squat and of medium height, and the Tutsi were slender and tall. Since the Twa were insignificant numerically, hardly a few percentage of the total population, attention focused on the Hutu and the Tutsi. The data physical anthropologists gathered confirmed the visual evidence: on average, Tutsi tended to be taller than Hutu. Working in the early part of this century, a German anthropologist found a 12-centimeter difference in average height between Hutu and Tutsi.1 Over a decade after independence, Jean Hiernaux, the director of research at the Paris-based National Center for Scientific Research (CNRS), confirmed: “The Tutsi are taller than the Hutu by nearly ten centimetres.”2 From the observation that Hutu and Tutsi were indeed different in some respects, colonial scholars concluded that they were indeed different peoples who must indeed have come from different places. Thus was born the migration hypothesis, that the ancestors of the Hutu and the Tutsi migrated as different peoples into the region of the African Great Lakes.
The critique of colonial anthropology came in two rounds. The first was also the more polemical: it denied the very possibility of migration and argued that physical differences originated in social selection, that is, the tendency of privileged elites to breed and feed selectively. The second round came as more evidence was gathered, this time from natural scientists showing the plausibility of migration as a historical fact. Rather than deny the possibility of migration, critics questioned the contemporary significance of this historical fact. Walter Rodney best represents the first round of response, a militant nationalist critique of colonial racial ideology. His writing was standard reading for RPF cadres in the late 1980s and early 1990s. In his influential work, How Europe Underdeveloped Africa, Rodney explained physical differences in Rwanda as evidence of different levels of social development. The Twa, he claimed, remained “pygmies” because “they wandered around in small bands, hunting and digging roots, thereby failing to assure themselves of plentiful or rich food.” The Hutu “were more socially advanced than the Batwa” because “they did not live entirely on the whims of nature.” But compared to the Tutsi, the Hutu remained “short and stocky” because “the quality of their food fell short of the protein-rich Tutsi diet.” And so the pastoralist Tutsi, “subsisting on a constantly accessible and rich diet on milk and meat” turned out to be “one of the tallest human groups in the world.”3 Writing after the genocide, a captain in the RPF repeated that same argument: “Science can account for the differences in physical traits which are explained in terms of diet and natural selection.”4
While Walter Rodney emphasized social selection in the form of selective feeding, others have coupled it with selective breeding. More recently, Dominique Franche, a French social geographer, had pointed out that a 12-centimeter difference in average height said to differentiate Hutu from Tutsi was “exactly the same difference that existed in France between a conscript and a senator in 1815.” On this basis, he concludes: “The difference in height can be explained by their different lifestyles and eating habits, and by the fact that Tutsi noblemen, unlike Hutus, did not till the land.” To this, he adds the reinforcing influence of sexual selection: “Ideals of beauty vary amongst different social groups.”5
More recent studies pay less attention to phenotype such as body height and width of nose, and more to genotype: blood factors, the presence of the sickle cell trait, and the prevalent ability among adults to digest lactose, a milk sugar. A 1987 survey, “Genetics and History of Sub-Saharan Africa,” concluded that “though surrounded by Bantu populations,” the Tutsi and Hima are “closer genetically to Cushites and Ethiosemites.”6 Another study concluded that while the sickle cell trait was “about as common” among Rwandan Hutu “as [it was] in neighbouring populations,” it was virtually absent among Rwandan Tutsi. Previously taken as a marker of “race”—a point of view now discredited—the presence of the sickle cell trait is now considered evidence of survival in malarial environments through natural selection over centuries, even millennia. This finding thus reinforced the first: that the ancestors of Tutsi had indeed moved from a relatively malaria-free environment.7 A third genetically determined characteristic said to differentiate Rwandan Tutsi from Hutu—and both from all surrounding peoples in the Great Lakes region—is the prevalent ability among adults to digest lactose, a milk sugar.8
Lactose is “a biologically unique sugar” which occurs “as a free molecule only in milk.” The ability to digest lactose is limited in most human populations, except in “milk-dependent nomadic desert populations” which, through natural selection over millennia, have a gene (allele) that accounts for their high lactose absorption capacity.9 Studies in the Great Lakes region bring out three contrasting sets of empirical findings. They highlight, at one end of the spectrum, the Tutsi of Rwanda and Burundi among whom three out of four adults display a high ability to digest lactose; and at the other end, the neighboring Shi people of eastern Congo, among whom only 5 percent of adults display this ability. In the middle are the Hutu, among whom one out of three adults shows that same ability. Could the explanation for the high prevalence of lactose digestion among as many as a third of the Rwandan Hutu be centuries of intermarriage between Hutu and Tutsi?
The migration hypothesis was further reinforced by regional myths that predated the colonial period and were recorded by early colonial anthropologists and explorers. They have recently been strung together and framed into a single grand hypothesis by Archie Mafeje in a recent work, The Theory and Ethnography of African Social Formations.10 The central myth concerns the Bachwezi dynasty in the kingdom of Bunyoro in western Uganda. The Bachwezi are said to have “migrated from south eastern Ethiopia and southern Somalia with their long-horned cattle,” but moved on after “a few generations” when “chased out by Babito invaders” from the north. Following the myth, Mafeje suggests a migration in “a southwesterly direction where ecological conditions are ideal for cattle-keeping.” Mafeje thus links the Bachwezi of Bunyoro with the Bahima of Ankole and the Tutsi of the Great Lakes. While many may be reluctant to accept the restatement of myth as historical fact, few would dare dismiss it as outright fiction.11 At the same time, one needs to beware that public memory—in this case, myth—also changes and that this change is not entirely unrelated to official discourse. A context in which official discourse privileged some because they were said to have migrated from elsewhere was certainly an incentive to those concerned to embellish stories about their having come from elsewhere.
Writing in the UNESCO General History in 1988, Bethwell Ogot, the leading East African historical authority on the subject, both accepted that pastoralists and agriculturalists had long inhabited the region and noted that the number of pastoralists increased “sharply” from “about the fifteenth century.” Whence did these pastoralists—among the ancestors of contemporary Tutsi—come? This is where views differed among those who supported the thesis of separate origins. The first group comprised a combination of colonial anthropologists, explorers, and missionaries. Anthropologists were led by John Seligman, explorers by Hohn Hanning Speke, and missionaries followed Father Léon Classe; all subscribed to the “Hamitic hypothesis.” Discredited with the first onslaught of militant nationalism, this point of view nonetheless animated the political institutions of both colonial and revolutionary Rwanda. The second group comprised several trends in nationalist and postnationalist historiography. The “older view” (Roland Oliver) was that the Tutsi—and the Bahima—came “from the north-east, probably from southern Ethiopia.” A later theory (Chris Ehret) put forward the possibility that the Bahima/Tutsi could have come from “the east rather than the north,” testifying to the “late continuation of Southern Cushites as important pastoralists in the southern half of the lacustrine region.”12 The most interesting amongst these was Jean Hiernaux, who argued on the basis of genetic and archaeological evidence that the Tutsi may be ancient East Africans—“elongated East Africans”—whose physical distinctiveness attested to successful adaptation to and survival in a dry arid climate over millennia.
