FgtQuery

1

Working Paper no.77

RWANDA’S ORDINARY KILLERS:
INTERPRETING POPULAR
PARTICIPATION IN THE RWANDAN
GENOCIDE
Omar McDoom
Development Studies Institute
LSE

December 2005

Copyright © Omar McDoom, 2005
Although every effort is made to ensure the accuracy and reliability of material published in this Working Paper, the Crisis
States Research Centre and LSE accept no responsibility for the veracity of claims or accuracy of information provided by
contributors.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by
any means without the prior permission in writing of the publisher nor be issued to the public or circulated in any form other
than that in which it is published.
Requests for permission to reproduce this Working Paper, of any part thereof, should be sent to:
The Editor, Crisis States Research Centre, DESTIN, LSE, Houghton Street, London WC2A 2AE.

1

Crisis States Research Centre
Rwanda’s Ordinary Killers: Interpreting popular participation in the
Rwandan genocide
Omar McDoom
Abstract
This paper examines the question of why so many ordinary Hutu participated in genocidal
killing of Tutsi in Rwanda in 1994. I find that mass mobilisation was contingent on the
fulfilment of two main conditions. Firstly it required a mindset – the internalisation of a set of
historical and ideological beliefs – within the Hutu population. These were predominantly
beliefs in a historical Hutu oppression at the hands of Tutsi and in an ideological definition of
the ongoing civil war as an ethnic one, a Tutsi attempt to reinstate this historical order.
Secondly, it required the commitment of State institutions to the genocidal project. This
commitment provided the initial trigger, legitimacy and impunity for civilian participation in
an anti-Tutsi programme. However, once triggered the degeneration into genocidal violence
was the product of a complex interaction of other motives ranging from coercion,
opportunism, habituation, conformity, racism, and ideological indoctrination.
At the time of writing, just over ten years have elapsed since the world stood by and
witnessed a genocide unfold in Rwanda. Since then new scholars from disparate fields of
research have been drawn to Rwanda seeking explanations of this macabre phenomenon.
With the passage of time their research has born competing hypotheses concerned in various
hues with the perplexing questions of ‘How and why did this happen?’ Beyond the simple
human tragedy of the violent nature of so much death, these questions are also driven by the
terrible peculiarities of these killings. The main agents of the Rwandan genocide were
ordinary civilians, wielding crude instruments of death often against those they personally
knew. Moreover, this genocide was not confined to specific regions of Rwanda – massacres
committed by a few – but took place on a national scale at a terrifying speed, and with the
active participation of a sizeable part of the population.1 It was not possible for any Rwandan
not to know what was taking place in his/her country. This empirical paper presents part of
the results of a research project whose primary objective is to examine exactly how Hutu
farmers were mobilised into committing acts of genocidal violence against their Tutsi
neighbours.2
In summary, I argue that mass mobilisation, including collective violence, during the
genocide was contingent on two main conditions, fulfilled in different ways in different parts
of Rwanda. Firstly, mobilisation required a mindset: the prior internalisation within the Hutu
population of a particular set of historical-ideological beliefs. The historical component,
deeply rooted in Hutu collective memory, was of Hutu oppression at the hands of Tutsi that
the Hutu Revolution had ended in 1959. The ideological component defined an external

1

The terms ‘active participation’ and ‘sizeable’ both require definitions that appear subsequently in this paper.
This research is being conducted for a doctoral dissertation at the Development Studies Institute of the London
School of Economics. I am grateful to Drs. David Keen and Peter Uvin for supervising it as well as the UK
Economic and Social Research Council for providing the funding for it.

2

2
threat, the ongoing war, in historical and ethnic/racial terms and rested on twin beliefs:3 one, a
belief that the war was an ethnic/racial one, a collaboration between Tutsi outside the country
(the enemy without) with Tutsi inside the country (the enemy within) against the indigenous
Hutu majority; two, that military defeat would reverse ‘les acquis’ of 1959, and return the
Hutu to a position of subordination and inferiority. The second condition necessary for mass
mobilisation was the commitment of state institutions, which still had authority and/or power
in the eyes of the population, to the genocidal project.4 This commitment provided the trigger
point for mobilisation, the initial legitimacy to the actions of the mobilisers and the mobilised,
and the impunity necessary for these actions to rapidly and ultimately degenerate into
genocidal violence.5 However, this subsequent degeneration was the result of a complex
interaction of individual motives and not only the consequence of the State’s continuing
authority and power.
Scholarship on the genocide to date has tended to focus either on the role played by important
individuals, elites acting within the key institutions of the State’s civilian and military
apparatus, or on the context of specific structural forces, macro-political and macro-economic
trends at play in Rwanda at the time, or on some combination of both.6 As a result, meso- and
micro-level variations in the causal mechanisms behind the violence in Rwanda are not
properly captured and the genocide is instead misleadingly conceived of as a homogenous,
indivisible event.7 Grand theories of genocide based on research or evidence at this level
cannot fully address the complex task of isolating those conditions that were conducive to
genocide, and those conditions that were not. In different parts of the country there were
different routes to genocide. There is also a paucity of research that examines the actions and
beliefs of the principal instruments of the genocide, ordinary rural Hutu, and the specific local
contexts in which they operated. Through micro-level research in four communities (cells)
and a meso-level survey in two prefectures in Rwanda conducted over 2002 and 2003, I
undertook to investigate the causes of popular participation in the genocide by identifying the
mechanisms through which ordinary Hutu farmers, the majority in-group, were mobilised
against the minority Tutsi out-group ultimately into committing genocidal violence.

3

By race I refer to the belief that the Tutsi were alien to Rwanda, originating outside of the country, and by
ethnicity I refer to the belief that the Tutsi were Rwandan but formed a subgroup distinct to the Hutu. The former
view was predominant during the first Republic of Rwanda (1959- 1973) and the latter predominant during the
second Republic (1973-1994), and were reflected in the terms used in official documents from those periods (e.g.
Les Rapports Annuels for each Prefecture). However, as we shall see later, the divisionist media in the years
leading up to the genocide revived the notion of ‘race’.
4
I distinguish authority from simple power by defining it here as legitimate power. For example in Rwanda I
would prima facie consider a civilian burgomaster to have more authority than a military soldier and a military
soldier to have more authority than an interahamwe militiaman, though all may be powerful.
5
As I shall explain later, genocidal violence is only one and the final choice of action during this mobilisation.
Other actions that took place on a massive scale as part of the evolution of violence included conducting night
patrols, manning roadblocks, espying Tutsi, controlling Tutsi movements, looting Tutsi property, destroying and
burning Tutsi homes, killing Tutsi men, raping Tutsi women, and killing Tutsi women and children.
6
Excellent syntheses of the roles played by civilian and military elites and macro-economic and macro-political
forces in the genocide can be found in G. Prunier, The Rwanda crisis, 1959-1994: history of a genocide, London:
Hurst and Company, 1995; and A. L. Des Forges, Human Rights Watch (Organization), et al., “Leave none to
tell the story”: genocide in Rwanda, New York and Paris: Human Rights Watch and International Federation of
Human Rights, 1999.
7
By ‘meso-level’ I am referring to prefectures and communes; and by ‘micro-level’, I am referring to sectors
and cells. In 1991, at the time of the last census before the genocide, Rwanda was organised administratively
into 11 prefectures, 145 communes, and 1490 sectors. A sector, whose average mean population was
approximately 4800 residents, was then subdivided into cells. In the 20 sectors where I conducted the survey
were a total of 87 cells ranging in size in from 174 to 2902 residents and whose mean population was 902.

3
As research sites I chose two cells from the North (Ruhengeri prefecture, Kinigi and Nkuli
communes) and two from the South (Butare prefecture, Maraba and Shyanda communes) of
Rwanda.8 In each ‘pair’ I selected one that had experienced a relatively high level of in situ
killings and one that had experienced a relatively low level of in situ killings.9 My primary
source of information was oral testimony, collected using semi-structured interviews and an
enumerator-administered survey.10 The semi-structured individual and focused-group
interviews captured a cross-section of each community comprising of Tutsi survivors,
imprisoned Hutu perpetrators, bystanders, figures of state and social authority, and older
residents with historical knowledge of the community.11 The broader survey in turn sampled
prison inmates and free citizens not accused of genocide-related crimes,12 and who were all
resident in those cells in April 1994. In other words, it was an attempt to sample perpetrators
and non-perpetrators from the same communities. The second main source of information
was the hill gacaca committees in each of these four cells, as well as the corresponding
gacaca committees established in the prisons.13 As part of the record of the genocide, each
gacaca committee at the cell-level established the overall population of the cell in April 1994,
the number of those accused of genocide-related crimes, and the number of victims. In other
words, I knew who stood accused, what they were accused of doing, against whom they did
this, and when and how they were said to have done this.14 Finally, I was able to access a
limited amount of archival data for the chosen prefectures and communes, though much had
been destroyed or lost during and after the genocide.15 These records consisted mainly of
prefectural annual reports, commune monographies, correspondence between burgomasters of
the communes and the Prefect, and birth, death, and marriage registers that included data on
ethnicity.
8

I have chosen to keep the names of the cells anonymous and refer to them only by the name of the commune in
which they are located.
9
I took as a measure of the level of violence the number of Tutsi who were resident and killed within the
territory of those communities (that is in situ) as a proportion of the overall Tutsi population of the cell. In the
Maraba and Kinigi cells, these were 8.3% (7 out of 84) and 2.9% (1 out of 35) respectively. In the Shyanda and
Nkuli cells these were 56.9% (41 out of 72) and 77% (46 out of 60) respectively.
10
The survey was conducted in Ruhengeri and Butare prefectures involving 296 respondents and was
representative at the prefectural level. The survey instrument was first translated into Kinyarwanda and then
translated back into French by a different individual. Questions in the interviews were asked in Kinyarwanda
while the answers to open-ended questions were recorded in French. Altogether the questionnaire comprised 223
questions of which 63 were open-ended. The non-perpetrator stratum of the survey was conducted by 19
Rwandan enumerators each of whom had previous knowledge of the communities they were working in and at a
minimum who had completed secondary school education. The enumerators were a mix of Hutu and Tutsi. I
also asked them to complete the survey questionnaires personally first to see whether there were any worrying
biases in their backgrounds and attitudes. I also interviewed them each individually. The perpetrator stratum of
the sample was conducted by myself with the aid of an interpreter, as a personal authorisation was required to
access prison inmates.
11
In total the project conducted 42 semi-structured and structured interviews of residents of the cells: 10
respondents from the Maraba cell, 11 respondents from the Shyanda cell, 10 respondents from the Kinigi cell,
and 11 respondents from the Nkuli cell.
12
The prisoners interviewed include both those who had confessed and those who had not confessed but stood
accused by others. The hill residents were chosen randomly (where possible every one hundredth name) from
the gacaca lists of the cell population as it existed in April 1994, while ensuring that they did not also figure on
the gacaca list of the accused.
13
In 2002 the Rwandan government instituted a pilot system of community-administered justice, based on a
traditional dispute-resolution mechanism called gacaca, one of whose objectives was to try and convict members
of those communities for genocide-related crimes. Gacaca had also been instituted in the prisons prior to this
and I used information from the prison gacaca to verify the information gathered from the hill gacaca.
14
It should be borne in mind that at the time of my research the gacaca programme was still in its pilot phase
and those accused had not yet been tried and convicted.
15
Prefectural level archives existed in both Butare and Ruhengeri, but communal level data only existed for
Nkuli. Records for Kinigi, Maraba, and Shyanda had been lost or destroyed.

