Burundi: Descent Into Chaos or a Manageable Crisis?
|Publication Date||1 March 1995|
|Cite as||WRITENET, Burundi: Descent Into Chaos or a Manageable Crisis?, 1 March 1995, available at: http://www.refworld.org/docid/3ae6a6c00.html [accessed 28 July 2015]|
1. THE WEIGHT OF THE PAST (1850-1966)
1.1 Tutsi and Hutu in Pre-Colonial Burundi
The recent Rwandese tragedy has caused a strong and not unreasonable anxiety concerning the fate of Burundi, especially since its promising experiment in democratization was violently interrupted in October 1993. In this perspective, given their similar social structures and parallel histories, the two countries are seen as twins, with the sickness of one easily infecting the other. To a large degree, this is true. But to a degree only. For if these two countries are indeed twins, they are dissimilar twins, not identical ones. And the fact is evident as soon as one looks at their pre-colonial history. Although the famous dual social structure of Tutsi and Hutu existed in Burundi, its nature and functioning were from the start somewhat different from the Rwandese case
Burundi was, like Rwanda, an old and centrally-organized kingdom dating back to at least the sixteenth century. But it had grown according to a different pattern. While Rwanda grew from a royal centre which kept adding to its territory in a rather homogeneous fashion and carried out an iron-fisted centralization, Burundi grew in a more supple, more 'organic' manner. Although the Mwami (king) was, like his Rwandese counterpart, a sacred and absolute monarch, his role was subtly different. He was the 'Father' of the Nation, almost more a religious than political figure, in whose mystique everybody shared. And, more importantly, the Tutsi segment of the population did not 'rule' politically. This role was devolved to the abaganwa (sing. muganwa ), a group of high-ranking nobles who dominated both Tutsi and Hutu. They were the provincial governors, ruling the various areas in the name of the Mwami. In turn, at the court, the king chose his close advisers among what was called the abanyarurimbi, 'those who can judge'. Abanyarurimbi were both Tutsi and Hutu by origin, but not abaganwa. And finally, in everyday life, the men who counted were the abashingantahe, 'those of the small stick', generally older gentlemen who were recognized as social referees and common law judges, arbitrating most of the ordinary quarrels and social problems
We had thus a situation rather different from the Rwandese case. A king with a 'softer' political definition and stronger religious and social roles. A specialized feudal aristocracy of abaganwa ruling over Tutsi and Hutu alike, both categories being considered as abanyagihugu (subjects). A royal court where influential courtiers were both Hutu and Tutsi. And a society where, through the abashingantahe system, the whole population was drawn into a common judicial practice. This did not mean that Tutsi and Hutu were equal. The Tutsi definitely formed an aristocracy. But because of the very closed elite circle of the abaganwa, this aristocracy was not exclusively political and the social distance between the two groups was less than in Rwanda. The capacity for social mobility between the two groups was also greater than in Rwanda and the general social homogeneity was stronger than in the northern kingdom. Although a distinct social domination of Hutu by Tutsi was evident, the cohesive nature of what could without anachronism already be called a nation-state, was stronger than the divisive potential of its social structure.
1.2 The Colonial Impact
Belgian policies in the two mandate territories of Ruanda and Urundi were generally similar, notably in the general administrative reform started in 1929 which led to the 'tutsification' of the native civil service. But there was a definite difference of emphasis. For the Belgians, Rwanda was the 'perfect case'. The whole system of a 'higher' race, of a protected king who acted as 'modernising ruler', of catholicism as a religious and moral extension of colonial rule and as a vector of europeanization, was always more complete, more absolute, in Rwanda than in Burundi. In Rwanda, it rose to the level of an ideology which was later to be taken over and turned around (but not destroyed) by the leaders of the Hutu 'democratic revolution' of 1959. The situation was never that extreme in Burundi. The king was protected by but not a direct tool of the Belgians. The action of the Rwandese King Mutara III Rudahigwa dedicating his country to Christ the King in 1946 was something which would have appeared bizarre to the Burundi court at the time, in spite of the national importance of the Catholic church. Also, at the level of the native administration, the 'tutsification' was in fact rather a 'baganwaization', most of the post-1929 chiefs being abaganwa rather than 'ordinary' Tutsi.
Socially, Belgian policies had a more limited impact than in Rwanda due to the fact that the main client/patron system of contract, the ubugabire, was both more equalitarian and more resilient in the face of foreign regulations than the Rwandese form of ubuhake. As a result, the political tensions in Burundi towards the end of the colonial period did not take at all the same shape as those in Rwanda. In both cases, the Belgians were shocked and panicked at the 'betrayal' of their erstwhile allies, the Tutsi. But while in Rwanda the word 'Tutsi' meant a well-defined and socially identifiable group which had developed serious problems during the colonial period with the Hutu masses, in Burundi the word 'Tutsi' was much less precise, and there was not a gaping chasm between Tutsi and Hutu.
Proto-nationalist political parties regrouped both Hutu and Tutsi, and their divisions were rather along the lines of former abaganwa lineage rivalries. There were two main lineages among the abaganwa traditional 'professional politicians', the Batare and the Bezi. The royal family had sided with the Bezi early on during the colonial period, because the Germans had supported the Batare. Later, during the Belgian Mandate, favours ebbed and flowed according to political tactics. At first, the Belgians had sided with the Bezi. But the Batare had in the person of Chief Pierre Baranyanka an extraordinarily clever politician who had managed to ingratiate himself with Resident Pierre Ryckmans, the greatest of Ruanda-Urundi's colonial administrators. Later Baranyanka had played on the detestation Resident Robert Schmidt (1944-1954) felt for the Mwami Mwambutsa IV in order to position himself politically. During the late colonial period, Chief Baranyanka certainly had more influence over the Belgians than the light-headed playboy-king Mwambutsa. Thus the colonial administration had to tolerate and even favour Baranyanka's nationalist party, the Parti Démocrate Chrétien (PDC) because it was considered a lesser evil than the main nationalist group, the Union Pour le Progrès National (UPRONA). The vocabulary used by the last Belgian colonial Resident, Jean-Paul Harroy, about both the PDC and the UPRONA, is eloquently clear:
There was a certain connivance and even a direct complicity between our Authority and the PDC.... The main point of their program we caught upon was their refusal of immediate independence.... The PDC quickly became the bulwark we hoped to use in order to stop the cancerous metastasis of UPRONA's progress
At the time--around 1960--such an attitude on the part of colonial authority was extremely damaging to an African political party and Resident Harroy's policy (fully supported by the Belgian Government) had, of course, the opposite effect of considerably strengthening UPRONA and weakening the PDC. UPRONA was led by a remarkable politician, Prince Louis Rwagasore, eldest son of King Mwambutsa IV. Far from being his father's tool, the young man was very much his own master and had managed to develop an original brand of nationalist politics. As the king's son he commanded considerable respect. But he had been educated in Belgium where he had acquired a then fashionable taste for radical left-wing politics. As a result, the 'Red Prince' could afford to play on several levels at the same time; as a Prince he could play on traditionalism; as a young radical intellectual, on fiery nationalist and socialist rhetoric; and as a person on a very enlightened approach to ethno-social politics. In September 1959 he had married Marie-Rose Ntamikevyo, a young Hutu, and the marriage had carried a powerful political and ethnic message. The young Prince, the firebrand nationalist leader, had married a 'low caste' girl, thus embodying personally the concept of national unity. And his two closest advisers in the UPRONA power structure, Paul Mirerekano and Pierre Ngendandumwe, were both Hutu. Thus UPRONA turned into a nightmare for the Belgians because the policy of divide and rule which they had applied in Rwanda, so successfully at first and so disastrously in the long run, could not work in the Burundi case Here, the main division was between the various abaganwa families vying for power and influence as Belgian authority receded. Both main groups, the administration-supported PDC and the ultra-nationalist UPRONA, carried with them a full range of abaganwa, of 'ordinary' Tutsi and of Hutu. In despair, the colonial administration resorted to sponsoring a newly-born purely Hutu party, the Parti du Peuple (PP or People's Party). The PP remained marginal in pre-independence politics, although this attempt at ethnic division was a sinister portent of things to come.
The elections of 18 September 1961 were a triumph for UPRONA which received 82 per cent of the vote and entered Parliament with 58 of the 64 MPs. All the other parties together, allied within the Front Commun (Common Front) could only muster 18 per cent of the popular vote. Ethnically, out of the 58 UPRONA MPs, 25 were Tutsi, 22 were Hutu, seven were abaganwa, a figure which showed a pronounced collapse of the old aristocratic elite, and four were 'uncertain', that is people of mixed parentage who cared little about displaying their ethnic tag.
Unfortunately, the abaganwa rivalries, socially and politically obsolete, were to have one last and enormously harmful effect before disappearing from the political forefront. On 13 October 1961, Prince Louis Rwagasore, UPRONA leader and logical future Prime Minister of independent Burundi, was shot dead by the young Greek settler Ioannis Karageorgis, while sitting at an outdoor cafe The murderer had acted as a hired gun on behalf of his employer, a Greek trader who hated UPRONA and was close to the sons of Pierre Baranyanka. It rapidly became clear that the whole murder conspiracy was a revenge plot by ousted Batare abaganwa who anachronistically saw UPRONA's victory as a triumph for the rival Bezi family. This shortsightedness was to have catastrophic consequences.
Rwagasore had personified a trans-ethnic form of nationalism. A prominent member of the royal lineage, an anti-colonialist, intimately linked with the Hutu community, he was a living incarnation of national unity. His violent death shattered the image, especially since the 'Hutu revolution' then taking place in neighbouring Rwanda carried a divisive message to Burundi. UPRONA Tutsi cadres immediately started working on 'tutsifying' the party.
1.3 The End of the Monarchy (1962-1966)
Burundi became independent on 1 July 1962. The festive occasion was to a degree marred by the memory of the dead Rwagasore, the man everybody had expected to lead the country wisely into its first years of autonomous existence. In a situation typical of the absence of real ideological or ethnic division in the country, parliamentary life was sharply divided between the so-called 'Monrovia Group' and its 'enemy', the 'Casablanca Group'.
These two groups had pretences at divergent economic theories, not always very clear. The Monrovia group was supposed to be more pro-western while the Casablanca group was identified with the 'progressive' countries whose leanings were more towards socialism. In the Burundese context, these divisions were purely subjective and artificial.
To make matters worse, the Monrovia Group had rallied 32 MPs out of 64 and the Casablanca Group the other half Politics became paralyzed in byzantine rivalries and personal conflicts between the various politicians. Nothing serious was undertaken and even business as usual became inordinately difficult.
The break--in itself an unhappy one--with that sterile situation came in 1965 after Pierre Ngendandumwe, Rwagasore's old lieutenant, briefly became Prime Minister. The various prime ministers between July 1962 and January 1965 had been a succession of rather lacklustre characters, with a brief period during which Ngendandumwe himself had been at the head of the government An accumulation of economic, diplomatic and administrative problems seemed to call for a firmer leadership. Ngendandumwe, a member of the 'Monrovia Group', but a national figure and an independent person, was called upon by the King to form a new cabinet after the preceding 'Casablanca' administration of Albin Nyamoya had accumulated a number of blunders. But on the very day he announced his new cabinet (15 January 1965), Pierre Ngendandumwe was shot dead. The event was to initiate a tragic course of events.
Although the circumstances of the murder were never fully clarified, it now seems reasonably certain that the killers were Rwandese Tutsi refugees with a deep hatred of the Hutu. The effect of that ethnic motivation for the killing was catastrophic. The Hutu, who had previously felt they were disadvantaged but that reasonable channels of redress were open to them, suddenly felt that they had become political and social outcasts, and that any means, including murder, would be used to stop 'them' from participating in the power structure. The elections of October 1965 contributed to a deepening of ethnic antagonism. UPRONA won with 73 per cent of the seats. But 70 per cent of the new MPs were Hutu, both within UPRONA and among the PP-linked 'independent' candidates. Nevertheless, the King chose a Tutsi Prime Minister in the person of Leopold Biha, his Personal Secretary and a particularly hapless politician. The radical Hutu leader Gervais Nyangoma, who was secretly hoping to be chosen for the post in spite of not even being an MP, experienced deep frustration at this nomination. Since Biha was so unpopular, even among the Tutsi, Nyangoma and his friends thought they could resort to violence The Nyangoma coup, aimed at killing the King and taking power, quickly fell through (18-19 October 1965). But the Hutu insurrection which followed within days in the province of Muramvya killed an estimated 500 Tutsi before being crushed at the cost of around 2,000 Hutu lives. A new political pattern--Hutu against Tutsi--had emerged in Burundi. It was to cost thousands of lives and has yet to be changed into a more constructive framework.
An associated development related to this new violent ethnic pattern took place in the role of the Army. A young Army Captain, Michel Micombero, had personally directed the battle against the coup-plotters in Bujumbura. In the following days he was the one who 'restored order' in a rather violent way in Muramvya. He emerged from the crisis as a 'strong man' on whom the Tutsi extremists quickly pinned their hopes.
King Mwambutsa IV, who was spending his time between Geneva and the Spanish Costa del Sol, living in luxury hotels, looked completely unable to play his role in the increasingly tense situation. He was deposed on 9 July 1966 and replaced by his son Prince Ndizeye who adopted the regal name of Ntare V. But the real power behind the throne was more and more the military might of Captain Micombero who became Prime Minister on 23 July. The enfeebled monarchy continued until 28 November 1966 when Captain Micombero simply declared it abolished.
