Burundi: Ethnicity and the Ethics of Responsibility

Myths and Misperceptions

Patterns of Violence: 1972 and 1988

The Politics of Ethnic Amnesia

The Social Impact of Development Efforts

The mind of man is far from the nature of a clear and equal glass, wherein the beams of things should reflect according to their true incidence: nay, it is rather like an enchanted glass, full of superstition and imposture... Francis Bacon

'Contrary to the Marxist 'adage that 'history re 1 peats itself, first as a Itragedy and then as a farce, the Yecent'history of Burundi has been 'tragically repetitivé. 'Nowhere else in Africa, with the qualified exception of Ethiopia and the Sudan, have human rights been violated on a more massive scale, and with more brutal consistency than in the course of the 1972 and 1988 killings of Hutu peasants. The 1972 b!oodbath took the lives of an estimated 100,000, in what must be seen as the closest approximation of an ethnic genocide recorded in post-independence Africa. In 1988 informed estimates suggest that as many as 20,000 may have been killed. As in 1972 the vast majority are Hutu, and so are the thousands of refugees that have fled their homeland.

There is another side to the Burundi tragedy, which calls to mind what Max Weber called the "ethi"cs of responsibility", as distinct from "the ethics of conscience. Imperfect as it still is, our knowledge of the Burundi situation prior to the events of 1972 and 1988 was not so limited as to rule out creative solutions designed to minimize the risks of renewed ethnic confrontation. Without in any way trying to prejudge the feasibility of such solutions, what is at stake here is the responsibility of scholars and observers in coming to terms with the dual imperative of establishing the facts of a situation, and, in the light of such facts, of assessing the range of possibilities of remedial actions, beari-ng in mind their costs and benefits, their short-term as well as thoir long-term consequences.

There is no simple or immediate explanation for the general indifference surrounding the Burundi siaughter, other than its

- 1 -

marginal position on the map of Africa. Smallness of scale (20,000 sq. klms.), difficulty of access beyond the capital city, not to mention the arcane character of its social system and its relative isolation from cold war issues, have contributed to relegate the country to that residual category of micro-polities, such as Equatorial Guinea, Gambia, Swaziland , Cape Verde, Sao Tome and Principe, whose minuteness and eccentricities are generally seen as sufficient reasons for neglect.

This , however, is at best a partial explanation, and at worst an alibi for more more fundamental factors and circumstances, having to do with the misperceptions, prejudices and highly questionable policy assumptions that have shaped much of our vision of Burundi society. The misperceptions are the consequence of the myths that have gained currency, both in the media and in scholarly jouffials, about the structure of social conflict in contemporary Burundi. The prejudices stem from the biases and preconceptions of certain observers of the Burundi scene about the claims of one group against another; insofar as they can be given an aura of academic respectability s uch biases find expression in a political discourse, or, better still, a form of apologetics designed to preserve a certain professional image while insuring continued access to the field. The policy assumptions are traceable to the uncritical endorsement of the notion, especially prevalent among development economists, that there exists a necessary trade-off between equity and economic growth.

Myths and Misperceptions | MENU

One of the least controversial propositions about Burundi is that it is a plural society consisting of a majority of Hutu (85 percent), with the Tutsi accounting for 14 per cent and the pygmoid Twa for 1 per cent of a total population of approximately 5 million. Where disagreement persist is on the characterization of each group and the nature of their relationships with each other.

That they can hardly be called "tribes", in the sense in which some commentatois describe ethnic communities elsewhere in the continent, is fairly obvious, except to the correspondents of the New York Times. Burundi simply does not fit into the conventional "tribal" mold. Assumptions that it does can only obscure the f undamental affinities -- linguistic, symbolic and socio-economic that bind one group to the other, and, in the present context,

- 2 -

suggest the wrong solutions. Thus to argue, as former US Ambassador to Burundi does (New York-limes, Sept. 1, 1988), that '"separation (i.e. partition) is the only solution that will prevent further ethnic slaughter" overlooks the continuing social interdependencies between Hutu and Tutsi, and the extent to which they- intermingle in most rural settings.