4

In this paper I reject two propositions: that popular participation was simply the product of a
longstanding tradition of obedience to state authority;16 and that genocidal violence was the
inevitable outcome of longstanding and overt Hutu racism or ethnic prejudice against Tutsi.17
I argue that people did not kill simply because they were told to kill, nor did the Hutu kill the
Tutsi simply because their hatred finally had an opportunity to express itself. The first part
provides a brief description of the genocide in each of the four communities. This is followed
by an examination of the two central questions. An appendix is included that summarises the
most important findings in each of the four communities.
Describing the genocide
Maraba cell, Butare, southern Rwanda
The genocide in the Maraba cell began thirteen days after Rwandan President Habyarimana’s
plane was brought down on 6 April 1994. Of the 84 Tutsi who were resident, only 26 were to
survive. The cell was close to the frontier with neighbouring Gikongoro prefecture where the
killing had already started. Tutsi fleeing these attacks passed through the Maraba cell seeking
refuge at the nearby Simbi Parish, pursued by their assailants. Initially the burgomaster and
the conseiller appealed for calm and urged villagers not to join the attacks.18 Meanwhile,
Tutsi from the sector joined their Gikongoro compatriots at the Parish having either been
forced to flee under threat from looters or having fled out of fear. Finally, the visit to Maraba
of the newly-appointed pro-genocide President, Sindikubwabo, on 17 April, unleashed attacks
on the 3000-5000 Tutsi who had gathered at the Parish; and with the removal to Butare of the
Tutsi prefect opposed to the genocide, the burgomaster committed himself to the genocidal
project that same evening by radio. Two days later, attack groups were organised and
roadblocks erected in the cell itself by longstanding racists, former soldiers, and MDR-Power
leader E.R.19 Meeting almost every morning, the attack groups hunted down Tutsi who had
either hidden or chosen to remain in their homes, looting homes in their wake. Tutsi
continued to be hunted and killed throughout May and June, though the majority had already
been massacred at the local parish and health dispensary in April. The groups continued their
operations until the Rwandan Patriotic Front (RPF) arrived on 2 July.20
16

For examples of the ‘obedience’ argument being advanced, see Prunier (1995), p.248; H. Hintjens,
‘Explaining the 1994 Genocide in Rwanda’, Journal of Modern African Studies, 37:2 (1999), p.271; Philip
Gourevitch, ‘After the Genocide’, The New Yorker, 18 December 1995, pp.84 & 93. For examples of the
‘ancient, tribal hatreds’ explanation, see CNN broadcast, ‘Burundi and Rwanda sites of ethnic violence’, 7 April
1994 (transcript obtainable through Lexis-Nexis); R. Dowden, ‘Rwanda’s Twins locked in eternal war’, The
Independent, 10 April 1994; W. Schmidt, ‘Strife in Rwanda’, The New York Times, 10 April 1994.
17
Given the space constraints of this chapter, I treat only two of the five questions that were under investigation
in the project here. The other three questions under investigation were: To what extent did the advent of
political parties in 1991 promote the ethnic division and consequent violent mobilisation along ethnic lines of
the rural poor? To what extent and through what mechanisms did the context of the ongoing civil war influence
participation rates in the genocide? To what extent was violence the product of poverty, that is a structural
decline in socio-economic living conditions through the mechanisms of demographic pressure, land degradation,
and land scarcity? The answers to these questions appear in my forthcoming dissertation.
18
As explained earlier Rwanda in 1994 was organized on four principal administrative levels. In descending
order of size these were: Prefectures headed by a Prefect, Communes headed by a Burgomaster, Sectors headed
by Conseillers, and finally Cells headed by Responsables.
19
The Mouvement Démocratique Républicaine (MDR) had strong roots in the ParmeHutu party credited with
the Hutu Revolution of 1959, and enjoyed strong support in Gitarama prefecture and the northern communes of
Butare prefecture. It was split between a moderate and extremist wing, MDR-Power.
20
The RPF’s army, comprised mainly of the descendants of Tutsi exiles from the 1959 Revolution, invaded from
Uganda in October 1990. It is also this army that put an end to the genocide.

5

Shyanda cell, Butare, southern Rwanda.
The remotest of the cells, Shyanda, was also the last to descend into the genocide, some
eighteen days after Habyarimana’s assassination. Initially the sector conseiller and cell
responsable, along with their constituents, were very concerned as to who was behind the
looting and burning of homes they could see on the neighbouring hills. It was not yet
apparent that Tutsi were the targets. This uncertainty impelled the sector conseiller to
continue to organise night patrols and also establish roadblocks, activities in which both Hutu
and Tutsi participated, to ensure the security of his sector against unwelcome intruders.
Meanwhile burgomaster T.S. had been summoned to a meeting in Butare town with the new
Prefect. The substance of this meeting is unknown. However, on 22 and 23 April, following
his return, assistant burgomasters D. and I., along with veterinarian G., accompanied by
villagers, conducted the first looting and burning raids on Tutsi homes in neighbouring
sectors, during what were ostensibly night patrols.. The following morning they came to the
centre of the sector in which the Shyanda cell was also located, and demanded the
whereabouts of all the Tutsi. Eyewitnesses report that they told the crowd they should not eat
the cows until their work was done. In response E.K., a local butcher, a well-known and
longstanding Hutu extremist, immediately mobilised a group of men and went to search the
Tutsi homes, but found them empty. The Tutsi in fact had suspected something was afoot and
had spent the night hiding in the sorghum fields. The next day, however, a group gathered at
7 a.m. at the local cabaret (bar), as E.K. had ordered, and several parties were organised and
set out to hunt down the Tutsi where they were hiding. Almost all the 41 Tutsi killed from
the cell met their end that day. The bodies were buried by villagers the following day.
Having exterminated the Tutsi, no more attack groups were conducted in the cell, but
villagers subsequently maintained night patrols and participated in groups that raided other
areas. This continued until the RPF arrived in July, when the surviving 11 Tutsi of the cell
dared to return to their homes.

Nkuli cell, Ruhengeri Prefecture, northern Rwanda
The Nkuli cell boasted a population of 60 Tutsi, a sizeable minority (7.6%) for a community
in Hutu-dominated Ruhengeri. Interethnic troubles began in the cell long before Habyarimana
was killed. The proximity to the war led to early suspicions that the Tutsi were acting as
accomplices of the invading RPF army, and in 1992 two Tutsi men were arrested from the
cell. From that point the remaining Tutsi men began to flee, leaving their families behind as it
was initially thought that only the educated men were being targeted as collaborators.
However, in 1993, in reprisal for the RPF violation of the ceasefire, a family of seven Tutsi
was burned alive in their home in the neighbouring cell R. by soldiers working together with
villagers. By this time the movements of the remaining Tutsi in the cell were extremely
restricted by the vigilance of their suspicious Hutu neighbours. On 7 April 1994, immediately
after it was known that Habyarimana had been killed, the conseiller of the sector, aided by
villagers, Coalition pour la Défense de la République (CDR) extremists,21 and soldiers from
the resident army camp, led a systematic campaign of attacks on all the Tutsi left in the cell.
Unable to hide or flee, within two days the extermination was all but complete, with 88% of
the Tutsi population killed. Forty-six met their end within the cell, and seven in the adjoining
military camp.
21

La Coalition pour la Défense de la République, the CDR, was a political party that ran on an open anti-Tutsi
platform and was often identified as the extremist arm of Habyarimana’s MRND party.

6
Kinigi cell, Ruhengeri Prefecture, northern Rwanda
The seven Tutsi homes that were occupied at the beginning of the war were deserted by April
1994, their 35 occupants having either fled or been killed in 1991. Located on a frontline of
the war, bordering the then Zaire, the villagers lived in a state of permanent insecurity, fleeing
their homes twice following RPF incursions into the area. The burgomaster of Kinigi
commune, T.G., chose to exact revenge for the war on his Tutsi population. In the cell this
first led to the arrest in October 1990 of three members of a Tutsi family by villagers from a
neighbouring sector, K., as part of a wave of 8000 arrests that had swept the country,
following the FAR’s simulated attack on the capital.22 They were released a week later,
having been beaten. Several days following the RPF raid on the Ruhengeri prison, on 23
January 1991, a group of villagers from the same K. sector came and killed the male head of
that same family. As in Nkuli, the movements of the Tutsi were closely monitored by the
population. In February 1991, FAR soldiers took and killed two more Tutsi men. Shortly
thereafter villagers from the same K. sector dealt the same fate to another Tutsi man. The
women and children who were remaining fled. By July 1991 the cell was empty of Tutsi.
When Habyarimana died on 6 April 1994 there were no Tutsi left to kill in the cell. In any
event, the cell was occupied the next day by the RPF who began their advance, causing the
majority of the population of the cell to flee to Zaire.
State Authority and the Legitimisation of Violence
To what extent was participation in ibitero (attack groups) and other activities during the
genocide, such as manning checkpoints and conducting irondo (night patrols), simply the
result of villagers following the orders of state authorities? At one extremity of this argument
lies the belief that rural Rwandans, largely uneducated, had grown up in a culture of
obedience, and thus responded unthinkingly to the call of the authorities to kill the Tutsi
enemy: they killed because they were told to kill.23 At the other pole lie those who seek to
credit the ‘masses’ with individual agency and autonomy in making strategic choices in their
own interest: they killed because they wanted to kill.24 However, my research indicates that it
was only once a commitment to genocide had been made by state authority figures who were
still authoritative in the eyes of the population that mass mobilisation in an anti-Tutsi
programme of action was triggered. Until that point, the pro-genocide position had support
only among a minority of extremists, and it was not yet a popular or predominant position.
However, it was not enough for the State to simply give the order. The massive response that
was eventually seen additionally depended on the prior internalisation within the population
of a particular set of historical and ideological beliefs. Also, although it was the above signal
that triggered an anti-Tutsi programme, it was not the State’s authority that lay behind the
programme’s subsequent degeneration into genocidal violence. There were other forces at
work here. Thus popular participation was not simply the product of a longstanding Rwandan
tradition of obedience to authority.
I present the analysis behind these conclusions in three timeframes. Firstly, I examine the
relationship between the state authorities and the population in the war years leading up to the
genocide. Secondly, I look at what happened immediately after Habyarimana was killed on 6
22

See Prunier (1995), p.109 for a description of this event in which the government pretended that the RPF had
launched an attack on the capital, Kigali, to legitimise the wave of arrests that followed.
23
Prunier (1995), p.57; P. Gourevitch, We wish to inform you that tomorrow we will be killed with our families:
stories from Rwanda, New York: Picador USA, Farrar Straus and Giroux, 1999, p.23.
24
The strongest and best case for this explanation is J. P. Kimonyo, Revue Critique des interprétations du
conflict rwandais, Vol.1, Butare: Université Nationale du Rwanda, Centre de gestion des conflits, 2000.

