INTRODUCTION
Early in 2001, voters in a poor, rural region of southern Ethiopia were presented with a unique referendum on ethnicity. The Siltie, previously considered a sub-clan of the Gurage ethnic group, were asked quite simply, ‘Are the Siltie Gurage or not?’ Their answer was overwhelmingly that the Siltie were not Gurage. This unique exercise in voting for an ethnic identity is one of the most dramatic modern experiments in injecting direct political competition into what has traditionally been regarded as a social or cultural matter. It followed years of campaigning, both for and against Siltie separation, by members of both the ruling political party and opposition parties, and prompted debate and eventually arbitration at the highest political authority on questions of nationality in Ethiopia, the House of Federation, which ordered the referendum.
The complex and contested meanings underlying Siltie and Gurage identities, and ‘the question of nationalities’ in Ethiopia,Footnote 1 as the debate over ethnicity and inclusion is typically called, provide a fascinating real-life test of critical assumptions of theory and public policy with respect to ethnicity and political institutions in sub-Saharan Africa. Vesting a decision regarding the boundaries and content of ethnicity in the hands of ordinary citizens is a somewhat unusual political manoeuvre. Some would call it inherently democratic, since democratic procedures were followed and the results were accepted by all major parties. This is, for instance, the argument of Laitin and Reich's (2003) ‘liberal democratic approach’ to language policy. Others would decry it as the inappropriate politicisation of ethnicity resulting from the flawed and politically divisive institutional structures of ethnic federalism,Footnote 2 and sure to lead down a slippery slope to further ethnic conflict. Even the ‘success’ of the referendum itself can be contested: Was it successful because it was peaceful – no riots or clashes broke out at the time, or have done so up to the present? To what extent is a conflict simmering just below the surface? How are we to interpret the motives of the Siltie in pursuing this separation? What was the role of the ruling party in this process?Footnote 3 Finally, how representative is the Siltie case of the conflict management potential of ethnic federalism?
The Siltie referendum is a critical test of the power-sharing potential of federalism in the context of ethnic conflict and contested identity claims. This paper asks several questions relevant to the situation in present-day Ethiopia, and to democratisation projects across sub-Saharan Africa: What is the connection between communal identities, particularly ethnic identities, and the formal political institutions which are designed to foster cooperation and create the enabling conditions for democracy? How successful are political institutions in adjudicating between the competing claims for recognition, inclusion and equality based on communal identities in multiethnic societies? The focus of analysis is the use of socio-political institutions for the purpose of creating a common sense of national citizenship among Ethiopia's disparate ethnic populations, and for creating the conditions of dialogue and participation central to democracy.Footnote 4 Conflicts over the most appropriate political institutions in a multiethnic state frequently indicate competing visions of citizenship, not just competing political interests. Different ethnolinguistic groups in Ethiopia have strikingly different visions of what a democratic Ethiopia should look like, and at least some of these conflicting visions represent distinct and competing institutional arrangements. A central claim of this study is that the failure of the present institutional arrangement in Ethiopia has been to resolve competing citizenship claims, while also arguing against the castigation of all forms of politicisation of ethnicity.
I begin by reviewing the role that citizenship is understood to have in democratisation processes, and the possible institutional remedies to multiple citizenship identities, particularly ethnic identity. I then analyse the constitutional and institutional features of ethnic federalism in Ethiopia, which is arguably one of the most important cases of institutional innovation in Africa in the last 15 years. In the final section, the Siltie referendum of 2001 is described and the outcome assessed. The conclusion considers the limited potential of formal institutional models, and suggests other areas of policy intervention that may support or supplement institution-building exercises such as federalism.
CITIZENSHIP AND ETHNIC IDENTITY
The challenge of building democracies in multiethnic states has long been a point of meaningful debate among scholars and political leaders of Africa. Post-independence African leaders experimented with electoral and political party systems which they argued to be more appropriate to African contexts. More recently, scholars have suggested new institutional models for designing democracy in non-Western and highly diverse societies (Horowitz Reference Horowitz1985; Lijphart Reference Lijphart1977; Reynolds Reference Reynolds2002). For instance, in assessing the state of democracy in Africa, Joel Barkan (2002: 74) counts among the more optimistic outcomes the fact that there have been ‘new experiments with federalism … to enhance governmental accountability to the public and defuse the potential for ethnic conflict’. While there is some evidence that institutional arrangements such as federalism and other forms of power-sharing can foster cooperation, particularly among political elites, it is increasingly clear that institutions are context-bound, and the results of contemporary institution-building exercises are quite mixed. There are several cases where well-designed formal institutions have failed to meet the expectations of the designers (Laponce & Saint-Jacques Reference Laponce and Saint-Jacques1997; Roeder & Rothchild Reference Roeder and Rothchild2005; Spears Reference Spears2002).