Among the physical anthropologists who wrote on the Hutu/Tutsi difference, Jean Hiernaux occupies a notable position because he combined explicitly antiracist convictions with the view that Hutu and Tutsi had separate origins. Based on studies of blood factors and on archaeological evidence, Hiernaux argued that the Tutsi were one extreme of humanity as it developed under African conditions, just as pygmies were the other extreme. He thus disagreed sharply with the Hamitic hypothesis—identified in anthropology with Seligman, a formulation we shall encounter in detail in the next chapter—that Tutsi were a civilizing Caucasian influence in Negro Africa. He began with the observation that, though Tutsi were on the average taller than Hutu, they were in most respects more different from Europeans than were Hutu: after all, they were darker than the Hutu in skin color, had thicker lips, while their hair was almost as “spiraled” as that of the Hutu.13 Hiernaux maintained that what had struck European explorers was a feature specific to Tutsi and related populations. This was “a tendency towards general elongation of the physical features: long and narrow heads, faces and noses, narrow thorax and shoulders relative to the stature; even the limb diameters are small when related to limb length.” He described this population as “elongated East African,” and located “the area of differentiation of these people in the interior of East Africa,” not outside it: “Fossil record tells of tall people with long and narrow heads, faces and noses who lived a few thousand years B.C. in East Africa at such places as Gambles Cave in the Kenya Rift Valley and at Olduvai in northern Tanzania. There is every reason to believe that they are ancestoral to the living elongated East Africans.” He then concluded: “Neither of these populations, fossil and modern, should be considered to be closely related to Caucasoids of Europe and West Asia, as they usually are in the literature.” Similar to the contrast between Hutu and Tutsi, Hiernaux pointed out, were the differences found in Kenya between “the elongated Masaii herders of the dry plains and the Kikuyu farmers of the well-watered mountainous areas nearby.” Along with the Masai of Kenya and the Tutsi of Rwanda and Burundi, Hiernaux classified the Fulani of West Africa in a single group—“elongated Africans”—and postulated “that they owe much of their constitution to a peculiar evolution in the semi-arid or arid crescent which caps sub-Saharan Africa to the north and the northeast,” and added that their most conspicuous features represent “genetic adaptations to dry heat.” It is not “an extra-African ‘Caucasoid’ element in their gene pool,” nor selective feeding or breeding, but genetic adaptation that explained for Hiernaux how such strikingly different physiques as those of the “elongated Africans” and the pygmies could be produced in tropical Africa: “These two opposite poles of morphological differentiation correspond to the two climatic opposites of tropical Africa,” one “hot and dry, with well-marked seasons,” the other “uniformly hot and wet.”14 The adaptation had to have occurred not only over centuries, but over millennia. The ancestors of Hutu and Tutsi, Hiernaux remained convinced, had to have had separate origins: “The Tutsi are evidently not Hutu transformed by selection.”15
I have argued that the critique of colonial anthropology came in two rounds. Written in direct response to colonial racial ideology, the earlier critique was the more polemical. It simply generalized backwards from the commonsense observation that—whatever the physical differences between them—Hutu and Tutsi lived in a single cultural and economic community and thus argued that they had always been a single people. A later response—the second round, so to speak—shifted ground. Without denying the possibility that ancestors of Hutu and Tutsi may have had separate origins, it questioned the significance of this historical fact in explaining contemporary realities. Historians suggested that contemporary cultural facts such as pastoralism, statecraft, or a common language could be the outcome of entirely different processes, with contradictory outcomes. Four kinds of studies made this point. The first proceeded by making a distinction between migration and conquest: that, whereas both can lead to cultural integration, they involve dramatically different political consequences. Migrations can be peaceful and protracted, but invasions are often a bloody and dramatic affair leading to political polarization. A second type of study questioned the presumed link between the Tutsi migration and the origin of pastoralism and statecraft in the region, thereby devaluing the significance of pastoralist migrations. A third type questioned the presumption that cultural exchange, in this case the spread of a common language, necessarily has to involve a movement of peoples, either via conquest or migration. An exchange of ideas or practices is possible without physical displacement of humans involved in the exchange. A fourth and final type questioned the presumption that contemporary Hutu and Tutsi are necessarily “pure” offspring of Hutu or Tutsi ancestors, whether these migrated or not, given widespread practices of cohabitation and intermarriage over generations, a historical fact otherwise erased by the patriarchal transmission of cultural identity.
Jan Vansina was the first historian to draw the distinction behind migration and conquest. Gathering his evidence from oral testimony, Vansina countered hitherto court-centered explanations of cultural assimilation based on conquest with popular accounts stressing a process of two-way cultural integration rather than a one-way assimilation. While we shall later see how this altered the prevailing account of the founding of Rwanda, it suffices to point out here that Vansina’s contribution almost immediately revived the migration controversy in the writing of African history. The context of the debate was the essay on Bantu expansion in volume 3 of the UNESCO General History, coauthored by Jan Vansina with the Ugandan historian S. Lwanga-Lunyiigo. While the authors had no trouble agreeing on devaluing the conquest-and-assimilation thesis, they could not agree on an alternative explanation. The body of the article cited the majority view among linguists and historians—advocated in the essay by Vansina—that “the Bantu languages originated in the West” and spread to most of subequatorial Africa through migrations. This, in other words, was the “migration-and-acculturation” view—that cultural transmission was the consequence of migration. Articulated by linguists and buttressed by genetic evidence, it gave primacy to the movement of peoples over linguistic borrowing. Instead of earlier notions implying one or several dramatic waves of migration, Vansina argued that migration must have occurred in dribs and drabs over two or three thousand years.16
In an appendix he penned under his name, Lwanga-Lunyiigo disagreed. Questioning whether the spread of language had necessarily to be the result of migration, he wondered whether it could not just as easily be the result of the borrowing of language. His conclusion called into question the entire migration hypothesis: “Basing myself on archaeological evidence, I suggested recently that the speakers of Bantu languages occupied from very early times a broad swath of territory running from the Great Lakes region of East Africa to the shores of the Atlantic in Zaire and that the supposed movement of Bantu-speakers from West Africa to central, eastern and southern Africa did not take place.”17 Whatever the merits of his specific argument, the important point about Lwanga-Lunyiigo’s contribution was that he questioned the prevalent tendency among historians to assume that every “development” within Africa had to be the result of an external impulse. In the process, even if unwittingly, he exposed the original and persistent sin of Western history writing when it came to Africa: the search for origins. The important question he raised was: Why presume that cultural development was the result of migration, of peoples, rather than the exchange of ideas?
In any specific instance, such as Rwanda, the question could not be answered without adequate historical information. The more historical information was gathered, the more the significance of earlier migrations was brought into question: both in explaining the origin of pastoralism and statecraft, and in the presumption that those who emigrated separately must have remained biologically separate and pure over centuries, so that those identified as Tutsi today must naturally be the offspring of those who may have come in as Tutsi centuries earlier. Even if historical research buttressed the notion that ancestors of Tutsi and Hutu had separate origins, it was forced to contend with a far more contemporary fact: that Tutsi and Hutu did not live as separate cultural communities in Rwanda, but spoke the same language, practiced the same religion, and lived on the same territory. This, too, was an outcome of historical processes, one that no historian could ignore.
A COMMON CULTURAL COMMUNITY: RECONSIDERING THE MIGRATION HYPOTHESIS
Recent historical research tends to deemphasize dramatic ruptures and to bring to light more integrative processes over the past five centuries. In this context, the tendency is to diminish the exaggerated significance attributed to migration by earlier historians, but without denying its possibility. J. K. Rennie’s summation of the state of research in the 1960s is still a consensus: while we do not know “whether the Bantu-speaking agriculturalists settled in Rwanda before the pastoralists,” we do know that “pastoralist groups began immigrating in considerable numbers from at least the 15th century.”18 It is research bearing on the five centuries since, and not the migration hypothesis, that has yielded results that can provide a truly fruitful starting point for further work. The focal point of this research is the formation of two communities: one economic, the other cultural.
An Economic Community
When the ancestors of today’s Hutu and Tutsi came together, they created different types of communities: economic, cultural, and political. The economic community was that of pastoralists and agriculturalists. The simple notion that Hutu were agriculturalists and Tutsi pastoralists is no longer sustainable in light of recent research which challenges the equation of Tutsi with pastoralism: more and more evidence has come to light that the predecessors of the Hutu had cattle long before the Tutsi appeared on the scene.19Such a finding questions the long-held assumption that the Hutu were always agriculturalists and the Tutsi always pastoralists. Agricultural and pastoral activities were hardly exclusive; they tended to be carried out jointly in most regions. Many Hutu had cattle, and many Tutsi farmed the land.20 Certainly, if Hutu had in fact had a natural and timeless aversion to cattle rearing and had always been cultivators, it would have made little sense for those in power to put restrictions on Hutu owning cattle.21 One needs to question the tendency to equate the origin of cattle keeping and pastoralism with the arrival of the Tutsi. These identifications—of Tutsi with cattle and Hutu with land—need to be understood less as mere facts unrelated to power than as historical artefacts created alongside the institutionalized power of the Rwandan state. The division of labor observed between the two at the onset of the colonial period is better thought of as a division enforced through the medium of political power rather than as a timeless preoccupation of two separate groups. The economic community was less a natural than a historical artifact, less a biological predisposition than a political creation.
