Beyond De-Ethnicisation in Post-Genocide Rwanda
Deconstruction of Identity Discourse, Reconstruction of a Political Community
Andrea Purdeková
Doctoral Student
Department of International Development
Oxford University
Abstract. ‘Groupist’ and unidimensional construals of collective identityunderpin not only much ‘conflict action,’ but also conflict theorizing and
policy prescriptions for resolution. The present paper exposes the simplistic
assumptions that underlie the mainstream approaches to building unity in postconflict
situations (identity ‘abolitionist’ versus identity ‘accomodationist’)
and seeks to offer an alternative that would be built on a more nuanced
understanding. Two interconnected themes pertaining to Rwanda are explored.
First, I challenge the ‘bi-polarity thesis’ as the (still) dominant conflict
narrative. Second, I outline what I call the ‘extension fallacy’ in order to
explain how and why ‘false whole’ groups are constituted in our imagination
and in practice. Rwanda as any other country emerging from conflict requires
solutions that are more nuanced than those currently in the stockpile of
standard recipes. To be able to offer them, we need to deeply rethink
‘difference,’ ‘diversity,’ and ‘division’ first.
The present inquiry builds on my previous work, which studied de-ethnicisation as the
answer of the Rwandan government to the challenges of reconciliation and repatriation postgenocide
(Purdekova 2008). De-ethnnicisation’ as the new nation-building project in Rwanda aims
to replace ethnicity and other potentially ‘divisive’ sub-state loyalties with an undifferentiating
‘rwandanness.’ De-ethnicisation is thus composed of the narrative of ‘unity’ and ‘dissolubility of
difference,’ both rooted in a re-reading of the already much embattled history. The attempt is to
cultivate patriotism (“rukundo rw‘igihugu”) and “rwandanicity” (Rwandan values exemplified by
the concept of “ubupfura”), while suppressing the expression of other forms of identity.
The discursive frame of the de-ethnicizsation project is ‘inclusive’, setting itself in
opposition to previous nationalist myths aiming to ‘exclude’, but this merely obscures the top-down
dissemination and tightly policed boundaries of the discourse, the intolerance to dissent, as well as
the forced and merely surface-deep nature of “unity and reconciliation” activities, which, as my
fieldwork research suggests, are in fact openings to strengthened social control. Most importantly,
de-ethnicisation does not preclude the perseverance of identities or discrimination based on these,
and merely remains blind to the multiplicity of old and new difference and the transforming
divisions.
If de-ethnicisation presents a sub-optimal identity politics in Rwanda, what are the more
suitable alternatives? How should one create a sense of unity Rwanda while not eclipsing
pluralism? In order to develop such a model, a theoretical position regarding corporate or collective
identity needs to be devised first. Without it, an assessment of ‘difference,’ ‘diversity’ and
‘division’ in Rwanda and an appropriate policy of its management can hardly rest on solid ground.
The concept of “identity” is exceedingly intricate. Some reject it outright. According to
Malesevic (2007:21), identity is a ‘thin and unarticulated concept despite the astonishing variety of
uses to which it has been put.’ I rather follow Brubaker’s diagnostics: Identity is a semantically
1
overburdened concept, which needs to be disaggregated and treated as dynamic (Brubaker 2004).
Identity needs to be studied in depth, albeit perhaps selectively, if other questions which intimately
tie to it are to be answered. For the purposes of my project, a critical reconceptualization is essential
in order to i) understand what ‘unity’ in Rwanda can mean and in which way Rwanda is a ‘diverse’
and ‘divided’ society; ii) why and how current theoretical models are unsatisfying; and iii) to start
devising an alternative identity politics based on the critique.
The analysis shows that a number of aspects in fact join the discourses on identity in the
global north and south-- that the framework of “multiculturalism” (posed as a solution to diversity
in the north) and the framework “tribalism/ethno-nationalism” (posed as a problem of diversity in
the south) in fact share some common problems. Hence the reworking that results from my analysis
should be applicable outside of post-conflict societies. Further to this, by re-conceptualizing
identity as such, the paper is able to offer an integrated critique of seemingly “opposite” postconflict
identity politics, which I call “abolitionist” and “accomodationist,” as well as a critique of
the prevalent identity metanarrative of Rwandan society, which not only survives today but which
directly determines and constrains how we understand its conflicts and what we consider as
available “post-conflict” solutions.
1. Identity: Escaping the landscape of Multiple Reductions
Simplistic construals of collective identity or what Sen (2006) calls the ‘appalling
miniaturization of people’ permeate the sphere of political action and academic thought alike: From
the ‘clash of civilizations’ thesis to the world of battling ethnonationalist groups or the enclosed
individualist worlds of the liberal economic man, we often build our understanding and direct our
action based on one-dimensional and departmentalized models of human diversity. These facile
models constitute little-examined and often tacit elements of broader frameworks of theory and
action. Representations of this kind might, predictably, lead to simplistic and sub-optimal
prescriptions in the sphere of policy.
The particulars of identity reductionism in the non-western world are easy to make out. The
story of clashing tribal or ethnic groups has become a true metanarrative of conflict. The problem
with this portrayal reaches beyond the “ethnicization/tribalization” of social reality. At the core of
this specific conflict discourse lies an almost universal unquestioned essentialism, which views the
world as a set of clashing bounded groups, whether these groups are civilizations, ethnies, races,
nations, or else. The presumption might be as obscurantist as it is dangerous. The notion that the
world is wrought by group conflict ‘misrepresents the genesis of conflict and ignores the ability of
diverse people to coexist’ (Bowen 1996:3).
While a hidden essentialism defines the current metanarratives of conflict, the popular and
openly professed constructivism has a shallow base. Theorists of nationalism and ethnicity often try
to determine their position in regard to the priomordialist-constructivist continuum rather than to
answer fundamental questions regarding change and permanence of collective forms. How are
categories created and how and why are they propagated? Why do categories persist when
identifications do not and why at times they engender action when at others they do not? Why are
collective forms believed to be or portrayed by some at certain times as ever-present? In other
words, when and how do different collective identities come to matter?
To take an example, ‘Hutu’ and ‘Tutsi’ are ancient terms, extending as far back in time as
oral histories allow us to glance. Yet Hutu and Tutsi are ‘ancient words with changing meanings’
(Vansina 2001:2), diachronically (across time) evolving and synchronically (at a point in time)
2
polyvalent (Eltringham 2004:13). ‘The salience and meaning of Rwanda’s ethnic categories have
changed over time…European rule did not invent the terms…but the colonial intervention changed
what the categories meant and how they mattered’ (Strauss 2006:20). Colonialism ‘institutionalized
and stabilized categories that have been more fluid…and intensified the connection between race
and power (ibid:22).’ After the genocide, although these categories are banned from public
discourse, they survive as identifications in private, and they continue to structure day-to-day lives.
It can be said that the act of genocide and its memory have strengthened the boundaries and the
self-identification on either side of the divide even as ethnic categories have disappeared from
identity cards and official social and political engineering.