More recently, scholars have argued that the citizenship lens is a more robust explanation for ongoing conflict and failed, stalled, or incomplete democratic transitions. The fact that there are different and competing forms of citizenship in the nation-state and in ethnic communities goes a long way in explaining the persistence of communal identities in many parts of sub-Saharan Africa. At a minimum, citizenship is a legal and formal position based in law. It is the ultimate expression of a state's sovereignty, in that it alone has the right to define who is and who is not a citizen (Heater Reference Heater1999: 80). Citizenship is central to the democratic principle of self-rule, and the distinction between citizens and subjects. Subjects of a state have no ‘voice in the way in which political power is exercised’, but citizens make laws and subject themselves to those laws through their own participation (Tully Reference Tully, McKinnon and Hampsher-Monk2000: 213).
Citizenship also encompasses a set of rights and duties enjoyed by individuals, by virtue of their belonging to a particular national community. For some, ‘citizenship requires … a direct sense of community membership based on loyalty to a civilization which is a common possession’, as opposed to kinship and sentiment (Marshall & Bottomore Reference Marshall and Bottomore1991: 40). The concept of citizenship captures the essence of belonging to an organised political community to which one feels an allegiance (Beiner Reference Beiner1995). Therefore, in modern times, citizenship has been equated with national identity, while in the global and postmodern world, scholars increasingly emphasise citizenship understood as participatory, and less related to a national community or a juridical link to a particular state (Benhabib Reference Benhabib, Joppke and Lukes1999; Martiniello Reference Martiniello, Goldberg and Solomos2002).
James Tully (Reference Tully, McKinnon and Hampsher-Monk2000: 215) has noted that the liberal view of citizenship as only a set of rights and duties is limiting, and that citizenship is better seen as ‘an identity that members acquire through exchanging reasons in public dialogues and negotiations over how and by whom political power is exercised’. Tully's view of citizenship as the ‘intersubjective and dialogical’ engagement in the ‘institutions of self-rule of a free people’ points to the importance of recognition and participation of members of diverse citizen groups in the practice of politics (ibid.: 214). This view privileges the struggles over recognition rather than a particular end-goal of achieving a sense of belonging. Similarly, Charles Tilly (Reference Tilly1995: 6) points to the fact that ‘scholars have come to think of citizenship as a set of mutual, contested claims between agents of states and members of socially constructed categories: genders, races, nationalities and others’.
Citizenship is, therefore, conceptually tied up with equality, inclusion, representation and participation in democratic states. These more nebulous elements of citizenship constitute perhaps the greatest democratic challenges for the multicultural, multiethnic and multiracial societies of the West, and a great deal of scholarly attention has been directed to these topics in recent years. Because of the types of groups and group conflicts in Western states, this has led some to call for ‘differentiated citizenship’, ‘special representation rights’, and other legal and institutional provisions recognising group-based rights (Kymlicka Reference Kymlicka1995; Young Reference Young1990).Footnote 5
Increasingly, citizenship scholars have come to emphasise that citizens are members of, and often active participants in, two substantively different political communities, which are not able to extract similar obligations from or grant similar rights to their members. This view, referred to as ‘dual citizenship’ or the ‘two publics’ by Africanist scholars, is related to the more general theoretical distinction between two conceptions of citizenship (Ekeh Reference Ekeh1975; Ndegwa Reference Ndegwa1997). The first, the liberal form of citizenship, focuses on citizenship as status, and holds that rights exist outside and prior to community, and are held by individuals because they are ‘both logically and morally prior to the society and the state’ (Oldfield Reference Oldfield1990: 179).Footnote 6 In contrast, the civic-republican form of citizenship tends to emphasise duties over status, and considers rights not as inherent but as acquired through civic practice and upholding obligations to the community. Because the community is primary, this view of citizenship associates the fulfilment of citizenship duties with the very identification as citizens (ibid.).Footnote 7
It is generally agreed that in the West these two conceptions of citizenship have competed for prominence at different historical periods, or through the intellectual and political activity of scholars and political elites. Some argue that liberal individualism is ascendant in the West, though communitarians and others wish to reverse this trend because they point to a lack of social solidarity and social cohesion which is created when citizenship is understood as merely or even primarily as status (Etzioni et al. Reference Etzioni, Volmert and Rothschild2004; MacIntyre Reference MacIntyre1981; Oldfield Reference Oldfield1990; Sandel Reference Sandel1982; Taylor Reference Taylor1995; Walzer Reference Walzer1983). There are also distinct historical and geo-political reasons why citizenship has developed differently throughout the world.
However, the nature of contested citizenship in sub-Saharan Africa is distinct from that of the West, both politically and historically. Rather than these two forms of citizenship – the liberal and the civic-republican – clashing and competing in the social and political realm at any given time, those writing about dual citizenship have argued that these operate simultaneously and compete within the individual and larger social community in the modern period. In this view, African citizens are members of two types of political communities in the same temporal and spatial world. The first is their civic-republican community, which is most often their ethnic or communal group. This community demands participation and fulfilment of duties to the collectivity. But at the same time, these individuals are citizens of a modern nation-state, guided by a liberal conception of status and a focus on rights rather than duties.Footnote 8 Modern constitutions and formal political practice vest citizens with a variety of rights that are individual and inherent rather than collective or earned (Heater Reference Heater1999). The clash of identities represented by these two forms of citizenship is unresolved in modern political and social life.Footnote 9 In his study of Kenyan politics, Ndegwa (Reference Ndegwa1997: 613) concludes that central to the failed transition to democracy throughout the postcolonial period is that liberal and majoritarian institutions' ‘presumption of autonomous individual actors is at odds with the reality of individuals fulfilling republican obligations to their subnational community’.Footnote 10 Therefore, citizens may be most engaged in their civic-republican community, which, by its exclusion from formal political practice, is unable then to transform these community goods into public goods for advancing democracy. Put simply, dual citizenship is not supporting democratisation processes.