A Community of Language
If we are to grasp fully the significance of historical research on processes of cultural integration, we need to think of the cultural community, too, as distinct both from this economic community of agriculturalists and pastoralists and from the political community of those living within the boundaries of a single state. From this point of view, the cultural community of those who speak a single language, Kinyarwanda, can be thought of as separate from the political community of those who have lived within the boundaries of the state of Rwanda since sometime in the sixteenth century. The parameters of the cultural community that speaks Kinyarwanda are much larger than the domain of the state called Rwanda.22 The disparity in their respective sizes was even greater in the precolonial period than it is today. The cultural community of Kinyarwanda speakers is substantial by any reckoning. Outside of Kiswahili, Kinyarwanda is said to have the largest group of speakers among the Bantu languages in the region. Speakers of Kinyarwanda number over ten million. Put alongside mutually intelligible languages, the pool expands to nearly 20 million.23People speaking variants of Kinyarwanda had settled widely in the region long before the consolidation of the Nyiginya dynasty as the state of Rwanda in what is known as the central court complex. Today, the Banyarwanda—the speakers of the language, Kinyarwanda—are spread over Rwanda, Burundi, Uganda, Congo, and Tanzania. If we understand an ethnic group to mean a cultural group, comprising those who speak a common language, then the Banyarwanda must be considered East Africa’s largest ethnic group.24
As members of the cultural community of Kinyarwanda speakers, the Banyarwanda are distinct from members of the political community framed by the state of Rwanda, variously called Rwandans or Rwandese. The cultural community of Kinyarwanda speakers long predated the political community framed by the state called Rwanda. Thus, we come to the point that the people called Tutsi, and those who came to be called Hutu, spoke the same language, lived on the same hills, and had more or less the same culture,25 depending on the cultural zone in which they lived. But they had yet to become one people.
Central Rwanda was the historical location of the precolonial Rwandan state. Outside it, there were at least two major zones that were culturally—but not politically—Banyarwanda. The first of these is today divided between northern Rwanda and western Uganda, settled by a people known as the Bakiga—“the people of the mountains”—who shared the same language but not the same social and political institutions with those who lived within the ambit of the Rwandan state. Not only did they have different settlement patterns, clan categories, and marriage forms, their political life was also highly decentralized and community based, in sharp contrast to the centralized hierarchy of the state of Rwanda. Here there were no Hutu and no Tutsi, at least not until German colonialism integrated part of this area into the Rwandan polity.
The second major cultural zone that lay outside the precolonial state of Rwanda is today divided between western Rwanda and eastern Congo. These speakers of Kinyarwanda live south and west of Lake Rweru (Edward), and north and south of Lake Kivu, in Congo. Unlike the Bakiga of the north, the Banyabwisha and Banyarutshuru of the west had long accepted delegates from the central court and thus their social institutions closely resembled those in the central region. Here, there were Hutu and Tutsi.
Two persons may possess the same broad cultural identity—Banyarwanda—and yet be marked by different political identities: Hutu or Tutsi. The cultural identity—Munyarwanda in the singular, Banyarwanda in the plural—exists alongside and in tension with the political identity, Hutu/Tutsi, whether in the singular or the plural. There is no Chinese Wall separating them. The middle ground is structured by several institutions that both contain and express the tension between cultural and political identities. The most important of these institutions are the family, the lineage, and the clan. All three have come to be shaped by the political power of the Rwandan state over centuries. We can see this in the social identity produced in Rwanda, through both intermarriage and clan affiliation. In some instances, there is evidence to contrast the outcome with practices among the Banyarwanda in the cultural Diaspora. The difference will illuminate some of the ways in which politics has come to shape culture since the sixteenth century.
Cohabitation and Marriage
Hutu and Tutsi lived together, not just as neighbors but also intimately, often through cohabitation, sometimes through intermarriage.26 The history of cohabitation and of intermarriage spans centuries. And yet, that history cannot be glimpsed from contemporary social identities. If you go to Rwanda or Burundi, the purity of social definition is striking: everyone you meet identifies as either Hutu or Tutsi; there are no hybrids, none is “Hutsi.” When cohabitation takes the form of marriage, the wife takes on the identity of the husband.27 The social identity is passed on through patrilineal descent.28 If the father is a Tutsi, then the child will be socially identified as Tutsi; and if the father is a Hutu, the child will be identified as Hutu. As the child takes on a unidimensional identity, that of the father, the identity of the mother—whether Hutu or Tutsi—is systematically erased. So it happens that the child of generations of intermarriage and cohabitation between Hutu and Tutsi comes into this world unequivocally Hutu or Tutsi.
One begins to understand the puzzled reaction of those new to the region, such as a visiting Sudanese intellectual:
I had come to know, more or less, the stereotypical description of the short Negroid Hutu and the tall, fine-featured Hamitic Tutsis. As I looked at my audiences, I saw a few who were clearly Tutsis and a few who were clearly Hutus. But most were somewhere in between, and I could not identify them. I later asked the Burundese, including senior government officials and ministers, whether they could tell a Tutsi from a Hutu. The response of the Foreign Minister, which represented the general tone, was a confident “Yes,” but “with a margin of error of 35 percent”—a remarkable margin given the confidence of the affirmative answer.29
“There’s been so much inter-marriage over the years that you often cannot tell who’s who,” said a presidential aide from Burundi to a Western reporter, and then added as an afterthought, “but everybody knows, anyway.”30
I have been unable to find comprehensive data on the extent of inter-marriage. Yet, all accounts I have heard of or read speak of considerable intermarriage: anywhere from a significant minority to a majority of contemporary Rwandans are likely to be children of Hutu and Tutsi intermarriages over the centuries. This means that we cannot equate the identities Hutu and Tutsi with those identified as Hutu and Tutsi when this process set in motion. Rather than being biological offspring of Tutsi of centuries ago, today’s Tutsi need to be understood as children of mixed marriages who have been constructed as Tutsi through the lens of a patriarchal ideology and the institutional medium of a patriarchal family.
Both Hutu and Tutsi recognized patrilineal kin groups, such as the inzu, the umulyango, and the ubwoko. The smallest of these was the inzu (lineage), and the largest was the ubwoko (the clan). The difference was this: the lineage had a shallow depth and comprised no more than four or five generations with a link to a recognized original ancestor; in contrast, the clan was supposed to have a common ancestor many more generations ago. In Rwanda, however, clan members “were definitely unable to trace their relationship” to a common ancestor. Although the lineage was either Hutu or Tutsi, never mixed, the clan—which comprised different Hutu and Tutsi lineages—came to be the only mixed kin group in Rwandan society: all eighteen major clans in Rwanda include Hutu and Tutsi (and Twa). While the lineage head had definite powers—such as presiding over its collective activities, judging internal disputes, supervising tax contribution, and meeting obligation for military service—the clan had neither a head nor any collective activities. Unsurprisingly, while members of a lineage behaved as kin with mutual obligations, Hutu and Tutsi clansmen “did not exhibit any solidarity at all and behaved towards each other as complete strangers.”31
The existence of mixed Hutu/Tutsi clan corporations, each composed of exclusively Hutu or Tutsi (or Twa) lineages, has led to an ongoing discussion seeking to explain their genesis. To explain the paradox of mixed Hutu/Tutsi clans in an otherwise divided society, Maquet asked his respondents (all Tutsi) whether the absence of a purely Hutu or Tutsi clan meant that “Tutsi and Hutu of the same clan descended from the same ancestor.” He noted that “the Tutsi answered that it did not. They explain it by the relationships which have linked Hutu to Tutsi as clients or servants. After some time the Hutu were identified with the group of their master. Such identification was particularly easy since Tutsi frequently emigrated from one region of Rwanda to another with their Hutu clients and servants.”32 In other words, Hutu clients simply adopted the clan identity of their Tutsi patron. Since the original clientship (umuheto)—whereby Hutu lineages gave a cow to a Tutsi patron in return for protection—was lineage based, it made sense that the incorporation of the client into the clan of the patron would also be lineage based, thereby explaining the existence of exclusively Hutu or Tutsi lineages inside otherwise mixed Hutu/Tutsi clans.
David Newbury has reformulated this hypothesis on the basis of comparative research that shows differences in clan identity among populations who lived on two sides of Lake Kivu, one on the island of Ijwi and the other on the Rwandan mainland.33 The importance of the Ijwi community is that while it is part of the cultural community of Kinyarwanda speakers, it is not part of the political community living within the borders of the state of Rwanda. It is significant that “no group on Ijwi—nor anywhere else west of Lake Kivu—retains a clan identity presently found among the eighteen largest clans of Rwanda.” Differences between Ijwi and Rwanda, he suggests, are likely to have arisen as a result of different developments in the two places since the clans of Ijwi migrated “from areas now part of Rwanda but formerly autonomous of direct Rwandan political penetration.” Newbury links the expansion in the size of clans in Rwanda to the process of state formation: “Centralized state penetration in the Rwandan case appears to have encouraged, maintained and perhaps extended broader identities.” The hypothesis makes sense of data gathered by d’Hertefelt in his 1971 study of clans in Rwanda: it explains why very small clans constitute “a much higher proportion of the population” precisely in those parts of the country where Rwandan state penetration has been “most recent and least intensive.”