We can conclude that ‘Hutu’ and ‘Tutsi’ are everlasting categories with contingent
meanings. What it means to be a “Hutu” or a “Tutsi” varies not only across time but also across
people. As will be shown later on, identity in Rwanda is also much more complex than a
unidimensional Hutu-Tutsi duality suggests. The task ahead then is to offer a critique of both of
these two broad types of reductionism, the “vertical” reduction of the very nature of identity
reflected in the tendency to “groupism,” and the “horizontal” reduction of multiple and crosscutting
identities through representations of unidimensionality. The exposé and critique will serve
to de-construct identity politics, particularly as applied today to post-conflict states, and thereafter
to reconstruct it in a more nuanced manner.
i. Vertical Approach: Beyond “Groupism”
The unquestioned adoption of a ‘group’ as a basic building group of collective identity has
been recently critiqued by Rogers Brubaker (2002, 2004). He calls ‘groupism’ the tendency to see
groups as fundamental units of analysis, as internally homogeneous and externally bounded entities,
basic constituents of the social world that can be cast as actors (Brubaker 2004: 1-7). Groupism is
the tendency to ‘to represent the social and cultural world as a multichrome mosaic of monochrome
cultural blocks (2004:8).’ While people might have a tendency for ‘psychological essentialism,’
social scientists should not adopt categories of practice as categories of analysis. Their interest in
‘folk sociology,’ while constituting a good epistemology, is bad metaphysics (83).
‘Groupness,’ the nature of and degree to which something constitutes a group, is a variable,
not a constant (ibid:4). It is what we want to explain, not what we want to explain with. Shifting
attention from groups to groupness as variable and contingent, rather than fixed or given, ‘allows us
to account for phases of extraordinary cohesion without implicitly treating such high levels of
groupness as constant, enduring and definitionally present (2004:12).’ The degree of groupness
then determines a position on the continuum starting with pure ‘category’ to a ‘group.’1
Groupism is not only bad metaphysics, its adoption as analytical lens can lead to reification
of false wholes in name of which conflicts are often sanctioned by political entrepreneurs.
Reification is a ‘social process, not just an intellectual bad habit, and as analysts we should account
for the conditions under which this process…can work. But we should avoid unintentionally
doubling or reinforcing the reification of ethnic groups in ethno-political practice with a reification
of such groups in social analysis (Brubaker 2004:10).’ We can operate with ‘practical categories;’
we can study ‘race, nationality and ethnicity but without assuming that people have a nationality, a
race, an ethnicity’ (Brubaker).
1 Group is a mutually-interacting, mutually-recognizing, mutually-oriented, effectively communicating, bounded collectivity with a sense
of solidarity, corporate identity, and capacity for collective action (Brubaker 2004:12). ‘Category’, on the other hand, might and might
not refer to an existing group as defined above. It might just refer to people sharing a certain characteristic. In our analysis, we can relate
the two, asking about the degree of groupness associated with a particular category in a particular setting.
3
The dangers of scholarly reification are nowhere more apparent than in the history of
Rwanda: ‘There is a chilling traffic back and forth between the essentialist constructions of
historians, anthropologists and colonial administrators, and those of Tutsi and Hutu nationalists.
‘Types’ and ‘traits’ incautiously and sometimes fancifully ascribed by social scientists and
bureaucrats have often taken on a terrible social reality as the reification and essentialization of
cultural difference have been harnessed to deadly political visions (source).’ As will be shown later,
most analysts still operate on a highly simplistic and reifying vision of social reality in Rwanda --
the “bi-polarity” thesis-- an assumption that ‘Hutu’ and ‘Tutsi’ form the basic building blocks of
society and that division between them holds the greatest explanatory power vis-à-vis prospects for
reconciliation or potential future conflict.
Brubaker is right-- certain scholarly representations should be precluded because they
might feed into harmful political practice. But this is a premature ending to the discussion. Are
there not certain representations that should be promoted because they preclude such practice?
Fallacious groupist thinking animates political practice; violence often continues in the name of
‘false wholes.’ Is it possible to weaken these mobilizations by actively replacing the map identity
they rest on? If groupness can coagulate, it can subside too2 and if academics can unwittingly aid
the former, can they also purposefully aid the latter, by promoting nuance and complexity in our
understanding of identity and its manipulations? Much rests on the verdict we pass on
‘psychological essentialism:’ Is it really ‘natural’ for people to divide their world into neatly
separated groups, producing a particular ‘lay’ categorization? Or is such a presupposition merely
another patronizing and imposed essentialism, which downgrades people’s capacity for complexity
and change?
ii. Horizontal Approach: ‘Differentialism’ and ‘Similation’
Whereas the above could be considered a vertical cut into the questions of identity, getting
into the very analytical questions pertaining to any named collectivity, the present section offers a
horizontal and aggregate look at ‘diversity’ and ‘sameness’. The first part sketches a more nuanced
approach to diversity. It does not do this exhaustively because the focus lies on the method of
recognizing or engendering this plurality rather than on ‘naming’ it. The second part proceeds in the
opposite direction, by asking what are perhaps the wrinkles in a particular plane of sameness (e.g.
certain inequalities) that should be ironed out.
Sen suggests that ‘it makes a difference how we choose to see ourselves (Sen 2006:148)’
because ‘reductionism of high theory can make a major contribution, often inadvertently, to the
violence of low politics’ (2006:xiv). Yet reductionism pervades narratives of identity.
Unidimensionality, the tendency to neatly divide the world into different sets of groups whether
religious, civilizational or ethnic/national, thus needs to be not only recognized but superseded. But
how to eclipse unidimensionality?
Sen’s proclamation that we are ‘diversely different’ (2006) starts to problematize the
unidimensional models and highlights identity as multi-referential (Hale 2004). According to Sen
‘the hope of harmony in the contemporary world lies to a great extent in a clearer understanding of
the pluralities of human identity, and in the appreciation that they cut across each other and work
against a sharp separation along one hardened line of impenetrable division (2006:xiv).’ Benhabib
(2002:16) seems to agree: ‘Strong multiculturalists’ often overlook how ‘most individual identities
are defined through many collective affinities and many narratives.’
2 Naturally, this is not positive in all cases, oppressed people benefit from organized action.
4
Identity thus needs to be ‘complexified.’ Sen and Benhabib introduce a second-- a ‘crosscutting’--
dimension. An important third dimension should be added- the possibilities of identity
transformation. People switch and acquire affiliations, social action can engender new difference,
and as highlighted above, groups, far from being static unchanging entities, are defined by different
degrees of groupness. People are not completely free to guide identity change, but if politically,
economically and socially empowered, they have the potential to be its agents. Change is not only
what happens (i.e. Hutu and Tutsi come to signify different things at different periods due to social
and political developments in the Rwandan state). Change can be guided. The main concern then
becomes who and how should guide change.
In contemporary political theory, the question of agency over identity change resurfaces in
the debates on embeddedness. In the 1980s, ‘communitarian’ thinkers launched a critique of the
‘disembedded’ man of liberal theory (i.e. individualist, rational, largely undetermined by contexts).