The new regime threatened repression but did not have to use it. It was well known that the new Army Chief of Staff, Major Albert Shibura, was a Micombero supporter and would not hesitate to use force. Another Micombero ally, Arthémon Simbananiye, became Public Prosecutor. His very presence at the head of 'justice' was enough to intimidate all opponents for the time being
The emergence of an Army dictatorship sharpened ethnic antagonisms. Captain Micombero was himself a Tutsi extremist and he promoted like-minded people around him. Both Hutu and moderate Tutsi politicians were sidelined. But the regime was using 'ethnicity' for reasons which were far from being abstractly ideological. In fact, the new military course reflected, somewhat similarly to the Idi Amin period in Uganda, a marginalization of the old elites and a rise of new groups of parvenus, among whom the military were the most prominent. Mediocre personalities benefitted from lightning promotions while capable administrators were forced to resign. The old abaganwa elite became irrelevant, now that the monarchy was gone, thus removing a buffer between Tutsi and Hutu. And worse, the new brand of Tutsi politics Micombero was ushering in was, as we will see, divisive of the Tutsi themselves. The fierce Tutsi-Hutu confrontation which was to become typical of what had to pass for politics in Burundi between 1966 and 1987 was in fact rooted more in social and economic rivalry than in supposed 'ancient tribal hatreds'. The best summary of this tragic situation had probably been given by Burundese Army Commander Martin Ndayahoze, one of the last Hutu officers remaining in the Service, who had said in 1968, four years before being murdered during the 1972 massacres:
We can safely say that it is the elite, the bourgeoisie, which carries the virus of tribalism. The disease comes from the top.... Mediocre civil servants need gimmicks to survive in their position or to get promoted. Greedy politicians use ethnic divisiveness as a political strategy. So if they are Tutsi, they denounce a 'Hutu peril' which must be fought, even by violence. And if they are Hutu they clamour against a 'Tutsi apartheid' which must be ended.
2. FROM MILITARY DICTATORSHIP TO DEMOCRATIZATION
2.1 The Micombero Years: The Establishment of a Tutsi Dictatorship (1966-1976)
Captain Micombero's coup had several dimensions. First and foremost, it was a move by the Army as a social group, to take power from the hands of a confused and divided civilian regime and to empower a new and less educated elite. But given the structure of the country's elite, it was also a new step towards the ethnicization of politics. Not only in the sense that Captain Micombero was a Tutsi, but because of the fact that he was a Hima Tutsi from Bururi, i.e. a member of a group which had been held in low esteem by the abanyaruguru, the high-ranking Tutsi clans of Muramvya Province. Micombero was a 'small' Tutsi using his Tutsi identity to persuade the 'big' Tutsi to support him. Of course, in order to achieve this--imperfect--Tutsi unity around his person, he had to brandish the real or imaginary threat of a Hutu insurrection. Thus the regime was to be permanently plagued by an ethnic double bind. On the one hand, the President-Dictator had to watch out for the partisans of the old monarchy and of a 'real' Tutsi regime who disliked him as an upstart. And on the other hand he had to face a largely self-manufactured Hutu 'danger', used as a Tutsi coagulant, but which in the long run would of course tend to become real.
The tension between the Micombero-led 'Bururi mafia' and the rest of the political class became increasingly dangerous. By late 1971 the Muramvya abanyaruguru circles were contemplating a coup d'état. The President pre-empted them and a series of arrests and rigged trials in late 1971 and early 1972 decapitated the neo-monarchist opposition This resulted in a double process: on the one hand the Muramvya 'high Tutsi' group felt it had to precipitate a monarchist restoration in order to eliminate the 'Bururi mafia', and on the other hand the marginalized Hutu elite felt it had to side, at least tactically, with the neo-monarchist plotters. For some of the extremists in the Micombero entourage, this was in fact a blessing18, an occasion to strike both at their Tutsi rivals and at the potentially dangerous Hutu mass which had so far showed almost infinite patience.
Their enemies fell into the trap. Former King Ntare V flew to Uganda from Europe where he had lived in exile since his deposition in 1966. For some of the Hutu extremist circles, this was the sign of a major showdown among the Tutsi which could be exploited for their own ends and they prepared an insurrection. King Ntare negotiated with President Micombero who guaranteed his safety. Trusting in his remaining popularity, the young king then flew from Entebbe to Bujumbura in late March and was arrested as soon as he stepped off the plane. After one very tense month the most extreme members of the 'Bururi mafia' prevailed: on 28 April 1972 President Micombero dismissed his whole cabinet and transferred de facto authority to the military and to a handful of civilian extremists. The next day, while the Hutu started--too late-- their long-awaited uprising, Interior Minister Shibura shot King Ntare dead and gave overall orders for a general massacre of the Hutu
The result was appalling. At least 100,000 people and possibly up to 300,000 died The action seemed at first to be aimed only at stopping the first wave of Hutu killing Tutsi. But once this had been achieved (within less than 48 hours) the killings went on, with a distinct social slant. The Hutu who were targeted were those possessing anything above the level of primary education. Army teams led by members of the State Security went around not only in the towns but even in the smallest villages, combing them for Hutu 'intellectuals'. They were all mercilessly slaughtered. The horror was such that everybody in Burundi, Tutsi or Hutu, still calls 1972 the year of ikiza (the catastrophe).
In the short run the regime had been superficially consolidated by this violence. But at a deeper level it had been contaminated by a sort of 'death fascination'. Since the massacres had no special East-West dimension, for the international community they were negligible. Soon, the violent extravagances of dictator Idi Amin Dada in neighbouring Uganda were to put the region back into the limelight. Since the massive and violent expulsion of the Asian minority in Uganda a few weeks after the Burundi massacres had definite and serious international consequences, world attention quickly turned away from Burundi to the new crisis. President Micombero could conclude that his genocidal policy had been a success and abandoned any form of overall national policy-making to sink into increasingly parochial or even personal politicking. Darkly misanthropic, given to prolonged bouts of drinking, he retreated into a world of his own. On 1 November 1976, Colonel Jean-Baptiste Bagaza, another Tutsi officer from a Bururi Hima clan, who was born in the same Rutovu commune as the President and some say was even a distant cousin of his, deposed the sombre and withdrawn Micombero who seemed to have lost touch with reality
2.2 The Bagaza Regime and the Continuation of Tutsi Rule (1976-1987)
The first Manifesto of the new regime (30 November 1976) seemed to address real isssues when it talked of the 'dark years of 1965, 1969 and 1972'. At first, the new regime seemed to try to play the card of a social and political opening, calling upon the UPRONA Party Youth, the Jeunesses Révolutionnaires Rwagasore (JRR), to denounce the social abuses they felt needed redressing. There was a moment of short-lived elation. But this was mere window-dressing and the Bagaza 'revolution' was soon seen for what it really was, i.e. a simple palace coup, a change of the guard within the 'Bururi mafia' There was no apology or attempt at reconciliation concerning the 1972 quasi-genocide. The Army became even more of a closed and all-powerful elite and its recruitment remained as narrow as ever. Civil service recruitment was broader in intra-Tutsi terms (in order to soften the impact of the almost exclusively Bururi officer corps recruitment) but it remained largely discriminatory towards the Hutu majority. By 1985 there were still only four Hutu cabinet ministers out of 20, 17 Hutu MPs in the designated 'Parliament' (out of 65), and two Hutu members in the UPRONA Central Committee (out of 52). Only one ambassador out of 22 was a Hutu and two provincial governors out of 15, while members of the majority social group represented only 10 per cent of the teachers and 20 per cent of the students at the 'National' University As time went on, even the Tutsi elite felt the weight of the despotism it had to support in order to keep the Hutu at bay and retain its privileges.
With the help of such devoted allies as Interior Minister Charles Kazatsa, Education Minister Isidore Hakizimana and Security Chief Lieutenant Colonel Laurent Ndabaneze, President Bagaza built an iron-fisted dictatorship compared to which even the Micombero years seemed somewhat relaxed. Security men were everywhere, the press was tightly controlled, any form of meeting was spied upon and reported to the Secret Service and students abroad were subjected to regular police reports. Ordinary peasants were commonly thrown in jail for lack of identification, an offense normally punishable by a fine equivalent to US$ 15 or less, but which could cause them to lose their freedom for six months and at times could cost them their life when jails were not properly supplied with food By 1986 the President's power base had narrowed not only to the Tutsi group, but to the Tutsi of one province (Bururi), among the Bururi Tutsi to one precise group (the Hima) and among the Hima to the three small sub-clans of Rotovu, Matana and Vyanda.
Given the extreme tightness of the political space, the Catholic Church became the last bastion defending a minimum of freedom of speech. As such it soon fell into the sights of the regime which started to silence prelates and close down religious establishments. Even the Bishop of Bururi, Mgr. Bernard Bududira, was not spared because he objected to the government policy towards the Church. He was submitted to constant harassment and attacked in the controlled press, and a nephew of his who was in the Army, Commander Léonce Majanja, was detained. Seminaries were nationalized in 1986 and local charismatic community meetings forbidden. Catechists were detained and at times tortured under the vaguest of pretexts. The prisons started to fill up and several detainees died under torture
Sensing Army opposition, by mid-1987 President Bagaza was preparing to expel large numbers of officers who opposed his authoritarian rule. He also wanted to force into retirement numerous NCOs in order to make room for young boys of his clan. Given the growing discontent, he ended up having to detain several prominent members of the Tutsi elite who increasingly questioned his despotic attitude So when on 3 September 1987 a group of NCOs led a bloodless coup to overthrow Colonel Bagaza and replace him with Major Pierre Buyoya, relief was almost universal.
2.3 The Buyoya Regime and the Attempt at Democratic Transformation (1987-1993)
Major Buyoya was in many ways a traditional Burundi military ruler. Like former Presidents Micombero and Bagaza, he was a Tutsi Hima from Bururi. He was also a pure product of the military establishment. But he was younger and intellectually more open. He also resented the heavy atmosphere of suspicion, palace plotting and constant backbiting which had made up the general political culture of the Burundi military elite during the last 20 years.
One could say at this point that there was a 'Burundese political pattern' just as there was a Rwandese one further north. But while the 'Rwandese model' was one of systematised ethnic antagonism, with the abstract but ever-present threat of military revenge from the Tutsi in exile acting to keep the Hutu masses in line and convince them that their dictatorship was democratic because it was run by members of the so-called 'democratic majority', the 'Burundese model' was rather different. First of all, it was much less 'tight'. Contrary to its Rwandese counterpart, it never had the benefit of ideological formalization and it consisted more of a set of practices than of a coherent collection of rules and values. Although similarly based to a large extent on ethnicity, it was more complex. Most of the time, the main political tensions and rivalries had been not between Hutu and Tutsi but inside the non-Hutu elites, abaganwa families at first and then various Tutsi clans. The Hutu provided a sort of mute background, something like the extras in a costume movie. When the situation was favourable and the main Tutsi leader was a man of quality (the epitome having been the national hero Prince Louis Rwagasore), prominent Hutu would be allowed to play real political, social and even economic roles. But their situation always remained precarious, as the murder of Prime Minister Pierre Ngendandumwe had shown in 1965. In less favourable situations, the Hutu were downgraded to the role of passive onlookers, with a few token members of their community being given a handful of symbolic positions.
Tension between major Tutsi factions tended to be 'solved' through a 'confrontation' with the Hutu in order to tighten up the ranks of the ruling minority, the major and most horrifying example of this policy having been the 1972 ikiza. When the leader was violent or started to decline, as Bagaza did in his later years, the oppression of the Hutu masses could rise from mild to nearly intolerable. And even in the best of cases, ethnic discrimination was an everyday fact of life.
The 1959 Rwandese 'revolution' played a terrible role in shaping this repressive system. At the back of every Burundese Tutsi's mind there was always the fear of 'what would happen if they would all decide to rise and kill us to the last one'. The successive dictators had played on that fear, especially to keep the Army a nearly 100 per cent Tutsi preserve. The presence of several hundreds of thousands of exiled Rwandese Tutsi in Burundi had also considerably contributed to this atmosphere of latent paranoia.
The main problem lay in the development of a sick political culture, made up of spite and fear on the Tutsi side, of inferiority and hateful resentment on the Hutu side. But when all was said and done, the Burundese syndrome did not have the character of machine-like ineluctability of its Rwandese counterpart. There were always men who were placed on the 'wrong' side of the fence, Hutu members of UPRONA, liberal Tutsi who refused a black form of apartheid, and a hope that the ideals of Rwagasore were not dead. In a way, with limited ideological means, probably too late and with too little real help, President Pierre Buyoya was going to try to use that existing window of opportunity.
At first, the new regime, although more liberal in its everyday political practice and at the human rights level27, was still functioning largely on the model of the 'traditional' Burundi military dictatorships. The first Buyoya cabinet, announced on 1 October 1987, had only five Hutu ministers out of a total of 20. There were only four Hutu out of 15 provincial governors and there were no changes either in the officer corps, the judiciary or the civil service to open them up to greater Hutu participation
But the ethnic situation had become very tense. In May 1988, an aggressive communiqué, issued by the extremist Hutu organization PALIPEHUTU, posed again directly the question of ethnic discrimination in terms of violence And Mgr. Bernard Bududira, the Bururi Bishop persecuted during the last year of the Bagaza regime, called attention to the situation in a vigorous pastoral letter issued practically at the same time and trying to defend a spirit of peaceful reform
During the night of 14 to 15 August 1988, a sudden explosion of inter-ethnic violence started in the two neighbouring communes of Ntega (Kirundo Province) and Marangara (Ngozi Province). For two days bands of Hutu peasants led by PALIPEHUTU activists scoured the collines, killing the local Tutsi. On the third day, the Army arrived and massacred indiscriminately all the Hutu they could find without bothering to try to first determine who was guilty and who was not. The violence caused 5,000 casualties according to the government, probably nearer to 20,000 according to foreign observers Over 60,000 refugees fled to Rwanda. What had happened? In fact, there was a mixture of causes all meshing into each other to produce a bleak scenario:
1. The tensions between various sectors of the Tutsi elite, notably the arrests of people who had been closely associated with the Bagaza regime up to a few months before, were perceived by the Hutu as a prelude, as in 1971-1972, to a new spate of massacres. PALIPEHUTU agents played on that fear and managed to trigger the ill-fated Hutu insurrection.
2. Some Tutsi elements in the Buyoya government were greatly alarmed by the liberalizing intentions of the President and wished to put a stop to these before they could have time to get implemented. The best way was interethnic violence. The Minister of Education, Lieutenant Colonel Jean-Claude Ndiyo, was probably the ringleader of a conspiracy aimed at frightening the Hutu through harassment and rumour, hoping that PALIPEHUTU would respond to the challenge.