7
April 1994. Finally, I consider the three months that followed, during which most of the
killings took place in each community, until the RPF took control of most of the country.

The Impact of War and Multi-partyism (1 October 1990 – 5 April 1994)
In both Maraba and Shyanda, from the time that opposition political parties established
themselves in 1991, there was an erosion in the authority and power of the Mouvement
Révolutionnaire National pour le Development (MRND) at the local level in the South. The
vulnerability of the MRND was visible and felt at the cell level, but the situation of the
burgomasters at the commune level was largely unknown, and their authority remained intact
for the ordinary population. In the Maraba cell the struggle was between the MDR, the most
popular opposition party, and the MRND. Eventually, with the eclipse of the moderate wing
of the MDR at the national level, it was the extremist wing, MDR-Power under the leadership
of a former soldier, E.R. that was to establish itself as the chief challenge to the authority of
the MRND in the community. The responsable of the Maraba cell found himself sidelined in
the power struggle:
I did not join any political party and I stopped going to the MRND meetings.
There would be abuse hurled between the parties. The PSD was full of young
delinquent men (ibarara). But the real problem was between E.R. who was in
MDR-Power and his elder brother who was in the MRND. There was tension and
they exchanged bad words after meetings.25
In the Shyanda cell, the responsable found himself in a similar predicament, though his
position was being threatened not by an opposition party leader but by the local butcher,26
well-known for his racism and feared for his skill with a machete, and whose extremism
became increasingly vocal as the war progressed. A genocide survivor reports how he would
always pinch her nose whenever he saw her saying he could not wait for the day that he could
shave it.27 Another villager remembered how the same butcher would come and sell banana
beer directly in front of the cabaret (bar) that the responsable owned to show to everyone that
he was more powerful than the responsable.
In the North the situation was quite different. In both communities the MRND was
predominant and the survey shows it enjoyed overwhelming grassroots support within
Ruhengeri, among Hutu and even Tutsi initially. The popularity of the MRND was not
surprising, as it was well known (and the survey corroborates this)28 that Ruhengeri, along
with Gisenyi prefecture, had been favoured by Habyarimana as the region of his birth.
Opposition was weak and parties were closely monitored and intimidated by the MRND.29
Opposition party activists were labelled as accomplices (ibyitso) of the enemy RPF, and the
survey shows they had few professed supporters.30 Instead it was the war that preoccupied
25

Interview with former member of the cell commitee, Maraba cellule, Butare, Rwanda, April 2003
E.K. was a supporter of the PSD party, the party dominant in this area, but was not a political leader.
27
Long noses were a stereotyped characteristic of Tutsi.
28
86% of all respondents thought that the northern prefectures of Ruhengeri and Gisenyi were favoured by
President Habyarimana.
29
In Nkuli’s commune records, I discovered a copy of an almost verbatim report of the proceedings of an MDR
meeting in 1991 written by an MRND infiltrator and submitted to Ruhengeri’s Prefect, the most senior regional
MRND representative. In addition an MDR office building had been destroyed and there was correspondence
from the burgomaster denying his authorisation of this action.
30
88% of the respondents in the North who admitted to supporting a political party stated that this party was the
MRND. The next most popular party was the MDR, scoring 10%.
26

8
the state authorities and distrust of the Tutsi ‘enemy within’ was deep. The state authorities
instituted anti-Tutsi campaigns in response. In the Nkuli cell in 1993, soldiers aided by
villagers burned a family of Tutsi alive in their home in reprisal for the RPF violation of a
ceasefire. Similarly, in Kinigi the burgomaster ordered the arrest of Tutsi following the RPF
invasion in October 1990. Then again, following the RPF attack on Ruhengeri prison in
January 1991, the same Tutsi family was the target of an attack leaving the head of the
household dead. Ordinary villagers were co-opted into these campaigns from early on too. It
was villagers, albeit from a neighbouring sector, who arrested and brought three Tutsi to the
commune for interrogation following the alleged RPF attack on the capital in 1990. It was the
same villagers who killed the first resident of the cell in the reprisal attack for the RPF raid on
the Ruhengeri prison in 1991; and it was villagers in Kinigi and Nkuli who curtailed the
movements of Tutsi in their cells, manned roadblocks, and conducted night patrols in the
years before the genocide. There was a commitment by the State to an anti-Tutsi position
already in place, and it was a commitment that the population accepted and was complicit in.
So we see quite clearly a sharp contrast in the years leading up to the genocide in the
relationship between the population and the state authorities in these two regions. Although
the MRND was under pressure and unpopular in the South, the war was distant and the Tutsi
threat appeared negligible. In contrast in the North, the power and authority of the State went
unchallenged, but the war provoked a strong anti-Tutsi backlash from the State and the
population. As we shall see in the next section, these differences determined both the trigger
point and the configuration of forces involved in the killings after Habyarimana was
assassinated. However, they did not affect the eventual outcome: a massive and popular
effort to exterminate the Tutsi.

Choosing Genocide (6 April 1994)
Once the news of Habyarimana’s assassination on 6 April 1994 became known, two very
different responses were seen in the North and the South. In the Nkuli cell, killings began the
following day, but in Maraba and Shyanda in the South the killing did not start for another
two weeks.31 The difference lies in the moment of commitment of the meso-level (commune)
authorities to anti-Tutsi violence. In the North this occurred long before Habyarimana was
killed, but in the South it did not occur until mid-April. In addition, once the burgomasters
had signalled their pro-genocide position, the violence was organised and led at the local level
by different figures in the North and South. It was the state conseiller in the Nkuli cell who
gave the signal, while the state officials giving the signal were an extremist and a political
opportunist in the Shyanda and Maraba cells respectively. This reflects the regional
difference in the authority and power of the state in the years leading up to the genocide.
In the North, the commitment from the State to an anti-Tutsi position was in place before
1994. The constellation of forces was already perfectly aligned in the North. The meso- and
micro-level civilian authorities, the military authorities, the extremist politicians, militia, and
the general population had all been complicit in anti-Tutsi activities from as early as 1991. It
was unsurprising, then, that on 7 April we saw in the Nkuli cell that the sector conseiller
immediately took the lead, assisted by villagers, CDR supporters, militia, and soldiers from

31

The killing in the Maraba cell began on 19 April and in the Shyanda cell on 24 April. In the Kinigi cell there
were no Tutsi remaining, having either been killed or having fled in 1991. The day after Habyarimana died the
RPF advanced from its position to occupy the Kinigi cell and the majority of the population fled.

9
the adjacent military camp, in attacks against the Tutsi in his area. In other words, it was the
same configuration as before.
The chronology of events was different in the South. With Habyarimana’s death, the situation
became fluid and ambiguous, and the burgomasters found themselves at the centre of
competing forces. There was no precedent for killing Tutsis as there had been in the North.
On the one hand, the Tutsi Prefect of Butare, Jean-Baptiste Habyalimana, the most senior
civilian authority in the region, remained opposed to the genocide. On the other hand, the
new central authorities in Kigali were calling for blood, as were extremist elements from
within the commune and from neighbouring communes.
In Maraba, burgomaster J.M.H. found himself in a dilemma over what to do with the several
thousand Tutsi refugees gathered at the Simbi parish. However, his mind was made up on 17
April when the interim President, Sindikubwabo, visited Maraba, and the Butare Prefect was
removed from office over the radio. The massacre at the Parish immediately commenced, in
which J.M.H. participated. He had signalled his pro-genocide position. Two days later, on 19
April, the first ibitero (attack groups) were organised in the cell. In Shyanda, burgomaster
T.S. held out longer, but immediately after returning from a meeting in Butare town with the
new Prefect, his assistant burgomasters led looting raids with the population. Upon their
arrival in the Shyanda cell on the morning of 24 April in a commune vehicle, the first ibitero
were organised by the racist butcher, E.K. Their arrival and address to the population sent the
unequivocal signal of the Shyanda commune’s pro-genocide position.
This reversal of position by the burgomasters effectively neutralised the power of the microlevel authorities, the conseillers and responsables, to stop the killing in the southern
communities. Overtaken by events and wrong-footed by their superiors, they could now
either join in the genocide or be left aside. Their initial uncertainty and ignorance of the
genocidal project resulted in their eclipse by their rivals from before: the racist in Shyanda,
and the extremist/political opportunist in Maraba. In the general instability and fear following
Habyarimana’s death, the gap in authority and power at the micro-level was quickly filled by
these two figures. In any event, the authority of the conseillers and responsables had been
waning in the years leading up to the genocide. Unlike in the North, it was the non-state
actors influential from before who organised and led the anti-Tutsi campaign. One
eyewitness, who became the responsable after the genocide, described this marginalisation of
the micro-level authorities in the Shyanda cell:
They [the responsable and the conseiller] did not stop the genocide because it was
supported by the high authorities. They were saying that the high authorities had
more powers than them and that they would be killed too. Once the responsable
asked a group of people why they were going to kill and loot at a certain home.
They responded that the leniency you have will sabotage our programme. So he
was afraid and was even told by the leader of this group that he was no longer the
responsible.32
Thus it was only once the higher, meso-level authorities signalled their pro-genocide position
to the villagers that the population could be mobilised on a massive scale, rapidly climaxing
in genocidal violence. It was clear from the interviews that ordinary villagers were oblivious
to the vulnerability of the burgomasters and the pressures they had been under from Kigali, as
well as pressure from political rivals and extremists within their communes. The meso-level
32

Interview, Shyanda, Rwanda, May 2003.