They proposed a more deterministic vision of man formed by and attached to particular physical
and thus moral communities. While communitarianism ‘grounds’ the abstractness of liberalism, we
should avoid that vision of embeddedness which leads us to a groupist and over-culturized vision of
man. As Benahbib (2002:81) suggests ‘individuals are capable of cultural narration and
resignification… [and] through their actions [they] re-appropriate and transform their cultural
legacies.’ In response to communitarianism, Sen claims that ‘an enriching identity need not, in fact,
be obtained through discovering where we find ourselves, it can also be acquired and earned’ (Sen
2006:36).
This vision has important implications as regards policies of recognizing or managing
difference. Multiculturalism with its purposeful attention to plurality and difference has become
popular in recent decades. It has been accompanied by fundamentally differentialist concepts such
as ‘identity politics,’ ‘politics of recognition,’ and strengthened by the postmodern celebration of
the multiple, the fragmented, the fluid and the ‘mongrel’ in us (Said). According to Sen (2006:114),
however, multiculturalism has gained ground perhaps not as much as an important value, but rather
an ‘important slogan, since the underlying values are not altogether clear.’ Should we, for the sake
of cultural diversity, support cultural preservation or cultural freedom? Unfortunately, it is precisely
a mosaic preservationist multiculturalism (Benhabib 2002:8) that often comes out of state policies.
The multiculturalism of today is ‘plural monoculturalism,’ the development of many cultures in
separation (Sen 2006:157).
Clearly, a ‘differentialist’ or ‘multicultural’ turn in policy is insufficient. ‘The merit of
diversity depends precisely on how that diversity is brought about and sustained’ (Sen 2006:116).
‘Celebration of diversity’ is thus empty term because it is not a value in itself. Differentialism can
appeal either to traditionalism/preservationism or the freedom to take decisions and express oneself.
If freedom is the value of choice, then diversity is celebrated only to the extent that it results from
people’s choices. I opt for liberty because not only because it gives people the ability to either
follow or escape the past, to reinvent themselves, but also because it heeds to the fact, albeit
indirectly, that we are diversely different, have the right to a personal hierarchy of attachments, that
we change.
However, there needs to be a move away not only from the simplistic visions of what
difference is, how it arises and how it should be expressed; there also needs to be a sway away from
the placid differentialist politics accentuating recognition. Sen’s ‘diverse difference’ is a superior
metaphor of human identity, but it disregards horizontal inequalities, the instances when many
aspects of identity are not conveniently intermeshed but are ‘stacked up’ like layers of a cake (e.g.
those of specific religion are also those most disempowered). If identity politics is to conform to a
broader goal of social justice (however conceived), then not only all difference is not to be valued
5
but sameness/convergence is to be pursued at its expense. According to White (1991, in Harvey
1996) ‘at some point, one must have a way of arguing that not all manifestations of otherness
should be fostered; [that in fact] some ought to be constrained.’
The attempt to resurrect some general principles of social justice makes us navigate away
from both the universalisms of modernism with its marginalization of otherness, and the often
unquestioning postmodern tolerance of multiple and hybrid ethics. It makes us reject the
‘romancing of the geographical stone’ such as Waltzer’s ‘radically particularist’ theory in which
every substantial account of social justice is a local account. It signifies a turning away from the
communitarian emphasis on place, the ‘dwelling on dwelling’ posing as a ‘reconquest of human
immediacy from the tyranny of space’ (Meier 2007:73). ‘Place’ can license the xenophobic, the
authoritarian (Meier 2007). ‘Placed ethics’ builds on false wholes. It remains largely backwardlooking,
static and reificationist. It is blind to power relations producing oppression and
inequalities.
Further to this, liberal multiculturalism, Zizek argues, leads to ‘culturalization of politics’
whereby differences conditioned by political and economic inequality are naturalized/neutralized
into ‘cultural differences, different ‘ways of life’ which are something given, something that cannot
be overcome, merely tolerated.’ The flipside of this neutralization-through-culturalization effect,
which is relevant to Rwanda, is an uncritical acceptance of “sameness” in a situation where no
“cultural difference” manifests itself-- If there are no large cultural differences (don’t all Rwandans
speak the same language?), it appears there is no sanction for political “difference,” although there
have been powerful stratifying factors at work in form of mobilization using hate propaganda and
particular portrayals of history, actual categorical killing and resulting trauma, powerful exclusions
from politics or rising rural-urban inequality, among others. In context such as Rwanda, it is the
ethnicized and groupist discourse that directly moulds the way we perceive and thus approach
political claims making (Fraser; in Benhabib 2002:49). It attunes us to issues of cultural
difference/similarity, making them salient, and it eclipses issues of social equality.
To sump up, current multicultural models represent an insufficient ‘identity politics.’ First,
multiculturalism often follows groupist thinking, generating false wholes, which, from an
instrumental point of view, are more intelligible to the state and thus easier to control and govern.
Second, multiculturalism accentuates differentialism. But the ‘what’ and the ‘how’ of identity
politics should subsume under itself both differentialism and similation. We need both the
recognition and accommodation of our (not unchanging) differences and the search for
universalisms that are (i.e. cosmopolitan accentuation of human connection, sharing of a political
project) and should be (procedural political rights, social rights). Such identity politics builds on the
possibilities of transformation. Some aspects of such transformation lie in
reconceptualization/reframing, others require measured and concrete investments.
New Meanings for “Unity” and “Division”
At this time we might ask where does the confusing and ideologically-charged discourse of
‘unity’ and ‘division’ fit in all of this? What are the so-called ‘deeply divided’ societies; and are
there ‘deeply united’ ones? When does difference become division? Does the absence of diversity
mean unity? Can unity exist other than top-down nationalism? The Rwandan state both i) attempts
to create a unified official (his)story and suppress public manifestations of particular groupness; and
ii) opposes unity to difference (Purdekova 2008). Paradoxically, even with its top-down attempt at
nationalism/de-ethnicisation, it still operates according to a groupist logic, operating on categories
of ‘Hutu’ and ‘Tutsi’ and ‘Twa.’ Although it might seem paradoxical, the rise of nationalism after
6
the end of the Tito era in former Yugoslavia or after the break-up of the Soviet Union can be
explained not so much by the ‘survival’ of tangible and everpresent ‘nations,’ but rather by the very
perpetuation of national categories (although action on their behalf was forbidden) and the lack of
alternative civic identities (Ignatieff 1994, Brubaker 1996). In other words, by no means does the
present project stem out the possibility of conflict along ethnic lines, although perhaps due to other
than ethnic reasons.3
Difference does matter but only inasmuch as it matters to people, representing their
particular needs and/or values. Division is that kind of difference that matters from the viewpoint of
social justice and relates to marginalization, oppression, prejudice and other similar kinds of
perceived inequalities. This is the type of difference that matters from the viewpoint of conflict
emergence as well. The recognition that such special kind of diversity is more likely to lead to
conflict has been translated into models of ‘horizontal inequality’ (Stewart) or grievance models
(where inequality in access to resources or political representation matter). ‘Deeply divided
societies’ are societies defined by horizontal inequalities. ‘Deeply united’ societies might not only
be homogenous societies (if such exist), but ‘diversely different societies,’ where different kind of
identities comfortably intermesh. Diversity recognition and division recognition are separate issues,
and both are important. But this is not all. There are other ways in which a society can become
more united or less divided. Attacking inequalities is important but changing perceptions and
building connections is equally so. As I will show in the next section, categorical extermination or
simply perception of division follow a cognitive fallacy, an ‘extension fallacy,’ which itself feeds
off simplistic groupist models of human identity.