The Newbury hypothesis gives greater weight to a number of claims put forth by Alexis Kagame, the historian of the central court in the colonial period. Kagame considered Rwandan clans “purely political” (“clans purement politiques”), and claimed that all Rwandan clans were Tutsi in origin; as Newbury points out, it makes better sense to think of them as artifacts of state construction, and thus as “dynastic” rather than Tutsi in origin. The point is brought home brilliantly with the case of the Renge,34 a group variously mentioned in Rwandan tradition both as a part of the Singa clan and as an autonomous group that was neither Hutu nor Tutsi. Whereas the two claims have hitherto been seen as contradictory, Newbury’s historical hypothesis dissolves the contradiction by showing the two identities as having existed at different points in history. Once we recognize that Renge was “a general term applied to formerly autonomous populations which were later incorporated within a system of state (in this case Rwandan central court) identities,” the contradiction ceases to exist. This underlines the point that we need to understand Hutu and Tutsi aschanging political identities by linking both to the history of the Rwandan state.
We shall see that the Rwandan state was a powerful political engine that restructured social relations wherever its tentacles took hold. The tendency was for social relations to follow rather than to precede or accompany the spread of political authority. The outcome was testimony to the primacy of politics. A whole array of institutions—from the army to clientship—enforced and undergirded the reproduction of Hutu and Tutsi as binary political identities. If we are to understand Hutu and Tutsi as changing political identities, we need to move away from notions of an unchanging Rwandan state in the “precolonial” period, and instead draw a historical outline of its institutional development.
Let me sum up this review of the existing literature as it bears on the question: Who is a Hutu and who a Tutsi? We have gathered the many answers to this question under two broad heads: one point of view claiming “no difference” (or “normal difference”) and the other a “distinct difference” between Hutu and Tutsi. Each is identified with a distinct political tendency, “no difference” with Tutsi power and “distinct difference” with Hutu power. Each is also anchored in one of the two broad perspectives that has guided scholarship on Rwanda: the “no difference” point of view explaining that the Hutu/Tutsi difference is an outcome of social selection characteristic of privileged classes throughout history, and the “distinct difference” perspective holding that the difference actually began with separate migrations into the African Great Lakes region.
The “no difference” point of view holds the Hutu/Tutsi difference to be socioeconomic, either a class difference or a division of labor.35 Yet, the Tutsi comprise different classes. The identification of Tutsi noble families—the “well-born Tutsi,” as colonial and Church officials used to say—with all Tutsi was erroneous, for it ignored the poor Tutsi to whom these officials usually referred as “petits Tutsi,” and who clearly belonged to a different class. Similarly, the notion that the Hutu/Tutsi difference is really a division of labor also does not hold in light of evidence that pastoralism was really a local development in the region and that the equation of pastoralism with a Tutsi migration needs to be rejected.36
At the other extreme is the “distinct difference” point of view, which holds the Hutu/Tutsi difference to be one between sociobiological groups.37 We have seen that this point of view tends to freeze Rwandan history in its misty beginnings. In so doing, it discounts the entire history of physical mixing (through cohabitation and intermarriage) and of cultural integration that spanned subsequent centuries. It thus innocently equates Tutsi of a few centuries ago with the patriarchal construct of Tutsi in spite of—or maybe because of—the widespread practice of social cohabitation. There is undoubtedly much truth in the refrain that RPF cadres were fond of repeating to every foreign visitor to postgenocide Kigali: “We speak the same language, have the same culture, and live on the same hills; we are the same people.”
The two points of views—one stressing separate origins as the source of the Hutu/Tutsi difference and the other highlighting the cultural integration that created a single Banyarwanda cultural identity from the diverse groups that migrated into the region at different times—need not be seen as incompatible. They can be seen as complementary rather than alternative accounts, each highlighting a different aspect of history. While neither is able to account for the history underlined by the other, each is incomplete without the other. In considering the plausibility of the two hypotheses before us—migration and social selection—I suggest we move away from an unequivocal embrace of either by distinguishing between a “strong” and a “weak” version of each. My point is that a weak version of the migration hypothesis is entirely compatible with a weak version of the social selection hypothesis. In a context where the historical evidence in both cases is slim, such a resolution has the merit both of taking the available evidence into account and of leaving the door open for future and even incompatible evidence.38
Let me begin with the migration hypothesis. While the strong version suggests an invasion or a mass migration in one or a few dramatic waves, the weak version disassociates the notion of migration from that of invasion. It then conceives of migration as a gradual infiltration, a cumulative result of a movement that unfolded in so many dribs and drabs over centuries and that had multiple outcomes. Jan Vansina, for example, has pointed out that there was evidence of peaceful coexistence between pastoralists and agriculturalists in the northeast, northwest, and west of Rwanda: “There was little raiding, no system of vassalage and no state formation to incorporate both groups.” On this basis, he conjectured that this represented “an early, continuing and quite stable relationship between the two cultures.” Noting that this hypothesis “corresponds with what we know of other areas such as Nkore,” J. K. Rennie argued that it made greater sense to think of the migration hypothesis as a series of separate propositions. The first of these, he suggested, would be that of a relatively peaceful coexistence between cultivators and pastoralists,39 giving way to a tension-ridden relationship only “with competition for land, or raiding and feuding if pastoralists moved in and dispossessed the cultivator of some land.” Similarly, he suggested that an “economic integration of the two groups by the exchange of cattle in a vassalage arrangement” be thought of as a distinct third set of relationships. All three possibilities, he suggested, can be thought of as distinct from a fourth—“the incorporation of the peoples into a centralized state”—that they “might follow or precede.”40 Such a hypothesis suggesting numerous migrations stretched out over centuries in no way rules out a weak version of social selection whereby rulers inbreed and reproduce on the basis of notions of social selectivity and physical beauty.
While this kind of account highlights the grain of truth in each hypothesis, it still fails to address the issue at hand. That issue arises from a third commonsense observation about contemporary Rwanda. It is, after all, the very political conflict and political violence that pit Hutu against Tutsi that has in the first place focused attention on the question: Who is a Tutsi and who a Hutu? Faced with the genocide, the social demographer Dominique Franche, previously cited, suggests that we think of Hutu and Tutsi as different communities: “The best term is ‘community.’ What we have here are two recently constituted communities, one Hutu, the other Tutsi, united by their hatred and fear of each other and thirst for revenge. What is now going on is a civil war between elites, who are fighting for power.”41 While Dominique Franche gets closer to understanding the Hutu/Tutsi divide, he still begs the question: What type of communities?
To answer this question, one needs to go beyond a simple aggregation of insights found in the existing literature. That simple aggregation brought us to the realization that even if we accept that Hutu and Tutsi have different and distinct historical origins, we still have to take into account a subsequent history that made of them a cohesive cultural group: one that not only lived on a common territory, but also spoke a common language and practiced a common religion. This, in turn, posed a question: Even if we start with the recognition that Hutu and Tutsi belong to a commoncultural group, we still need to explain that they have yet to create a common political community, one based on consent. Is it possible to arrive at a third point of view, an overarching one, that can take the first two into account by incorporating their respective insights, but without reproducing the limitation contained in each? To do so, I suggest two propositions.
First, we need to make an analytical distinction between three different kinds of identities: market-based, cultural, and political. For Hutu and Tutsi are best understood, not as market-based or cultural identities, but as political identities reproduced first and foremost through a form of the state.
Second, political identities—and the state institutions that undergird them—need to be historicized so they may also be understood as changing identities. There has not been one single and constant definition of Hutu and Tutsi through Rwandan history. Rather, the definitions have shifted as a consequence of every major change in the institutional framework of the Rwandan state.
In what follows, I will show the usefulness of differentiating political from cultural identities by contrasting two tendencies in Rwandan history. Having traced the tendency to cultural integration from separate origins, I will now highlight the countertendency: to a political differentiation, even polarization, notwithstanding the cultural integration. Later, when I deal with the colonial and the postcolonial periods, I shall show how the meaning of Hutu and Tutsi has shifted every time a new power has seen it fit to reorganize the institutions of rule. The real challenge, I will argue, is to go beyond understanding Hutu and Tutsi as political identities, to grasping the process whereby they have turned into polarized identities with no middle ground between them.