10
authority of the burgomasters was still intact in the population’s view. The burgomasters’
eventual endorsement of the genocide consolidated the positions of the racist in Shyanda and
extremist/political opportunist in Maraba, and legitimised them as the new authorities in their
communities. In these new roles they quickly embarked on the, by now, state-sanctioned
programmes to eliminate the Tutsi that I shall describe in the next section. Thus the defection
of the burgomasters, and the rise of new authority figures at the micro-level were critical to
mass mobilisation. Until this point, there had been growing pressure from a minority of
extremists within the commune to adopt the genocidal position, but there was not as yet a
popular commitment to the project.33

The Evolution of Violence (April – July 1994)
I have tried to show that the State’s authority was necessary to trigger and popularly
legitimise a programme of action that was to culminate in the elimination of the Tutsi. In the
Nkuli cell in the North, genocidal violence occurred immediately after Habyarimana was
killed. It took only two days to kill 46 Tutsi. The speed of this response was due to the
alliance of necessary forces already existing, having collaborated with impunity in an antiTutsi campaign long before April 1994. It was therefore not a large step to total elimination
in this already-established campaign.
The South presents a bigger puzzle. There had been no earlier state-sanctioned anti-Tutsi
programme of activity. Once Habyrimana died, there was not an automatic move to violent
elimination. It was not until the moment that a representative of the State, whose authority
was still intact in the eyes of the population, signalled a pro-genocide position that we see the
start of an anti-Tutsi programme in the South. This programme comprised a spectrum of
increasingly anti-social activities that ended in genocidal violence. In this section I argue that
it was not the simple authority or brute power of the State alone that sustained and intensified
popular participation in this programme. In Maraba, for example, the population was engaged
(without the intervention of figures of State authority) for three whole months in the hunt for
Tutsi before the RPF arrived in July 1994. The rapid degeneration to genocidal violence in
the South was, instead, the result of the complex interaction of a number of dynamics.
In the South, the population was progressively but rapidly co-opted into the new programme
for action. There was a broad similarity between the two communities, showing a pattern of
increasingly anti-social events. First, Habyarimana’s death created fear and uncertainty.
Extremists began a campaign of anti-Tutsi rhetoric. Homes could be seen burning on
neighbouring hills, but the cause was not yet known. As a result, night patrols and roadblocks
were organised, in which both Hutu and Tutsi participated, to ensure the security of the areas.
Second, it became clear that the Tutsi were the targets. Reports came in of anti-Tutsi violence
in neighbouring communities. The meso-level authorities signalled their defection to an antiTutsi position in their own communities. As a result, the looting of Tutsi property and
destruction of Tutsi homes began. Third, the micro-level authorities were eclipsed by the
extremist in Shyanda and the political opportunist in Maraba, who mobilised and led part of
the population in the first ibitero. As a result, the killing began. Tutsi men were initially
33

Ideally one would have the counterfactual situation among the selected research sites to see whether mass
mobilisation would not have occurred if a burgomaster continued to hold out and the extremists did not come to
the fore. However, it is my view that in such a situation a burgomaster would simply have been removed or
killed, as the Prefect Jean-Baptiste Habyalimana had been, and replaced by someone willing to toe the genocidal
line.

11
targeted, but the number of participants in ibitero quickly increased and the violence became
genocidal. Tutsi women and children were also killed. Fourth, the ibitero multiplied,34 and
moved to other communities. They continued until the RPF arrived in July 1994.
In summary, there were different routes to commitment to genocide in the North and the
South. In the North, the State’s authority at all levels was undisputed, as opposition parties
were weak or non-existent and the MRND (the party synonymous with the State) was
popular. In response to the outbreak and the continuing proximate threat of the perceived
‘Tutsi’ war, anti-Tutsi activities had been initiated from as early as 1991. All elements,
namely the State institutions (civilian and military, meso- and micro-level), the non-state
extremists, and the general population were already aligned in an anti-Tutsi constellation.
When Habyarimana was assassinated in April 1994, the response was swift and unequivocal.
In Nkuli, Tutsi began to be killed the following day; while in Kinigi there were already no
Tutsi left: they had been eliminated or had fled since 1991.
In contrast, in the South the war was distant, but the authority of the State had been under
threat at the local level since the introduction of opposition political parties in 1991. In
Maraba, the sector conseiller was challenged by an MDR-Power leader, and in Shyanda the
cell responsable was being undermined by a local racial extremist. Once Habyarimana had
been killed, the two commune burgomasters, under pressure from the new central command
in Kigali and a minority of extremists from below, succumbed, and committed themselves, to
an anti-Tutsi programme. It was at that precise moment that violence erupted at the
grassroots level. However, unlike in the North, the violence in the South was organised and
led at the local level by the two figures who had previously challenged the State’s authority.
Once the burgomasters defected to the pro-genocide position, these two were able to establish
themselves as the new legitimate authority figures at the local levels, completing the eclipse
of the sector conseiller and cell responsable. In this sense, state authority was necessary to
legitimise the extremists’ position, and thereby their programme of action into which they coopted the population (with its particular mindset). However, once triggered and legitimised,
this anti-Tutsi programme of activity was sustained and intensified by forces independent of
the state’s authority and power. The violence had a momentum of its own to which the State
continued to lend its influence. In the absence of any State sanction for their increasingly
anti-social actions, the mobilised population rapidly degenerated into genocidal violence.
It is difficult to convey, let alone analyse, the villagers’ perceptions of events during these
different stages in the degeneration towards genocidal violence. Their own words best
illustrate the diversity and contradictions in motives. The first account is that of a confirmed
non-participant in the genocide from the Maraba cell:35
When villagers went to the places where the Tutsi fled to, they said they could not
leave them alive there and take their things as they knew each other. And the
community had been sensitised that one group was the enemy and this gave them
the energy to do what they did. It was clear in the gacaca meetings that some
people participated because of sensitisation and so most people killed to get rid of
the enemy and those people who were being killed were the supporters of the
enemy. I would say that in general there was a personal liking of each other and it
was the authorities who are to blame as they turned them against these people.
34

In Shyanda the ibitero only lasted two days in the community but irondo continued and people participated in
ibitero in other communities.
35
This interviewee was a Vice-President of the gacaca committee for Maraba. One of the qualifications for this
position is that a person not be accused by the community of having participated in the genocide.

12
Some of the villagers were saying that because the authorities told them to do this
they thought that they would not have to answer for anything. But now people see
it was wrong. Now it is the villagers who are suffering and those in the prison
who are in suffering while those who organised the genocide are in Arusha where
they get television and complain of not having enough to eat.
If they were going to attack a rich person, more people would join. If they were
going to attack a poor person, there would be less people. For example in A.’s
case, her father had many cows and so many people, more than 20, went. There
were no women in our cellule who did this. In other parts of the country, there
were women. They would do something that matches their strength like uproot
sweet potatoes but they would not kill. Ibitero were everyday. When they
attacked a home, somebody would say that “so and so had cows” and they would
plan to meet there. In the first days people went by force to go and fight in the
war (gutabara) and after that they went voluntarily because they could get
property.36
This observer’s account is rich in contradictory motives. His initial account explains that
Hutu reasoned that they had to kill the Tutsi if they wanted to keep their possessions. In the
same breath he ascribes the killing to the ideological messages that the Tutsi were the enemy.
He goes on to say that people felt emboldened, and that they could act with impunity since the
authorities did not stop them. In his second account, he explains how initial coercion was
quickly replaced by material incentives for participation. Evidence from the broader survey
supports the coercion argument to some extent. In-group sanctions did exist, but they were
overwhelmingly monetary: a cash fine or property looted. Although many believed in the
threat of physical harm, “putative duress”,37 there was not a single instance of anyone being
killed for not participating or even for hiding or helping Tutsi. His use of the Kinyarwanda
word, gutabara (to go and fight a war, to go to save) also affords us a little insight into an
ideological mindset of war.
The next account is of a confessed killer from the Shyanda cell:
The next house was G.’s. There was no one in the house. I was in this attack
group. I just came across these people and followed as I was coming back from a
bar and was drunk. I did not go in earlier attack groups as I did not know what
was happening. The house was already looted. They had taken everything. I
don’t know who did this. As it was a tiled house it was not burned. Instead they
looked for the occupants of the house and killed them. When they found them
they killed 8 people including G.’s family. 5 members of his family were killed.
The others were just hiding with them. I stoned one of them but it was my
neighbour who finished him. There were many in the attack group—between 20
and 30. About 10 had pangas. It was not all 10 who used their pangas – only
about 4 did. The others watched only and were shouting words corresponding
with what they were doing. The bodies were buried the next day. We all
participated in the burial. We were over 100. It was to prevent the dogs from
eating them.38
36

Interview with Vice-President of gacaca committee, Maraba commune, Butare, Rwanda, April 2003.
This phrase is used by Browning to describe the fear of members of the German battalion responsible for the
massacre of Polish civilians in World War II if they did not obey orders (C. R. Browning, Ordinary men:
Reserve Police Battalion 101 and the final solution in Poland, New York: HarperCollins, 1992).
38
Interview with confessed prison inmate, Butare central prison, Rwanda, May 2003.
37

13

In his account, this perpetrator attempts to explain that his participation in this killing was
accidental. He did not plan to kill anyone as he only met the attack group on his way home.
Also, the killing came as a surprise to him. He did not know that there was a plan of attack
against the Tutsi. There is also evidence of group de-individuation, the absence of a sense of
individual responsibility. Everyone participated in this kill either as machete-wielders or as
pro-active bystanders. They all had a role to play, but no single person was responsible. In
addition, a clue to the extent of mass mobilisation for the genocide is given in the number,
more than one hundred, of villagers who buried the bodies the following day.
The following account is of a genocide survivor from the same cell as this perpetrator:
When the President died on the 6th, it took a long time for such things to happen
here in [Shyanda cellule]. On Friday April 24th, one of the boys coming from the
market said that the President said all Tutsi should be killed. The following day,
Saturday, the conseiller told us not to flee and we even went out to dig in the
fields. That same day at 2 p.m. we saw houses burning and we saw it was only
the Tutsi houses. We did not sleep at home that night. We slept on the hills.
When I was hiding in the sorghum, I heard gunshots but later we heard that they
were just making a noise so that people would come out to be killed. In the
morning we came back and then E.K. came and wanted to take the cows of my
husband. My husband refused and then he took the cows by force. My husband’s
brother gave E.K. one free cow and so he left him alone. That was in the
morning. Then at around 3 pm that is when all my family members were killed
except my husband who was killed 2 days later on Tuesday.
I saw one attack group that killed my family members (my in-laws). I was hiding
in the house and I heard them being killed. There were a lot of people. I heard
them drumming and whistling and after they had killed them they came in the
house. I had locked myself in one of the rooms with a child I was breast-feeding.
The leader sent in 2 men and they took away my 2 children and when they
reached outside, they undressed them to see whether they were boys or girls.
They were arguing among themselves as to whether they should kill them. My
elder child was begging for forgiveness. One of the men said let us leave them as
they are only girls. I was not killed as I was Hutu married to a Tutsi. They only
started hunting the women when the Inkotanyi were approaching. Then they said
we should start killing all those married to Tutsi so that they cannot tell them what
happened here.39
This genocide survivor’s account illustrates the uncertainty in Tutsi minds, and the initial faith
they placed in their sector conseiller after Habyarimana’s death. It was not clear even to them
that they were the targets. As she is from the same cell as the perpetrator above, her
testimony possibly corroborates his claim that he did not know that the Tutsi were being
targeted until he joined an attack group. The second account confirms that it was initially
only Tutsi men and boy children who were being hunted in this community. Women were
not targeted, as ethnicity is patrilineal in Rwanda. However, women became targets later on
when Hutu feared reprisals from the approaching RPF. The degeneration into genocidal
violence, that is the total elimination of Tutsi, was complete.

39

Interview with genocide survivor, Shyanda commune, Butare, Rwanda, May 2003.