Again, what results is the need for transformation in our perception of difference and
transformation in the politics of its expression and accommodation. Both Brubaker and Sen urge us
to re-imagine our identities based on a more complex and accurate vision of them. Rather than
perceiving us and others as participants of separate groups, we should see each of us as diversely
different, our allegiances connecting us to others at different levels. Young (in Kymlicka 1995:165)
calls us to create politics that treats difference as variation and specificity, rather than categorical
otherness. Transformation needs to occur in the direction of similation too, both in terms of
building interconnections between people and reducing vital inequalities. Unity should be seen as
connectivity, rather than abrasion of difference. The goal is not ‘sameness’ but rather creation of
threads of interconnection. It can be conceptualized as sharing of particular goods (rights), and as
convergence in particular aspects of life (Brubaker’s ‘selective assimilation’4 thesis).
2. The Rwanda ‘Bi-polarity’ Thesis - Challenging a Dominant Narrative
Rwandan society and its cleavages have long been framed as an ethnic “bi-polarity.” This
approach contains in itself both the vertical and horizontal reductionisms that I have discussed
above. As I hope to show, Rwanda is not simply “a country of Hutus and Tutsis” as the dominant
metanarrative of this particular conflict suggests. The diversity and the dividing and connecting
lines have always been more complex, and they have been changing over time. They are perhaps
even more complex today after the genocide and the vast post-genocide returns from multiple
countries of exile. In what follows, I want to demonstrate this complex diversity and division in
3 Conflict is often sanctioned along ethnic lines and happens along ethnic lines but this does not mean it happens due to ‘ethnic reasons’
such as a political or economic grievance of a particular ethnic group; it might happen due to human reasons, such as to survive or to get
ahead.
4 Brubaker calls for a reformed, ‘selective’ concept of assimilation. ‘As a normatively charged concept, assimilation is opposed not to
difference but to segregation, ghettoization and marginalization (Brubaker 130).’ Certain aspects of sharing are highlighted as beneficial
by the new assimilation literature (e.g. speaking of English while not loosing the language of origin; convergence of incomes).
7
order to undermine the dominant bi-polar vision of Rwandan society. My aim is not to enumerate
all differences, to exhaust the whole variety, as this is both impossible and fruitless. After all,
‘capturing’ diversity or division is always already stifling to that which is a dynamic and multidimensional
process. I merely want to probe the simplistic identity maps presently used to draw
vast conclusions.
The analysis does not seek in any way to deny that the genocide happened across clearly
marked ‘ethnic lines’ (although moderate Hutus and Twas were also targeted). But a bi-polar view
simply cannot fully explain events and social processes in Rwanda, including the genocide. Other
identities, besides simply ethnic ones, have to be taken into account. Furthermore, the fact that the
genocide operated along ethnic lines does not suggest existence of two homogeneous and separate
groups with unwavering levels of groupness over time. The genocide was contemplated and
ordered by a minority of Hutu, only after they managed to wrestle power from the opposition, it
operated along a categorical logic and rhetoric which overshadowed the diverse difference of
Rwanda, and represented a temporal rise in groupness along categorical lines which had to do more
with fear (resulting from war and information asymmetry) than ethnic difference (Straus 2006).
What describes a conflict (conflict happening along more or less clearly demarcated lines such as
religious lines in Sudan or ethnic lines in Rwanda) and what in fact causes it – at a broad structural
level as well as at the level of individual participation— are two different things.
Today as in the past, Hutu were separated from other Hutu, and Tutsi from other Tutsi by
regional, lineage, clan and clientelistic affiliations, among others, which developed previous to
class and ‘ethnic’ distinctions (Newbury 2001, Vansina 2004). Yet these diverse distinctions have
been little studied. According to David Newbury (2001:263), this has been largely a result of the
focus on dynastic integrity and colonial boundaries, rather than on the process of incorporation of
diverse cultural areas by the expanding Rwandan state. This has consequently ‘blinded outside
observers to both the broader continuities and internal diversity [of Rwanda and Burundi]…and
compromise[d] our understanding of their histories and identities (ibid:263).’
Within three broad geographical and cultural areas (eastern grasslands, western highlands
and Tanganyika shoreline), David Newbury (2001:263) identifies at least nine ‘smaller regions
[which] developed their own outside fields of action and forged individual identities in opposition
to each other as well as to the dynastic core.’ Some areas, especially in the northwest and northeast
of present-day Rwanda, ‘were conquered only in the past century and even then imperfectly
controlled’ (Prunier 1995:21). Newbury’s case study of Kyniaga reveals that pre-conquest areas
falling outside the purview of the Rwandan state operated primarily along clan lines. While Tutsi
was likely an ‘ethnic’ identity at the time, Hutu was not (Mamdani 2001:101; Vansina 2001:3).
Hutu started as a transethnic, political identity of subjects, becoming relevant through the forced
incorporation of a marginal area into the Rwandan state.
The dynamic historical process of expansion of the Nyiginya monarchy, originating in the
Nduga area (roughly in the center of today’s Rwanda), shows that at the time before colonization ‘it
was a centre versus periphery affair and not one of Tutsi versus Hutu [emphasis added]. If the
king’s chiefly agents in this process were mostly (but not all) Tutsi, their ‘victims’ in the newly
controlled situation were both Tutsi and Hutu, and they were defined by their geographical location
(Prunier 1995:21).’ Hence the center-periphery divide was one of center-Tutsi versus all those on
the particular regional periphery who did not coopt (both Tutsi and non-Tutsi) and by cooptation
benefit from the expansionary state. Besides regional identities and the center-periphery divide,
divides existed within the center itself among lineages and these often engendered violence (Prunier
1995:39).
8
Regionalism and lineage politics played an important role also in the run-up to
independence, under the two Republics, even during the genocide. On the verge of independence,
‘that political mobilization of the Hutu masses was infinitely more arduous in the core area of the
kingdom than in the peripheral marches’ has been argued by Felicien Kamashabi and Jean Rumiya
in their analysis of peasant protest (or rather lack of protest) in the Bufundu region, which they refer
to as ‘une region non prepare a la revolution’ (Lemarchand 1990:3). Furthermore, ‘the amazing
proliferation of Hutu-led parties other than the Parti de l’Emancipation du Peuple Hutu
(Parmehutu) at the time of the 1960 elections seems to suggest that regional and clan ties (focused
on the so-called bakuru bimiryango) provided at least as important a basis for social cohesion as
ethnicity’ (ibid).