3. In fact, PALIPEHUTU was almost sure to oblige because the Hutu extremists also knew about the liberalizing plans of President Buyoya and, just like the Tutsi extremists, knew that they would risk losing their constituency if the regime opened up.
For President Buyoya, the Ntega-Marangara events constituted a violent eye-opener. He realized that after 25 years of independence, the life of the country was poisoned by ethnic discrimination, fear and hatred, and that he had to try to solve the problem before the problem destroyed the country. But it was obvious things were not going to change immediately. Twenty-seven Hutu intellectuals who had addressed an open letter denouncing ethnic discrimination to the President (22 August) were all dismissed from their jobs and many had to flee the country Former Hutu MP Cosme Bibonimana, who in the past had criticised the practice of ethnic discrimination in examination marking, was summarily executed. The problem for Major Buyoya was the stubborn clinging to Tutsi supremacy on the part of a political, administrative, judiciary and military establishment used to 25 years of unbroken ethnic privilege. Any effort to democratise the country was going to be a two-way fight: first, to create the positive conditions for Hutu participation, i.e. to go beyond the distrust and entrenched radical hostility of the Hutu elite; and second, to avoid the obstacles a fearful and privileged Tutsi power structure was going to put in the way of any effort at liberalization.
But the President's decision had been made: he was going to try. On 19 October 1988 he formed a new cabinet. The new Premier, Adrien Sibomana, was the first Hutu to occupy this post since the murder of Pierre Ngendandumwe in 1965 and he had been encouraged to select a significantly bi-ethnic team. Soon after, a Special Commission was created to study the question of national unity. On 13 May 1989, it made public a special report on the question of national unity in an effort to understand the reasons for the deep-seated national disunity Professor Filip Reyntjens offers the following evaluation:
In itself, the document was disappointing. It simply offered once more the old Tutsi arguments on the plurisecular unity of precolonial Burundi, on the ethnic divisions being a pure product of Belgian colonial policies and on the Hutu responsibility in all the massacres perpetrated since the mid-1960s
While perhaps agreeing with this assessment, one must nevertheless recognise that the very fact that the central problem of Burundese society was thereby acknowledged, that the unspoken reality everybody knew about but nobody dared to mention publicly was finally brought out in the open, had an enormous collective therapeutic effect The Hutu opposition attacked the report as 'too timid' and the cabinet changes as 'cosmetic', but lost no time in organising itself in order to be ready for the day political parties would be allowed to operate freely. In 1990 a group of militants headed by Melchior Ndadaye, a young bank employee who had returned three years earlier from exile in Rwanda, created the Front pour la Démocratie au Burundi (FRODEBU) in semi-clandestinity And since he believed in free political competition, the President decided to overhaul the UPRONA single party and try to turn it back into the truly national party it had been in the days of Rwagasore. The December 1990 UPRONA congress elected a new Central Committee comprising 41 Hutu, 38 Tutsi and one Twa Nicolas Mayugi, a Hutu, became the new Secretary-General. Under this new political direction, Civil Service recruitment was opened up and school and university entrance exams began to be graded more and more according to merit and less and less according to ethnic origin. But the one bastion President Buyoya did not dare touch, because its recruitment reached so deeply into the unspoken fears of the Tutsi minority, was the Army. This timidity, understandable as it may have been, was to have tragic consequences later on.
The adoption by referendum of the Charte de l'Unité Nationale (Charter for National Unity) on 6 February 1991 was the first public test of the new political orientation and it was ambiguous. On the one hand, with 89 per cent of favourable votes, it constituted an approbation of the liberalization process. But on the other hand, the whole new direction appeared as too 'clean', as too well organized and too controlled by the Government to be really representative of deep-seated and pent-up Hutu popular feeling.
The clandestine PALIPEHUTU which was worried about losing its support wasted no time in exploiting this 'emotional gap'. It organized various public demonstrations, especially in the North where its support was greatest (April 1991), tried to put together a tax boycott and finally when it realized that it was losing support anyway, organized a series of terrorist attacks (November 1991) in the hope of provoking the Army into a bloody repression. In spite of the fact that President Buyoya was abroad at the time (the terrorists had cleverly scheduled their action) and that the Prime Minister was a Hutu, the cabinet managed to keep Army violence under control, while the Hutu masses did not rally to the uprising attempt. PALIPEHUTU drew the logical conclusions from this failure and decided from then on to change tactics. Since the organization's aim of a Hutu-dominated state remained the same, since it was unlikely to be legalized under any circumstances and since the Hutu masses seemed more interested in the moderate FRODEBU tactic of playing along with the government and trying to remain within a legal framework, PALIPEHUTU cadres used their sympathisers to start infiltrating FRODEBU and the nascent democratic movement. This, too, was to have dire consequences for the future.
A new democratic constitution was adopted by referendum with a 92 per cent vote in favour on 9 March 1992 and independent political parties became legal by the following June. FRODEBU quickly asserted itself as the leading opposition force among the ten or so different parties which had asked to be registered.
The year which elapsed between the advent of completely free political activity and the general election of June 1993 showed the limits of the President's policy of political voluntarism. He had taken the proverbial horse to the water but making it drink was proving to be difficult. In spite of its 'new' overhauled Central Committee, UPRONA carried out a generally slanderous campaign of ethno-political innuendoes against FRODEBU, accusing it among other things of being 'another version of PALIPEHUTU with a legal tag'. The Hutu who adhered to FRODEBU were described as 'subversives' and the Tutsi who did so (there were some) were branded as 'traitors'. This led Christian Sendegeya, a prominent Tutsi member of FRODEBU to attack the UPRONA leadership in an open letter in which he wrote:
The great weakness I would reproach your government is to tend to portray any Hutu who does not agree with you as a subversive sympathiser of PALIPEHUTU and any Tutsi who thinks differently from you as a misguided soul
On the other hand, PALIPEHUTU militants and sympathisers did join FRODEBU and some of the smaller Hutu-identified parties such as the Rassemblement du Peuple Burundais (Rally of the Burundese People or RPB), the Parti Libéral (Liberal Party or PL) and the Parti du Peuple (People's Party or PP). They obviously had ulterior motives which were hardly of a democratic nature.
Thus the parties went to the election with a mixture of democratic openness tempered by a belief in their own unshakable 'right to rule' on the part of President Buyoya and UPRONA, and an honest desire for a democratic alternative tainted by visions of ethnic revenge on the part of candidate Ndadaye and FRODEBU.
The first part of the election was the presidential contest which took place on 1 June 1993. There were three candidates: President Pierre Buyoya for UPRONA, Melchior Ndadaye for FRODEBU and Pierre-Claver Nsendegeya who ran in the name of the small monarchist party, Parti de la Réconciliation du Peuple (Peoples Reconciliation Party or PRP). Over 97 per cent of the 2,360,096 registered voters went to the polls. The election was scrupulously honest and its results were a surprise for many observers who had expected President Buyoya to receive the reward for his spirit of democratic openness But he got only 32.47 per cent of the vote against 64.79 per cent to his FRODEBU rival. The PRP candidate got only 1.4 per cent, which showed how irrelevant the monarchic question had become. Everybody behaved very responsibly. Disappointed President Buyoya sportingly congratulated Ndadaye on his victory. Colonel Michel Mibarurwa, the Army Chief of Staff conferred with Prime Minister Adrien Sibomana 'to coordinate our action and see how we can manage this transition period so as to avoid problems'. The President-Elect immediately moved to reassure the vanquished, saying:
There has been an attempt to make the population believe that our party is set on revenge, that it is fighting against the interests of a certain category of the population. None of this is true.
Knowing how explosive the Rwandese question was in ethnic terms, the new President also declared that the (Tutsi) Rwandese refugees would be allowed to stay in the country and that he would even use his influence to try to convince President Habyarimana that they should be allowed to recover their lost citizenship, something which he called 'a justifiable ambition'
On 8 June, the new President had declared a general amnesty both to free prisoners and to allow political exiles to come home. All prisoners and exiles were included, whatever their political persuasion. The PALIPEHUTU commandoes arrested in November 1991 were freed, but so were the (Tutsi) soldiers who had attempted to overthrow President Buyoya in March 1992 in the hope of stopping the democratization process. Exiled Hutu radicals came back from Rwanda and so did Tutsi supremacist former President Bagaza who had been living in Libya since 1988.
There were of course a few disturbing notes. Tutsi students marched through the streets chanting: 'No to the victory of division! No to the violation of the Unity Charter!' In the interior, the Hutu peasantry took the FRODEBU victory as a personal victory over the 'State' which was perceived by them as a Tutsi matter. So they refused to pay taxes any more and started to cut down the communally planted trees to use them for firewood. It took some persuasion to explain to them that the State was now theirs as well as the Tutsi's, and that unfortunately the continued existence of the State meant they still had to pay taxes and refrain from breaking the law.
The second (legislative) round of elections took place on 29 June and its results confirmed those of the presidential vote. Apart from UPRONA and FRODEBU, the PRP also ran, together with three other small parties, one, the Rallye pour la Démocratie et le Développement Economique et Social or RADDES, which was linked with Tutsi supremacist circles and was disgusted at UPRONA's 'softness', and the two others, the Parti Populaire (PP) and Rassemblement du Peuple Burundais (RPB) who were purely Hutu Participation, at over 91 per cent of the registered voters, was almost as high as for the presidential election. The results were as follows:
Number of seats
1. Twelve of the 16 UPRONA MPs were former ministers in the Buyoya regime, which shows that there was a 'political visibility premium'.
2. Twelve of the 16 UPRONA MPs (although not the same 12 as the ex-ministers) were Hutu.
3. Eight of the 65 FRODEBU MPs were Tutsi.
4. Although the vote was obviously mainly ethnically motivated, it was not fully so. A non-negligible section of the Hutu electorate had voted for UPRONA.
5. The small parties did not represent any sizeable portion of the electorate. But they represented an extremist opinion (whether Tutsi or Hutu) which felt that the two mainstream parties were too moderate for their taste.
By early July, Burundi seemed set to try to turn into economic, administrative and social reality what looked like a particularly successful political transition.
3. THE BIRTH OF THE PRESENT CRISIS (1993)
3.1 The Events Leading to the Murder of President Ndadaye (July-October 1993)
The general atmosphere of hope accompanying President Ndadaye's election had left a bitter aftertaste for the various extremist factions. The first to act were the Tutsi extremists in the Army. On 3 July, four days after the legislative polls, elements of the 2nd Commando Battalion from Muha Barracks tried to take power by force The movement was led by Lieutenant Colonel Sylvestre Ningaba who had been an ADC to President Buyoya But it was quickly stopped by another officer, Major Isaie Nibizi, who managed to talk the men and the NCOs out of following their mutinous officers. The reaction in Army circles was ambiguous. Many officers criticised Ningaba not for attempting a coup but for doing so with little serious planning. Many in the officer corps seemed to be paying only lip-service to democratic principles in spite of the clear verdict of the polls. The feeling was clearly racist in tone: 'those people' (meaning the Hutu) were described as not being capable of actually governing the country
Unfortunately, this criticism was not entirely devoid of foundation, even if its basis had nothing to do with 'race'. FRODEBU cadres were largely inexperienced for the simple reason that, firstly, there had always been a marked anti-Hutu bias in Civil Service recruitment and, secondly, the 1972 massacre had achieved its purpose by decimating the Hutu elite. The result was that many of the newly-nominated FRODEBU administrators at the regional level, and many of their men then entering the central administration were tragically incompetent. And this at a moment when the Hutu peasantry was expecting wonders from them, and many in the Tutsi administration were discreetly doing their best to complicate their work in the hope of seeing them fail. Furthermore, FRODEBU leaders were prompted to adopt too quick a changeover from the old personnel to the new by the feeling of having to deal with an enormous backlog of discrimination and the fear of disappointing their electorate. One of the main problems was the question of the refugees, most of whom had been living in Tanzania since 1972 although smaller groups had fled in 1988 and 1991 to Rwanda, who were watching the situation in the hope of being able to come back During his inauguration speech President Ndadaye had mentioned the question, saying that he was going to send 'delegations to foreign countries in order to assess the numbers of the refugees, to find out how many wanted to come home and what their needs were' This rang a danger bell for the Tutsi minority which had taken over the lands and other properties left behind by those refugees.
The new cabinet had been announced on 10 July 1993. Led by Sylvie Kinigi, a liberal female Tutsi UPRONA member, it was ethnically and politically balanced. It immediately drew fire from the Hutu extremists who had hoped for a 'radical Hutu' cabinet and found the new government much too moderate for their taste. From exile, extremist leader Kabora Kassan threatened an armed attack on Bujumbura if the Hutu did not get more cabinet posts and his guerrilla troops were not allowed to join the national Army President Ndadaye was very anxious to reassure the Army and he declared that no officer would be fired from the forces
But the Tutsi extremists remained unconvinced. They knew there were plans to 'open up' the Army to the Hutu and that the President was discreetly arranging for his own (Hutu) presidential guard to be formed. Government was functioning according to the principle of intwaro rusangi (shared power). But this really worked only at the highest government levels. The expectations of the Hutu electorate were too great and in order to try to satisfy them, all the lower and regional echelons of the administration were being solidly 'frodebu- ised', with uneven results Even at the highest level, clumsy errors were made. One of the biggest ones was the dismissal on 25 September of national radio and TV Director Louis-Marie Nindorera by Information Minister Jean-Marie Ngendahayo for 'systematic sabotage'. Nindorera had been chosen by the Minister himself only a few weeks before and his only mistake had been to try to practice an open and vigorously investigative form of information, which spared neither the new government nor the opposition. Years of media control had taken their toll on peoples' minds, even among former opponents, and by this action the new government gave the impression not only of not tolerating criticism, but of having something to hide This was most unfortunate since not only the issue of Army democratization but also the question of the return of refugees gave public debate a rather heated tone. In late August, the Commission des Réfugiés had admitted the principle according to which land illegally acquired during the last 20 years could be open to legal proceedings to ensure restitution to the rightful owners
Thus the political situation was tense, but not overly so. This kind of debate seemed unavoidable with the advent of such great and radical social change, and former President Buyoya's moderation coupled with President Ndadaye's obvious goodwill seemed to guarantee a basic framework of political reason in which the experiment had its chance of developing peacefully. This is why when on the evening of 20 October 1993 President Ndadaye was warned by his Defence Minister Lieutenant Colonel Charles Ntakije of the possibility of a coup during the night, he did not seem unduly worried. Major Isaie Nibizi, the man who had foiled the 3 July putsch attempt and who had been made head of Presidential Security, only took minimal precautions. But at midnight elements of the 11th Armoured Battalion came out of their barracks and moved towards the Presidential Palace. A major crisis had started.