14
From these abbreviated anecdotes of the micro-dynamics of the genocide in the South, a
multiplicity of possible motives for popular participation in the various activities leading up to
and including genocidal violence can be inferred. It was clearly not simply because the
authorities had told them to do so, but seems to have been a complex interaction of racism,
ideological indoctrination, opportunism, habituation, conformity, and coercion.
There are obvious difficulties in attempting to quantify the determinants of free will or in
analysing human motivation generally. Individuals may have several motives at once. Their
motives may change as their participation in the activity progresses. They may remember
their motive differently after the event. Their conscious motivation may be different to their
unconscious motivation. They may choose to represent their motivation differently to third
parties. The research techniques used in this project are too crude to find conclusive answers
to such questions that really belong to a number of subfields of psychology. These risks are
highlighted by the different responses made to two questions. Respondents were asked why
they personally participated in night patrols or manned roadblocks. The most popular
response was that it was the law or an obligation (itegeko) (obedience to authority, 57%). The
second most common answer was that they were fighting for their country or defending
themselves against the enemy (ideological mindset of war, 33%). The third most popular
answer was that they or their loved ones were under physical threat if they did not participate
(coercion, 20%).40 However, when asked why other people joined the attack groups, the
responses were quite different. The most popular answer was that these people stood to gain
materially from the looting (opportunism, 40%). The second most popular answer was either
that they thought the Tutsi were the enemy, that they were fighting for their country or that
they were defending themselves against the enemy (ideological mindset of war, 22%). The
third most popular answer was that it was simply the law or an obligation (obedience to
authority, 19%). In other words, there is a tendency to explain personal motivation in terms
of having no choice. However, the same individuals tend to explain other peoples’
motivation in terms of opportunism, that is a self-interested choice.
From Chauvinism and Racism to History and Ideology
Although the authority of the state was necessary to trigger and initially legitimise an antiTutsi programme of action that very swiftly escalated into genocidal violence, once triggered
there was a complex interaction of dynamics, independent of the State’s authority and power,
laying behind the rapid increase in the number of ordinary participants in these activities, as
well as the genocidal direction that these activities took. A new and unstable norm,
legitimising and in some cases institutionalising anti-Tutsi violence, was becoming rapidly
established within these communities. In this section I argue that there was a deeper
historical-ideological basis, a set of beliefs internalised by Hutu in the North and South of the
country. This mindset was also a necessary pre-condition for mass mobilisation, and in part
explains the speed with which such a norm could be assimilated. Without it the commitment
of the state to the genocidal project could not have been implemented. The bulk of the
population must also have been primed and ready to respond to the right signal from state
authorities, even if subsequently driven by a complexity of motives.

40

As individuals were allowed to give more than one possible motive for their participation, these percentages
signify the proportion of individuals who gave that answer. They do not signify any particular answer as a
proportion of all possible answers.

15
Refuting racism and ethnic prejudice
To what extent can popular participation in genocidal violence be explained by longstanding
and overt racism or ethnic prejudice that Hutu had for Tutsi? 41 Behind this proposition is the
idea that the Hutu killed the Tutsi because they finally had the chance to do so, having always
hated them. I would argue against longstanding racism and ethnic prejudice being seen as the
basis of popular participation in an anti-Tutsi campaign.
Starting from a definition that a racist has an inherent sense of superiority, I find that Hutu
from the North and South of the country did not meet this criterion. Hutu respondents were
asked whether it had ever been possible for a Hutu to change his/her ethnicity and become
Tutsi. Those who answered affirmatively were then asked why a Hutu would want to do so.
The most popular response was that they had either become wealthier or acquired more cows
(superior economic status, 47%), while the second most popular answer was that they sought
the prestige or honour of being Tutsi (social status, 17%). Interestingly, half of those who
spoke of social status also equated this with having more cows or more wealth generally.
However, it was unclear whether respondents felt that this ability to change ethnicity only
existed under the Tutsi monarchy: that is to say, it was a historical sense of inferiority; or
whether it still existed as a contemporary sense of inferiority.
Before Habyarimana’s death in April 1994, Hutu and Tutsi were less integrated in the North
than in the South. Not only were there fewer Tutsi living in the North as a whole, but there
were proportionately fewer inter-ethnic marriages too. In the South, 21% and 12% of the
Tutsi population in the Maraba and Shyanda cells had married Hutu, compared with only 3%
and 0% in the Nkuli and Kinigi cells respectively.42 This statistic becomes even more telling
when one considers that the relative number of Tutsi living in the northern Nkuli cell (7.6%)
was higher than in the southern Maraba cell (5.6%). The absolute number of Tutsi living in
Nkuli, however, remains lower than in Maraba. Indeed, when I probed further into the Tutsi
families in the Kinigi cell, where there were no interethnic marriages at all, I discovered that
of the two distinct, unrelated Tutsi families resident, all the marriageable members had gone
to other sectors, and in one case had gone as far as Gisenyi, a neighbouring prefecture, to find
Tutsi partners.
In addition to looking at ethnic preferences expressed in individual behaviour when choosing
marriage partners, I also asked a series of questions designed to gauge opinions and attitudes.
The results again suggest a lower level of ethnic prejudice in the South than in the North. I
first asked a question intended to measure perceived ethnic inequality. Respondents were
asked whether the Tutsi in their communities generally had more land or more cows than their
Hutu neighbours. The data revealed that, in both the North and the South, Hutu respondents
felt that this was the case. However, when asked the follow-up question intended to measure
the legitimacy of this Tutsi success, the consensus was different. More Hutu in the North than
in the South perceived this advantage as a result of special treatment the Tutsi had received
historically rather than as a result of their own hard work. In other words, Hutu in the North
were more likely to feel aggrieved by perceived Tutsi socio-economic superiority. In
contrast, one Hutu respondent in the South recognised that while some Tutsi in their

41

One of the best cases made for the importance of longstanding racism in the genocide is made by P. Uvin,
Aiding violence: the development enterprise in Rwanda, West Hartford, CT: Kumarian Press, 1998, p. 31
42
It is interesting to note that the majority (76%) of these marriages were between Hutu men and Tutsi women,
and more rarely between Tutsi men and Hutu women.

16
communities did have more cows than the Hutu, this was customary and did not confer upon
the Tutsi a superior socio-economic status in his opinion.43
Ethnic distrust also increased as a result of the war in 1990, particularly in the North, and to a
lesser extent in the South, as a result of the rise in the extremist ‘Power’ factions within the
opposition parties in 1993. In the North the suspicion of the ‘enemy within’ was very
pronounced. Hutu suspected Tutsi, particularly educated Tutsi at the beginning, of
collaboration with the RPF. In my interviews I came across a variety of suspicions. They
were accused of knowing in advance that the invasion was planned, of providing intelligence
to the RPF, of sending their sons to join the RPF ranks, and even of stockpiling guns to use
against the Hutu. These suspicions were shared by the state authorities and the villagers, and
as a result Tutsi were arrested, harassed, intimidated and murdered by villagers working in
tandem with state authorities in the years leading up to the genocide. Distrust was therefore
deep and prevalent in the North, and Tutsi were targeted without proof of complicity. Such
indiscriminate targeting and distrust, I argue, is further proof of ethnic prejudice in the North.
In contrast, while the war did cause Hutu to ‘begin to look at Tutsi with a bad eye’ (batanjiye
kureba abatutsi nabi), it was not a suspicion that was very strong or very widely held in the
South. It was mostly confined to a few extremists. This difference in ethnic distrust between
the North and South is confirmed by the interethnic marriage rates during the war. I found
that they declined in the North but were maintained in the South. Between October 1990 and
July 1994, only two interethnic marriages were recorded in Nkuli commune, whereas between
January 1986 and September 1990 there were 7.44 This was at a time in which the Hutu-Hutu
marriage rate increased substantially. Thus in the North, inter-ethnic marriage rates did
decline during the war, confirming the rise in inter-ethnic distrust. In contrast, however, I
discovered that three mixed marriages had been contracted in the war years in the Shyanda
cell alone, and one as late as 1993, suggesting that distrust was much weaker in the South.
To verify this inference of the differential impact of the war on inter-ethnic relations in the
North and South, I also asked when respondents thought relations began to deteriorate
between Hutu and Tutsi. In the North, 54% of the respondents thought it was when the war
began in 1990 compared with only 22% of respondents in the South. In contrast, only 38% of
respondents in the North thought that relations were good up until the genocide began in 1994
compared with 56% in the South. Finally 18% of Hutu respondents from the South attributed
the deterioration in interethnic relations to the advent of multipartyism compared with only
5% from the north. When I eliminated Hutu respondents and sampled only genocide
survivors these results were corroborated,45 though there was a slight increase in the
proportion of respondents in the South who attributed the decline of inter-ethnic relations to
the start of multipartyism in 1991.
Interviews provide additional confirmation of this interpretation of the state of inter-ethnic
relations before Habyarimana’s death. Besides the obvious difference that Tutsi were killed
in the North but not in the South in the war period, there was also the testimony from
genocide survivors that confirms this difference in prejudice levels. In the Maraba cell, a
43

Interview with inmate formerly resident in the Shyanda cell, Butare central prison, Rwanda, May 2003. Note
that although he believes ownership of cows does not confer superior socio-economic status, this does not
exclude the possibility that this confers superior social status.
44
All these interethnic marriages were between Hutu men and Tutsi women. Nkuli communal registers.
45
Genocide survivors include respondents of Tutsi origin and Hutu women who were married to Tutsi men that
were killed during the genocide.

17
baptism of a Tutsi child took place on 10 April 1994 to which both Tutsi and Hutu were
invited, even though only four days had elapsed since the death of Habyarimana. In Shyanda,
one genocide survivor described relations in her cell as this:
At the start of the war, all people were afraid – not just the Hutu but also the Tutsi
as they heard there was war. There were, however, no problems between Hutu
and Tutsi as a result in my cell. Nothing bad was said about Tutsi at the time,
though perhaps they said it in their huts. After the political parties started there
was one person who was strong here and every time he saw me he would pull my
nose and say he wished he could shave it. There was no-one else in [the Shyanda
cell] like this man [E.K.] I don’t think there were others who disliked the Tutsi.
Perhaps in their huts they said things but he was the only to reveal it to me.46
In contrast, in the Kinigi cell a survivor spoke of her days in school under Habyarimana’s
regime (before the war started), when she would not linger after classes as she would be
teased and hounded by Hutu students because she was Tutsi. As these anecdotes suggest,
prejudicial feelings were much more overt and pronounced in the North than in the South,
even before the war started in 1990.
In summary, racism, when defined as a sense of inherent superiority, was contradicted by the
presence of a historical sense of Hutu inferiority. However, we do see ethnic prejudice,
though this was more overt and longstanding in the North than in the South, dating from at
least October 1990 when the war broke out and possibly earlier. While there was also some
ethnic distrust in the South (possibly due to the rise of extremist factions in opposition
political parties) it did not resemble the profound suspicion in the North that drove the state
authorities and villagers to arrest, chase and kill the Tutsi there. Given these differences
between the two regions, and that genocidal violence was to occur in both places, it would be
difficult to maintain the view then that longstanding or overt ethnic prejudice was thus a
necessary condition for mass mobilisation. This point is subtly underlined when one
considers that among the 111 perpetrators counted in the South, five of them had been
married to Tutsi wives before the genocide. Clearly, then, not all killers were established
racists or chauvinists. Having argued against a racist and chauvinist mindset, in the next part
I argue that the mindset that did prevail was an historical-ideological one instead.