The framework of dual citizenship has been usefully deployed to explain a vast array of political dysfunctions which impede democratic consolidation in sub-Saharan Africa, such as political despotism, corruption, ethnic voting and even ethnic violence. More recent scholarship on formal citizenship laws points to the critical contribution of legal and institutional structures to democratisation processes (Herbst Reference Herbst and 1999). In the wake of political liberalisation, African leaders have sought to invoke nationalist sentiments, using citizenship laws to narrow the realm of political competition (Aminzade Reference Aminzade2003; Whitaker Reference Whitaker2005; Woods Reference Woods2003). With respect to dual citizenship, Africanist scholars have also used the citizenship lens to explain stalled or failed democratic transitions (Mamdani Reference Mamdani1996; Ndegwa Reference Ndegwa1997). Others have used dual citizenship identities to theorise the persistence of racial, class and ethnic identities despite the ‘ascendancy of liberal constitutionalism’ in these contexts (Halisi et al. Reference Halisi, Kaiser and Ndegwa1998: 423). However, as the guest editors to a special volume on citizenship in Africa note, ‘the tendency to overemphasize liberal citizenship with emphasis on legal-rational factors while ignoring populist and communitarian conceptions of citizenship’ in sub-Saharan Africa is problematic, in part because it obscures the distinct ways in which citizenship is conceptualised in Africa, owing to the vagaries of history, culture and contemporary political practice (Halisi et al. Reference Halisi, Kaiser and Ndegwa1998: 342).
With respect to ethnic identity in particular, Osaghae (Reference Osaghae2004: 229) goes so far as to claim that in the context of ethnic domination, ethnicity ‘functions as a liberationist ideology, providing aggrieved groups a pedestal for seeking redress and state reconfiguration’. Similarly, Lonsdale (Reference Lonsdale and 2004: 75) in his normatively rich discussion of Kenyan politics, points to the paradoxes of ethnicity and politics in Africa, concluding that for Kenyans, ‘their imaginations of ethnicity, too often destructive, can nonetheless be among their most fruitful sources of nationally active citizenship’. Increasingly, the view of long-time African and Africanist scholars alike is that the study of ethnicity must be approached in radically new ways to account for its democratic potential. It would seem that a theory of citizenship can facilitate the identification of critical citizenship questions relevant to contemporary African democratisation challenges. This must extend beyond formal legal or constitutional provisions relating to citizenship, and into the realm of democratic participation and democratic outcomes.Footnote 11
Citizenship provides a theoretical lens by which to accomplish a re-framing of ethnicity as positively experienced, enacted and as having potential for expanding and deepening democracy in parts of sub-Saharan Africa. Citizenship also introduces questions of social solidarity which can be useful in explaining the failure of institutional reform packages to create the enabling conditions for democratic consolidation. In fact, citizenship is both a legal status and a practice, created by the actors – states and citizens (Tilly Reference Tilly1995: 9). The notion of contentious and engaged citizenship is almost entirely lacking in political analyses of African politics today. While there are certainly important cases where formal citizenship laws play an increasingly important role in restricting political participation by citizens, there are other countries on the continent where the competing notions of citizenship within and between groups and the state constitute barriers to democratisation efforts.Footnote 12 Rather than restricting our study of citizenship in Africa to formal, state-centric definitions of citizenship, there is a need to also consider citizenship as a ‘set of mutual, contested claims between agents’ as Tilly (Reference Tilly1995) urged, widening our lens to look at popular claims to recognition and inclusion based on ethnicity.
What a number of contemporary scholars of diversity and multiculturalism theory in the West offer is a dynamic and optimistic view of the role of identity politics in supporting democratisation. Adopting and modifying this theoretical approach would radically change the pessimistic and defeatist tone of what is typically studied as ‘ethnic politics’ in the African context. It will, I argue, go a long way in explaining the persistence of ethnic and communal identities in Africa far better than we have done to-date, in line with the admonitions of scholars like Osaghae to ‘mainstream’ ethnic politics in studies of transitional democracies.Footnote 13 One of the most important linkages between ethnic identity and political outcomes in democratising states is citizenship. The democratic potential of dual citizenship in one country, Ethiopia, suggests that neither extreme of ignoring or reifying ethnicity in formal politics achieves the requisite balance needed to promote democratic consolidation.
INSTITUTIONAL RESPONSES TO DUAL CITIZENSHIP: THE CASE OF ETHNIC FEDERALISM IN ETHIOPIA
The analytical framework of citizenship is an innovative and compelling conceptual lens for studying the tensions between identification with the modern nation-state, and the various communal identities which persist and give citizens' lives meaning across Africa today. Yet this theoretical work remains somewhat divorced from the practical exercise of democratisation, understood as the process of deepening and strengthening democratic institutions and outcomes. Because of the normative and institutional preference for non-ethnic solutions to ethnic and communal diversity, empirical and theoretical work in sub-Saharan Africa has not adequately explored the relationship between dual citizenship as just described and formal political institutions.