3.2 The October 1993 Putsch 
When Major Nibizi ordered the Presidential Guards to take defensive positions in order to stop the mutineers from penetrating inside the Palace grounds, they obeyed him but, according to his words later, 'dragging their feet'. Outside the putsch was being carried out not as the type of technically clear-cut action which had brought to power Colonel Bagaza in 1976 or Major Buyoya in 1987 but in a confused and unclear manner. The elements from the 11th Armoured Battalion had been joined by a motley of troops from a variety of units (1st Infantry Battalion, some Gendarmes). All these were under the command of a low-ranking officer, Lieutenant Jean-Paul Kamana. Shots were fired in a desultory way, mostly in the air. The only two soldiers wounded during this action were tank drivers who hurt themselves when crashing their vehicles through the Palace gates At 1:30 a.m. Lieutenant Colonel Ntakija, who was not present, advised the President by telephone to climb into an Armoured Personnel Carrier (APC). The vehicle remained parked on the Palace grounds with the President inside. Finally, at 6:00 a.m., the APC driver was ordered by unknown officers to drive to Camp Muha 'where the President would be safe'. Army Chief of Staff Colonel Jean Bikomagu was present at Camp Muha. He talked briefly with the President as he emerged from the APC and told him everything would be all right. The putschists had by then arrested the President of the National Assembly, Pontien Karibwami and taken him to Camp Muha. Colonel Bikomagu then walked away from the camp, apparently without leaving any orders. At 10:00 a.m. Lieutenant Kamana ordered the murders of Ndadaye and Karibwami. They were bayoneted to death but not mutilated Meanwhile, rebellious soldiers had searched the capital, killing the Minister of Territorial Administration, Juvénal Ndayikeza, Gilles Bimazubute, National Assembly Vice-President, and Richard Ndikumwami, head of the Secret Service. They had also tried to kill the Foreign Minister, Sylvestre Ntibantunganya. But he was not home and in frustration the mutineers killed his wife and a female visitor who happened to be present.
The situation appeared extremely confused. The rebels had freed Lieutenant Colonel Sylvestre Ningaba, the leader of the abortive July putsch. But he himself seemed lost and unable to decide what to do. The main public buildings had been occupied and the telephone cut off. A Comité de Crise (Crisis Committee) had been created, presided over by an UPRONA Hutu civilian, François Ngeze, who had been President Buyoya's last Minister of the Interior. Ngeze had called on four senior UPRONA members to advise him: Libère Bararunyeretse, Charles Mukasi, Jean-Baptiste Manwangari and André Kadege; but this bizarre Committee's legal or political standing remained extremely vague. It had announced a reorganization of the Army general staff, but at the same time stated that Colonel Jean Bikomagu would keep his position as its head. The remnants of the government had taken refuge at the French Embassy. A shadow military committee had been formed in the meantime, with Colonel Bikomagu at its head, comprising Lieutenant Colonel Jean-Bosco Daradangwe and Lieutenant Colonel Pascal Simbanduku. This self-appointed committee had given itself the job of 'managing the crisis' and was operating separately from Ngeze's Committee. On 22 October, the governments of France, Germany, Belgium and the United States, as well as the European Union authorities in Brussels, announced the suspension of all economic aid to Burundi as long as the coup leaders remained in power. All political parties, the churches and the various civil associations condemned the coup. On 23 October, Colonel Bikomagu finally decided to do the same thing over Radio Burundi. The putsch immdiately collapsed. But mass killings of Tutsi had started in the hills as early as 21 October when the news of the President's death had become public. Two days later, when the Army moved to stop these killings, it immediately started its own indiscriminate killings of Hutu.
3.3 The October-November 1993 Massacres
The first violent acts appear to have been spontaneous and to have been triggered by the news of President Ndadaye's arrest and death. But quickly FRODEBU local cadres 'organized the resistance', an ambiguous term since in the first two to three days nobody attacked them. In fact, they organized the indiscriminate massacre of ordinary Tutsi peasants who were collectively scapegoated for the murder of the President. Pro-UPRONA Hutu were also massacred along with the Tutsi as they were considered 'accomplices' of the 'UPRONA coup'. In a minority of cases, local authorities did their best to protect the Tutsi citizens from the lynch mobs chasing them.
Within two or three days, Army units moved in to protect the Tutsi. They regrouped them in towns and ensured their security. But they went beyond that. They entered the areas where the massacres had taken place and which were by then empty of Tutsi. And they started a violent and indiscriminate repression of ordinary Hutu peasants, who in some cases were indeed guilty of murder but who were often innocent of the massacres which had been carried out by more politicised FRODEBU supporters Later, everybody would try to occupy the moral high ground. But it was obvious that the violence came from the conjunction of mutual bad faith refusing to play the democratic game. It has been well summed up by two Burundese journalists who wrote:
During the electoral campaign, Tutsi extremists kept repeating that a Hutu was not fit to rule over Burundi. And then, after the failed coup of 2-3 July, the Hutu extremists started to prepare on their side and to arm the population in case something would be done against the President
It was because of this twin and mutually reinforcing extremism that the violence of October-November 1993 could occur.
But the political dimension of this catastrophe was equally appalling. On the one hand, UPRONA and the Army gave the impression of being in an extremely ambiguous position. They did not openly condone the putsch but it took them three days before they finally declared themselves publicly against it. And their rallying to 'democratic legality' gave the impression not to have occurred because they were moved by a real deep-seated democratic commitment but rather because the coup had been terribly organized and because the international community lost no time in rallying against it. On the other hand, the Government gave a terrible example. It remained secluded in the French Embassy in a state of utter confusion and irresolution. While Health Minister Jean Minani, who was in Rwanda, kept making incendiary proclamations on Radio Kigali, calling for the formation of a government in exile and the development of 'popular resistance' (which was understood to mean wider killings of Tutsi), Prime Minister Sylvie Kinigi kept floundering helplessly without establishing any kind of clear leadership. It took her government over one month to start working again on anything like a normal basis.
4. ABORTIVE NORMALIZATION (1993-1994)
4.1 The Social and Cultural Constraints
The social and cultural aspect is often overlooked by the sources that concentrate on the violence perpetuated against the Tutsi minority The Hutu had been progressively marginalized within the sphere of power as early as 1961, following the death of Prince Louis Rwagasore. Contrary to what had happened in Rwanda, and for political and cultural reasons outlined in the first part of this paper, they had taken part in the independence movement on an equal footing with the Tutsi. They did not feel that there was a gap between the two communities and it was only with the murder of Pierre Ngendandumwe that the split became apparent.
The abolition of the monarchy had marked the end of the hopes for an integrated polity in Burundi. The military regimes of Presidents Micombero and Bagaza had institutionalized a most violent and hypocritical form of social discrimination which had been enforced in 1972 by torrents of blood. Given this historical context, it was to be expected that President Buyoya could only be partially successful in his honest and genuine attempt at liberalization Pent-up feelings of rage, resentment and injustice remained general in the Hutu populace. President Ndadaye's victory had been a remarkable symbol of peaceful compensation, a fact that Hutu radicals had immediately understood. But if PALIPEHUTU and other radical groups had been disenfranchised almost overnight, their continued marginalization depended on the peaceful unfolding of the new democratic experiment. President Ndadaye was not only a Head of State, he was an almost Christ-like figure who had come to symbolically release his people from bondage. This feeling was due to history and symbolic politics, but it was also reinforced by the deepening economic crisis into which Burundi, like the other countries of the area, was gradually sinking after world coffee prices had started a rapid decline in 1987-1988. This has been remarkably well understood by a Burundese (Tutsi) college professor, writing in April 1994:
A growing part of the peasantry gradually realized that, through the system of export cash crops, it was caught in a situation which completely blocked the way of any social and economic promotion for its children. In turn, these children realized that they could not escape from an agricultural economy whose remuneration steadily decreased. ...The State remained the only hope. ...For these poorly educated youths, these low-ranking civil servants and their peasant families, Ndadaye was more than a President. He was a King, a God, he was the only hope. ...One should remember these women who took off their dresses to spread the cloth on the ground for Him to walk on. When one thinks of the sexual modesty of our Burundese women!
Thus economic interests and political symbolism reinforced each other. President Ndadaye would have been bound to disappoint such enormous hopes. But he was not given the time.
On the other hand, the motivations of the murderous political dinosaurs who confusedly tried to reverse the verdict of the polls were also linked to the exploitation of a social fear. And there we should quote again the same remarkable analysis by Ndarishikanye:
The nominations of Hutu in the administration after July 1993, followed by the replacement of both Tutsi and Hutu UPRONA Civil Servants, down to such low levels as Communal Secretaries and marketplace watchmen...frightened a lot of people into thinking that they were going to lose not only the symbols of their hegemony but their permanent sources of monetary income and of familial patronage. The press magnified this feeling and the UPRONA party played on it. ...Demonstrations such as those of the students after the FRODEBU victory and the two later ones organized by UPRONA to protest against losses of employment can be seen in the perspective of this organized panic. This led some members of the Armed Forces to think that the whole of Burundese society was in a state of upheaval.
This crisis has partly been a crisis of identity, of habits, of culture. Life had functioned in Burundi for the last 25 years, well or badly, but according to a certain pattern. The election of Melchior Ndadaye and the restructuring of the administration first, then of patterns of land tenure, of job opportunities and finally of Army structures, represented a tremendous leap into the unknown. Everybody had lost their bearings, positive or negative. Familiar reactions and past patterns of behaviour simply did not seem to operate any more.
In a way, the paralysis of the cabinet following President Ndadaye's murder was in itself typical of this aspect of the crisis: the FRODEBU cabinet, quite literally, did not know what to do. The Hutu could have faced another 1972 (in fact, this is what they thought would happen, hence the rush into 'defensive' massacres while they were not threatened) but they could not understand the incoherence of the quasi-putsch-cum-murder. This was also the case on the Tutsi side. The would-be coup-makers did not seem to really believe themselves that they would be able to turn back the clock through their action. Hence its confused and indecisive character. Hence also the ambiguous attitude of the UPRONA and Army power structures which neither supported nor condemned the rebels.
All around, fear had become the dominant motivation. Fear of losing their prestige and even their livelihood in the case of the Tutsi minority, fear of being victims of another 1972-like ikiza on the part of the Hutu; then minority fears for the Tutsi, i.e. the fear of total physical annihilation, along the pattern of what was to happen later in Rwanda. It was largely this pyramid of meshing fears which caused the mutual massacres of October-November 1993.
4.2 The Military Problem
The failed putsch of 21 October 1993 was in itself a crisis of the Army as an institution. Half in and half out of the putsch, it cannot be considered a neutral institution any more. For the Hutu (and not only the Hutu radicals, but even the moderate FRODEBU cadres), the Army is a purely Tutsi entity, highly suspect because of its behaviour during the coup and guilty of having 'restored order' in a most bloody way. And for the Tutsi extremists, it is a dubious ally because of its desire to become again a politically neutral national body. Thus it is trusted by nobody while having the formidable task of trying to maintain law and order not only against the constant attacks of PALIPEHUTU extremist guerrillas operating with the backing of ex-Forces Armées Rwandaises (FAR) but also in spite of Tutsi extremist militias in Ngozi and Kirundo Provinces who are always keen to 'punish' supposed Hutu extremists.
On top of the problem of having to maintain law and order in rural provinces despite its tarnished image, the Burundi Army also faces a major problem of maintaining law and order in the capital itself with the quasi-insurrection of the Kamenge area of Bujumbura. Kamenge was practically a personal fiefdom of Interior Minister Leonard Nyangoma, an extremist member of FRODEBU66, who through his contact with the Police de l'Air et des Frontières (PAF--Air and Borders Police), headed by Festus Ntanyangu, a famous Hutu extremist, managed to arm his followers. There were constant skirmishes throughout February and March 1994, until the Army finally moved in on 27 April 1994, occupying at the same time the Cibitoke, Kinama and Mutakura areas of the capital. Fighting lasted until early May and hundreds of weapons were confiscated while several dozen people were killed In a way, this only lead to a displacement of the problem, many of the extremists just moving to the hills surrounding Bujumbura and keeping tenacious guerrilla warfare going directly on the outskirts of the city Even the disarming of Kamenge was only partial: between 12 and 16 September fighting flared up again in that section of town and the Army had difficulty re-establishing control.
4.3 The Political Problem
Social, cultural and military, the present Burundi crisis is also, and perhaps overwhelmingly so, a political crisis. Its components are relatively simple, but they combine with each other to create an extremely difficult situation:
1. A radical Hutu fringe which is determined to go beyond 'victory at the polls' and to wrest total political control from the Tutsi, preferably by massacring them. For them, the murder of President Ndadaye has been a godsend since it started the crisis of the moderate Hutu opposition and gave them a marketable cause. They are represented both by illegal organizations such as PALIPEHUTU or the militias and by radical elements within FRODEBU itself.
2. A radical Tutsi fringe which believes that power has to be wrested back from the Hutu, whether extremists or moderates, by force if needed. They are represented by the micro-parties such as RADDES, PRP or PARENA as well as by some of the UPRONA mainstream opposition.