Collective historical memory and the ideology of war
For the State’s call to arms to resonate with Hutu, I argue that there must first have been a set
of popular beliefs, or mindset, in place: a set of ideological beliefs, instilled relatively recently
from above through State institutions and the extremist media in the context of an external
threat, the ongoing war; and a set of historical beliefs, a collective Hutu historical memory,
embedded from an early age from below through the school system and through the family. It
was the fusion of these two sets of beliefs that formed the Hutu mindset at the time of
Habyarimana’s assassination in April 1994, and allowed the progressive but rapid co-optation
of the population into the anti-Tutsi programme.
1. Collective historical memory
Prior to the genocide, the dominant discourse amongst Hutu, and to a lesser degree Tutsi, was
one of ethnic solidarity and collaboration in the South. Hutu and Tutsi were good friends,
46

Interview, Shyanda cell, Butare, Rwanda, May 2003.

18
shared local beer, and married each other; and this opinion seems to have been borne out
objectively in the inter-ethnic marriage rates in the South. However, perceived ethnic
differences reappear when looking at collective historical memory, and it is in the past that we
encounter a divergence of opinion between Hutu and Tutsi, but a remarkable unity of opinion
between Hutu of the North and South.
Hutu historical beliefs, in both North and South, were shared across generations, including
among those too young to have direct knowledge of these historical events, as well as across
the prison and non-prison populations. Overwhelmingly Hutu respondents believed that Tutsi
were treated more favourably under monarchic rule by the mwami (75%)47, and also under
colonial rule by the Belgians (68%).48 Similarly the quasi-feudal relationship of ubuhake
prevalent during the reign of the Tutsi monarchy was remembered as an oppression forced
upon the Hutu (85%).49 Tutsi were generally remembered as being at the top of the social
hierarchy, usually the shebuja (patron or master), while the Hutu was the mugaragu (vassal or
servant) (85%). It is widely believed that Hutu received unjustified and cruel punishment
from the Tutsi in the form of the ingoyi, the whip. Similarly, it emerged that there was a
widespread shared view of the socio-historical origins of the three ethnic groups: the Twa
were the indigenous hunter-gatherers, the Hutu were farmers, and the Tutsi were cattleherders. However, in the survey about half of the respondents were unwilling to answer the
controversial question about where exactly Hutu and Tutsi originated from. Of those who
were willing, 27% claimed the Hutu were indigenous to Rwanda and 50% claimed the Hutu
came from Chad but arrived before the Tutsi. Of these same individuals, 92%, however,
believed that the Tutsi came from outside of Rwanda and 62% specified Abysinnia/Ethiopia
as their country of origin. Perhaps the most sensitive question I asked was about the Hutu
memory of the events of 1959, the year of the Hutu revolution in which the Tutsi monarchy
was disempowered and the first Hutu Republic inaugurated. All respondents were aware of
what transpired. However, when asked whether the revolution was a good or bad event, only
38% were willing to call it good, while 55% called it bad. Part of this discrepancy lies in the
problem of self-censorship under a Tutsi-dominant regime. However, part of it is explained
by the follow-up question of ‘Why was it bad?’ – 22% of these individuals thought that it bad
because it was the origin of the ethnic conflict today between Hutu and Tutsi. In other words,
the genocide and its consequences had influenced their opinion of the revolution. Of those
individuals who described it as a good event, the overwhelming majority thought it was so
because it brought Hutu freedom from Tutsi oppression.
So a clear picture of ethnic grievance emerges from the collective Hutu memory. Hutu were
seen as having been oppressed, discriminated against, and the victims of injustices under the
Tutsi. Moreover, the Tutsi were seen as indubitably foreigners to the Hutu. This was an
important mytho-historical belief that the first Republic had instilled, and the extremist media
reinforced before the genocide.50 In their minds, the Tutsi did not belong to Rwanda.
Interestingly this memory is specific to the Hutu, and is not shared by the Tutsi respondents,
who overwhelmingly gave the opposite answers to these same questions. This underlines the
relative importance of the family and ethnic community in the formation of such beliefs, as
47

The traditional Tutsi king.
Note that these percentages become even higher (82% and 88% respectively) when we discount those Hutu
who chose not to answer or claimed not to know – unsurprising given the climate of the current regime.
49
92% if we do not include those who did not answer or claimed they did not know.
50
One of the most famous examples is the speech of Leon Mugesera, the extremist advocate of the Parti Liberal
in November 1992 in which he urged Hutu to send the Tutsi back to Ethiopia where they came from but this
time through the Nyabarongo river (that flows northwards).
48

19
both Hutu and Tutsi attended the same schools under the first and second Republics, in which
the history of Hutu oppression was taught. However, the role of the school system should not
be underestimated. Respondents from Butare and Ruhengeri who were in primary schools
during Kayibanda’s first republic (1959-1973) reported that teachers asked Tutsi children first
starting school to stand up to distinguish them from their Hutu friends. Those who did not
know their ethnicity were told to go home and ask their parents. Hence ethnic consciousness
was instilled from a very early age, typically less than ten-years old according to the survey.
The history of Rwanda continued to be taught in terms of Tutsi suppression of Hutu under
Habyarimana’s regime (1973-1994). One young Hutu man, now serving in the gacaca
committee for the Maraba cell, was a primary school student in the early 1980s:
They taught us that the Tutsi lived better than the Hutu and that the Tutsi had
come to colonise them. They said that the Tutsi had many cows and that for a
Hutu to get one cow he had to look after their cows for many years. So we grew
up thinking that the Hutu were oppressed.51
Similarly, in the North a Tutsi survivor from the Kinigi cell described her school days also
under Habyarimana in this way:
When I was in school they taught the history of Rwanda and this might have
caused suspicions among the students. They taught that the Tutsi were ruling with
an iron hand. The students would then tease me because I was Tutsi. After
classes you had to hide to avoid being teased or beaten. You would not linger in
case they got hungry and you would rush straight home.52
How can this collective Hutu memory of grievance, present in both the North and South of
the country, be reconciled with the earlier observation that southern Hutu had low levels of
prejudice against the Tutsi in their communities? Part of the discrepancy can be explained by
the difference that often exists between how one perceives individuals personally known, and
perceptions one has of the wider category or group to which they belong. The Tutsi in these
southern cells were indeed on friendly terms with their Hutu neighbours, but on another
mytho-historical level Hutu were also the aggrieved victims and the Tutsi their age-old
oppressors. This is best illustrated in the discrepancy between the answers to two similar
questions in the survey. A lower percentage of Hutu in the South felt that the Tutsi in their
cell (47%) had received special treatment compared with a higher percentage who felt the
Tutsi in general (73%) had been favoured by the mwami. Thus the individual-individual
relationship is perceived differently to the inter-group relationship. Part of the discrepancy
also lies in the relative importance we assign to behavioural outcomes over cognitive
processes. By this reasoning, simply having the memory of being wronged does not make a
person racist or prejudiced until the moment the person acts in some discriminatory way
towards members of this group.53 In the South we do not encounter any overt form of Hutu
discrimination against Tutsi, certainly not before the war broke out in 1990 and probably up
until Habyarimana’s death in 1994.54

51

Interview with gacaca committee member, Tare cell, Butare, April 2003
Interview with Tutsi survivor, Kinigi cell, Ruhengeri, June 2003
53
For a discussion of how political scientists engaged in explaining ethnic violence tend to ignore cognitive
processes while psychologists are preoccupied with the measurement of attitudes and beliefs without testing their
effect on real world situations, see Green & Seher, ‘What is the Role of Prejudice in Ethnic Conflict?’, Annual
Review of Political Science, 6 (2003), pp.509-531.
54
Confirmed by Tutsi survivors from those communities.
52

20

2. The Ideology of War and Justification
The messages contained in State and other extremist media in the years leading up to the
genocide have been well documented and analysed by scholars, and do not need to be
repeated here.55 However, what is missing from the scholarship is a sense of the extent to
which these messages were internalised by their intended audience. What exactly did Hutu
remember and believe of what they were being told? I argue there were two fundamental and
widespread messages that did take root among Hutu because they resonated with the
collective historical beliefs embedded from much earlier on: firstly, that this was an ethnic
war; and secondly, that this war, if lost, would reverse the gains, ‘les acquis’, of the 1959
Hutu revolution. An external threat was needed, that could be defined in these ethnic and
historical terms and thereby activate these latent mytho-historical beliefs in the Hutu
population.
Respondents were asked whether they thought the enemy during the war years was just the
RPF or all Tutsi, including those inside the country. Seventy-one percent of all Hutu
respondents, with no significant difference between the North and the South, affirmed that
they thought all Tutsi were the enemy compared with 28% who thought it was the RPF only.
Respondents from the North were also asked what the war refugees from other regions were
saying when they reached their communities.56 Fifty-one percent of Hutu respondents
answered that they were told that the RPF was killing and committing atrocities specifically
against Hutu; and an additional 9% described it as a war between Hutu and Tutsi. That this
was an ethnic war was a very strong belief in the minds of the Hutu population surveyed.
Respondents were also asked what they thought would happen if the RPF won the war.
Forty-seven percent of Hutu respondents from the North and South believed that the RPF and
Tutsi would kill or exact vengeance on the Hutu specifically, and an additional 24% believed
either that the Tutsi monarchy would be reinstated or Hutu would be oppressed and
disfavoured again. Here, again, we have evidence that it was perceived as a war of Hutu
against Tutsi, but also evidence of the belief in a link between the 1959 revolution and the
current war. Specifically, they feared that the political and economic gains of the 1959 Hutu
revolution would be lost. The Tutsi would be on top again, and the Hutu returned to their prerevolution position as the servile and oppressed majority. One interviewee in Shyanda
expressed the fear that the land forfeited by the Tutsi exiles in 1959 would be taken back as
well as “the air that people breathed”.57
Did people think that what they were doing was right during the genocide? When asked
whether they felt that other people thought their actions were justified during the genocide,
86% of Hutu respondents in the North who replied answered that people felt justified,
compared with 65% in the South. The results were broadly similar when broken down by
whether the respondent was in prison or not. The regional difference might be explained by
the North’s proximity to the frontlines of the war, and it would be reasonable to surmise that
55

One of the best analyses of the Rwandan media before and during the genocide can be found in J.-P. Chrétien,
J.-P. Rwanda : les médias du génocide, Paris: Éditions Karthala, 1995.
56
This question was developed in the North after I had administered the survey in the South. However, when I
returned to the South to ask this and other questions, I received a letter from the Rwandan government advising
me that my research was no longer authorised, as it was ethnically divisive. I would like to point out to others
considering research in Rwanda the danger that their research may place local research assistants in if their
research project falls foul of the administration’s sensitivities.
57
Interview, former member of the cell committee, Shyanda cell, Butare, Rwanda, May 2003.