There are competing perspectives on what might be the appropriate institutional forms to address the distinct challenges posed by dual citizenship, particularly in its ethnic form.Footnote 14 A common response has been to de-legitimise membership in the civic-republican community, at least formally. This can be explained in part by the whole-scale adoption of Western models of governing by most African leaders at independence (Jackson & Rosberg Reference Jackson, Rosberg and Kohli1986; Mamdani Reference Mamdani1996). In many cases this was partially based on a belief that civic-republican citizenship was irrelevant or inappropriate for national citizenship identities in Africa. The imperatives of nation-building and national integration were almost universally understood to mean a shift from sub-national, ethnic or communal loyalties to nation-state loyalties (Elaigwu Reference Elaigwu, Mazrui and Wondji1993). The dominance of one-party states across Africa was frequently justified on the grounds that it discouraged ethnic politics. It has also been convincingly demonstrated that time and again the formal or official policy and practice de-legitimising ethnicity facilitated its manipulation and deployment by savvy powerful ethnic actors in the political realm (Ndegwa Reference Ndegwa1997; Olukoshi & Laakso Reference Olukoshi and Laakso1996). The post-independence African leadership frequently engaged in what has been called ‘ethnic politics’, including consolidating political and economic power in the hands of their own ethnic group. However, the explicit and formal practice has almost entirely been non-ethnic from independence up to the present time.Footnote 15
Even if the tendency across most of sub-Saharan Africa has been to build explicitly non-ethnic political institutions, there is some precedent globally for building ethnic and communal identities into formal political structures. For the most part, this has been through a variety of supplemental political and social measures, notably affirmative action programmes, systems of proportional representation and reserved seats in legislative bodies based on ethnicity or other markers of communal identity, as well as various forms of regional autonomy or self-government rights for particular groups (Reynolds Reference Reynolds2002; Roeder & Rothchild Reference Roeder and Rothchild2005).
Few states in the world, and none in sub-Saharan Africa, have taken the formalisation of sub-national or ethnic identities to the same level as Ethiopia. Like many other sub-Saharan African nation-states, Ethiopia's multiethnic composition has been a significant source of conflict in the modern period.Footnote 16 Centuries of social inequality, most prominently manifested as ethnic inequality, emerged in the mid-twentieth century as one of the most critical political questions on the national agenda. Though never colonised by the Europeans, increasing internal politicisation and interaction with independence movements in other parts of Africa and movements for civil rights and equality in the United States and elsewhere facilitated the development of a movement for ethnolinguistic inclusion and equality in Ethiopia.Footnote 17 This movement was not unlike independence movements in the rest of colonial Africa in pursuing a more democratic and inclusive political structure. Yet neither the modernising reforms of Haile Selassie nor the socialist agenda of the Derg were able to resolve the ‘question of the nationalities’ (Bahru Reference Bahru1991; Clapham Reference Clapham2002; Kidane Reference Kidane1997; Marcus Reference Marcus1994).
Therefore, after decades of state-driven centralisation and assimilation, Ethiopia embarked in the early 1990s on a controversial experiment with decentralisation and federalism explicitly organised along ethnic lines. Formally, the institutions of federalism are exciting African innovations in democracy and power-sharing. Even as ethnic federalism in Ethiopia is perceived as a uniquely African institutional solution to a persistent and vexing source of political conflict and instability, it raises pressing questions about the proper place for politicised ethnicity. Critics and hesitant supporters alike have watched the last decade of federalism in Ethiopia with great interest because of this. While receiving tentative support from Western countries such as the United States, there were many sceptics about ethnic federalism from the outset. They feared that the departure from commonly accepted political practice in Africa of non-ethnic political institutions was a grave mistake. Because of this, the implications of the success or failure of ethnic federalism reach across the entire continent.
Federalism in Ethiopia is structured ethnically, at least on paper. Most of the largest regional states under the federal system are ethnically delineated and primarily mono-ethnic.Footnote 18 A few others have two main ethnic groups and at least one is multiethnic, the Southern region, with at least 45 groups. City-states are considered distinct (Addis Ababa and Dire Dawa).Footnote 19 The Constitution (1995: art. 46) explicitly states that Regions should be delineated based on ‘the settlement patterns, language, identity and consent of the people concerned’.Footnote 20 The preamble to the Constitution vests power in ethnic groups, by opening with the words: ‘We, the Nations, Nationalities and Peoples of Ethiopia’. Similarly, Article 8 states that sovereign power ‘resides in the Nations, Nationalities and Peoples of Ethiopia’, through their elected representatives and ‘their direct democratic participation’. While recognising a variety of individual rights, the Ethiopian constitution takes the radical position that political power resides with ethnic and nationality groups. This unique formalisation of communal identity as the basis for formal citizenship in Ethiopia has vast political implications (Abbink Reference Abbink1997; Minasse Reference Minasse1996).Footnote 21 Ethnic federalism is a radical departure from the Western model which constructs citizenship on a liberal/civic/individual foundation. In this view, ethnic federalism is a framework for constructing a constitutional and political role for dual citizenship as discussed above.