3. A moderate opposition (UPRONA) which is always pushed to make unreasonable demands because of pressure from its extremist fringe, and because the Army tells it that unless tremendous pressure is put on the FRODEBU Government, the Army might have to stage a coup. Colonel Bikomagu and his men present themselves as 'moderates' who are pushed by younger more radical officers, like those who carried out the confused action of 21 October 1993.
4. A moderate FRODEBU Government which has had to give in time and time again to the demands of the Tutsi opposition because of the constant threat of another military coup, and which now begins to lack credibility with its own Hutu political base because it is seen as being feeble.
One could therefore say that the whole of the political game played for the last years has been a steady confrontation between the FRODEBU mainstream and the UPRONA mainstream, each one operating under the pressure of its own extremist fringe, but also using the threat represented by these extremist fringes to scare the other side into a better bargaining position.
The first episode of this multi-faceted confrontation was the opération ville morte (Operation Dead City) organized by the opposition in January 1994 when they forced Bujumbura and other major towns to a total standstill. It started as a 'protest' against the choice of Agriculture Minister Cyprien Ntaryamira as the new President on 5 January 1994 Then the confrontation grew when five (Tutsi) judges from the Supreme Court who had refused to accept the choice of the new President (in spite of a vote from the Assembly) were dismissed from their positions The opposition immediately launched another opération ville morte, with dire consequences. Rioting broke out leaving 12 people dead Eventually President Ntaryamira was confirmed in his post on 5 February 1994 after an agreement with the opposition which meant a small reduction in FRODEBU's power. Two days after being confirmed, President Ntaryamira chose as Prime Minister Anatole Kanyenkiko, a moderate UPRONA Tutsi with a Hutu mother and married to a Rwandese Hutu In the Kanyenkiko cabinet inaugurated on 10 February, two-fifths of the ministries went to opposition members. This did not prevent Joseph Nzeyimana, President of the RADDES Tutsi extremist micro-party74, from protesting against the 'lack of concertation' of FRODEBU for its ministerial choices, and to threaten the new cabinet with further opérations ville morte. Tutsi monarchist extremist Mathias Hitimana simply accused the government of 'treason' and asked for the resignation of the Kanyenkiko cabinet.
In retaliation, Justice Minister Fulgence Dwima Bakana, a Hutu hardliner, ordered the release from prison of André Baryimare, one of the organizers of Tutsi massacres in Ryansoro commune (Province of Gitega). He also invited the Director of the Kibimba Secondary School, notorious for having burnt alive his Tutsi pupils, to the inauguration of President Ntaryamira
The FRODEBU majority and the opposition were inextricably linked in government. They were most of the time without a shred of good faith or genuine desire to collaborate in solving problems. Every event, every occasion was seen only as an opportunity to accuse the other side of various evils. And most of these accusations were indeed true since plenty of evil had been committed and plenty kept on being committed.
Interior Minister Léonard Nyangoma armed the Hutu bastions of the capital and encouraged them to insurgency The Army used its role as keeper of law and order to try breaking the back of opposition militias and to help the extremist Tutsi gangs In March 1994 alone, 30,000 people fled Bujumbura to escape the street fighting between the various militias But in spite of this constant violence, events did not quite go to the bitter end. When three officers and 50 paratroopers attempted a coup on 24 April, they did not manage to get any sizeable Army unit to follow them The UN Special Representative in Burundi, Ahmedou Ould Abdallah, a rather outspoken and very courageous diplomat, expressed best the exhaustion felt by a number of observers at this perpetual game of brinkmanship, where the (relatively) safe politicians kept playing games while the population suffered, when he said:
The extremist elements do not want any solution. They play for time, one does not know what for.... I did not see any goodwill in June when the Government--the largest party--was dragging its feet, and I did not see it in July either when it was the turn of certain fractions of the opposition to drag their feet. Currently [late July 1994] there is a deadlock...this is childish behaviour, when the population is in such a desperate situation. The Security Forces are tired. Since October  they have been trying to hold the floodgates; they have had to provide security in the country and ensure security at the border ; and in the meantime all the politicians on all sides are just sitting around in Hotel Novotel, talking.
And yet, this evil, dangerous and bloody political face-off slowly moved towards some kind of a solution, perhaps partial and temporary, but a solution anyway. By mid-August, Charles Mukasi, the new Secretary-General of UPRONA, was coolly asking for three-fifths of government portfolios (while his party had received 32 per cent of the vote in the Presidential election and 22 per cent in the legislative one), his argument being that 'FRODEBU has by now showed proof of its incompetence and technical incapacity to manage the crisis' Haggling went on for another month, and on 12 September a power-sharing agreement which for the first time looked serious was signed between the government and the opposition Fifty-four precise and carefully-worded articles detailed the workings of the prolonged crisis government, defined what was and what was not acceptable, and outlined peace-keeping mechanisms. Everything gave the impression of having been thought out and of being considered realistically. There were no abstract principles. Institutionalized defiance was the keyword. The whole document looked as if its draftsmen realized that 'peace was not around the corner', that they were going to have to live with the crisis and that the best way to survive it was not to deny its existence but, on the contrary, to design 'permanent crisis mechanisms'. The contending parties, in a sense, 'agreed to disagree' and to go on living in that state of tense but perhaps less violent confrontation.
On 30 September, Provisional President Sylvestre Ntibantunganya was 'elected' by the Assembly He confirmed Prime Minister Kanyenkiko in his position (3 October) and a new cabinet was assembled by 5 October. Everything had been quick and relatively trouble-free. Extremists gave a kind of reverse proof that the agreement was serious this time when they tried, without too much success, to cause an uprising in Kamenge in the hope of breaking it up
5. THE RISE OF EXTREMISM (1994-1995)
5.1 The Slow Erosion of the September 1994 Agreement
The September 1994 power-sharing agreement had been a necessary evil. But it was also a bulwark against extremism. It soon became apparent that neither of the two extremist sides wanted to accept it. On the Hutu extremist side, it was made clear in a letter written by Leonard Nyangoma, former Minister of the Interior and organizer of the Conseil National pour la Défense de la Démocratie (CNND--National Council for the Defence of Democracy) In that letter, addressed to the Zairian President Mobutu Sese Seko, the former Minister of the Interior not only reiterated the CNDD opposition to the agreement but also denounced the creation of the Conseil National de Securité (CNS--National Security Council) as an unconstitutional body which had been made 'in reality the supreme institution of the Nation' and which was nothing but a tool for the opposition to dominate the political process
One must admit that the CNS, which had been conceived of as a form of emergency 'super government', seemed indeed, because of its composition, biased in favour of the opposition. Created on 10 October 1994, it was presided over by Interim President Sylvestre Ntibantunganya and its Vice-President was Premier Anatole Kanyenkiko. Its seven other members were Foreign Affairs Minister Jean-Marie Ngendahayo, Interior Minister Jean-Baptiste Manwangari, Defence Minister Colonel Firmin Sinzohiyeba, Pierre Barusasiyeko who sat for the small parties allied with FRODEBU in the majority, Lieutenant Colonel Jérome Sinduhije who represented the small parties allied with UPRONA in the opposition, Zacharie Bukuru who was supposed to represent 'Civil Society' and Sylvestre Ntambutso who represented the Civil Service. The problem was that only three of the nine (President Ntibantunganya, Pierre Barusasiyeko and J.-M. Ngendahayo) were members of the majority, while the six others either were members of or sympathised with the opposition. And since the CNS had extraordinary powers (capacity to modify the constitution, capacity to override the Assembly's decisions, capacity to veto war-and-peace decisions made by the President), the role of the opposition in that body could prove vital.
The problem was that Nyangoma and his allies had decided to 'correct' that institutional imbalance by direct armed actions88, and these armed actions tended not to confine themselves to members of the Burundi Army but to target civilian Burundese Tutsi as well. In early October, they had massacred a number of Tutsi civilians in the Tangara commune of Ngozi Province During a bout of intense fighting with the Army between 15 and 20 October in Cibitoke Province around 30 people had been killed and the subsequent Army repression had caused 15,000 Hutu refugees to flee to Zaire Later that month armed Hutu gangs began operating on the Bujumbura-Bugarama road, stopping cars and buses, robbing travellers and killing the Tutsi This of course caused violent reactions on the Tutsi side. On 24 October, 50 Hutu Rwandese refugees were killed in Gitwa Camp (Ngozi Province) by the Burundese Army who considered them as accomplices of the FDD raiders92, and on 3 November a commando of 'displaced' Tutsi killed about 30 Hutu in Cyanzu commune (Ngozi Province) But if the Hutu extremists had obviously no intention of abiding by the September 1994 power-sharing agreement (they denounced President Ntibantunganya as a 'traitor'), the Tutsi extremists had no desire either to respect the document although, unlike the FDD group, they had in fact signed it They would 'respect' it as long as they could play by their own rules and violate its spirit, if not its letter, when they felt events were not going their way
The month of November was marked by a seemingly endless series of reciprocal murders and massacres. Emile Ntanyangu, a FRODEBU militant who was Director of Planning at the Ministry of Planning and Development, was shot in his hospital bed on 10 November, together with his son who had come to attend him. The same day Fridolin Hatungimana, former Secretary of State for cooperation and head of the Buyoya campaign during the 1993 presidential election was shot dead at his home Over 16,000 Hutu, both Burundese and Rwandese refugees in Burundi, fled to the Ngara District of Tanzania after joint operations by Rwandese Patriotic Front and Burundese troops resulted in massacres in the North On 20 and 21 November Tutsi militiamen killed 42 Hutu in Muruta commune (Kayanza Province) and Vumbi and Bwambarangwe communes (Kirundo Province). The next day, FDD commandoes attacked the Army in Kirundo and Ngozi Provinces and the Army retaliated by killing a number of local civilians A week later, somebody threw a grenade in broad daylight at the Central Market in Bujumbura, killing five and wounding 20. The act was not claimed by any group It is against this background of continuing violence that one has to look at the election of Jean Minani to the position of Speaker of the Assembly.
5.2 The Minani Crisis
Jean Minani was a controversial personality. As Health Minister in the Kinigi cabinet of 1993, he happened to be in Kigali at the time of the October putsch. He had made some rash statements on the Radio des Mille Collines, calling for 'popular resistance' at a time when FRODEBU militants and sympathisers were systematically massacring Tutsi, and proclaiming himself head of a 'provisional government in exile'. He eventually had come back to Burundi after the putsch had collapsed and some semblance of order had been restored. He was somewhat impetuous, but not more so than a number of Tutsi extremists who were sitting in the cabinet and in the CNS. But the opposition chose to focus on his election to the position of Assembly Speaker as a casus belli. In fact, the probable reasons for choosing a showdown were two:
1. Minani was a tough operator who was not likely, as Assembly Speaker, to accept the backseat role given to the Parliament since the creation of the CNS.
2. He wanted to set up a serious Commission of Inquiry into the October 1993 putsch
Immediately after Minani had been elected Speaker, the UPRONA MPs walked out of Parliament. Prime Minister Kanyenkiko who was away in Europe cut short his trip and came home to an emergency CNS meeting It soon became apparent that the UPRONA hardliners such as the Party's Secretary-General Charles Mukasi or former putschist François Ngeze wanted Prime Minister Kanyenkiko to resign so as to bring the cabinet down, enabling them to impose a new one where the opposition would have a clear dominance. But Kanyenkiko, a moderate UPRONA Tutsi, refused to resign, preferring to try to negotiate the Minani issue rather than use it for confrontational politics. That meant directly challenging his own party's extremist group The small Tutsi extremist groups (PARENA, ABASA, PIT, PSD) who wielded considerable influence in the capital, especially in the Civil Service and among the students, were organising demonstrations and the Tutsi gangs had gone on the rampage This was, of course, a blessing for the Hutu extremists of CNDD whose support grew among previously moderate FRODEBU sympathisers, who began to realize that the Tutsi extremists had no intention of abiding by the September agreement, no matter how much was given to them Foreign governments began to fear a repeat of the Rwanda catastrophe. On 17 December U.S. National Security Adviser Anthony Lake arrived in Bujumbura to denounce extremism105, while French Minister for Cooperation Bernard Debré came two days later for talks with the government Their presence did not deter the extremists: in three days (18 to 21 December) more than 50 people were killed in street clashes in the capital On 21 December Prime Minister Kanyenkiko had to proclaim a curfew. Six days later, President Ntibantunganya announced a 'solution' to the crisis: Minani would step down and be given another position The Tutsi extremists had achieved another success.
However, this was definitely a Pyrrhic victory. Given the propensity of the two extremes to feed on each other, Hutu extremists also came out of the crisis with added strength. The Kamenge area of Bujumbura, under the control of militia leader Pascal Gashirahamwe, popularly known as 'Savimbi', had become practically 'off limits' for the Army and Police109, and FDD forces were now launching increasingly daring attacks closer to the city centre. During the night of 22 to 23 January, they attacked Mutanga Military Hospital, causing considerable damage and killing three people
The UPRONA hardliners were not satisfied with Minani's resignation. They now wanted their revenge on Prime Minister Kanyenkiko for not agreeing to collaborate with them in late December. On 29 January they managed to engineer his expulsion from the party. The next day when the Prime Minister called a cabinet meeting, two ministers, Libère Bararunyeretse, Minister for Resettlement and former UPRONA Secretary-General in the single-party days, and Claudine Matuturu, Minister for the Civil Service, refused to attend. The reason given was that since the government pact which followed the failed October 1993 putsch required the Prime Minister to be UPRONA, and since Prime Minister Kanyenkiko had just been expelled from the party, he, in fact, was no longer Prime Minister. Kanyenkiko responded by dismissing the two from their ministerial posts for insubordination In effect the Minani crisis had rebounded and become a Kanyenkiko crisis.