21
ideological messages about the war would resonate more strongly with inhabitants there. In
general, though, the majority of Hutu clearly thought that their compatriots felt justified in
their actions during the genocide. I then asked those respondents who answered affirmatively
why these people thought themselves justified. A slightly stronger ideological mindset in the
North was again confirmed by these results. The most popular answer in the North was that
they thought that the Tutsi was the enemy (40%), compared with only 18% in the South. In
contrast, the most popular answer in the South was that it was the law or that the authorities
had ordered it (25%), compared with only 4% in the North. The answer on which both
northern and southern Hutu seemed strongest and closest to each other was that it was
because the authorities did not stop or punish them for their actions (18% and 29%
respectively). In other words Hutu in the North had a bias towards justifying the genocide in
terms of fighting the wartime enemy, while Hutu in the South had a bias to justifying it in
terms of the legitimacy conferred by the State’s order to kill, and both groups felt that
impunity was an important factor.
This ideology was indeed a justificatory one for the Hutu I interviewed. However, I believe
it was not only to justify their particular degree of participation to third parties like myself, or
even to themselves ex-post. It was also needed to justify it to themselves ex-ante, at the
trigger moment for the anti-Tutsi programme. Indeed it was also needed during the
degeneration into genocidal violence, even though at that point they may really have been
acting under an amalgam of forces, such as conformity or habituation or opportunism, as
described earlier. Self-acceptable answers to the questions of ‘why should I do this?’ and
‘why am I doing this?’ were needed as much as for the question ‘why did I do this?’ Their
own interpretation of the anti-Tutsi actions they acquiesced to and participated in was
necessary at a cognitive level to justify and also drive their behaviour. The ideological
messages that were being hammered from above provided ready-made, acceptable
justifications that more importantly sounded right when laid on top of their own collective
historical memory of grievance.

Counting the victims and killers
When one looks at the data that the gacaca process has revealed for the four cells, at first
glance they seem to fall far short of mass mobilisation. In the southern cells of Shyanda and
Maraba the percentages of adult men who stood accused were 16.8% and 24% respectively.
Counter-intuitively, given the higher level of ethnic prejudice in Ruhengeri, in the northern
cells of Nkuli and Kinigi the percentages were even lower at 4.6% and 0.8% respectively.
The reason for the regional difference in participation rates is relatively straightforward.
Most of the Tutsi had fled the Kinigi cell in 1991 before the killing had reached genocidal
proportions. There was no one left to kill in 1994. In the Nkuli cell, much of the killing was
also said to be done by FAR soldiers based in the adjacent military camp. This interpretation
is supported by the data that show that while only 9 individuals from the cell itself stood
accused, 24 individuals from areas outside the Nkuli cell were credited with crimes
committed within the cell.
Nonetheless, the bigger puzzle remains concerning how one explains the still relatively low
proportion of perpetrators in all four communities. The problem is mainly definitional. The
overwhelming majority of the prison inmates whom I personally interviewed, as well as those
others for whom I collected information on their criminal charges, were accused of either
actually killing/wounding or being present at that time (that is as a member of ibitero).

22
However, as I described earlier, there were many activities other than participation in attack
groups that were part of the anti-Tutsi programme of action. We saw in both the northern
and southern cells that Tutsi attempts to flee their communities were thwarted by the general
population. In Nkuli, Tutsi who fled were captured and returned to the conseiller. Tutsi who
were in hiding in others’ homes were denounced. In addition to this popular surveillance
there was the considerable verbal abuse of Tutsi after Habyarimana was killed, as the RPF
was widely believed to be behind his assassination. In Shyanda more than 100 people
participated in the burial of a Tutsi family. However, participation in any of these acts did not
make one a criminal accomplice in the interpretation of the 1996 Rwanda Genocide Law.
None of the perpetrators in my sample were accused of committing such relatively minor acts.
The gacaca judges’ training manual, which explained among other things what kinds of
behaviour are criminalised under the Genocide Law, was silent on the definition of what
constitutes an accomplice.58 Yet these acts of complicity occurred on a wide scale, and do
indicate a broader, but ultimately unquantifiable, level of involvement or mobilisation of the
population. In this sense it is possible to speak of the active participation of a sizeable part of
the population.
The data on victims also need explaining. Having selected the communities on the basis of
the proportion of victims killed in situ, I found that these numbers were the result of
idiosyncratic factors, and that as a ‘dependent variable’ they did not provide a meaningful
contrast from which to infer possible determinants of popular participation. For example, in
the Shyanda cell, the last to descend into genocide and where political activity and ethnic
prejudice were relatively low, I counter-intuitively found had one of the highest proportions
of Tutsi killed (56.9%). It was precisely because it was the last cell to receive the signal from
above that so many Tutsi were killed in situ. The delay in mobilisation meant that the
populations of the surrounding cells had already begun to patrol the perimeters of their
communities making it impossible for the Tutsi to flee. Conversely, in Maraba very few
Tutsi were killed in situ (8.3%), as the warning signals of an impending threat were clear.
Due to its proximity to the border with Gikongoro, where the attacks had already started, the
fleeing Tutsi passed through the Maraba community pursued by their attackers, alerting the
Tutsi of Maraba. These then fled their homes and congregated at the Parish of Simbi where
most of them (60.7%) were massacred. In both Shyanda and Maraba, then, the number of
Tutsi killed was largely the accident of geography: Shyanda’s distance from the communal
authorities and Maraba’s proximity to the border of Gikongoro.
Finally, I should point out that not all the victims were Tutsi and not all the perpetrators were
Hutu. Firstly, in the Maraba cell I discovered that 2 out of the 7 residents killed within the
community were in fact Hutu women who were not married to Tutsi husbands. One was
perceived as the local madwoman who had insulted a group manning a roadblock that then
pursued and killed her. The second woman, allegedly, was killed by her stepbrothers in
relation to a land dispute. Thus I have a very small amount of evidence for personal scoresettling occurring under the cover of the genocide. Secondly, regarding perpetrators, in the
Shyanda cell I discovered one Tutsi adolescent among the 48 persons accused of violent
crime there, and in the Maraba cell 7 Twa from among 63 accused by other inmates.59 Both
58

The manual follows the broad principle of the 1996 Genocide law that crimes against the person fall into
categories I-III, while crimes against property fall into category IV. All the perpetrators I counted fall into
categories I-III.
59
These individuals have been accused by their fellow inmates but have not confessed to these crimes
themselves. From among the 7 Twa, 2 still reside in the community, 1 is dead, and 4 are in prison. The Tutsi
adolescent continues to reside in the community and has also not confessed to his crimes.

23
of these stories provide a very small amount of evidence that there was a multiplicity of
motives, other than simply Hutu racism or the State’s coercive power, behind participation in
the genocide.
Conclusion
I have argued for an interaction of top-down and bottom-up factors in explaining mass
mobilisation. Each was a necessary but insufficient condition. First, the authority of the State
had to be committed to initially trigger and legitimise an anti-Tutsi programme of action in
each community. There were local variations in how and when this commitment was reached
and in how it was then implemented. These variations were a function of the differential
impact of war and multi-party politics on the authority of the State and inter-ethnic relations
in the years leading up to 1994 in each of these communities. However, for the involvement
of the population, more was needed than simply an order or signal from above. Second, the
population needed a particular mindset. I have argued this comprised of part-historical and
part-ideological beliefs that allowed them to accept and participate in the anti-Tutsi
programme. Once the programme was triggered, however, it degenerated through a complex
interaction of forces independent of the State’s power and authority into genocidal violence.
In making this argument I reject the pure statist, elitist position that the genocide was
organised at the top and simply implemented locally.60 This argument assumes an
unquestioning obedience to superior authority that I did not find. There was already activity
afoot at the grassroots level that interacted with this order from above. In the North it was
deep ethnic prejudice, reinforced by the war, which facilitated the co-optation of the
population into anti-Tutsi violence. In the South, it was the gap in the State’s authority at the
micro-level that allowed the pro-violence elements to come to the fore. However, I have not
rejected the statist argument entirely, as I have argued that its authority was needed to initially
legitimise these pro-violence elements from below before mass mobilisation was possible.
I have also rejected the purely instrumentalist argument of ethnicity. I do not accept that Hutu
were simply manipulated into hating and fearing Tutsi by a media campaign of extremist
rhetoric from above. For these messages to ring true I have argued that there had first to be a
collective Hutu memory, deeply embedded, of historical, essentialist differences. However,
this Hutu memory is not a racist or prejudicial one. I do not accept that Hutu were
longstanding racists or ethnic chauvinists. Hutu and Tutsi could peacefully cohabit with these
different collective memories, as the years of inter-marriage in the South confirm.
I have also argued, perhaps more intuitively than scientifically as I rely heavily on anecdotal
evidence, that there are psychosocial explanations for what happened once the signal from
above was given. The forces at work here cannot be explained simply by concepts such as
prejudice and authority alone. They have something to do with group psychology and bear
further investigation.
However, having argued this case I nonetheless feel that the units of analysis must be
multiplied to different parts of Rwanda, in order to identify in a more systematic way the
different means by which the state and population both came round to genocide. We need a
more robust model of the genocide. One particularly opaque area that has yet to be properly
understood is what happened precisely after the signal for violence was given. What I loosely
60

The best example of this view is Des Forges et al. (1999), p.8.

24
termed a complex interaction of individual motives that degenerated into genocidal violence
needs to be deconstructed. This will require more evidence and different methodologies. I
believe scholars from subfields of psychology would be best placed to conduct fieldwork and
use their research techniques to understand better both individual and collective thinking at
this relatively un-researched stage of the genocide.

25
Appendix 1: Summary of Selected Results for 4 Communities
Butare (southern Rwanda)
Maraba cell
Shyanda cell
Demographics
Total population (Hutu, Tutsi,
Twa)
Total resident Tutsi population
Victims
Total number of resident Tutsi
killed
Total number of Tutsi resident
and killed within the cell
Total number of Tutsi resident
but killed oustide of the cell
Total number of Tutsi from
outside cell killed within cell
Total number of Tutsi
survivors
Total number of Hutu killed
(not by RPF)
Perpetrators
Number of perpetrators
accused of committing crimes
in categories I-III65 by prison
gacaca
Number of perpetrators
accused of committing crimes
in categories I-III by hill
gacaca
Total number of people in
cats. I-III from outside cell
who committed crimes within
cell

Ruhengeri (northern Rwanda)
Kinigi cell
Nkuli cell

1497

803

969

785

84 (5.6%)

72 (9.0%)

35 (3.6%)

60 (7.6%)

58

6161 (84.7%)

8 (23%)

53 (88.3%)

7 (8.3%)

41 (56.9%)

1 (2.8%)

46 (77%)

51 (60.7%)

2062 (27.8%)

7 (20%)

7 (11.7%)

10

963

4

0

26

1164 (15.3%)

27 (77.1%)

7 (11.7%)

2

0

0

0

63 (16.8%)66

48 (24%)

2 (0.8%)

9 (4.6.%)

Unknown

Unknown

1 (0.4%)

11 (5.6%)

Unknown

Unknown

19

24

18

Attacks started in
1991. There were no
Tutsi left in the
Kinigi cell in April
1994

1 day, but attacks on
Bagogwe Tutsi
started as early as
1992 and continued
in 1993

Raid led by assistant
burgomasters in
neighbouring cells

Not applicable. No
Tutsi resident in
April 1994.