Many scholars have noted that discrimination under the two republics has taken the form of
regional favoritism. Under the 1st Republic, Kayibanda’s rule benefited disproportionately Hutus
from his home region of Gitarama in the south-central region. Under the 2nd Republic,
Habyarimana’s rule benefited the northerners, especially those from the northwest. As Newbury
and Newbury (1999:299) suggest ‘the main axis of conflict [under the 2nd Republic] was based on
region and class, and for most part these were conflicts between diverse Hutu factions—again a
reminder of the internal heterogeneity within these broad ‘ethnic’ categories.’
‘Within each ‘regional mafia’’, however, there would further ‘be a tendency to create subunits
according to precise, more narrowly defined geographical origin (Prunier 1995:85).’ Under
the 2nd Republic, it was the members of the President’s wife’s family and close associates who
benefited because the President himself did not come from a ‘respectable’ family (Prunier
1995:85). Hence the name ‘le clan de Madame’ and later ‘akazu’-- the little house-- a name
historically given to the inner circle of a King’s court. By 1989, when coffee and tin prices started
falling, the dividing lines within the northern support group became apparent. ‘The elite started
tearing each other apart to get at the shrinking spoils, Abakiga against Abanyanduga, then among
the victorious northerners Abashiru against Abagoyi, and within the top Abashiru people between
the various affinity groups or families (Prunier 1997:350).’
Regional identities, especially the recent experience of conquest and incorporation into the
Rwandan state ruled by a Tutsi elite seems to account for different degrees of violence in Tutsi
purges of 1959 and the genocide in 1994. Prunier (1995: 51) mentions that in 1959 ‘the harshest
violence was in the north-west, where the Hutu principalities had made their last stand against the
Belgian-Tutsi forces in the 1920s and where hatred of the Banyanduga ran high. Although it was
the area with the smallest Tutsi population, it was where the Tutsi were most relentlessly hunted
down.’ In regard to the 1994 genocide, it was again northwest Rwanda that was considered the
‘breeding ground’ of ethnic extremism (US Committee for Refugees 1998:19).
There are other divisions that directly mattered as regards inter-ethnic violence and that
undermine the picture of two homogeneous groups of antagonists. Allegiance to different parties,
different sections of parties (moderates versus extremists; the Hutu Pawa circle), and hence
different ideologies, even class difference (les petits Tutsi versus the urban-dwelling elite) also
mattered. Not all Hutu political elites subscribed to extremist ideology. In fact, these were only a
small sub-section. It was the elimination of opposition, and consolidation of power not only at
central level but in regional centers that was determinant of the onset of mass killing in 1994
(Straus 2006). The powerful show of groupness on behalf of the Hutu during the genocide could
happen only at the time when the political and ideological diversity was eliminated, silenced or
side-tracked. Rather than a conflict of Hutus and Tutsis, the genocide followed a victory of a
particular ideology on behalf of a sub-group in Rwanda desperately trying to remain in power.
9
Obscuring this process means misunderstanding and simplifying the history of Rwanda, and
eventually surrendering oneself to a little too fatalistic image of its future.
After the genocide, Rwanda is even more diverse, but more importantly, differently divided
than before. The pure disruptive force of genocide, the diverse experiences of it, the mass returns
from various countries of exile, and the changed power structure all point to new constituencies and
new inequalities that have to be taken into account. The bi-polarity thesis never fully explained the
events in Rwanda, and if employed again, might miss perhaps the most important divider today.
Rwanda’s diversity today is reflected through the various experiences of genocide not only
as victim and participant, but also bystander, absentee or saviour (Buckley-Zistel 2006b:1). None of
these completely fit ethnic categories. The genocide hit moderate Hutus, Twas, as well as Hutus
who looked like Tutsis (Prunier 1995:249). In fact, ‘the first victims were political opponents of the
Habyarimana regime; many were Hutu. Furthermore, significant numbers of Hutu resisted the
genocide, often by hiding and protecting those at risk’ (Newbury&Newbury 1999:295). Muslims,
concentrated in Kigali’s Nyamirambo quarter, disproportionately refused to kill Tutsis and appealed
to the universalistic nature of their religion, while the same cannot be said of the Catholic Church. It
is for this reason that both Muslim faith is the fastest growing in Rwanda, and that there has been a
proliferation of new churches.
The post-genocide returns further complicate the social landscape (Berman 2004:4) in ways
that suggest tensions but are far from known and ‘do not necessarily cut along the familiar divides’
(Minow 2003:40). Since 1994, about 2.1-3.4 million refugees have been repatriated or resettled, a
‘record in world history’ (Africa Church Information Review, 18 March 2003). The result is that
25-40% of the present population5 is constituted by post-genocide returnees, coming back from
Uganda, Burundi, Tanzania, DRC, and many other countries in Africa and beyond. Most returnees
‘had lived their entire lives abroad, never having seen their mother country until return’ (Pitsch-
Santiago 2003:7). Despite this, it is returnees, especially Ugandan returnees, who are perceived as
those who hold political and economic power in today’s Rwanda (to be explored below).
Tutsi is hardly a unified group today. A returnee explains: ‘I feel close to other people like
me, people who have returned after the war. Somehow I feel free with them, not like people who
lived here before and during the war. With these people, it is more difficult. They are more cautious
and closed, I don’t know where they stand’ (Pitsch-Santiago 2003:7). Other returnees said, ‘I do not
feel close to survivors because they misinterpret us being the root cause of genocide; though I do
sympathize with them’ and ‘rapport does not exist [between survivors and returnees] because we do
not know each other well enough; we do not have the same habits (ibid).’ Besides cultural barriers,
Pitsch-Santiago cites language barriers as well. ‘In the past, most Rwandans spoke Kyniarwanda
which is no longer true. Many returnees did not speak Kyniarwanda as their first language when
they returned, and Ugandan returnees are not fluent in French (ibid).’ The government has recently
changed all instruction in schools from French to English. Hence tensions today exist not only
along returnee-non-returnee lines, among different returnee groups, but more largely among the
“Francophones” and “Anglophones” as the latter are perceived as being privileged by the current
system.
Naturally, new diversity that defines the Tutsi as well as Hutu community does not
automatically imply a diversity of division. As Straus (2006:130) notes, ‘awareness of difference
alone does not cause strife.’ But is there more than awareness? Are there fundamental new
inequalities between certain categories of people or grievances that might represent or lead to new
5 In some areas, a third of the population is new and in two prefectures half of the inhabitants arrived recently (Delame 2004:4).
10
divisions? There are indeed signs that new divisions of this kind are emerging. Power is perceived
to be concentrated among the ‘Tutsi’ or the ‘Ugandan Tutsi.’ As a corollary, tensions are fuelled by
‘the feeling among the Burundian and Congolese diasporas that they are marginalized from power’
(ICG 2002:18).’ The process of reconciliation does not include Hutu victims of RPF crimes and
some claim there is collective blame being placed on the Hutu. Even more worrying is the growing
income inequality between town and rural areas, documented by An Ansoms (2009).