On 31 January, UPRONA Secretary-General Charles Mukasi called a public meeting to ask for the Prime Minister's resignation. Regardless of anybody's merits or 'legal' standing, in the tense situation then prevailing in Burundi this was an irresponsible form of brinkmanship, and elicited a warning from the U.S. Department of State which declared on 1 February 1995, 'Those who seek to overthrow a democratically elected government in order to replace it with a minority group are leading the Burundese people on a dangerous path'. In New York, UN Secretary General Boutros Boutros-Ghali issued a similar statement On 1 February 1, Prime Minister Kanyenkiko declared that he would not step down and UPRONA immediately launched a general strike, while the extremists resorted to grenade attacks in various parts of the capital Former Prime Minister Adrien Sibomana, a moderate UPRONA member who as Premier had been a force for stability during the late Buyoya period, criticised Mukasi for deliberately creating a crisis situation by first promising Minani that he would not object to his election, then turning against him after he had been elected. Kanyenkiko's resignation would be no solution to the crisis On 7 February, President Ntibantunganya backed down once more and declared that he would seek a solution to the crisis by accepting Kanyenkiko's resignation. This further accelerated the descent into instability, since by giving in to the demands of the Tutsi extremists the President only played into their hands as well as into those of the Hutu extremists who accused him of being weak and irresolute. In Paris, former UPRONA moderate Prime Minister Sylvie Kinigi issued a statement critical of her own party's extremists
This vacillation at the centre encouraged all armed groups to launch fresh actions in various parts of the country. On 9 February 12 people were killed in Rumonge commune (Bururi Province); the following day over 30 were killed in clashes in Karuzi Province in the East and in Buruhikiro in the South, and on 12 February Hutu guerrillas attacked a military outpost in Giteranyi commune (Muyinga Province) in the North-East In a most alarming development, armed clashes which until then had mostly occurred in the northern provinces were now spreading to the rest of the country. The general strike was by now absolute in Bujumbura where Charles Mukasi and his ally, the extremist PRP leader Mathias Hitimana, coolly declared that their action was in the spirit of the September 1994 power-sharing agreement because 'the opposition parties have withdrawn their support for Anatole Kanyenkiko'
Abandoned by President Ntibantunganya, attacked by his own party, powerless to act independently, Prime Minister Kanyenkiko had to resign on 15 February, thus giving another boost to all extremists His parting declaration on Radio Burundi had a tragic ring:
Today, a handful of politicians from all sides are refusing to draw lessons from the Rwandese tragedy which happened on our doorstep. They want to lead our country towards a similar tragedy. ...What has been referred to as 'the Minani crisis' and what is today referred to as 'the Premiership crisis'...are simple pretexts
5.3 On the Brink of Catastrophe
The selection of a new Prime Minister followed a disturbing pattern. UPRONA wanted to nominate Aster Girukwigomba, but the seven small parties which made up the rest of the opposition imposed their own candidate, Antoine Nduwayo, a former Bagaza appointee to the Executive Secretariat of the Communauté Economique des Pays des Grands Lacs (CEPGL-- the EU-sponsored Great Lakes Area Economic Cooperation Organization) Whether Antoine Nduwayo was a better or worse candidate than his rival is irrelevant. What mattered was that seven mini-parties, which together had polled 2.68 per cent of the popular vote in the legislative election of June 1993, were allowed to pick a Prime Minister of their own choice. Although the opposition parties were not uniformly made up of extremists, the extremists dominated the group, and the process clearly exemplifies the degree to which the mainstream political parties (and mainstream political opinion) had become their hostages.
These developments led the CNDD forces to renew their attacks on the government with an added sense of righteousness, as for instance CNDD Spokesman Jerome Ndiyo the day after Nduwayo had been named by the opposition as the new Prime Minister:
We have run out of all peaceful means. ...The international community has done absolutely nothing to protect the civilians who are massacred regularly. ...It is a self-defence strategy. ...We are fighting against the September 1994 Agreement which sets itself above the Constitution and we are fighting against some figures from within our party who have strayed and given moral backing to the signing of the Agreement .
Fighting flared up all over the country: violent repression by the Army in Muyinga Province led to the flight of 25,000 Hutu refugees to Tanzania ; armed clashes in the Kamenge suburb of Bujumbura involving the use of a tank on the Army side left six killed and two wounded; an ambush in Ruhororo commune (Ngozi Province) caused the Army to retaliate against the civilians, killing 30 The only international presence on this increasingly violent scene were 38 OAU observers who, as an Italian journalist commented 'only played a decorative role'
On 1 March Prime Minister Nduwayo introduced his new cabinet which, given the circumstances surrounding its creation, was surprisingly moderate. Apart from a number of rather technical permutations between several Ministries (Lands, Communal Development, Agriculture) the main politically significant changes went in the direction of moderation. If Claudine Matuturu was reintegrated in the cabinet (as Minister for Resettlement), the former holder of the Ministry, extremist Libère Bararunyeretse, was out. So was fellow extremist Jean-Baptiste Manwangari who was replaced by Gabriel Sinarinzi as Minister of the Interior. The Ministry of Commerce was left without an appointee after Tutsi extremist Joseph Nzeyimana had been dismissed. The Prime Minister was aware that his cabinet might not please UPRONA's hardliners (and even more the fringe extremists from the seven allied small parties) when he declared, 'It is possible that many people will be disappointed by the new government because most of the members of the former team are still there'
But the extremists had no intention of adopting a low profile. On 11 March Ernest Kabushemeye, Minister of Land and Mines and head of the small Rassemblement du Peuple Burundais (RPB), a political ally of FRODEBU within the government coalition, was shot dead. The murder was in fact quite a complex affair since it obviously had several motives. As Minister of Mines, Kabushemeye had been involved in the so-called 'Gold Wars' which had been pitting against each other the three precious metal companies operating in Burundi, AFIMET on the one hand and IMAG and Ashons Gemstones on the other. AFIMET had won and its rivals had seen their trading licences withdrawn by the government. All of them were dealing in smuggled Zairian gold, in quantities which far exceeded the small Burundese production. But it was rumoured that Kabushemeye, a Hutu hardliner, had links with Leonard Nyangoma and that part of the AFIMET money was in fact going to finance the CNDD. This was known to the Tutsi extremists who consequently supported IMAG and Ashons Gemstones. With Kabushemeye's murder, they therefore disposed of both a political opponent and a commercial rival In retaliation, Hutu extremists kidnapped Colonel Lucien Sakubu, a former Mayor of Bujumbura, a few days later. His body was found the next day, crucified, eviscerated and with his eyes gouged out
6. CONCLUSION: POINTERS TOWARDS DESCENT INTO CHAOS
6.1 The Deleterious Role of the 'Elites'
Internal and external observers agree on the perverse role played by the country's political 'elite'. The President himself recently declared, 'The politicians have to achieve peace among themselves because the problems originate directly from them' Some observers can be even more pointed, including usually politically cautious UN and diplomatic personnel. UN Special Representative Ahmedou Ould Abdallah recently stated that 'there are only 75 to 100 people who are all the time inciting to violence', while one of his diplomatic colleagues said that 'Burundi today is like Chicago in 1930. A handful of people gunning for power and money are terrorising a whole population' A recent UN mission which had warned of the 'explosive' situation in Burundi was quoted as combining its warning with a suggestion that the UN should study the possibility of 'sanctions against these people', such as freezing their financial assets abroad and refusing them visas for foreign travel The Burundese political 'elite' is criminally blind and monstrously selfish. Closeted in its comfortable villas, protected by squads of bodyguards, many of its members are waging a relentless ethnic war either to keep their privileges (in the case of the Tutsi) or to eliminate the rival elite in order to replace it (in the case of the Hutu). The horror is that they wage their struggles by proxy, using simple-minded peasants in the countryside and rootless youths from the urban slums in the capital. These can at times realize that they have been manipulated and leave their gangs, but they mostly blindly follow their 'leaders' into a vortex of violence and death
In order to keep a proper perspective, it is necessary to remember that at the elections in June 1993, all the extremist parties together, from either side, did not reach 10 per cent of the popular vote. Admittedly, there were clandestine supporters of extremist policies inside both UPRONA and FRODEBU. But one can reasonably argue that they constituted small minorities in their respective parties. Today, as the violence grows, these extremist groups are growing ever stronger as a culture of fear takes root.
One of the most disturbing aspects of the situation is that the worst excesses and the most violent extremism come not from the ordinary people (the Church can report many examples of silent popular refusal to enter the deadly spiral of ethnic violence ), but from the most educated strata of society. Why should this be the case? There seem to be three reasons:
1. The elite has been constructing a theory of racism through often superficial reading of the European classics of colonial history and anthropology. The ethnic myths which are so strong in this part of Africa have mostly been digested, reorganized, vulgarized and propagandized by members of the elite. They are often convinced, just as their early twentieth century European counterparts, of the 'scientific' validity of their racial prejudices. This makes them more coherent, more radical and more relentless in their ideological views
2. The elite stands to gain the most from either a continuation of the present hegemony (if they are Tutsi) or an ethnic revolution (if they are Hutu). In either case, ordinary people are likely to remain spectators of the political and economic show their superiors will stage-manage. Hence the greater determination of the elites in the ethnic struggle.
3. The elite is the social group most threatened by ethnic violence. Doctors, teachers, lawyers, Army officers, politicians and even students are the prime targets of ethnic killings, the idea being either to 'make an example' or, in the case of massive killings, to deprive the other camp of its most able members. Therefore, part of the violence is defensive, based on a strategy of pre-emptive attack The cynicism of the active extremist politicians is almost total. Both the CNDD (clandestinely) and former dictator Jean-Baptiste Bagaza (openly) discuss genocide. They of course attribute the genocidal will to the other camp, but the 'defensive' measures they advocate constitute in fact a preparation for counter-genocide.
6.2 The Deteriorating Economic Situation
Part of the sharpening of the ethnic confrontation comes from the fact that the economic base of the country--already poor before the cycle of violence restarted in October 1993--has deteriorated further during the last 18 months Last year the disruption of the agricultural process caused by civil violence resulted in an estimated 200,000 tonnes shortfall in foodstuff production, and Burundi (as well as Rwanda) is now threatened with food shortages that could verge on famine In such a situation, both rich and poor feel increasingly in need of political control or protection. The rich want to monopolise whatever resources the State can still offer (among others, the control of foreign aid) and the poor feel that in a time of shortages, they will need political 'pull' in order to gain access to aid if their own crops are not sufficient.
6.3 A Culture of Violence
All the various aspects of violence have now become an ordinary accepted fact of life in Burundi, thus tending to perpetuate themselves since they have evolved into a coherent whole forming a distinct new subculture. Guns are freely available because the Army has been arming the Tutsi militias while the fleeing Rwandese interahamwe and FAR have armed the Hutu radicals. More guns are constantly being bought by both sides on a free international market where, since the disintegration of the former Soviet Union, the price asked for ordinary light weapons has become very low.
Violent groups have gained a form of social acceptability. They have names with almost neo-tribal connotations such as intagoheka (those who never sleep) for the FDD militants, inziraguhemuka (those who are not traitors) for the FRODEBU activists or imbogaraburundi (those who will bring Burundi back up) for the partisans of former dictator Jean-Baptiste Bagaza. Militias of all kinds proliferate, mostly made up of desperately poor urban youths who operate at the street gang level: on the Tutsi side the Sans échec (those who never fail), the Sans défaite (the undefeated), and the Sans pitié (the pitiless ones). On the Hutu side, the JEDEBU (Jeunesse démocratique du Burundi--Democratic Youth of Burundi), the Jeunes 72 (Youths of '72--an allusion to the 1972 massacre of the Hutu by the Tutsi Army), the FDD, and even the Chicago Bulls in Kamenge. Most of these gangs were bi-ethnic before October 1993 and operated as minor delinquents. After the failed putsch and the following ethnic radicalization, they began receiving weapons and money from political 'sponsors' and split along ethnic lines. Today they subsist mostly on contract killings and looting. They recognise some vague political allegiances (the Sans échec are supposed to support former dictator Jean-Baptiste Bagaza while the Sans défaite are rumoured to follow PRP extremist leader Mathias Hitimana), but they are totally undisciplined and very hard to control, even for their political patrons. Their wild actions have added a dimension of near anarchy to the political violence.
Since the October-November 1993 reciprocal massacres, the ethnic pattern of the population has changed radically. The Tutsi have been regrouped in various locations (trading centres, public buildings) where they live under the protection of the Army, while most of the hills are now entirely populated by Hutu. This has had several consequences. First of all, the rural economy is in complete disarray. The traditionally cattle-owning Tutsi have not been able to take their herds with them. The cows have been randomly slaughtered and eaten. There is now a lack of meat and milk products. The Hutu peasants are constantly on guard and many, in areas of severe Army repression, have taken to hiding during the day in the low-lying swampy areas, coming out at night to cultivate the crops. Trading is extremely difficult since the Hutu fear being killed at Army roadblocks when entering the towns, while the Tutsi fear being killed if they venture unescorted into the countryside. Vehicles are frequently attacked, more or less at random, such as in a recent incident when 17 people (including three Belgians) were machine-gunned to death on a road near Bujumbura Even the capital is now ethnically divided, with some areas such as Kamenge, Cibitoke, Kinama or Gasenyi purely Hutu while some others such as Musaga or Ngarara are purely Tutsi. Those who tried to remain in the 'wrong' areas of their previously integrated neighbourhoods have been killed.
Another fundamental problem in this 'culture of violence' now bred in Burundi is the role of the media. Paradoxically, the detrimental role of the media is a product of the democratization process. The press in Burundi is free and its freedom is pushed to the point of irresponsibility and to calls for violence. Le Carrefour des idées, Le Patriote, L'Etoile and La Balance in the Tutsi camp, L'Eclaireur, Le Témoin and Le Miroir on the Hutu side are violent and politically and ethnically slanted. They contribute greatly to political destabilization Just as an example, Le Carrefour des idées offered a prize of one million Burundese Francs to whomever would bring them the severed head of either Léonard Nyangoma or Festus Ntanyungu. The offer was splashed across the front page No legal action was taken against the paper for fear of upsetting the already fragile political situation. The press feels no responsibility towards such abstract concepts as truth, information or even reporting. The papers are made up of long diatribes interspersed with lies and slanted news items. There is no distinction between what passes for reporting and 'editorial material'. Assertions, bare statements and prejudices are presented as facts and personal slander is common. Although some papers struggle valiantly to maintain some objectivity (one can mention Le Renouveau and Ubumwe published by the government, Ndongozi published by the Catholic Church and the independent La Semaine) the extremist papers play a particularly poisonous role in that they are more 'fun' to read and that their violent and slanted style appeals more to the crudest prejudices.