Scarcely more accurate is the notion of 'caste" recently offered as a corrective to "tribe" by a distinguished roster of academics, consisting -- of ail people! -- of the Executive Committe of the African Studies Association, "several presidents emeritii of the Association and other distinguisheà Africanists". Burundi, we are told, was "stratified into castes of different rank and privilege, the primarily pastoral Tutsi minority closely linked to the ruling group and the agricultural majority of Hutu under their domination"( ASA News, Oct.-Dec. 1988, 8). While reasonably usefui, albeit with appropriate qualifications, to an understanding of precolonial Rwanda, such a portrayal is entirely at odds with what we know of traditional Burundi society, where the ruling group, the so-called ganwa, though never identified with either Hutu or Tutsi, did rely heavily on the former to administer the royal domains. 'Furthermore, social interactions between Hutu and Tutsi were far too close, ' intermarriage far too frequent, and the division of labor between them far too fluid, to provide the basis for a caste-like systern of stratification. The case for caste -- ' as Catherine Newbury and Alison Des Forges have recently reminded readers of the ASA Newsletter -- is simply unarguable.

While the concept of caste grossly exaggerates the socio-culturai distance between Hutu and Tutsi, scarcely more illuminating is the "tribal" label as it conjures up an image of "ancestral hatreds" rooted in the circumstances of the precolonial past. Contrary to much of the conventional wisdorn projected through the media the Hutu-Tutsi conflict is a relatively recent phenomenon. Ethnic polarization in Burundi is traceable to a self-fulfilling prophecy inspired by the Rwanda model. Despite the greater complexity of its ethnic map, and the absence of rigid divisions of the kind that characterized Rwanda society, the Burundi social systern could not but be profoundly affected*by the political message of the Hutu revolution in neighbouring Rwanda (1959-62) . While providing the nascent Hutu elites of Burundi with the "model politym that some tried to emulate, it also gave the Tutsi elites ample justification for their incipient fears of Hutu domination.

- 3 -

Rather than treating ethnicity as a historical datum, a far more fruitful perspective is to recognize the critical role played by aspiring politicians in moulding collective self-images. Only through a conscious effort at political mobilization on the part of political entrepreneurs, Hutu and Tutsi, could a sense of "us against them" develop within each group. As one Hutu observer who grew up in the sixties admitted to this writer: "I remember that at home and at school Hutu and Tutsi, while aware of belonging to different ethnies, generally lived on good terms with each other... The game of politics was frequently unnoticed in the countryside: thus the repression of 1965 and the purge of 1969 --programmed and orchestrated by Tutsi incumbents - had virtually no impact upon the populations concerned, which to-day would be unthinkable" (Miburo, 1988). Until the early sixties Burundi society was remarkably free of ethnic tensions. Whatever forms of conflict the kingdom experienced prior to and during the colonial interlude involved princely factions (Batare vs. Bezi) whose followings were recruited among both Hutu and Tutsi. One might indeed argue that the structure of factional conflict in the traditional society was an important factor in maintaining inter-ethnic harmony. To speak of "ancestral enmities" in the contgxt of the present crisis is patently at variance with the historical racord.

Yet another myth thàt needs to be dispelled has to do with the ethnic identity of the ruling group. To speak of Tutsi supremacy is sdmewhat misleading. A closer look at the ethno-regional roots of the civilian and military elites prior to the August 1988 killings shows that they are largely recruited from among the Hima subgroup. Most of thern come of the province of Bururi in the south, which claims the largest concentration of Hima elements. Within the Tutsi community power rests within the hands of a clearly identifiable core-group, culturally distinct from the other -- and far less influential -- Tutsi sub-group, the so-called Tutsi- Banyarug u ru. Thus to hold the Tutsi community collectively responsible for the abuses and atrocities of the Second and Third Republics is to make unduly short shrift of the critical role played by what can best be described as a minority within a minority.

Patterns of Violence: 1972 and 1988 | MENU

It may be useful at this point to take a restrospective look at the events of 1972 and 1988, if only to identify the similarities and

- 4 -

differences between them. While there is no gainsaying the irreducible fact of Tutsi oppression underlying these events, the precipitating factors behind them, and the -resulting scale of violence, suggest significant variations.