Habyarimana’s
death. Conseiller
and soldiers mobilise
immediately

Racist butcher in the

Not applicable. No

Conseiller and

Genocide dynamics
Number of days between
Habyarimana’s death and the
day the first person killed

Genocide trigger(s)

Main leadership figure(s)
61

13

Visit from the
President of the
Republic to Simbi
parish where Tutsi
gathered
Ex-FAR who was

This is the prison gacaca figure. The hill gacaca figure was 59 victims.
This is the prison gacaca figure. The hill gacaca figure 18 victims.
63
This is the hill gacaca figure. The prison gacaca figure is 7.
64
This includes the 6 Tutsi wives of Hutu men.
65
The 1996 Rwandan Genocide law established four categories of crime. For the purposes of this table I have
placed together all the individuals who committed crimes against the person (categories I-III) to distinguish them
from perpetrators of merely property crimes (category IV).
66
The percentage is expressed as a proportion of adult Hutu male. To obtain the number of adult Hutu males in
each community, I had to assume that it was one quarter of the community’s overall population.
62

26
during the genocide at the
cell/sector level

also
MDR
representative at the
communal level

cell

Tutsi resident in
April 1994.

businessman who
was also Vicepresident of CDR at
the Prefectural level.
Aided by FAR
Captain of the
military camp
Conseiller
imprisoned. Also
CDR leader believed
to be in D.R.C.

Present situation of main
leadership figure (s)

Imprisoned (though
release pending as of
July 2003)

Dead

Not applicable

Date of RPF arrival in
community

July 4th 1994

July 2nd 1994

April 7th 1994

July 15th 1994

Hutu

Hutu

Hutu

Hutu

Yes

No

No

No

At liberty in cell

Dead (natural causes)

No

Yes

Hutu

Hutu

Dead (suicide)

At liberty in Kinigi

Imprisoned

Yes

Yes

Yes

Yes

Hutu

Hutu

Hutu

Hutu

7

1

0

11

6

0

2 (but unknown
when they married)

State Authorities
Ethnicity of Responsable
Responsable accused of
genocide-related crimes
Present situation of
responsable
Conseiller accused of
genocide-related crimes
Ethnicity of conseiller
Present situation of conseiller
Burgomaster accused of
genocide-related crimes
Ethnicity of burgomaster

Imprisoned
Contradictory
evidence
Hutu
At liberty. Believed
to be in DRC.

Dead (natural
causes)
Contradictory
evidence
Hutu

Ethnic cohesion
Interethnic marriages before
genocide: Tutsi husband.
Interethnic marriages before
genocide: Tutsi wife
Interethnic marriages continue

0

Yes

Yes

No

Unknown for cell but
for commune as
whole there is a
decline

State-led anti-Tutsi activities
before Habyarimana’s death?

No

No

Yes

Yes

Multipartyism
Most popular political party
before death of Habyarimana
Political party associated with

MRND and MDR

MRND and PSD

MRND

MRND
and CDR

PL and PSD

PSD

MDR

MDR

after the war begins in October
1990?

the enemy

27
Bibliography
Browning, C. R., Ordinary men: Reserve Police Battalion 101 and the final solution in
Poland, New York: HarperCollins, 1992
Chrétien, J.-P., Rwanda : les médias du génocide, Paris: Éditions Karthala, 1995
CNN broadcast, ‘Burundi and Rwanda sites of ethnic violence’, 7 April 1994 (transcript
obtainable through Lexis-Nexis)
Des Forges, A. L., Human Rights Watch, et al., “Leave none to tell the story”: genocide in
Rwanda, New York and Paris: Human Rights Watch and International Federation of Human
Rights, 1999
Dowden, R., ‘Rwanda’s Twins locked in eternal war’, The Independent, 10 April 1994
Gourevitch, P., ‘After the Genocide’, The New Yorker, 18 December 1995
Gourevitch, P., We wish to inform you that tomorrow we will be killed with our families:
stories from Rwanda, New York: Picador USA, Farrar Straus and Giroux, 1999
Green and Seher, ‘What is the Role of Prejudice in Ethnic Conflict?’, Annual Review of
Political Science, 6 (2003), pp.509-531
Hintjens, H., ‘Explaining the 1994 Genocide in Rwanda’, Journal of Modern African Studies,
37:2 (1999)
Kimonyo, J. P., Revue Critique des interprétations du conflict rwandais, Vol.1, Butare:
Université Nationale du Rwanda, Centre de gestion des conflits, 2000
Prunier, G., The Rwanda crisis, 1959-1994: history of a genocide, London: Hurst and
Company, 1995
Schmidt, W., ‘Strife in Rwanda’, The New York Times, 10 April 1994
Uvin, P., Aiding violence: the development enterprise in Rwanda, West Hartford, CT:
Kumarian Press, 1998

Other Working Papers in Series
WP7
WP10
WP11
WP14
WP15
WP16
WP17
WP19
WP20
WP21
WP23
WP34
WP43
WP44
WP48
WP49
WP50
WP53
WP54
WP55
WP58
WP59
WP67
WP68
WP69
WP70
WP71
WP72
WP73
WP74
WP75
WP76

Hugh Roberts, ‘Co-opting Identity: The manipulation of Berberism, the frustration of democratisation and the
generation of violence in Algeria’ (December 2001) – Also available in Spanish
Jo Beall, ‘The People Behind the Walls: Insecurity, identity and gated communities in Johannesburg’
(February 2002) – Also available in Spanish
Jo Beall, Owen Crankshaw & Susan Parnell, ‘Social Differentiation and Urban Governance in Greater Soweto:
A case study of post-Apartheid reconstruction’ (February 2002) – Also available in Spanish
David Keen, ‘Since I am a Dog, Beware my Fangs: Beyond a ‘rational violence’ framework in the Sierra
Leonean war’ (August 2002)
Joseph Hanlon, ‘Are Donors to Mozambique Promoting Corruption?’ (August 2002)
Suzette Heald, ‘Domesticating Leviathan: Sungusungu groups in Tanzania’ (September 2002)
Hugh Roberts, ‘Moral Economy or Moral Polity? The political anthropology of Algerian riots’ (October 2002)
Hugh Roberts, ‘From Segmentarity to Opacity: on Gellner and Bourdieu, or why Algerian politics have eluded
theoretical analysis and vice versa’ (December 2002) – Also available in French
Jonathan DiJohn, ‘Mineral-Resource Abundance and Violent Political Conflict: A critical assessment of the
rentier state model’ (December 2002)
Victoria Brittain, ‘Women in War and Crisis Zones: One key to Africa’s wars of under-development’
(December 2002)
Giovanni M. Carbone, ‘Emerging Pluralist Politics in Mozambique: the Frelimo-Renamo Party System’
(March 2003)
Hugh Roberts, ‘North African Islamism in the Blinding Light of 9-11’ (October 2003)
Jacklyn Cock, ‘Rethinking Militarism in Post-Apartheid South Africa’ (June 2004)
Debby Bonnin, ‘Understanding the Legacies of Political Violence: An Examination of Political Conflict in
Mpumalanga Township, Kwa-Zulu Natal, South Africa’ (June 2004)
Jo Beall, Sibongiseni Mkhize & Shahid Vawda, ‘Traditional Authority, Institutional Multiplicity and Political
Transition in KwaZulu-Natal, South Africa’ (July 2004)
Sarah Mosoetsa, ‘The Legacies of Apartheid and Implications of Economic Liberalisation: A Post-Apartheid
Township’ (July 2004)
Laurie Nathan, ‘The Absence of Common Values and Failure of Common Security in Southern Africa, 19922003’ (July 2004)
Andries Bezuidenhout, ‘Postcolonial Workplace Regimes in the Engineering Industry in South Africa,
Swaziland and Zimbabwe’ (November 2004)
Jo Beall, ‘Decentralisation and Engendering Democracy: Lessons from Local Government Reform in South
Africa’ (November 2004)
Laurie Nathan, ‘Security Communities and the Problem of Domestic Instability’ (November 2004)
E. A. Brett, ‘From Corporatism to Liberalisation in Zimbabwe: Economic Policy Regimes and Political Crisis
(1980-1997)’ (January 2005)
Jo Beall, ‘Exit, Voice and Tradition: Loyalty to Chieftainship and Democracy in Metropolitan Durban, South
Africa’ (January 2005)
Antonio Giustozzi, ‘The Ethnicisation of an Afghan Faction: Junbesh-i-Milli from the Origins to the
Presidential Elections (2004)’ (September 2005)
Andrew Fischer, ‘Close Encounters of an Inner-Asian Kind: Tibetan-Muslim Coexistence and Conflict in
Tibet Past and Present’ (September 2005)
Jenny Kuper, ‘Law as a Tool: The Challenge of HIV/AIDS in Uganda’ (September 2005)
Fabio Sánchez & Mario Chacón, ‘Conflict, State and Decentralisation: From Social Progress to an Armed
Dispute for Local Control, 1974-2004’ (October 2005) – Also available in Spanish
Dennis Rodgers, ‘Urban segregation from below: Drugs, consumption and primitive accumulation in
Managua, Nicaragua’ (October 2005)
Dennis Rodgers, ‘Subverting the Spaces of Invitation? Local Politics and Participatory Budgeting in Post-crisis
Buenos Aires’ (November 2005)
Giovanni Carbone, ‘“Populism” Visits Africa: The Case of Yoweri Museveni and No-party Democracy in
Uganda’ (December 2005)
Marcela Ceballos, with the collaboration of Iván Romero, ‘The Country Behind the Ballot Box: The Impact of
Political Reform in Colombia during a Humanitarian Crisis’ (December 2005) – Also available in Spanish
Neera Chandhoke, ‘Of Broken Social Contracts and Ethnic Violence: The Case of Kashmir’ (December 2005)
Jonathan DiJohn, ‘The Political Economy of Anti-Politics and Social Polarisation in Venezuela, 1998-2004’
(December 2005)

It is our intention for all Crisis States Working Papers eventually to be available in English, Spanish and French. Some
in the series have already been translated. For further details, and an up to date list of Working Papers, and other Crisis
States publications, please consult our website (www.crisisstates.com).

Fiche du document numéro 34450