6.4 A Fundamental Institutional Anomaly
Burundi political society suffers from a fundamental institutional anomaly or, as Professor Reyntjens calls it, 'a creeping coup' Basically, the problem lies with the refusal of certain sectors of the Tutsi to accept that a Hutu could be elected President of the Republic. Hence the demonstrations denouncing the perfectly free and fair election as 'an ethnic referendum', the strikes and eventually the assassination of the President. Feeling 're-legitimised' by the ethnic violence directed at innocent Tutsi peasants in the aftermath of the President's murder, these same sectors of the Tutsi community decided to achieve by gradual erosion what the putsch had failed to accomplish immediately. They used a number of approaches:
1. De-legitimising FRODEBU by accusing it of incapacity or of 'having planned a genocide' in October 1993.
2. Paralyzing the Presidency and the National Assembly through the Constitutional Court.
3. Using street violence as reinforcement of political manoeuvring.
4. Re-creating a new quasi-constitutional order aiming to achieve the objectives of the coup (i.e. removing power from the elected FRODEBU members to give it back to the unelected UPRONA or other Tutsi politicians) but without the stigma of an overt armed putsch.
From the beginning, that is, during the period when the country was painfully trying to recuperate from the effects of the failed putsch, it became obvious that even though the government parties were ready (without admitting it) to compromise on the one-man-one-vote principle in order to assuage the communal fear of the Tutsi minority, mere 'accommodation' would never be enough. The process of 'creeping reconquest' would not stop before overturning the results of the June 1993 election. One example is that of the Negotiation Forum, created on 16 August 1994, which was made up of 33 representatives of eight opposition parties (who had polled a grand total of slightly more than 24 per cent of the vote) facing 16 delegates from the four government parties (who had polled 76 per cent of the vote). In spite of this very generous over-representation, the opposition never stopped arguing that it was being discriminated against
The violent confrontations leading to the various 'agreements' culminating in the September 1994 power-sharing agreement, the creation of a politically unbalanced CNS, the Minani crisis and the Kanyenkiko crisis are all examples of this relentless institutional guerrilla war. The aim of the whole operation appears to be to increase the instability (while at the same time regaining maximum control over various institutional sectors) in order to push the Army into a coup. This strategy, although effective in the short term, has had a most harmful overall effect:
1. The FRODEBU moderates, constantly forced to bow to the will of the minority, are seeing their credibility slowly but surely eroded. In parallel, as the moderates go down in Hutu public opinion, extremists such as Léonard Nyangoma are gaining greater and greater support.
2. The UPRONA moderates have been forced into aquiescence. UPRONA is now becoming if not a party of extremists, at least an extremist-led party. Simultaneously, former dictator Jean-Baptiste Bagaza is slowly siphoning off UPRONA's militants because he offers more open acceptance of strong arm tactics than the still semi-democratic former single party.
3. The Army is in serious crisis. Defence Minister Sinzohiyeba is a legalist who is desperately trying to contain his lower ranks. Chief of Staff Colonel Jean Bikomagu is irresolute. High-ranking officers are conscious of the international problems a putsch would pose and they want to remain within the law if at all possible. But on the other hand, they have had enough of trying to keep law and order while being sniped at by guerrillas, ridiculed by armed extremists of their own camp, hated by the peasants, criticised by the international media and subject to parliamentary control. Younger officers and NCOs would welcome a coup and a 'good cleanup', a new 1972-type massacre. Strangely enough, part of the pressure comes from the Hutu rank-and-file soldiers who have been recruited since 1987 and whose families were wiped out by FRODEBU extremists in October 1993 for being 'traitors'. Although Hutu, they thirst for revenge and would follow orders in the event of a coup.
It is hard to say at this point in time and while the crisis is still ongoing what will finally happen. Burundi is now situated as Rwanda was between the signature of the Arusha Agreements in August 1993 and the start of the genocide in April 1994. The legal government is paralyzed, extremist groups hold the country virtually hostage, violence is rampant and the media publish incitement to murder. Failing international intervention through preventive diplomacy of a rather vigorous and unconventional kind, catastrophe looks inevitable and the question is only one of scale. Given the degree of disorganization of the country and the capacity for resistance of the Hutu, the number of victims in case of an explosion is likely to be substantially lower than in Rwanda, but could still reach 100,000 or 150,000. Such an outcome would put a heavy burden of responsibility on the international community.
 1850 is the approximate date of the beginning of the reign of Mwami Mwezi Gisabo, the last ruler of independent Burundi. 1966 is the date of abolition of the monarchy.
 The best comparative treatment of what it means to be Tutsi or Hutu in the two countries can be found in Jean-Pierre Chrétien, "Hutu et Tutsi au Rwanda et au Burundi" in Jean-Loup Amselle and Elikia M'Bokolo (eds.), Au coeur de l'ethnie, (Paris: La Découverte, 1985), pp. 129-165. For a very detailed analysis of Burundi pre-colonial society see Emile Mworoha (ed.), Histoire du Burundi des origines à la fin du XIXème siècle, (Paris: Hatier, 1987), especially chapter 10 for a description of the social and political order on the eve of colonization.
 Bantu languages are divided into word classes, distinguished through their varying prefixes. Living creatures are prefixed m- in the singular and ba-, wa- or aba- in the plural. Hence the correct expression for 'Tutsi' would be mututsi in the singular and abatutsi in the plural. For the sake of easy comprehension we will nevertheless keep using the grammatically incorrect but more easily identifiable forms 'Tutsi' and 'Hutu' (correctly muhutu and abahutu).
 Their name came from the small stick they carried and which they planted in the ground before them before speaking, causing everybody to fall silent.
 See on this subject the remarks by one of the first German travellers who entered the kingdom: Hans Meyer, Die Barundi, (Leipzig: Otto Spamer Verlag, 1916).
 The most complete analysis of Belgian colonial policies in Burundi can be found in Joseph Gahama, Le Burundi sous administration belge, (Paris: Karthala, 1983).
 See the remarks on ubugabire by Emile Mworoha in his Peuples et rois de l'Afrique des Lacs, (Dakar: Les Nouvelles Editions Africaines, 1977), pp. 187-190.
 Jean-Paul Harroy, Burundi (1955-1962), (Brussels: Hayez, 1987), p. 399
 From that point of view, the two subtitles given by former Resident Jean-Paul Harroy to his two-part memoirs on the decolonization of Rwanda and Burundi are remarkably candid. The volume on Rwanda is triumphantly subtitled: "Memories of a companion of Rwanda's march towards democracy and independence" while the volume on Burundi is sadly subtitled: "Memories of a fighter in a lost war". Jean-Paul Harroy, Burundi (1955-1962), (Brussels: Hayez, 1987) and Rwanda (1955-1962), (Brussels: Hayez, 1984).
 For a good account of the complicated murder case, see Jean-Paul Harroy, Burundi, pp. 576-593.
 M. Manirakiza, La fin de la monarchie burundaise (1962-1966), (Brussels: Le Mat de Misaine, 1990), p. 43
 The six non-UPRONA MPs had not chosen to create an opposition group but rather to join either 'Monrovia' or 'Casablanca' as they were popularly known. The style of the rivalry between the two groups (there was no ethnic connotation) was reminiscent of the rivalry between famous opposing soccer teams rather than between political tendencies.
 For a description of these events see M. Manirakiza, op. cit., pp. 54-78 and René Lemarchand, Rwanda and Burundi, (New York: Praeger, 1970), chapters XIII and XIV.
 Simbananiye is the man who was later to acquire a most sinister reputation after he became credited with having drafted a plan for general genocide of the Hutu. The 'Simbananiye plan', which nobody has ever seen but in which many people in Burundi believe almost as an article of faith, was a major element in sparking off the 1972 massacres.
 Commander Martin Ndayahoze, Radio Broadcast on Radio Burundi, 25 November 1968. Quoted in Jean-Pierre Chrétien, "Les massacres de 1972", in Burundi: l'Histoire retrouvée, (Paris: Karthala, 1993), p. 431.
 On this crisis see M. Manirakiza, Burundi: de la révolution au régionalisme (1966-1976), (Brussels: Le Mat de Misaine, 1992), pp. 47-108
 Their main leaders were the Minister of Information, André Yanda, the Minister of the Interior, Albert Shibura, and especially the Minister for Foreign Affairs, Arthémon Simbananiye.
 Déo Hakizimana, Burundi: le non-dit, (Vernier: Editions Remesha, ), pp. 21-24
 For an evaluation of the 1972 massacre, see R. Lemarchand and D. Martin, Génocide séléctif au Burundi, (London: Minority Rights Group, 1974) and B.F. Kiraranganya, La vérité sur le Burundi, (Sherbrooke: Editions Naaman, 1977), pp. 76-81.
 Convinced of the reality of the ideological propaganda which since the colonial days had presented the Tutsi as 'Egyptians' or 'Ethiopians', he went to finish his life in exile in Somalia, 'to be near his ancestors'.
 See Filip Reyntjens, L'Afrique des Grands Lacs en crise, (Paris: Karthala, 1994), pp. 39-40.
 Jean-Pierre Pabanel, "Statistiques tribales au Burundi en 1986", Politique Africaine, No. 43 (December 1988), pp. 111-115
 Déo Hakizimana, Burundi: le non-dit, p. 41
 C. Carral, "Burundi: l'Eglise sous surveillance étatique", La Revue Nouvelle (Février 1986); Amnesty International, Prisoners of Conscience and Political Detainees Held in Burundi, (London: Amnesty International, May 1987)
 Many were very well-known people, such as businessman Didace Nzohabonayo, Térence Nsanze (the former Burundi Ambassador to the United Nations), Dr. Dominique Gacukuzi (Director of Bujumbura's Medical Services), the President of Bujumbura's Court of Appeal Bernard Rukingamubiri, and several others (Personal recollections of the author).
 One of the first measures of the new government was to rescind all the anti-Church legislation and to free all the political prisoners.
 Filip Reyntjens, L'Afrique des Grands Lacs en crise, p. 49
 PALIPEHUTU, "Communiqué No. 6", (May 1988). PALIPEHUTU is the acronym for Parti de la Libération du Peuple Hutu (Party for the Liberation of the Hutu People), an extremist party created in 1980 by Rémy Gahutu, a Burundese Hutu living in exile in Rwanda. This party was based on a strict racialist ideology strongly inspired by the 'Rwandese model'.
 On these events, see J.P. Chrétien, A. Guichaoua and G. Le Jeune, La crise d'août 1988 au Burundi, (Paris: Centre de Recherches Africaines, 1989) (Cahiers, No. 6)
 See Déo Hakizimana (who was among the signatories), Burundi: le non-dit, pp. 84-159.
 Burundi. Commission Nationale, Rapport de la Commission Nationale chargée d'étudier la question de l'unité nationale, (Bujumbura : Commission Nationale, 1989)
 Filip Reyntjens, L'Afrique des Grands Lacs en crise, p. 70
 It is for example typical that the (moderate) Hutu opponent Déo Hakizimana had chosen as a title for the book he published in exile, Burundi: le non-dit (Burundi: the 'unsaid' or the 'unspoken').
 Although political parties were in theory not allowed, the Government's attitude was quite relaxed and the author was able at the time to meet FRODEBU cadres in Bujumbura without any hindrance. Party literature was also regularly printed and distributed without police interference.
 The Twa are the pygmoid populations who were the original inhabitants of Rwanda and Burundi, before the arrival of either the Bantu Hutu or the Cushitic Tutsi. Today they represent only about one per cent of the population in either country.
 The local correspondent of the French press agency, Agence France Presse, had predicted that President Buyoya would be elected in the first round of voting with about 60 per cent of the vote.
 BBC Summary of World Broadcasts, 3 June 1993, quoting Radio France Internationale, interview with President-Elect Ndadaye
 BBC Summary of World Broadcasts, 25 June 1993, quoting Radio Burundi, interview with President-Elect Ndadaye
 This throws us back to the question of the 'emotional gap' mentioned earlier. The catchword of 'Unity' had been so used and overused during the democratization process since 1988 that the word had become associated with UPRONA and Tutsi domination. The feeling, on both sides of the ethnic divide, was that 'Unity' meant continued Tutsi rule with a mostly cosmetic opening up. This was what the Tutsi supremacists who had accepted Buyoya's political course had hoped for and this is what the Hutu population had feared. FRODEBU's victory was quickly denounced by the Tutsi supremacists as 'ethnically divisive'.
 One should keep in mind that in spite of the ethnic polarization, both UPRONA and FRODEBU made a point of having in their ranks members of the 'other' ethnic group. UPRONA was definitely more ethnically pluralistic (its Secretary-General Nicolas Mayugi was a Hutu) but its Hutu members did not have much of a real say in the workings of the party. UPRONA had fewer Tutsi but they tended to play a more influential role in the mostly Hutu party. One of UPRONA's prominent (and founding) members was Jean-Marie Ngendahayo who was from an old and distinguished abaganwa family and who was to play a key role in the political developments of 1993-1994.
 Due to the difficulty of communicating with some isolated parts of the country, the results had not yet been proclaimed. They were published only on 9 July and it is probable that the mutineers were hoping to take the yet unformed government by surprise.
 President-Elect Ndadaye was quick to exonerate former President Buyoya of any involvement in the plot.
 Author's interviews with Burundese Army officers in Addis-Abeba (July 1993).
 BBC Summary of World Broadcasts, 10 July 1993, quoting Radio Burundi
 François Misser, "Democrazia assassinata", Nigrizia (December 1993). Kabora Kassan had been the PALIPEHUTU 'Chief of Staff' in Rwanda. After falling out with his erstwhile friends, he moved to Tanzania where he started a new armed movement, the Front de Libération Nationale (FROLINAT), which was not considered a serious military threat at that time.