The structural constraints of Tutsi rule, including the use or threatened use of force, is of course the crucial common denominator in the background of the massacres of 1972 and 1988, and the extent to which violence was used in 1972 provides an important clue to an understanding of what happened in 1988. In the context of ethnic tension that developed in the wake of a local incident (of which more in a moment) mernories of the gruesome .killings of 1972 led a number of Hutu to anticipa te violence on a similar scale. In the popular consciousness of the Hutu masses forebodings of an impending bloodbath, comparabie to what happened in 1972, played a key role in the "grande peur" phenomenon that led to the indiscriminate killing of Tutsi elements in the Ntega commune in August 1988. An isolated act of violence by a single individual of Tutsi origins thus immediately took on ominous proportions. Widely perceived as the premonitory sign of a reenactment of the 1972 killings, panic. spread through the Hutu community. Partly out of an uncontrolled fear of what might happen next, partly out of a sense of outrage, some Hutu extremists unleashed their fury against innocent Tutsi, killing scores of them, thus bringing unto themselves a devastating retribution.

Although many of the basic facts about the circumstances of the 1988 holocaust are still a matter of speculation, there is little question that it differs in several important respects from that of 1972. For one thing the nature of the challenge faced by the Buyoya regime in 1988 was entirely different from that faced by Micombero in 1972, when coordinated attacks against Tutsi elements were launched in several localities, at the instigation of Hutu leaders and for the specific purpose of capturing power. In 1988 the initial attacks against Tutsi civilians were more in the nature of a spontaneous outburst of rage, triggered by the provocations of a local Tutsi personality, and fueled by rumors of an impending massacre of Hutu peasants. The precipitating event took place in the commune of Ntega on August 14, when, according to one observer, "a r-ich coffee merchant, -a Tutsi said to have been involved with the killings in 1972, refused to pay sonne Hutu peasants money he owed them. He taunted them, then shot and kilied five" (Watson 1989, 53).

- 5 -

Vengeance quickly ensued. After killing the trader and his family, a veritable manhunt got underway. Armed with clubs, spears, machetes, bows and arrows, bands of Hutu proceeded to set fire to Tutsi homes and kill the occupants.

Meanwhile, a few miles away, in the commune of Marangara, HutuTutsi tensions had already reached their peak, but for different reasons. At issue was the presence in this commune of four Tutsi civil servants, including the mayor, a judge and a medical assistant, whose ethnic prejudices were seen with enormous distaste by the local Hutu population. On June 28 the mayor made an inflammatory speech, that left many Hutu convinced that a plan was afoot to wipe them out. »You are preparing your knives", the mayor is reported to have said, "but ours are sharper and cut more than yours"". Hutu efforts to appeal to Bujumbura to have the mayor dismissed proved unavailing. And when three army vehicles appeared on the scene, on August 5, following what some described as "a mild uprising", panic broke out. "We knew what they were going to do", said one refugee interviewed in Rwanda; "everyone was saying: 1972! 1972!"(Watson 1989, 53). In the climate of intense fear that spread through Marangara and the neighbouring communes, almost any incident could trigger a violent reaction. The incident came with the provocations of Reverien Sarushingura, the Tutsi merchant from Ntega. Thus, while originating from different arenas, and out of separate issues, ethnic hatreds suddenly surged in the form of a blind fury directed against every Tutsi in sight. As many as 500 may have been killed in Ntega and its vicinity. The restoring of "peace and order" by the army proved even more horrendous. Assisted by helicopters and armored vehicles, Buyoya's troops unleashed their retribution with appaling brutality in the three northern communes of Marangara, Ntega and Kamya. According to Amnesty International, "the scale of resistance offered to the soldiers is not known, but the inclusion of large numebrs of women and children among thoir victims suggests that troops were engaged not just in quelling armed resistance or indeed in searching for those who had participated in killings of Tutsi, but rather in reprisais aimed at the Hutu civilian population as a whole and carried out to punish and eliminate them rather than just to restore public order" (Amnesty International 1988, 3). In short, the 1972 repression came about in response to-an organized conspiracy against the state; the 1988 repression followed in the wake of a highly disorganized display of rural violence directed against local officiais.