Ethiopia provides the ideal setting for study of the intersection of ethnicity, citizenship and political institutions, since its federal experiment, now entering its second decade, has had mixed outcomes. Certainly Ethiopia is more democratic and less politically violent today than it was prior to the introduction of federalism. For instance, Ethiopia has moved from ‘not free’ in Freedom House's categorisation to ‘partly free’ in the years since the new regime took control and the constitutional framework developed. Despite serious charges of ruling-party dominance, and growing problems of human rights violations, it is hard to argue that political or social repression under the EPRDF is at the same level as during previous regimes. In some cases, the specific institutions of the federal arrangement have arbitrated ethnic conflict peacefully. Closer inspection reveals that at least some of these institutional successes were incomplete or flawed. In other instances, violent low-level ethnic conflict persists, and the formal institutions have been either underutilised or ineffective in promoting a peaceful resolution to some communal conflicts. More recently, the 2005 national and regional elections have led most thoughtful scholars to conclude there has been a significant tightening of control by the ruling party and a deterioration of democratic gains.Footnote 22
Generally, assessments of ethnic federalism in Ethiopia based on various indicators of democracy have focused on four types of flaws in the federal arrangement: ruling-party dominance (Aalen Reference Aalen2002; Medhane & Young Reference Medhane and Young2003; Samatar Reference Samatar2004), resource and human capacity shortages (Keller & Smith Reference Smith2005; J. Young Reference Young1999), the persistence of authoritarian traditions (Merera Reference Gudina2003), and the politicisation of ethnicity (Mesfin Reference Mesfin2003; Poluha Reference Poluha, Salih and Markakis1998). Each of these has some role in terms of explaining the incomplete democratic transition in the country. Most scholars cite a mixture of these variables in their assessments of Ethiopian democracy today. Almost all of them take a national approach in analysing the impacts of ethnic federalism, whereas this study focuses on a specific and local ethnic conflict and the way in which the institutional and procedural parameters provided by the 1995 Constitution addressed it. In the next section, I consider the case of the Siltie and Gurage, to assess the role of the institutions of ethnic federalism in arbitrating between two ethnic groups with competing visions of both ethnic and national citizenship identities.
ETHNIC SELF-DETERMINATION: THE CASE OF THE SILTIE AND GURAGE ETHNIC GROUPS
The granting of constitutional rights to ‘Nations, Nationalities and Peoples’, rather than to individuals, inherently makes the boundary-drawing exercises between ethnic groups highly politicised. The most controversial section of the Ethiopian Constitution, Article 39, which grants to all of the country's constituent ‘Nations, Nationalities and Peoples’ the right to self-determination ‘including the right to secession’, is arguably more about rights to ethnic self-determination than rights to secession, though the symbolism of the secession provision is profound.Footnote 23 In the April 2001 referendum mentioned at the start of this paper, members of the Siltie ethnic group voted overwhelmingly in favour of declaring themselves a distinct ‘nationality’ or ethnic group. In addition to being an administrative and political move to separate the Siltie from the Gurage group of which they were previously a part, the referendum is a fascinating test of the role of institutions in managing conflict, promoting democratisation, and arbitrating between disparate identity groups. Other ethnic groups have tried to do the same, but not all attempts have resulted in the hoped-for grant of self-determination rights.
Background to the referendum
Located in the Southern Nations, Nationalities and Peoples regional state (SNNPR), the Siltie are predominantly Muslim, and are economically poorer and less likely to migrate than the Sebat Bet Gurage, who live to the west.Footnote 24 Few studies of the Siltie people have been completed, though political and administrative changes in recent years are likely to lead to more focused research on them.Footnote 25 In the modern period, the Siltie were considered a sub-clan of the Gurage ethnic group (Shack Reference Shack1966). Though the Gurage were understood as a loose collection of related clans, the term Gurage came to be operative at the national level, and subsumed the identities of ethnic groups which were distinct in certain ways, and some of which had little interaction with others.
The commonalities among Gurage groups include high levels of political fragmentation, the cultivation of ensete or false banana, and cultural factors such as housing patterns. The so-called sub-clan distinctions mark off different languages, dialects and religions, in addition to other social and cultural characteristics. Bahru Zewde (Reference Aminzade2003: 20, 21) notes that the three categories of the Gurage were formed largely on the basis of linguistic studies, and that ‘in spite of the strong tradition of their common identity, these are mutually unintelligible categories’. Similarly, Grover (2003: 95) has pointed out in his study of the 1994 Ethiopian Census that ‘Gurage is not a single language but at least five languages.’