Despite these alarming developments, no academic study offers a satisfactory theory of
division in post-genocide Rwanda, one that would translate the observed ‘complex orders of
difference’ (Berman 1998) into predictions regarding conflict-prone cleavage. Could this be a
consequence of the unwavering hold that the ‘bi-polarity’ thesis has in any analysis of Rwanda?
Although it is not my aim to build a more complex theory, in the next section I will try and analyze
what I believe is one important factor causing division today and will discuss how that division
might be diffused through a better politics, including better identity politics.
3. Rwanda and the ‘Extension Fallacy’ – An Outcome of Groupism, An Ingredient in
Reification
Identity reductionisms not only misrepresent societies and prevent our understanding of
them, they also contribute to what I call ‘extension fallacy,’ a conceptual operation that has a real
impact on the ground, as the case of Rwanda will again show. The ‘extension fallacy’ is a term I
use to refer to the purposeful or unwitting extension from a sub-group in a category onto the whole
of the category. It can be both i) the simplifying cognitive tendency to attribute actions of a subcategory
to a greater category; and ii) a purposeful extension fulfilling specific aims, in which case
it is hardly a fallacy for those who employ it. Extension of this kind always represents a logical
mistake of ‘pars pro toto’ whereby part is substituted for a whole. The first category mistake is
committed by scholars/observers and public figures alike and can be unwitting, the second is
committed purposefully in order to further political agendas. In both cases it obscures the real
cleavages in society and reifies false wholes.
What do we extend towards? This depends on i) the dominant construal of diversity; and ii)
the dominant taxonomy or ‘metanarrative of division’. As mentioned above, diversity is often
construed as unidimensional; a mosaic of fastly separated homogeneous groups (e.g. religious or
ethnic). Such groups are easily ‘fillable,’ a part-for-whole mistake rests on equally mistaken
assumptions regarding identity (‘groupism’). Furthermore, if groups in a particular society are often
portrayed as divided along ethnic lines, then the extension fallacy would lead from a sub-group
within an ethnic group or a sub-group that cross-cuts ethnics onto an ethnic group. In fact, the
metanarrative of division along ethno-national lines is so strong in journalistic but also some
scholarly accounts that it often acts as a quick heuristic explanatory device even in cases, such as in
Rwanda post-genocide or present-day Kenya, where the dynamics are more complex. In what
follows, I will demonstrate both a purposeful and an unwitting extension fallacy and their
repercussions. The first example comes from the time of the genocide. This is an extreme case,
which led to literally ‘categorical’ extermination, but it is no more than an example of a broader
phenomenon of an extension fallacy, when all stand in for actions of a few (Straus, 2006:199). The
second example comes from scholarly accounts post-genocide, which try to force-fit the old Tutsi-
Hutu divide to explain new realities.
During the genocide, the extension was purposeful and it led from the group of invading
inyenzi to all Tutsi as enemy. According to Straus, ‘the main rationale that perpetrators consistently
11
gave for genocidal violence is the following: the RPF killed President Habyarimana, RPF soldiers
had invaded to kill Hutus; all Tutsis were RPF supporters or potential supporters; ergo, Hutus had
to kill Tutsis to prevent being attacked by them’ (Straus 2006:153). Straus calls this particular
extension fallacy a ‘collective ethnic categorization’ (2006:173) Brubaker would highlight those
cognitive processes, which at a specific moment transforms diversely different people into singlelayered
individuals (tutsi-non-tutsi= enemy-non/enemy), as important ingredient in the rise in
groupness (the enemy has been defined) and observes that ‘over and over again, perpetrators speak
of the Tutsis as a unit, as ‘one.’ Over and over again, Tutsis stop being neighbours, friends and a
disparate population: they become a single entity with identical—and permanent—intentions
(Straus 2006:173).’
But is it really all the perpetrators who ‘committed’ an extension fallacy? Perhaps we
would commit an extension fallacy of our own if we claimed so. The actions of most Hutu seemed
to be animated by other factors and informed from above. All Tutsis were targeted because killing
Tutsi was understood as an act of war, relayed as a preemptive action to protect oneself in a state of
generalized insecurity. Evidence gathered by Straus suggests that ‘acute insecurity and orders from
above ignited a categorical logic of race and ethnicity (2006: 173).’ It is more accurate to say that
the common executioners ‘carried out’ the extension fallacy.
The extension fallacy was a purposeful and desperate bid engineered by the extremist
sections of the government willing to remain in power . They thought that if they could not prevail
over the invading force militarily, they could do it politically, by destroying the whole Tutsi ‘power
base.’ Even if the inyenzi forced their way into a shared government, the logic went, they would
have no one to represent. In July 1994, facing military defeat, the extremists pushed this logic even
further. Even if the inyenzi won, they would have noone to govern over. Those Hutu who were not
fleeing on their own, were forced to move and prevented from returning by the government on-theway-
to-exile (Prunier 1995). The Tutsi returnees might gain their home-land, but the population
would meanwhile settle across the east border in the DRC provinces of North and South Kivu
terrorized by remnants of the old government and its militias. Extension fallacies feed off or try to
separate out, and in turn reify, political constituencies. Extension fallacy has always been a useful
political gambit.
Today, the extension leads from concentration of power in a network of returning refugees
to all Tutsi. It is done perhaps unwittingly, and often by observers of the situation in Rwanda. I
came across this particular extension fallacy while writing and researching on de-ethnicisation as a
‘minority vision’ of a Rwandan future, originating from and promoted by the authoritarian
government whose kernel hails from exile. What caught my attention were the divergent
attributions regarding the present quite-undemocratic concentration of power. While some viewed it
as a Tutsi minority rule, other saw it as the rule of the Ugandan returnee group. I found Rwanda’s
political system described as a ‘Tutsi monopoly of military and political power’ (Buckley-Zistel
2006:11; Reyntjens 2004:187), representing the ‘Tutsification of power’ and the nominal Hutu
presence (Mamdani 2001:271), the ‘manipulation of ethnicity through ‘ethnic amnesia,’ resulting in
discrimination against the Hutu’ (Reyntjens 2004:38), ‘ethnocracy but with just enough powersharing
at the top to enlist a measure of Hutu collaboration’ (Lemarchand 1997:10).
Yet as much as the elite in power are predominantly ‘Tutsis,’ they are also predominantly
‘returnees’ and, more specifically, predominantly ‘Ugandan returnees’ supporting the RPF (which
itself is fragmented). Today, there is no homogeneous ‘Tutsi’ category. Tensions exist among Tutsi
12
returnees and survivors (Mgbako 2005) and among the Tutsi returnee community as well6 (HRW
2000). As Oomen (2006:19) writes ‘military, bureaucratic and political power is consolidated in the
hands of a minority of Tutsi linked to the RPF, of whom most returned to Rwanda in 1994 after
having been exiled in Uganda for decades [emphasis added]’. Reed (1996:498) further adds that
‘while the RPF maintained virtually no civilian administrative structures [in exile], nearly all the
skilled manpower now available in Rwanda has come from the exiled community which gave birth
to the movement.’ Pitsch-Santiago (2003) and Zorbas (2004) also relay the perception among the
rest of the population that a sub-set of the Ugandan returnees are more politically connected and
have more opportunities.