The most striking thing about the African Great Lakes region is the contradictory nature of cultural and political developments. The very people who came to be integrated into a common cultural community—the speakers of Kinyarwanda—became polarized into two distinct and even antagonistic political identities, Hutu and Tutsi. While the historical origin of the Hutu and the Tutsi may be shrouded in mystery, the nature of the state they built is not. The history of the encounter between Tutsi and Hutu is important, not because of where their ancestors came from, but because in their coming together they created certain political institutions which outlived that history and shaped a tragic future.
The Early Rwandan State
Rwandan historiography has been so caked with orthodoxy that it has usually taken a profound political crisis—such as the 1959 Revolution or the 1994 genocide—to crack the crust. Following the 1959 Revolution, the key challenge to orthodoxy was to understand the process of state formation. Whereas pre-1959 historians gave a “court interpretation” from the point of view of the centralized power that emerged at the end of this process, post-1959 writers tended to view the expansion of Rwanda from the standpoint of the societies that were incorporated into it. Whereas pre-1959 studies tended to emphasize a more or less complete assimilation of conquered peoples by a system of vassalage in which they received the use of cattle in return for services and loyalty, post-1959 writing paid more attention to pre-Rwandan kingdoms and gave a much less functionalist or harmonious interpretation of the vassalage system in the state of Rwanda.
The crowning text in the pre-1959 corpus came from the historical sociologist Jacques Maquet. The first major post-1959 reinterpretation to challenge the accepted version came from the pen of Jan Vansina, the second from that of Catharine Newbury. Jan Vansina reinterpreted the process leading to the founding of the Rwandan state; Catharine Newbury historicized the institution of clientship rather than presume it to be a transhistorical Rwandan or African affliction. Both checked the tendency to read history backwards from the present. I shall first deal with the significance and consequence of Vansina’s work, and then turn to Newbury. Following in the footsteps of the Ibadan historians, Vansina subjected oral histories to a criticism of form and transmission, thereby subjecting them to rules of evidence while establishing their credibility.42 His conclusions challenged the notion that the Rwandan state was an artifact singularly constructed by the Tutsi aristocracy. Instead, he suggested an understanding of state evolution in which ritual institutions were borrowed from the earlier pre-Rwandan states and military organization from the Nyoro. He argued for a more protracted process of expansion and a less complete assimilation of subject peoples into Rwanda. In my view, this aspect of Vansina’s research has an important methodological significance. As I have already pointed out, it both discredits the view which sees the making of Rwandan state and society as the outcome of a process of one-sided assimilation and invites us to see this as the outcome of a process of two-sided integration.
The immediate consequence of Vansina’s research was to devalue the significance of the conquest-and-assimilation hypothesis: that state formation in the region began with conquest by a group external to the region, the pastoralists who began migrating into the region in large numbers since at least the fifteenth century.43 As others followed Vansina’s pioneering leadership, the regional history of states and statelets constructed by surrounding agriculturalists before the rise of Rwanda came to light. The post-Vansina consensus was modified by J. K. Rennie in a major interpretive essay,44 and it was adopted by Bethwell Ogot in his essay on the Rwanda kingdom in volume 4 of the UNESCO History of Africa. According to this account, “the first inhabitants were almost certainly forest hunters and gatherers, represented by the Batwa.” The agriculturalists arrived later and began to clear forests for permanent settlements. By the fifteenth century, “many of the Bantu-speakers were already organised into small states.” They included at least three: “the oldest state in Rwanda” for which posterity has preserved no name but, which “was probably established by the Renge lineages of the Singa clan” and covered “most of modern Rwanda except the eastern section;” the Mubari state of the Zigaba clan, “which apparently covered an extensive area”; and the powerful state of Gisaka in southeast Rwanda, which managed to maintain its independence until the middle of the nineteenth century. Only the strongest of these found a place in Tutsi tradition as recorded by Kagame. Although Ogot spoke of the foundation of the Rwanda kingdom sometime in the fifteenth century, we should note that the founding date, established mainly on the basis of an oral tradition referring to a solar eclipse at the time of a royal coronation, has been subject to much debate, from 1312 (Kagame) to 1468 (Nkurikiyimfura) to 1482 (Vansina) to 1532 (Rennie).45
No matter the precise date of its expansion, it is clear that Abanyiginya clan expanded “not only into areas populated by fiercely independent lineages of Kiga hill-dwellers, but also into a myriad of states, some tiny and some quite powerful.” These states had institutions of kingship and regalia (drums, royal hammers, etc.), and they had “developed ritual power over the land and over rain,” features that scholarly orthodoxy had come to identify with only or mainly the state of Rwanda.46 The more the local history of state formation came to light, the more it deflated the significance of the conquest hypothesis associated with the rise of Rwanda. The more the presumed link between pastoralism and power was challenged, the more it became possible to put the history of the state of Rwanda within a broader regional context. Instead of a singular rupture, the Tutsi conquest, it now became possible to investigate elements of continuity in state development before and after the expansion of the Abanyiginya dynasty. In the words of one born-again historian, “It looks as if the newcomers found nothing better to do than to lie in the bed which had already been made by their predecessors.”47
The balance of scholarly opinion today is that the state of Rwanda emerged as did many a state in the region, through the amalgamation of several autonomous chiefships into a single nuclear kingdom, under the leadership of a royal clan. This royal clan was the Abanyiginya clan. The location of early Rwanda was near Lake Muhazi, in open savanna country between Lake Victoria and Lake Kivu. This was pastoral country, and the various states which gradually formed in that region “were originally built on the alliance of pastoralist groups.” The period was probably the fifteenth century. From this location, and through “several centuries of turbulent political history,” the center of gravity of the state gradually shifted westward to the forested highland area near the Nile-Congo divide, quite different from its original open savanna homeland in the east.48
The Rwandan state had a distinctive ideological and institutional characteristic. A mark of the very circumstances of its birth was that it associated “Hutu supernatural powers with Tutsi military powers.” This ritual basis was institutionalized “at the very heart of the state.”49 A narrative associated with Cyilima Rugwe, considered the founder of the state, testifies to its historical significance. According to this narrative, Cyilima Rugwe was advised by his counselors that the way to rid his territory of “rebels” was to go to a famous “Hutu” diviner and take him gifts of butter, goats, and honey. The diviner, however, demanded, before anything else, a blood brotherhood pact, which Cyilima refused on grounds that he could not enter into one with a “Hutu.” In the face of continuingdifficulties, however, he relented. In return, he got advice from the diviner, advice he followed and from which he benefited. Thus began the chain of events that led to the creation of the independent state of Rwanda—at least according to this account.
The ritual prescriptions that the king was periodically required to carry out for the welfare of the country were known as the ubiiru. The guardians of the ubiiru, those who laid down the principles of rule but did not themselves rule, were known as the abiiru. They advised the king on whether or not to conduct wars and designated the heir apparent and the family from which he was to come. The most important abiiru positions are said to have been created at the time of Rugwe (1559–86) and his successor, Mukobanya (1586–88). The three topmost abiiru were from the lineages of Tsobe, Tege, and Kono. The story of the origin of the Tsobe relates how its founder, Rutsobe, was a “Hutu.” J. K. Rennie suggests that the story of the origin of different abiiru is the most important clue we have of how predominantly agriculturalist (later “Hutu”) political units were incorporated into the state of Rwanda.
The formative period of this state, the period of its distinctive development, is associated with a series of wars that, beginning with the rule of Rujugira (1756–65), spanned several reigns.50 To understand the growing social polarization between Hutu and Tutsi, we need to focus on three ideological and social institutions—the court rituals (ubiiru) through which important Hutu lineages were incorporated into the court as ritualists (abiiru), the patron-client relationships through which the pastoralist hierarchy was organized, and the military and administrative systems that were the true backbone of the state—as they changed over the next century. Together, these changes suggest both a centralization of state power and a reorganization of society along hierarchically exploitative lines. We shall see that these changes not only happened simultaneously, they also reinforced one another. The political tendency to free the king’s power from restraint exercised by countervailing institutions had important social effects. The abiiru who set the rules of governance, but without themselves governing, were also the institution through which important Hutu lineages were incorporated into the Rwandan state. Then there was the parallel tendency whereby a new form of corvée clientship was imposed on newly subjugated Hutu populations. Both came to a head under the rule of Rwabugiri in the late nineteenth century. Together, they made for a double and related development: just as power was increasingly defined as Tutsi, the political and social position of Hutu was getting progressively degraded. Yet, the relation between Hutu and Tutsi was far from polarized even under Rwabugiri. It was mitigated by the military and administrative systems, both of which provided avenues—albeit limited—for Hutu participation in the state, while at the same time allowing for a form of organization that made for countervailing tendencies placing a check on administrative power at all levels.