One of the most important early studies of the Gurage was an ethnographic study of one group of the Gurage, who are now referred to as the Sebat Bet (seven houses) or seven clans of western Gurage, comprising the Chaha, Ezha, Geyto, Muher, Ennemor, Akilil and Walani-Woriro (Shack Reference Shack1966). Under the imperial administration, the Gurage formed their own awraja, within which there were three politico-administrative divisions – Chaha, Walani and Selti (ibid.: 37). Since then, several scholars have contradicted or clarified some of the assumptions and assertions made in this early work.Footnote 26
In addition to ethnographic and political counter-claims to the notion of a pan-Gurage ethnic identity, linguists have also re-assessed the relationships between the languages and dialects of the so-called Gurage languages. Early scholarship designated as Gurage some ‘fourteen “tribal” divisions in the Gurage cluster … each comprising a number of politically independent clan chiefdoms, [and] further distinguishable on the basis of the language or dialect spoken by each tribal unit’ (Shack Reference Shack1974: 94). Recent scholarship, rather than ignoring significant linguistic and politico-territorial differences, has corrected the historical tendency to cluster the Gurage linguistically. Hudson (Reference Hudson1994: 692) notes twelve to fifteen recognised varieties, ‘within which six or seven languages and/or dialect clusters may be distinguished, in three distinct groups’, the northern (Soddo), western (Chaha) and eastern. He also notes that the ‘eastern Gurage languages are more divergent from western and northern Gurage languages than, for example, Amharic is from Tigrinya’ (ibid.: 692).Footnote 27
The history of linguistic and ethnographic contributions to the construction of a pan-Gurage identity is only part of the story. More recently, John Markakis (Reference Markakis, Salih and Markakis1998: 130) has developed these insights in his argument for the ‘contextual, multidimensional and fluid’ nature of ethnic identity through a brief study of Gurage ethnic-identity formation. He argues that the clan (bet) provides the primary source of identity for the western Sebat Bet Gurage, territoriality (ager) for the Soddo or Kistani Gurage, and religion for the Siltie-speakers of the east. In fact, Markakis concludes that historically there was no people who self-identified as Gurage, and that the Siltie-speakers would traditionally have identified themselves as ‘Muslims’ (ibid.: 131), something consistently confirmed in focus groups throughout Siltie in 2003 (L. Smith Reference Smith2005). What is significant for our purposes is Markakis' conclusion that, while the notion of a cohesive and united entity known as Gurage did not come from within those groups who were classified as such, it suited the Gurage to develop this ‘pan-Gurage universe’. This was useful both to expand social networks in urban areas, and to provide the Gurage with political and economic currency when dealing with Ethiopian imperial authorities (Markakis Reference Markakis, Salih and Markakis1998: 134).
It is surprising that the Transitional Government of Ethiopia (TGE) and later, the EPRDF, unproblematically built ethnic federalism on ethnographic work of the Derg regime. This is a little-known fact, buried in the 1994 Census, nowhere mentioned in early TGE proclamations, the 1995 Constitution or other related government policy statements such as the National Education and Training Policy (NETP 1994), and never discussed by supporters or opponents of the federal arrangement. In fact, the EPRDF, despite a stated interest in the self-determination of all nations and nationalities in the country, has done little to contribute to either scholarly or public discussion on the content and nature of ethnic group identities. In the case of the Siltie, by using ethnic categories from previous regimes, the EPRDF itself laid the groundwork for the contentious issue of Siltie self-determination. In this, the work of the Derg's Institute for the Study of Ethiopian Nationalities (ISEN) represents the most systematic and thoughtful attempt to conceptualise and standardise the terms and categories of ethnicity and language in Ethiopia. This is recognised in the 1994 Census, which explicitly attributes the ethnic and linguistic categories it uses to those that were created by the ISEN.Footnote 28
The politics of self-determination for the Siltie
Scholarly research, state and nation-building imperatives, and a historic lack of precision with regard to defining ethnic or linguistic group, all colluded to create a murky and ambiguous terrain for ethnic federalism in Ethiopia. It is of little surprise that earlier regimes, particularly that of Haile Selassie, had intentionally subverted the clear development of indicators of ethnolinguistic identity and categorisation. The nationalist vision of modern Ethiopia has always been of one dominant culture, unified not only territorially but culturally. On the other hand, despite the contribution of the ISEN's work, especially the preparation of the Almanac of ethnic groups on which the present-day institutions of ethnic federalism rest, there had yet to emerge clear criteria for determining ethnic group boundaries, nor was there a procedure for arbitrating disputes or contested claims. The Siltie pursuit of recognition as a distinct ‘nationality’ under ethnic federalism has resolved procedurally, if not politically, essential political and constitutional questions.
Despite some precedence for a political designation which was pan-Gurage in content, the Siltie mobilised early in the transitional period to receive status as a separate Nationality. This is attributable to several factors worth considering here, including change and continuity in administrative and political categories of ethnic groups, and evolving social relations among Gurage groups, and nation-wide. The position of Siltie ethnic group members with regard to their earlier designation as Gurage is unclear. Some indicate that the experience of subjugation and exploitation under Emperor Menilek united the Siltie with their Gurage neighbours. ‘They saw themselves as the same because of this mistreatment’ (interview #85). The Siltie experienced military conquest and similar processes of labour displacement caused by their unequal incorporation into the Ethiopian empire to the Gurage, though they tended to remain more rural and considerably poorer than the Gurage. There are also historic patterns of interaction between the groups, particularly in border towns and regions. Yet in these same areas where Siltie came into contact with Gurage and Hadiya, they were also more likely to become aware of their difference, as they were often referred to as not Gurage or Hadiya by members of those ethnic groups (interview #85).