- 6 -

Which in turn points to further differences in the repressive responses of the Micombero and Buyoya regimes. Whereas in 1972 almost every region and locality, including the capital, was targeted for the physical liquidation of all educated Hutu elements down to primary school level, in 1988 repressive violence was both more indiscriminate and more localized. Again to quote from the Amnesty International report, "more and more eviudence has become available that members of Burundi's armed forces were respon'sible for massive numbers of executions of unarmed civilians, and that many of the victims were young children, some of them babies, who had played no part in a.ny of the preceding violence. They were selected for execution simply because they and their parents were Hutu and lived in an area in an area in which members of the Hutu community had attacked and killed Tutsi" (Amnesty International 1988, 1).

But perhaps the critical difference between the 1972 and '1988 killings is that they occurred in radically different political contexts. Although violence was in each case preceded by major discords among the ruling elites, there was no equivalent in 1972 for the move toward "liberalization" initiated by President Buyoya after seizing power from Bagaza in September 1987. As is now becoming increasingly clear, Buyoya's call for a more liberal stance on the issue of Hutu-Tutsi relations had little impact on local Tutsi functionaries, most of whorn stuck to a rigidly discriminatory posture. The circumstances leading to the events of Marangara, on August 6 and 7, make this abundantly clear. Whilé Hutu expectations were raised to an unprecedented level, causing some of thern to appeal to Bujumbura, their hopes remained largely unfulfilied, and their demands unmet.

During much of the Bagaza years (1976-1987) the Hutu-Tutsi problem wàs closely tied up. with the issue of Church-State relations. As a result of the restrictions imposed on catechism schools, tens of thousands of Hutu school children were in effect denied access to educational opportunities. When, for example, in October 1986 all catechisrn schools (the so called Yaka Mukama) were closed at the request of the government, an estimated 300,000 Hutu school children found îhemselves excluded from primary schools. Meanwhile between May 1979 and October 1985 about 280 missionaries were forced to leave the country. Given the fact that the Catholic Church runs about a third of the primary schools, and over half of the secondary establishments, it is easy to see why Bagaza's br'and of

- 7 -

anticlericalisrn should have been seen by many Hutu as beîng primarily directed against their own group interests.

The coming to power of Buyoya on September 3, 1987, led to a normalization of sorts with the Church. Most of the catechisrn schools were reopened. Some of the missionaries expelled by Bagaza were allowed to return. A n'umber of Hutu political prisoners were set free; and at least three high-ranking Tutsi officials accused of taking bribes from the French during the 1985 Bujumbura summit were thrown in jail. And yet, as noted earlier, while these were seen by most Hutu as unmistakable signs of an impending trend towards even greater liberalization, very little was done at the grassroots to adjust policies to expectations.

Thus if the notion of relative -deprivation has any meanîng, it is nowhere more apparent than in the gap between the rising expectations of the Hutu masses and their bitter disappointments upon discoveri'ng that, in spite of officiai statements to the contrary, nothing would alter the harsh realities of Tutsi supremacy.

The Politics of Ethnic Amnesia | MENU

For ail the differences noted above in the patterns of challenge and response that have accompanied the events of 1972 and 1988, in one fundarnental respect officiai responses to ethnic crises have remained remarkably consistent: even when confronted with the most dramatic evidence of ethnic conflict, the existence of separate ethnic identities has been consistently shoved under the rug. Officially, ethnic differences simply do not exist. Asked by a journalist what proportion of Hutu and Tutsi were killed in 1988, Buyoya laconically brushed aside the question" "We are all Barundi". The assumption, as far as one can tell, is that by eliminating all public references to ethnic identities, ethnic self-awareness will evaporate, and ultimately ethnic discrimination wili no longer matter either as a policy issue or a source of inter-group conflict. Proceeding from the axiom that ethnic labels and stereotypes belong to the dustbin of colonial historiography, the officiai attitude of the Burundi authorities is that ethnic references are, at best, a figment of the colonial imagination, at worst part of a neo-colonial strategem designed to play one group of citizens off against another. That ethnic identities may have taken on a new reality in the years that followed independence, through processes of political mobilization and class formation, is never seriously considered.