Let us recall that a distinctive ideological feature of the Rwandan state was that it associated Hutu supernatural powers with Tutsi military powers. The supernatural powers were said to be the preserve of the abiiru. The heyday of the abiiru as a courtly power seems to have been the seventeenth and the early eighteenth centuries. A corollary of their predominance was that the king’s powers were limited. The progressive emancipation of the king from ritual prescriptions of the abiiru began with the reign of the fighting king, Rujugira (1756–65). He first undermined the spiritual monopoly of the abiiru by bringing to court a possession cult (the mandwa) said to have emerged among politically and militarily defeated lineages. Thereby, Rujugira was able to coopt into the service of the central court a ritual authority that might otherwise have operated against it. By the end of the nineteenth century, the abiiru had been so weakened that King Rwabugiri could afford to demonstrate publicly how little he cared for their ritual prescriptions.
Clientship was the second distinctive institution characteristic of the Rwandan state. I have already pointed out that our understanding of clientship has been profoundly affected by post-1959 research, particularly the work of Catharine Newbury.51 The pre-1959 assumption was that everyone, except the king at the top and those at the bottom, was simultaneously a patron and a client in an unending and unvarying chain of patron-client relationships. This claim about clientship as a transhistorical institution subsequently came to dominate African Studies in the West. As a corrective, Newbury carried out a bottom-up examination of changes in clientship in a peripheral region as it was incorporated into the kingdom of Rwanda in the last quarter of the nineteenth century. Her work underlined the importance of historicizing institutions in the face of a widespread tendency to read history backward from the present. That tendency had been epitomized by the work of Jacques Maquet. Those who followed Maquet’s ahistorical and functionalist lead assumed that clientship was the key social institution holding Rwandan society together. After all, it made for a structure in which everyone but the king was the client of someone else.
To counter this claim, Newbury studied the changing forms of clientship in Kinyaga, a region that had been outside firm control of the central court until the last quarter of the nineteenth century. Going back a hundred years from 1960, she studied changes in three types of clientship—umuheto,ubuhake, and ubureetwa—particularly as these affected reciprocity and inequality in the relationship between patron and client. In the umuheto form in which it first existed, cattle clientship involved the periodic gift of cattle from client lineages to patrons in return for regular protection. This meant, first, that this relationship was necessarily confined to those who owned cattle, leaving out the growing number of the cattle-less poor. It meant, second, that in spite of the inequality inherent in it, the relationship was more like a historical version of a modern protection racket, which explained more the social cohesion of elites than any bond between the rich and the poor, the powerful and the disempowered. The larger significance of Newbury’s work is that it diminished the importance of cattle clientship as an explanation of social cohesion even in the kingdom of Rwanda.
It does this by allowing us to identify two key changes in the institution of clientship from the time of Rujugira. The first was a shift in the form of cattle clientship from umuheto to ubuhake. It was a shift whose effect was to erode reciprocity while intensifying the inequality in the relationship. Between umuheto and ubuhake, there were at least two important differences. Whereas umuheto had linked an entire lineage as a group to an umuheto chief or his delegate, ubuhake most commonly linked an individual to his patron. Also, while the umuheto clientship involved the gift of a cow at regular intervals from a client lineage to its patron, ubuhake involved exactly the opposite: a patron ceding the use of a cow to a client. This means that while the umuheto clientship was limited to cattle-owning lineages, for only they had the cows from which to give one as a regular gift to a patron, ubuhake clientship was more likely to involve families with no cattle. Ultimately, ubuhake exposed the clients to “more arbitrary forms of exploitation,” including possible confiscation of any personal cattle at the pleasure of the patron.
Changes in the nature of clientship were closely affected by changes in the nature of land tenure, particularly from lineage control over land (ubukonde) to control by the king who then decided to assign it as pasturage (igikingi) to his closest subjects through the administrative appointment of chiefs. Whereas the claim ubukonde emphasized the right of a lineage over the land it had cleared, the basis of igikingi was a political grant from the King. The latter system is supposed to have started in the first half of the nineteenth century, in the reign of Yuhi Gahindiro, and most certainly led to a decline in the social position of the Hutu. For so long as lineages controlled land, the lineage head—the person having effective right to land—was appointed as the land chief (or subchief), since the role of the land chief was to gather agricultural tribute. This is how one got to the point where many land chiefs in early Rwanda were Hutu.52 As control over land passed from lineages to the king, and thus its effective control from the lineage head to the administrative chief, there was no guarantee that the lineage head would continue to function as the land chief.
It is the loss of land rights by the mass of cultivators that explains the introduction of ubureetwa, a form of clientship, that was almost entirely without an element of reciprocity, in fact one that starkly underlined the serflike status of the Hutu population. In contrast to ubuhake, a form of clientship that attracted all those with an interest in access to a cow, usually Tutsi more than Hutu, ubureetwa was a form of clientship that was imposed only on the Hutu.53 Ubureetwa originated under the reign of Rwabugiri: it was imposed on Hutu lineages by hill chiefs, who replaced lineage heads and took their land by right of occupation. Ubureetwa entailed manual labor for the local hill chief, performed as “payment” for occupation of the land. While his regime imposed a harsh rule on the formerly semiautonomous Hutu and Tutsi lineages, Rwabugiri imposed corvée-type labor obligations only on the Hutu, thereby polarizing the social difference between Hutu and Tutsi.54 As a result, more than at any other time in its history, the state of Rwanda appeared as a Tutsi power under Rwabugiri.
THE MILITARY SYSTEM
While the effect of changes in court rituals and forms of clientship was to polarize relations between Hutu and Tutsi, we also need to note countertendencies that mitigated the polarization for a time. These countertendencies stemmed from both the military and the administrative systems in the state.
More than any other, it was the military system that began to develop during the reign of Rujugira (1756–65) that explains the distinctiveness of the state of Rwanda in the region of the Great Lakes. It was a time when Rwanda was faced with the combined military threat of three neighbors. The external wars led to a reorganization of the state internally. The internal reorganization was the real source of Rwanda’s growing strength in the region. Triggered by the spectacular military expansion that began under Rujugira, the military reorganization had administrative consequences of long-term significance.55 Armies were permanently posted to outlying areas, making it possible both to occupy a conquered area and to use it as a staging ground for cattle raids. Army members were drawn from geographically diverse areas and assigned for a number of years of common service outside their home region. This gave endurance to the umuheto groups that constituted the army and turned them into a key institution for socialization. In the standard literature, these groups are referred to as umuheto (“social armies”). At the outset, umuheto (meaning “bow”) incorporated mainly Tutsi and required of them some form of military service.56 Gradually, there was a blurring of the distinction between supplying the central court with spoils from military raids outside the territorial domain of Rwanda and supplying it with tribute from the population within the territory. As internal tribute gained in significance over external booty, the administrative functions of these umuheto groups came to predominate over their military functions. This was particularly the case by the middle of the nineteenth century.
Increase in state capacity was thus the result of both horizontal and vertical growth, the former through the incorporation of newly conquered territories, and the latter through a restructuring of state relationships in hitherto central areas. Army membership came to include all social groups. The first Hutu sections of the army were noted from the time of Rujugira in the second half of the eighteenth century. At that time, most Hutu were not yet warriors, but were associated with administrative structures of the army and were required to provide prestations on a permanent basis. Subjugation was characterized less by exclusion from the army, more by the differentiated manner of their incorporation into the army. By the end of the nineteenth century, every Rwandan male—Twa, Hutu, or Tutsi—was affiliated to the army. Every unit in the army had a distinctive name, evocative of a dynastic or historical event: “the tough ones” (abashakamba), “the first to be praised” (imbanzamihigo), or “the fearless ones” (inzirabwoba). As the army expanded, the forms of subordination of Twa and Hutu also changed. The more they were used in a fighting capacity, the less their participation was confined to nonmilitary activities such as herding. Contradictory claims by different writers about the social nature of the army make sense if we realize that their informants may be referring to different historical periods. Thus, Maquet’s claim that only the Tutsi had a fighting role in the army seems to refer to an earlier period.57 But it is in no way contradicted by De Lacger pointing out that the great conqueror Kigeri IV Rwabugiri preferred to recruit mainly Hutu armies when he went into battle.58 In the herding section, the army recruited entire lineage groups rather than individuals. In this arrangement, the head of the lineage performed an intermediary function between the army chief and individual members of the army coming from his lineage. The result was to tie every Munyarwanda to the state structure through the kinship structure, at the same time limiting the authority of the kin head by his incorporation into the military structure. The military system combined in a single organization economic, political, and military functions through a combination of state cattle keeping, provincial government, and fighting. It socialized Tutsi youth into dominant positions in the army at the same time as it gave the army a vested interest in state expansion.