Neither ‘Tutsi ethnocracy’ nor ‘Ugandan Tutsi domination’ is a completely correct
rendering of what is problematic with the ‘RPF-ruled,’ effectively one-party state. Yet both are
dangerously reificationist. Using such categorical attributes might later form the contours of
conflict. The way the division is described is not only groupist, but it builds on the Hutu-Tutsi
binary with which politics in Rwanda has been described for past 50 years. The discussion is
wrongly focused altogether. It is not a group or a sub-set of Rwandan population that is a problem
(this always risks an extension fallacy), but rather the political system that has come to place, which
favors a particular vision of the future of the country, a diasporan nationalism repatriated. All that
should remain of the discussion is that there is a monopoly of power in place, that there is a
minority vision of the future pursued, and that, hence, the political process should change. The
important lesson to draw here is that social markers can lose their salience if people feel that they
can access resources including power on an equal footing, or at least have the same chances to do
so. In other words, the solution to ‘authoritarianism’ might also connote the prevention of loading
of a certain social category with discriminatory meaning.
4. Conclusion: Alternative Approaches to Post-Conflict Political Unity
There are generally two mainstream and opposed approaches to post-conflict ‘unity’—the
abolitionist attempt (destroy/deny/delete difference, insert unity) and the accommodationst attempt
(recognize difference and balance it). De-ethnicisation belongs to the first camp. Consociational
democracy or power sharing belong to the other. Yet these seemingly ‘opposed’ approaches share
much in common. Both are top-down structures erected on a simplistic and groupist vision of
identity. Through the lens of their theory, societies are amalgams of separate groups. The two
theories differ only as to whether and how these groups can be modified. While consociationalism
proposes to recognize the set of groups that ‘already exist’ (through either pre-determination or
self-determination), de-ethnicisation proposes that these separate groups can be merged into one.
Neither theory is capable of challenging the dominant narrative of bi-polarity in Rwanda. Powersharing
would consider Hutus and Tutsis as the building blocks of a new political deal, deethnicisation
does not preclude perceived inequality in access to power narrated along this bi-polar
spectrum. Operating along a fundamentally groupist logic, none of the two approaches undermines
the extension fallacy. Both are fundamentally disconnected from the bottom-up response.
In today’s Rwanda, power-sharing is not the optimal scenario. As Mamdani suggests,
Bahutu and Batutsi should not be reproduced as dualities, which is the case in power-sharing, but
rather transcended (Mamdani 1996:30). Mamdani clearly operates on a simplistic and archaistic
notion of a ‘bi-polar’ society (the cleavage might not be between Hutu and Tutsi) but his general
conjecture holds. ‘If people of different identities can communicate only through the recognition of
6 E.g. ‘Tutsi who returned from exile in Burundi or Congo have found their hopes for rapid success blocked
by the predominance of those who returned from Uganda’ (HRW 2000).
13
their differences…how [can they] ever foster supra-ethic or universalist attitudes’? (Kebede
2001:280). The transcendence that Mamdani talks about though cannot be done by negation,
suppression and even criminalization of particular expressions of groupness, even if this sanction
goes hand in hand with broader “national” identity engineering, as is the case today. We can both
build unity (universalist attitudes, interconnection), and attend to difference without instituting it as
division/cleavage. The only question is how this should be done and how should it not.
Some basic contours of an alternative political system and identity politics should be
apparent by now. The political outline should follow deliberative democracy. The expression of
plurality should be in the hands of those who constitute a polity, although plurality (and thus liberty
of expression) should be minimally eclipsed from above to exclude extremist ideology. Such
approach fits within the wider waning of the ‘differentialist’ turn in social thought, public discourse
and public policy (Brubaker 2004:117) and builds on the resurgent interest in forms of civic
commonality. This is where the new identity politics of both difference and similation fits in.
Based on all considered above, in fact a three-layer approach emerges. First layer
represents the ‘proceduralism of plurality.’ Rather than building a political system on enumerated
diversity (such as in certain power-sharing systems in post-conflict contexts), we should build a
political process whereby difference/diversity can express itself (almost) freely and continually. It is
Young (in Kymlicka 1995: 271) who calls for an ‘openness to unassimilated otherness’ which is
‘perpetually being constructed and deconstructed out of the flows and shifts of social life.’ Only a
democratic system allows for such a continual negotiation. In such a model, the main goal is not
preservation but control over expression.
The second level represents the project of ‘common denominator’ creation. This attempt
does not strive for homogeneous expression but rather connectivity and equal respect for certain
givens. Such connectivity is to a certain extent achieved through the procedure of democratic
participation, but this is not sufficient (Snyder and Ballentine). In order for the new identity politics
to take root, certain things have to be equally shared among all-- respect for diversity, complex
visions of identity, dedication to dialogue and procedural rights, common building of a vision of
social justice. The third level involves ‘radical leveling’ or similation. The attempt is to recognize
the warping of space of difference by political economy and to address the resultant inequalities.
Altogether, the above three-layer approach seeks to overcome the simplistic attempts of
accommodation versus abolition, or assimilation versus separation (depending on which vocabulary
we choose). The model seeks to overcome the passive acceptance of as-is states; the groupist and
unidimensional visions of diversity on which policies of separation or accommodation operate. It
also shows that unity and diversity can coexist, which runs against the visions of nationalists who
‘attempt to create forced unity out of diversity, coherence out of inconsistencies, and homogeneity
out of narrative dissonance (Benhabib 2002:8).’ At the same time as we allow the free play of
(more complexly conceptualized) difference, and build unity as proceduralist connection and
cultivation of basic denominators, we have to understand certain inequalities as difference that
should not be perpetuated.
On a careful concluding note, the fact that a sort of deliberative democracy described above
presents a superior political system, which can accommodate a more nuanced view of human
identity, and new identity politics derived from it, does not mean that it will. Most of the world
seems to be cued to simplistic dominant narratives most of the time and to behave accordingly. The
renderings of conflict, and the repertoire of post-conflict solutions seem to hopelessly revolve
around groups as static and bounded beings. Many political futures seem to be fundamentally
eclipsed by the accepted narratives of their past. Yes, it is true that freedom and plurality of speech
14
and remembrance ‘open the domain of the possible to what is not yet thought or said’ (Arendt, in
Saunders 2007:2). The real challenge though does not lie in bringing freedom and plurality. The
real challenge is to make the freely spoken and remembered something else than divergent
positions on an accepted or acceptable, and many times distorted, narrative of the past.