- 8 -

Even more surprising are the efforts deployed by some social scientists not only to endorse the official position of the Burundi authorities on ethnicity, but to give it'an aura of respectability through their scholarly endeavors. Here the case of Jean-Pierre Chretien deserves special consideration as it emerges as the most egregious example of academic sycophancy in the annals of French historiography. After twenty years of sustained research into the history and politics of the country, perhaps no other scholar has a better grasp of the political realities of contemporary Burundi. Yet nowhere in his writings is there the slightest intimation of the ethnic crisis facing the country, much less of an awareness of possible solutions; nowhere is there anything like a recognition of the potential for ethnic violence generated by the discriminatory policies followed by every government since 1972. For Chretien, as for the majority of Burundi officials, including his former student, now Executive Secretary of the ruling party, Etienne Mworoha, the Hutu -Tutsi question is a ialse problem, created out of whole cloth by European missionaries.

Over the last decade Chretien has written extensively on Burundi, but not without generating intense controversy over what one critic, Roger Botte, described as an "enterprise of disinformation". (Botte, 1983). Botte's strictures were inspired by a joint article (published in 1983) by Chretien and-Gabriel Le Jeune purporting to show that the communal congresses held in 1981 under the auspices of the ruling party, the Union pour le Proares National (UprQD,11 captured the essence of a peasant democracy aimed at rural development. Although the authors admit that the proceedings of the congresses made no reference whatsoever -- "not even by allusion" -- to the " general political situation of the country", they cite as evidence of grass-roots democracy the fact that "the reports of ' the congresses are extremely precise on issues of local interest". How far issues of local interest can be dissociated from issues of ethnicity is what remains unclear. The article, like the proceedings of the congresses on which it is based, remains utterly silent on the subject of HutuTutsi relations. As the title of Bottes rejoinder cruelly suggests ("When the essential remains unsaid") the authors' discussion is fatally flawed by their refusal to acknowledge the centrality of ethnic issues. That these are indeed of critical importance to an understanding of local issues is what was made dramatically clear by the events of August 1988 in Ntega and Marangara.

- 9 -

Burundi demacracy, according to Chretien and Le Jeune, is nowhere more deserving of praise than when compared to Rwanda. The reasons why, on that score, Burundi compares favorably with its neighbour to the north, are set forth in a 1986 article, in which the authors attempt to deal comparatively with the 1981 and 1982 legislative elections held, respectively, in Rwanda and Burundi. From their sociological travail emerges the conclusion that unlike what can be observed in Rwanda, in Burundi "le principe affirme est celui du depassement des clivages ethniques" (Chretien and Le Jeune 1986, 337); but this is not meant to suggest that ethnicity is ignored, as shown by the fact that out of 104 candidates at least 20 were Hutu, and, more tellingly, by the presence of one fourth of Hutu deputies in the legislative assembly ( seven of them having been appointed by President Bagaza); an additional proof of the quality of democratic processes in Burundi is provided by the nomination procedure, which involved a grass-roots selection process -- "a kind 'of primary election within the Uprona" (Ibid. 337) -- wh - ereas in Rwanda the candidates were selected "under the controi of the presidential cabinet" (Ibid., 337). While the first quote does little more than reiterate the vacuuous rhetoric of the authorities of the Second Republic, the other points, concerning the nomination of candidates and the ethnic breakdown of deputies, are patently inaccurate, as anyone familiar with the evidence avaliiable from more reliaible sources wili admit; According to one such source, "of four Hutu dighitaries who presented their candidacy, three were elected" (Africa Contemporary Record 1982-83, B 106).