Even prior to the referendum and the separation, the Siltie pushed for the use of their own nationality language in their own administrative districts (woredas). At a symposium on the question of language policy, the Siltie decided in about 1992/93 to use their own language, and informed the Gurage zone administration (interview #67).Footnote 29 This was facilitated in part because the Siltie language was one of the languages of the Derg's literacy campaign, and therefore more educational materials were available in the language (interview #54).
Today, most agree that unequal development was a primary impetus for Siltie separation. It was argued that the location of Welkite, the Gurage zonal capital town, was too far away, and the isolation meant that the Gurage Zone government was not adequately addressing the development needs of the Siltie (Markakis Reference Markakis, Salih and Markakis1998; interview #67). When the economic development gains did not materialise, resentment built against Gurage leadership, and a sense of distinct Siltie suffering emerged. A political organisation called the Siltie Gogot Democratic Party was formed in the early 1990s, calling for separate Siltie representation and self-administration based on the claim that the Siltie were a distinct nationality. Later, there was a split and a new party, the Siltie Peoples Democratic Unity Party (SPDUP), was formed. Shortly thereafter, anti-separation Siltie formed the Siltie Gurage Peoples Democratic Movement, apparently funded by a wealthy Siltie merchant (Markakis Reference Markakis, Salih and Markakis1998: 142). Finally, the Gurage themselves mobilised to oppose the separation and the Gurage Peoples Revolutionary Democratic Movement (GPRDM), which controlled the Gurage zonal government, used its power and influence to work against the separation.Footnote 30
The political wrangling was played out in the early years over questions of ethnography and identity. The distinctions between the Gurage groups, including particularly the Soddo Gurage and the Siltie, were important from the very beginning of the federal arrangement. During the initial stages of designing and building ethnic federalism, only two Sebat Bet Gurage representatives attended critical early meetings of the Transitional Government (TGE), and intra-group representation concerns led to the development of procedures for determining which of the three main Gurage sections (east, west, north) would represent the ‘Gurage’ on the two seats they were allotted on the Council of Representatives. Particularly after the establishment of the EPRDF affiliate, the GPRDM, the splits between these groups became political, with the first important conflicts between Soddo Gurage in the north and their Oromo neighbours, known as the Soddo Jida.Footnote 31 Somewhat later, but still early in the transitional period, the Siltie began to agitate for separate status, based primarily on their Muslim identity, distinct language and relative underdevelopment.Footnote 32
The early period of Siltie mobilisation was characterised by attention to developing local ethnographies and a unified discourse of Siltie identity, followed by what could be called a civil-society building phase.Footnote 33 This involved the creation of associations and, eventually, political parties that could advance Siltie interests in the political realm. A short time later, these Siltie political parties began to pursue legal and constitutional recognition of the Siltie ethnic group, an unprecedented move in Ethiopian political history. The constitutional provisions of ethnic federalism concern themselves directly with questions of self-determination of ethnic groups, requiring the establishment of procedures for groups to be delineated as such, thereby qualifying them for self-determination. Because the constitutional and institutional framework attached primary political importance to ethnic groups, determining ethnic group boundaries emerged as a pivotal political exercise, and the Siltie case became the test for claims to autonomy and distinct identities. Until this point, it is unclear how groups were designated as separate ethnic groups or not. The ambiguous legal nature of ethnic group specifications, however, could not continue, since the distinctions between groups involved tremendous political and economic resources, not to mention social status. As one EPRDF document itself states, ‘without resolving such nationality issues the building of a stable democratic society and progress in the political and economic life of the people cannot materialise’ (EPRDF n.d. 2: 1). Interestingly, the position of the Party was that ‘in most cases there was no ambiguity in the ethnic identity of the people of Ethiopia … [but] that there has occurred one case which has special attention’ – the case of the Siltie people (ibid.: 1).
At this point, a petition was made to the House of Federation by those in favour of separation for the Siltie. The decision of the House of Federation initially was that this matter must be handled at the regional level, if at all possible.Footnote 34 Thereafter, a meeting was organised in the town of Butajira in 1997 with representatives from various parts of the Siltie community.Footnote 35 Rival political parties were present, as were regional and zonal officials of the dominant party, the EPRDF/GPRDM. A resolution at the end of the meeting opposed separation and the matter was considered closed by the ruling party (Markakis Reference Markakis, Salih and Markakis1998). But there is strong evidence that the ruling party's position against the separation inhibited free discussion and nullified the decision of the group assembled. As one respondent noted, ‘the people did not feel free because the EPRDF was campaigning strongly that the separation was wrong’ (interview #85). The EPRDF's own report on the matter concluded that ‘the fact that EPRDF took a position on the matter threw doubt on the sincerity of the ballot to many observers … This made the Butajira conference and its outcome undemocratic’ (EPRDF n.d. 2: 2).