Rwanda was an expansionist state. Starting from roughly the geographical center of present-day Rwanda, the Abanyiginya dynasty expanded its dominion aggressively and progressively. Its court history thus reads like an account of successive annexation wars. This was not only a fighting but also a looting army: besides going to battle to defend or to annex, the army also regularly raided neighbors for cattle. To the limited extent that the state managed to create one people of the Twa, Hutu, and Tutsi of Rwanda, it did so more on the battlefield than anywhere else.
THE ADMINISTRATIVE SYSTEM
While Rwabugiri’s reform centralized power, we need to bear in mind that power was nowhere near as absolute as it would come to be in the colonial period. In theory, the mwami (king) of Rwanda was said to be an absolute monarch, who was said to be both the supreme judge and the legislator with the right to change any custom.59 In practice, however, the country was administered through a threefold hierarchy: running from province to district to hill. Each province was entrusted to an army chief. Known as the chief of men, he was in charge of recruiting soldiers. The province was in turn demarcated into districts, each with two chiefs independent of one another. The chief of landholding was in charge of agricultural land and production; accordingly, he collected dues from agriculture. His counterpart was the chief of pastures; he ruled over grazing land and collected dues from stock. Finally, each district was divided into hills, each with a single hill chief.
The chiefs did not claim their position by virtue of right or inheritance; they were bureaucrats who were appointed to their position by a superior, either the mwami or a superior chief. The kings were considered sacred, and all were Tutsi. So were all army commanders, chiefs of pasture, and provincial chiefs. The occasional presence of Hutu chiefs could be noticed among chiefs of landholding, more so the lower one went down the hierarchy. The lowest ranks of administrators, the hill chiefs, could be Tutsi, Hutu, or Twa. The administrative system of the Kingdom of Rwanda toward the end of the nineteenth century was notable for two features. One, the rule of the monarchy was less absolute in practice than were its claims in theory. In practice, the monarch ruled through two sets of parallel hierarchies: at the level of the smallest administrative unit, the hill, every subject was linked to the monarch through the hill chief and the army chief; at the district level, the cattle chief and the land chief functioned as two parallel hierarchies. The existence of parallel administrative hierarchies made it possible for peasants to find breathing space by playing off one set of officials against another when the need arose. Two, while power was visibly Tutsi the higher one reached in the military and administrative hierarchy, it was still true that the lower ranks of administration, where officials were most involved in face-to-face contact with subjects, continued to include a significant presence of Hutu and Twa officials.
The Reign of Rwabugiri
The final and the most spectacular expansion of the boundaries of the Rwandan state took place under the reign of Mwami Kigeri Rwabugiri (1860–95), one of the most prestigious historical figures of the Rwandan court. Rwabugiri led a series of military campaigns that led to the incorporation of “Hutu” statelets in both eastern and western Rwanda; the northern and southwestern parts, however, remained largely autonomous.60 He further centralized the state structure, but through a series of reforms that had a contradictory outcome: at the same time as it expanded Hutu participation in the army from nonmilitary to fighting roles—and appointed Hutu to administrative positions while taking on the power of uppity Tutsi aristocratic lineages—these reforms debased the social position of the Hutu outside the army and administration and further polarized the social opposition between Hutu and Tutsi.
Research on the expansion of the Rwandan state during the reign of Rwabugiri and the early colonial period gives us critical insight into the transethnic nature of the Hutu identity. For Hutu, it appears, were simply those from a variety of ethnic backgrounds who came to be subjugated to the power of the Rwandan state. Take, for example, Catharine Newbury’s study of Kinyaga in southwestern Rwanda from 1860 to 1960. Its population became Hutu only in the last quarter of the nineteenth century, through gradual Tutsi military occupation and as a consequence of absorption into the institutions of the expanding state of Rwanda. This story of previously autonomous communities being absorbed within the boundaries of an aggressively expanding state focuses on the process of state expansion and its contradictory outcome. On the one hand, as local chiefs were dismissed and replaced by incoming collaborators, identified as Tutsi, land and cattle gradually accumulated into Tutsi hands. On the other hand, as those subjugated lost land and were forced to enter into relations of servitude to gain access to land, the “Hutu identity came to be associated with and entirely defined by inferior status.”61
The same can be said of the fiercely autonomous communities of Kinyarwanda speakers in northwestern Rwanda who were only brought under the fold of the state of Rwanda by a collaboration between German troops and the mwami’s soldiers. They, too, never really saw themselves as Hutu before their forcible incorporation into Rwanda. Before that, as we have seen, they were the Bakiga, the people of the mountains. And they used the term Banyanduga, not Hutu, to refer to the southern Hutu who joined the Tutsi in the war of conquest.62
One institution in precolonial Rwanda prevented the Hutu/Tutsi distinction from hardening into feudal-type orders, just as it prevented the formation of a Hutu counterelite that would in time challenge Tutsi domination. This was kwihutura: the rare Hutu who was able to accumulate cattle and rise through the socioeconomic hierarchy could kwihutura—shed Hutuness—and achieve the political status of a Tutsi.63 Conversely, the loss of property could also lead to the loss of status, summed up in the Kinyarwanda wordgucupira. Both social processes occurred over generations. Of little significance statistically, their social and political significance cannot be overstated. Noting that the process of “accession to the nobility” accelerated as did the expansion of the state, Jean-Népomucène Nkurikiyimfura astutely observed that “this ‘ennoblement’ prevented the birth of a distinct Hutu chiefly stratum which could have become a privileged intermediary between the court and the larger population.”64
The strongest proof of this is that when organized protest did emerge against Tutsi Power at the outset of colonial rule, it did not take the form of a struggle for Hutu Power. Instead, the protest, which drew its strength from the participation of those newly subjugated as Hutu, was led by none other than the Tutsi excluded from power. The protest arose on the basis of a joining of two forces: the Tutsi excluded from Tutsi Power, and the Hutu newly subjugated to it. The first was the consequence of a struggle for accession to the throne on the death of Mwami Rwabugiri in 1895. The struggle turned into a prolonged civil war between two clans, the Abeega and the Abanyiginya, and culminated in 1896 with the coup of Rucunshu in favor of the Abeega. According to Kagame, “countless members of the defeated party were massacred” and “new chiefs were appointed to fill the posts vacated by the death of the incumbents.” The pro-Abanyiginya legitimists sought refuge in the north and the east. The second ingredient that made for the protest, initially against Tutsi Power and then against colonial rule, was the dissatisfaction among the Bakiga people of the north, who had been formally incorporated into the kingdom scarcely a decade before the coup of Rucunshu.