The issue at stake here is not just one of shoddy scholarship (aggravated by personal biases). What is at stake is the responsibility of scholars in contributing to our understanding of crisis situations involving the fate and life chances of thousands of human beings. It brings into focus the basic choices that need to be made between the exigencies of scholarly objectivity and fairmindedness, on the one hand, and the temptations of apologetics, on the other, or, to put it differently, between the ethics of responsibility, and calculations of professional -- or should one say unprofessional"? -- self-interest.

The Social Impact of Development Efforts | MENU

The critical question raised by the 1972 and 1988 tragedies -- What can be done to promote democratic change in the midst of institutionalized ethnic inequality? has been consistently eluded

- 10 -

not just by academics but by development experts, with the result that development aid has tended to perpetuate, and in a sense legitimize existing power relationships.

The reasons for this are complex and go far beyond the obvious conclusion that Tutsi elements generally end up being the privileged recipients of aid benefits. A more relevant consideration has to do with the trade-off between efficiency and inequality that seems to govern the thinking of many development economists as well as their relationships with the host government. As one World Bank official recently stated, 'the level of education among the Tutsi is extraordinary; it makes for a damned good administrationm. So good indeed that Burundi ranked in 1988 as the largest per capita recipient of IDA loans ' in the worid. It is aiso one of the few states of Africa to have carried into effect the train of economic reforms advocated by the Bank. As Blaine Harden recently observed, "Burundi has devalued its currency, eliminated most important restrictions, raised prices paid to farmers and stripped the economy of regulations that inhibit trade" (Washington Post, Sept. 4, 1988). Compliance paid off handsomely. No sooner had Buyoya seized power than the Bank granted his government a three-year $ 90 million structural adjustment loan.

To put the matter in the mildest terms, the concept of project development as advocacy enjoys a very low order of priority among international donors. Neglect of the socio-ethnic pararneters within which aid projects are formulated and implemented has been a recurrent characteristic of most development efforts, with the consequence that there has been remarkably little concern over monitoring the social impact of such efforts.

Dismal as it may sound, the conclusion that suggests itself is that few people care, and fewer still know enough about the country to challenge public indifference about its destinies. Even among those who are most knowledgeable about the history and politics of the country something approximating a conspiracy of silence surrounds the subject of Hutu-Tutsi relations.

The most significant departure form the generalized public indifference about Burundi came in September 1988, in the course of a congressicynal hearing that led to the passage of a non-binding resolution by the House of Representatives. In it the House "urged the government of Burundi to maintain and greatly increase its

- 11 -

recent efforts at national reconciliation... condemned the recent violence reportedly carried out by the armed forces (...) and urged the President and Secretray of State to conduct a comprehensive reassessment of the United States bilateral relationship with the Government of Burundi with a view towards the immediate suspension of US assistance (other than humanitarian aid) unless within six months'after the date of this resolution (a) an -impartial enquiry (...) has been initiated to determine the causes of the outbreaks of violence; (b) the Government of Burundi has taken steps to investigate and prosecute those military and administrative officials and private individuais responsible for the recent atrocities ( ... ), (c) the Government of Burundi has made substantial progress in promoting the safe return to their homes of Burundi's refugee population ( ... ); (d) the Government of Burundi continues foreign journalists and international humanitarian relief organizations free access to the areas affected by recent violence".

The most significant aspect of the present situation is not that some of these recommendationbs have yet to be put into effect, but the extent to which they have been complied with. On October 6 a Consultative Commission of 24 members, including an equal number of Tutsi and Hutu, was created to investigate the circurnstances of the massacre and make appropriate recommendations te promote Nnational unity"; shortly thereafter, on October 12, Buyoya agreed to a major reshuffling of his cabinet, increasing the number of Hutu ministers from six to twelve, with the prime ministership now in the hands of a Hutu, former provincial governor Adrien Sibomana.

Whether, as a resuit of these reforms, significant changes will take place in the distribution of rank and privilege is hard to tell. While there can be little question about the trend towards ethnic depolarization, memories of.the 1972 and 1988 killings will persist for generations, and so will the mutual fears and hatreds they have instilled in the minds of the Barundi. For years to come the past will indeed continue to haunt Burundi's political future, shaping its ethnic destinies in ways that are as yet impossible to predict.

- 12 -