Wordcount: 8,314
15
REFERENCES
BENHABIB, S. (2002) The Claims of Culture: Equality and Diversity in the Global Era,
Princeton, Princeton University Press
BERMAN, B. (1998) ‘Ethnicity, Patronage and the African State: The Politics of
Uncivil Nationalism,’ African Affairs, July 1998
BESTEMAN, C. (1999) Unraveling Somalia: Race, Violence and the Legacy of Slavery,
Philadelphia, University of Pennsylvania Press
BOWEN, J. R. (1996) ‘The Myth of Global Ethnic Conflict,’ Journal of Democracy 7(4)
BRUBAKER, R. (2004) Ethnicity Without Groups, Cambridge, Harvard University Press
-------------. (2003) ‘Beyond Comparativism,’ UCLA, Available from URL
http://repositories.cdlib.org/cgi/viewcontent.cgi?article=1008&context=uclasoc
------------. (2000). "Beyond 'Identity'." Theory and Society 29, 1–47.
-------------. (1996) Nationalism Reframed: Nationhood and the National Question in the New
Europe, Cambridge, Cambridge University Press
BRUBAKER, R., FEISCHMIDT, M., FOX, J. and GRANCEA, L. (2006) Nationalist Politics
and Everyday Ethnicity in a Transylvanian Town, Princeton University Press
BUCKLEY-ZISTEL, S. (2006a) ‘Remembering to Forget: Chosen Amnesia as a Strategy for
Local Coexistence in Post-Genocide Rwanda’, Africa: Journal of the International
African Institute 76(2): 131-150
--------------. (2006b) ‘Dividing and Uniting: The Use of Citizenship Discourses in Conflict and
Reconciliation in Rwanda’, Global Society 20(1): 101-113
CALHOUN, C. (2007) ‘Social Solidarity? A Problem for Cosmopolitan Democracy,’ In Seyla
Benhabib, Ian Shapiro and Danilo Petranovic (eds). Identities, Affiliations and
Allegiances, Cambridge, Cambridge University Press, 285-302.
CASTELLS, M. (1997) The Power of Identity, Oxford, Blackwell
d'Hertefelt, M. (1971) Les clans du Rwanda ancien; elements d'ethnosociologie et
d'ethnohistoire. Reprint of the 1971 ed., published by the Musee royal de l'Afrique
centrale, Tervuren, in its Annales, Serie in -8o, Sciences humaines, no 70. ed, (Institut
national de recherche scientifique. Publication no 7). Butare: Institut national de
recherche scientifique.
d'Hertefelt, M. , A. Trouwborst, and J. Scherer. (1962) Les Ancience Royaumes de la Zone
Interlacustrine Meridionale (Rwanda, Burundi, Buha). Edited by D. Forede. Vol. East
Central Africa Part XIV, Ethnographic Survey of Africa. London: Inernational African
Institute.
16
ELTRINGHAM, N. (2004) Accounting for Horror: Post-Genocide Debates in Rwanda, London,
Pluto Press
ERIKSEN, T. H. (1993) Ethnicity and Nationalism, London, Pluto Press
FOX, J. E. and IDRISS, C. M. (2007) ‘Everyday Nationhood,’ Paper prepared for ‘Everyday Life
in World Politics and Economics’, Centre for International Studies, London School of
Economics, May 2007
HARVEY, D. (1996) Justice, Nature and the Geography of Difference, Oxford, Blackwell
HOROWITZ, D. L. (1985) Ethnic Groups in Conflict, Berkeley, University of California Press
IGNATIEFF, M. (1994) Blood and Belonging: Journeys into the New Nationalism, London,
Vintage
INTERNATIONAL CRISIS GROUP (ICG) (2002) ‘Rwanda at the End of the Transition: A
Necessary Political Liberalisation’, Available from
<http://www.grandslacs.net/doc/2555.pdf> (accessed 23 February 2007).
LIJPHART, A. (1995) ‘Self-Determination versus Pre-determination of Ethnic Minorities in
Power-Sharing Systems,’ In Will Kymlicka (Ed) The Rights of Minority Cultures, 275-287
-------------. (1977) Democracy in Plural Societies: A Comparative Exploration, New Haven, Yale
University Press
MAIER, C. S. (2007) ‘ ‘Being There’: Place, Territory and Identity,’ In Seyla Benhabib, Ian
Shapiro and Danilo Petranovic (eds). Identities, Affiliations and Allegiances, Cambridge,
Cambridge University Press, 67-84
MALESEVIC, S. (2007) Identity as Ideology: Understanding Ethnicity and Nationalism, New
York, Palgrave Macmillan
MAMDANI, M. (2001) When Victims Become Killers: Colonialism, Nativism, and the Genocide in
Rwanda, Princeton, Princeton University Press
MINOW, M. and CHAYES, A. (eds.) (2003) Imagine Coexistence: Restoring Humanity After
Violent Ethnic Conflict, Hoboken, Jossey-Bass
NEWBURY, D. (2001) ‘Precolonial Burundi and Rwanda: Local Loyalties, Regional Royalties’,
The International Journal of African Historical Studies 34 (2): 255-314
KAKWENZIRE, J. AND D. KAMUKAMA. (1999) ‘The Development and Consolidation of
Extremist Forces in Rwanda,’ in Suhrke, A. and Adelman, H., The Path of a Genocide.
The Rwanda Crisis from Uganda to Zaire, New Jersey, Transaction Publishers, 61-91
PITSCH SANTIAGO, A. (2003) ‘Returning to their Homeland: Rwandan Old Caseload Refugees
Motivated to Rebuild the Country they Never Knew’, University of Maryland Center for
International Development and Conflict (CIDC) Management Papers. Available from
<http://www.cidcm.umd.edu/ICT/papers/diaspora_paper_cahier.pdf> (accessed 15
February 2007).
PRUNIER, G. (1995) The Rwanda Crisis: History of a Genocide, London, C. Hurst & Co
PURDEKOVA, A. (2008) ‘Repatriation and Reconciliation in Divided Societies: The Case of
Rwanda’s Ingando,’ Oxford University Refugee Studies Centre Working Paper 43,
January 2008
SEN, A. (2006) Identity and Violence: The Illusion of Destiny, New York, W.W. Norton
17
SNYDER, J. and BALLENTINE, K. (1996) ‘Nationalism and the Marketplace of Ideas’,
International Security 21(2): 5-40
STRAUS, S. (2006) The Order of Genocide, Ithaca, Cornell University
U.S. COMMITTEE FOR REFUGEES (USCR) (1998) ‘Life After Death: Suspicion and
Reintegration in Post-Genocide Rwanda’, USCR February 1998
VANSINA, J. (2004) Antecedents to Modern Rwanda. The Nyiginya Kingdom, Madison, The
University of Wisconsin Press
------------. (2001) ‘History Facing the Present: An Interview with Jan Vansina’, with Karel Arnaut
and Hein Vanhee, H-Africa 1 November 2001. Available from http://www.hnet.
msu.edu/~africa/africaforum/VansinaInterview.htm
VERWIMP, P. (2000) ‘Peasant Ideology and Genocide in Rwanda Under Habyarimana,’
Journal of Genocide Research, November 2000
YOUNG, I. M. (1995) ‘Together in Difference: Transforming the Logic of Group Political
Conflict,’ In Kymlicka, W. (Ed.) The Rights of Minority Cultures, Oxford, Oxford
University Press
ZIZEK, S. (2007) ‘Tolerance as an Ideological Category,’ forthcoming in Critical Inquiry