The Cult of Nyabingi
Protest against Rwandan aggression and expansion at the time of Rwabugiri was widespread on the periphery of the kingdom. Two important sites of protest were the island of Ijwi and the kingdom of Ndorwa. Ndorwa is said to be the home of Nyabingi. Though the origins of the cult are obscure, its stamina is renowned.65 Beginning at the time of Rwabugiri, Nyabingi continued well into the colonial period. Though he “killed several leading mediums,” Rwabugiri “could not destroy the spirit,” for it “simply moved on to another host, infusing him or her with all the authority held by those fallen.”66 The cult survived into colonialism and “succeeded in immobilizing the administrative efforts of three colonial powers for nearly two decades, until its final suppression in 1928.”67
Nyabingi made an easy transition into the colonial period, for one reason: colonial rule simply added on to precolonial impositions. To the demands of occupying nobles were added the demands of European overlords. When first appointed in 1907, the German Resident for northern Rwanda wanted to build a European-style capital at Kigali. So he called for “hundreds of labourers” to work daily to construct a road into the north, and “thousands of days of labour” to cut northern forests and transport the wood. The missionaries, too, needed “800 labourers a day” to build the church at Rwaza, and more to build homes that would “reflect the scale and solidity of their civilization.” While missionaries and administrators paid a regular salary to regular workers, they “usually paid nothing” to “unskilled labourers whom they requisitioned through the notables.”68
When defeated claimants to the throne rose up against those they considered usurpers of royal power, there was no shortage of popular discontent from which to harness support. After the death of Rwabugiri in 1895, the resistance went through three phases, each associated with a different leadership. The best known of these was the opening phase under the leadership of Muhumusa, one of Rwabugiri’s wives, who took shelter in the rugged mountains of the north. In J. M. Bessel’s words: “Herself an outstanding personality, possessing great powers of leadership and organisation, and far more brains than probably any Tutsi woman before or since, she was in intelligence quite up to the standards of her late husband.” And then he adds: “Not only in intelligence but in ambition: in 1911 she proclaimed herself Queen of Ndorwa and promised her followers that she would soon liberate the country from the yoke of the Europeans.” Muhumusa was captured in 1911 by British authorities in Bufumbira, Uganda. This inaugurated the second phase of the resistance under Ndungutse, accepted by most authorities as a son of Muhumusa and Rwabugiri. Now the chief spokesman of the legitimist faction laying claim to the Rwandan throne, Ndungutse was nonetheless “viewed by the local populations as their saviour, as the prophet who would restore peace to the country and free the labouring masses from the servitude of the corvée (ubureetwa).” Lemarchand’s comment is worth noting: “Though himself a Tutsi, Ndungutse’s name became a symbol of anti-Tutsi sentiment, and by implication of anti-European sentiment as well.”69
Though German troops killed Ndungutse and his comrades in April 1912, the northern region remained the site of recurrent outbreaks against chiefs and the central administration for many years, right into the period of Belgian rule. The thread that knit together the protest movement, from its origins under Muhumusa and Ndungutse to its recurrence in subsequent years, what I call the third phase, was a possession cult of messianic proportions that went by the name Nyabingi. Nyabingi literally means “one who possesses great riches.” It is believed to have been the title of an eighteenth-century queen of Karagwe who ruled until she was murdered by a Hima chief of Mpororo named Ruhinda. One tradition had it that the Nyabingi sect was a vehicle through which the queen’s spirit enacted vengeance upon her murderers and disloyal subjects.
The Nyabingi cult presents a paradox. On the one hand, it was regarded as a powerful anti-Tutsi protest. J.E.T. Philipps, a British district commissioner of Kigezi who claimed firsthand knowledge of it, described it as “revolutionary in method and anarchic in effect,” and added: “The whole appeal is to fear and the lowest instincts, to the masses, Hutu, against the classes, Tutsi and Batwa.”70 Historians who have written of the fiercely independent spirit of the Bakiga of northern Rwanda and western Uganda have seen Nyabingi as a crucial ingredient in the making of this culture. Thus wrote P.T.W. Baxter: “The proud boast of the Kiga is that they never were, as a people, subjugated by either Tutsi or Hima.”71 And yet this resistance against the combination of Tutsi power and colonial rule—one that spanned decades and whose regional claim to being the most powerful rebellion in recent history is unrivalled—was for the most part led by disaffected elements from the Tutsi monarchy and aristocracy of Rwanda. The answer to this paradox lies in an understanding of the difference between the kingdom of Rwanda, even at its most repressive moment under Rwabugiri, and the colonial state of Rwanda under the Belgians. The answer lies in the changing nature of Hutu and Tutsi as political identities from the kingdom to the colonial state, even if both went by the name Rwanda.
There cannot be a single answer to the question we began with: Who are the Hutu and who the Tutsi? Not only do the identities Hutu and Tutsi have a history, they have also changed in the course of this history. Although seeds of an alternative hypothesis can be found in the writings of several authors, writings from which I will quote extensively, they have yet to be worked out fully.72 As part of that endeavor, I would like to suggest that Hutu and Tutsi be seen as political identities that changed with the changing history of the Rwandan state. This has two implications. First, if Hutu and Tutsi are historical identities, then we need to be open to the possibility that the definition of Hutu and Tutsi may have changed over time, and that there may therefore not be any single answer to the question asked so often: Who is a Hutu and who a Tutsi? Second, if Hutu and Tutsi arepolitical identities, then their history is likely to be coterminous with that of the institutions of power, particularly the state of Rwanda. While we may be able to speak of Tutsi as an ethnic identity preceding the formation of the state of Rwanda, we certainly cannot speak of Hutu with the same historical depth. For as a political identity, Hutu was constructed as a consequence of the formation and expansion of the state of Rwanda. If subject populations only came to be defined as Hutu after being incorporated into Rwandan state structures, we cannot speak of these as Hutu before that incorporation.
My historical overview leads to three conclusions. The first is that the search in migrations in dim history for the origins of Hutu and Tutsi is likely to be fruitless since Hutu and Tutsi are political, not cultural, identities. Ancestors of Hutu and Tutsi most likely had separate historical origins. Hutu did not exist as an identity outside of the state of Rwanda; it emerged as a transethnic identity of subjects in the state of Rwanda. The predecessors of the Hutu were simply those from different ethnicities who were subjugated to the power of the state of Rwanda. Tutsi, in contrast, mayhave existed as an ethnic identity before the establishment of the state of Rwanda. With formal mechanisms in the Rwandan state that allowed rulers to absorb the most prosperous of their subjects into their own ranks through intermarriage, Tutsi too became more and more a transethnic identity.
My second conclusion is that the predecessors of today’s Hutu and Tutsi indeed created a single cultural community, the community of Kinyarwanda speakers, through centuries of cohabitation, intermarriage, and cultural exchange. That cultural community is to be found today both within the borders of the state of Rwanda and outside of it. It is a regional community. From this point of view, the speakers of Kinyarwanda who today live outside of the borders of Rwanda can be considered a cultural diaspora.
My third conclusion is that Hutu and Tutsi emerged as state-enforced political identities. The context of that development is the emergence of the state of Rwanda. It is the history of that state that ultimately made of Hutu and Tutsi bipolar political identities. This definition happened over time. Its context was twofold. One was the process of state centralization, whereby the powers of the king grew at the expense of the ritual powers of the abiiru and the army emerged as the central administrative institution. The other was social processes, particularly changes in clientship, that led to the social degradation of Hutu.
With the Tutsi identity sufficiently porous to absorb successful Hutu through ennoblement and Hutu clearly a transethnic identity of subjects, the Hutu/Tutsi distinction could not be considered an ethnic distinction. Neither could it be considered a socioeconomic distinction, one between exploiters and exploited or rich and poor. This is because of the “petits Tutsi” who could not be told apart from many Hutu in their socioeconomic circumstances, and who were substantial in number and continued to reproduce through intermarriage. At the same time, the petits Tutsi could always be told apart from the Hutu socially, on account of both the petty privileges and the more substantial exemption from forced labor (ubureetwa) they were entitled to as Tutsi under Rwabugiri. It was also not a division of labor between pastoralists and agriculturalists, once again because the petits Tutsi were usually as cattle-less as the majority of Hutu, and the “moyens Tutsi” tended to combine herding a few cattle with cultivating a modest garden.
To be a Tutsi was thus to be in power, near power, or simply to be identified with power—just as to be a Hutu was more and more to be a subject. By subjecting only the Hutu to a serflike tribute, ubureetwa, and by eroding the spiritual powers of the abiiru, Rwabugiri’s reforms highlighted the growing bipolarity of the identity just as they did the social difference between a petit Tutsi and a Hutu. It was toward the end of the nineteenth century, as Rwabugiri’s rule was drawing to a close, that the Hutu/Tutsi distinction clearly began to appear as a political distinction that divided the subject population from those identified with power. Yet, when contrasted with Belgian rule, which was soon to follow, one is struck by two mitigating features. First, the Hutu continued to be present at lower levels of officialdom. Second, the boundary between Hutu and Tutsi was softened by a degree of social mobility; no matter how low its quantitative significance, this would prove to be a fact of great social and ideological importance.73 If Hutu/Tutsi evoked the subject-power distinction in the precolonial Rwandan state, the colonial state gave it an added dimension: by racializing Hutu and Tutsi as identities, it signified the distinction as one between indigenous and alien. By making of Tutsi and Hutu identities evocative of colonial power and colonial subjugation—and not just local power relations—colonialism made them more volatile than ever in history.