 BBC Summary of World Broadcasts, 5 August 1993, quoting Radio Burundi
 This is the most important factor singled out by the Human Rights Watch report on the violence of October-November 1993 as having been a major cause in the attempted Tutsi putsch which was to cost President Ndadaye his life. Human Rights Watch/Fédération Internationale des Droits de l'Homme, Rapport de la Commission Internationale d'Enquête sur les Violations des Droits de l'Homme au Burundi depuis le 23 Octobre 1993, (Brussels, July 1994), p. 11. [henceforth referred to as Burundi Report]
 La Libre Belgique, Marie-France Cros, "Première crise pour le pouvoir au Burundi", 4 October 1993
 Marchés Tropicaux, Jean-Pierre Chrétien, "Tournant historique au Rwanda et au Burundi", 1 October 1993
 All the information on the confused coup and on the President's murder comes from the above-mentioned Burundi Report and from conversations with one of its authors, Professor Filip Reyntjens of the University of Antwerp.
 This point is important because the military later argued that they had fought to defend the President, 'losing the lives of several of their men'. This is not true.
 This point is also important because the Hutu extremists later circulated stories of atrocious mutilations on the President's body in order to incite public violence. The Human Rights Watch Commission which authored the Burundi Report was able to disinter the President's body and examine it. There were no mutilations.
 Given the reciprocal nature of the massacres, sources often tended to contradict each other, some writers preferring to insist on the violence of FRODEBU and the Hutu, others on the Tutsi Army repression. Thus Le Monde, Jean Hélène, "Des réfugiés Hutu font état de massacres après le coup d'Etat contre le Président Ndadaye", 24/25 October 1993, presents a relatively 'pro-Hutu' case, while Libération, Jean-Pierre Chrétien, "Purification ethnique au Burundi", 28 October 1993, presents a 'pro-Tutsi' view of events. More objective accounts can be found in Le Monde,"Les massacres continuent alors que le pays est coupé du monde", 28 October 1993, and Libération, Gilles Millet, "Vengeances aveugles dans les campagnes Burundaises", 5 November 1993. While the already mentioned Burundi Report remains the best objective guide to these tragic events, a typewritten document issued
by the civil servants of Karugi Province entitled "Lumière sur les massacres d'Octobre-Novembre 1993 dans la Province de Karugi", 10 November 1993, constitutes a very serious indictment of local FRODEBU authorities' responsibility in the organization of the massacres. La Libre Belgique, Marie-France Cros, "Juger, pas lyncher", 13 November 1993, is a fair assessment of the need to make a distinction between the criminal activities of Tutsi extremists and the actions of ordinary Tutsi peasants.
 Le Renouveau du Burundi, A. Kwigize and C. Uwera, "Ce sont nos voisins qui nous ont poursuivis", 23 November 1993
 A perfect example can be found in the otherwise correct and well-informed article by Jean-Pierre Chrétien, "Burundi: pogromes sur les collines", Esprit (July 1994), which uses the term 'genocide' to describe the Tutsi massacres of October-November 1993. This piece, written in the emotional aftermath of the genuine genocide in Rwanda, tends to obscure the extremely deep and traumatic effect of President Ndadaye's murder on the Burundi Hutu population and the fact that, horrible as it may have been, the massacre of the Tutsi in Burundi was of a completely different nature from what was to happen in Rwanda six months later. Even if FRODEBU extremists aggravated the Burundi massacres, they could do so only because the feeling of the population was one of rage, shock and frustration after President Ndadaye's murder. This admittedly cruel and irresponsible use of popular feeling is nevertheless quite distinct from the cold-blooded and administrative planning of the Rwanda genocide by a government fully in control of the situation.
 Barnabé Ndarishikanye, "Quand deux clientélismes s'affrontent", Komera, No. 3 (March-April 1994)
 Le Monde, "Burundi: affrontements entre l'armée et les extrémistes Hutu", 18 October 1994
 After living in self-imposed exile for over six months in Belgium, Minister Nyangoma was replaced in the new cabinet formed on 5 October 5 1994 by Jean-Baptiste Manwangari, an UPRONA Tutsi.
 See L'Humanité, 3 May 1994; François Misser, "Senza uscita", Nigrizia (June 1994); and BBC Summary of World Broadcasts, 29 April 1994, quoting Kenya News Agency
 BBC Summary of World Broadcasts, 13 June 1994, quoting Radio Burundi: fighting in the Isale and Kanyosha communes (at least 15 killed); BBC Summary of World Broadcasts, 29 June 1994, quoting Kenya News Agency: fighting in the hills around Bujumbura (14 killed).
 Parti de la Réconciliation Nationale (Party for National Reconciliation), created in August 1994 by former dictator Jean-Baptiste Bagaza
 Le Monde, 7 January 1994; Libération, 7 January 1994
 BBC Summary of World Broadcasts, 30 January 1994, quoting Radio Vlaanderen International [Brussels]
 BBC Summary of World Broadcasts, 31 January 1994, quoting Panafrican News Agency [Dakar]
 In the tense and byzantine world of Burundese politics, descent, province of origin, marriage, marriages of your relatives, are all relevant factors in terms of one's position within the field of political forces.
 For himself Nzeyimana had obtained the Ministry of Commerce.
 Le Renouveau du Burundi, 12 February 1994
 Le Renouveau du Burundi, 9 March 1994
 BBC Summary of World Broadcasts, 30 March 1994, quoting Radio France Internationale
 BBC Summary of World Broadcasts, 25 April 1994, quoting Radio Africa No 1 [Libreville] and RTBF [Brussels]
 This refers to the protection of the northern border with Rwanda during the hectic and violent days of April to June 1994.
 BBC Summary of World Broadcasts, 30 July 1994, quoting Radio France Internationale
 Le Monde, 18 August 1994
 A previous power-sharing agreement signed on 12 July had been a pragmatic, one could almost say indecent, carving up of important positions, allocating nine Governorships to FRODEBU and its allies, seven to the opposition, 14 embassies to the government, nine to the opposition and so on. It did not solve anything institutionally.
 President Cyprien Ntaryamira had been killed together with President Juvénal Habyarimana of Rwanda on 6 April 1994 when their plane had been shot down as it was about to land at Kigali Airport.
 BBC Summary of World Broadcasts, 14 September 1994, quoting Agence France Presse
 The CNDD seems to have been created clandestinely as early as March 1994. But it was only formally announced in September of the same year when one of its leaders, Christian Sendegeya, denounced the power-sharing agreement and called for an internationally sponsored conference on Burundi under the aegis of the UN and the OAU (BBC Summary of World Broadcasts, 7 October 1994, quoting Radio France Internationale).
 Mimeographed letter, dated 30 October 1994
 The CNDD had an armed branch called Forces pour la Défense de la Démocratie (FDD) and in practice its members often operated in cooperation with PALIPEHUTU or even Forces Armées Rwandaises (FAR) elements. Most of them were based in the Kivu Province of Zaire from where they launched operations in Cibitoke, Kayanza and Ngozi Provinces. Later some groups started operating from Tanzania, attacking Muyinga and Kirundo Provinces.
 BBC Summary of World Broadcasts, 5 October 1994, quoting Radio Burundi
 Le Monde, "Burundi: affrontements" and BBC Summary of World Broadcasts, 22 October 1994, quoting RTBF [Brussels]
 BBC Summary of World Broadcasts, 24 October 1994, quoting Radio Burundi, and 27 October 1994, quoting Agence France Presse
 BBC Summary of World Broadcasts, 26 October 1994, quoting Agence France Presse
 BBC Summary of World Broadcasts, 4 November 1994, quoting Radio Burundi. The 'displaced' Tutsi were those Tutsi who had survived the October 1993 massacres and who had been regrouped by the Army in small trading centres, schools, and hospitals, where they could be defended from further Hutu attacks.
 The only Tutsi party which had refused to sign the September agreement was the ultra-extremist PARENA (Parti de la Renaissance Nationale--Party of National Rebirth) created by former dictator Jean-Baptiste Bagaza.
 There were Tutsi extremists in the CNS itself, since both Lieutenant Colonel Sinduhije and Interior Minister J.B. Manwangari had been involved in the October 1993 putsch and the subsequent murder of President Ndadaye.
 BBC Summary of World Broadcasts, 11 November 1994, quoting Radio Burundi and Agence France Presse
 BBC Summary of World Broadcasts, 11 November 1994, quoting Radio Tanzania
 BBC Summary of World Broadcasts, 22 November 1994, quoting Radio Burundi and 23 November 1994, quoting Radio France Internationale
 BBC Summary of World Broadcasts, 29 November 1994, quoting Radio Burundi
 There had been a Commission created already. But it had been given to the Army which had no intention of seriously looking into the matter and which had let it lapse.
 BBC Summary of World Broadcasts, 3 December 1994, quoting Radio Burundi and Radio France Internationale
 BBC Summary of World Broadcasts, 7 December 1994, quoting Radio Burundi; Le Monde, 11 and 12 December 1994
 That much had already been given up by FRODEBU was obvious to all objective observers. The UN Special Representative in Burundi, Ahmedou Ould Abdallah, a strong supporter of the September agreement, had remarked that 'the government has given the opposition more than its due' (BBC Summary of World Broadcasts, 11 October 1994, quoting Radio France Internationale.)
 United States Information Agency Dispatch, 19 December 1994
 Le Monde, 17 December 1994
 BBC Summary of World Broadcasts, 23 December 1994, quoting Radio Africa No 1 [Libreville]. See also Le Monde, 21 December 1994.
 BBC Summary of World Broadcasts, 27 December 1994, quoting Radio Burundi
 See Libération, Jean-Philippe Ceppi, "Dans un maquis Hutu du Burundi", 19 January 1995, and La Cité, François Misser, "Terreur à Bujumbura", 9 February 1995.
 BBC Summary of World Broadcasts, 23 January 1995, quoting Radio Burundi. Author's interview with a CNDD militant, Kampala, 25 January 1995. More than 20 vehicles were destroyed.
 BBC Summary of World Broadcasts, 30 January 1995, quoting Radio Burundi and Radio Africa No 1 [Libreville]
 Libération, 2 February 1995
 BBC Summary of World Broadcasts, 1 February 1995, quoting Radio Africa No 1 [Libreville] and 2 February 1995, quoting Radio Burundi
 BBC Summary of World Broadcasts, 6 February 1995, quoting Radio France Internationale
 BBC Summary of World Broadcasts, 8 February 1995, quoting Kenya News Agency
 BBC Summary of World Broadcasts, 9, 10 and 13 February, quoting Radio Burundi
 BBC Summary of World Broadcasts, 14 February 1995, quoting Radio Burundi
 Le Monde, 17 February 1995
 BBC Summary of World Broadcasts, 18 February 1995, quoting Radio Burundi
 BBC Summary of World Broadcasts, 20 February 1995, quoting Radio Africa No 1 [Libreville]
 BBC Summary of World Broadcasts, 21 February 1995, quoting Radio France Internationale
 Le Monde, 3 March 1995
 See two articles in La Lettre Afrique Energies, 4 February and 8 March 1995.
 Le Monde, 16 and 17 March 1995
 President Sylvestre Ntibantunganya, "Analyse rapide de la crise actuelle", Dialogue, No. 180 (February 1995). (Special issue: "Burundi: danger d'implosion à la Rwandaise".)
 Le Monde, Afsané Bassir Pour, "Situation explosive au Burundi", 21 February 1995
 BBC Summary of World Broadcasts, 18 February 1995, quoting Radio France Internationale
 See François Misser, "Terreur à Bujumbura"; Author's interview, Paris, 20 March 1995, with a Burundese journalist who had the opportunity to meet former gang members who had been 'conscientised' by church workers and who very bitterly denounced the way they had been used by the politicians who provided them with weapons (for the gang phenomenon, see further below 6.3 A culture of violence).
 See Emmanuel Ntakarutimana, "Une Eglise Catholique impuissante face au tragique chemin de la démocratie", Dialogue, No. 180. (Special issue: "Burundi: danger d'implosion à la Rwandaise".)
 Thus, the Tutsi extremist paper Le Carrefour des idées, No. 32, 27 May 1994, even asked in a headline: 'Do the Hutu have a soul?'
 For a discussion of the role of the elites in the process of ethnic radicalization, see Nicéphore Ndimurukundo, "Scolarisation des élites et renforcement de la conscience ethnique" in André Guichaoua (ed.), Les crises politiques au Burundi et au Rwanda (1993-1994), (Lille: Université de Lille I, 1995), pp. 125-135 and Barnabé Ndarishikanye, "Les intellectuels Burundais et la crise" in the same volume, pp. 155-166.
 For an assessment see Rapahël Ntibazonkiza, "Impact de la crise politique d'Octobre 1993 sur la situation économique du Burundi", Dialogue, No. 180. (Special issue: "Burundi: danger d'implosion à la Rwandaise".) This paper has to be read keeping in mind the strong pro-Hutu position of its author.
 The World Food Programme launched an appeal for US$ 385 million for Rwanda and Burundi on 11 March 1995, declaring that the lives of up to three million people could be at stake (Le Monde, 12 and 13 March 1995).
 Le Monde, 21 March 1995
 For more details on the deleterious role of the media, see Jacqueline Papet, "La presse au Burundi", Dialogue, No. 180 (February 1995). (Special issue: "Burundi: danger d'implosion à la Rwandaise"), pp. 75-77; and Reporters Sans Frontières, Rapport de mission au Burundi (17-25 November 1994), (Paris: Reporters Sans Frontières, December 1994).
 Filip Reyntjens, Burundi: Breaking the Cycle of Violence, (London: Minority Rights Group, March 1995), p. 16
 André Guichaoua, "Un lourd passé, un présent dramatique, un sombre avenir" in A. Guichaoua (ed.), Les crises politiques au Burundi et au Rwanda (1993-1994), p. 49, footnote 32