The initial position of the EPRDF was that the Siltie were indeed part of the Gurage ethnic group. According to one respondent close to the process, the feeling of the EPRDF was that they were the same people, only speaking different languages, and that they would ‘evolve as Gurage … [Whereas] the population [of Siltie] as a whole is more inclined to see themselves as different’ (interview #85). But even the ruling party acknowledged that the Siltie people had a distinct language, and since language was a primary designator of ethnic group identity under the ISEN studies which were foundational to the boundaries of ethnic federalism, it could not resolve this anomaly.Footnote 36 Therefore, the party admitted that they ‘mishandled’ the matter from the start. The EPRDF report on the matter states that the party ‘failed to appreciate the rise of nationalism and misconstrued this trend, fearing it might lead to the break-up of the Guraghe people … At the start of the campaign the EPRDF took a defensive position. It felt it had to take a stand on the identity of the Siltie and therefore supported the unity of the Guraghe people as one body. EPRDF has always struggled for the unity of people but in this instance this support was misplaced’ (ibid.: 2).
This ‘recognition’ by the EPRDF appears to have allowed or at least facilitated the legal process undertaken by separatist opposition political parties, particularly the SPDUP. Once the matter was brought before the House of Federation the second time, it was sent to the Council of Constitutional Inquiry, which is tasked by the Constitution to consider in depth any issues relating to the nations and nationalities, and make a recommendation to the House of Federation for a vote.Footnote 37 Their recommendation, in 1999, demonstrates that the Council saw the constitutional issues at hand to be two-fold: (1) ‘According to the FDRE Constitution, who has the power to decide about the identity of a given group of people?’; and (2) ‘What procedure should be followed to do that?’ (House of Federation, n.p.).Footnote 38 In the Council's decision is therefore to be found the most explicit procedural answer to the political question of determining ethnic identity under the new federal arrangement.
Since the Council understood the Siltie case as an issue of self-determination, particular provisions of the Constitution were considered for their relevance. Articles 39(4), 47(3), 52(2), 62(3) were considered to be most relevant to the ‘question of identity’ (mannennat teyyaqe). The recommendation of the Council was that ‘a demand for the determination of identity must be entertained by the Council of the State where the issue is raised’ because of the requirements of Article 52(2). But ‘if there is a procedural irregularity or complaint that the issue was not resolved in accordance with the constitution; or if there are similar reasons, or if the group and the state council do not agree, the case will be submitted to the House of Federation for a final decision’ (ibid.). However, Article 39(5) was seen to have some bearing with respect to determining the procedures for the group to make their wishes known. In particular, the ‘State Council to which the case is submitted is expected to conduct research that is cognizant of the conditions under Article 39(5) and then present the case for referendum. For the referendum to be democratic and fair, the members of the group which has raised the demand must participate directly, and the process should be free and held by a secret ballot’ (ibid.). The use of a referendum as the procedural answer was argued to be provided for by Articles 39(4), and 47(3) of the Constitution, which relate to secession and statehood, respectively.
As provided by the Constitution, this recommendation went for a vote by the House of Federation, which supported the Council's recommendation. Thus, the historic decision of the House of Federation was that the Siltie had made a sufficient case for their status as a distinct ‘Nationality’ to at least have the right to a referendum on the matter. The question on the ballot was simply: ‘Are the Siltie Gurage or not?’. A referendum held on 1 April 2001 in the Siltie area and other places where the Siltie lived resulted in a vote of over 99% for the separation of the Siltie from the Gurage. Siltie residents in Addis Ababa were reported to have taken to the streets to ‘express their happiness on the results of the referendum’ (Addis Tribune 13.4.2001).
The economic, political and administrative results have proven significant for the Siltie people. Moving from the status of several woredas within Gurage Zone, the Siltie now have their own Zone, with the attendant resources and direct access to higher levels of administration. Officials at zone administrative offices, some of whom had previously performed their duties for the Siltie area from the Gurage Zone offices in Welkite, reported that the location in Worabe, which is inside the Siltie area, and the greater financial resources, were both tangible signs that the referendum had given the Siltie a greater share of the political and economic power that they needed and deserved. By late 2003, a number of new government and private buildings were under construction along the main road. Several self-help organisations with links to urban entrepreneurs were advertising their work in schools, as well as other poverty-alleviation and development projects. A large mosque was under construction just as one entered Worabe from the north.
Echoing the words of the EPRDF report, though presaging them, the Council concluded that the important constitutional principle ‘is that the group itself is the decision-maker. Thus, the House of Federation and the State Council's role should be seen in light of that’ (House of Federation). This historic decision of the House of Federation and the outcome of the Siltie referendum mean that future decisions about who is an ethnic group and who is not will be decided procedurally according to processes somewhat similar to those for secession and statehood, notably a referendum of the people concerned (Baylis Reference Baylis2004).
It is worth noting a few critical political issues raised by the outcome of the Siltie case, as well as the House of Federation decision and the EPRDF's role in this. First, the Siltie were successful in using existing constitutional provisions and institutional processes to advance their claim. This success in a country with such a short history of democracy is itself monumental. The EPRDF concludes that this is because the Siltie are a distinct nationality, and that their inclusion with the Gurage was quite uncommon in Ethiopia. According to the party report, most ethnic groups are already properly identified (EPRDF n.d. 2: 1). The validity of this assertion will only be proven as other groups come forward (or do not come forward) claiming self-determination and identity rights.Footnote 39 But the precedent that this case has created, beyond the most obvious answer to procedural matters relating to identity questions, is for peaceful and democratic use of formal institutions and the Constitution to resolve the stickiest questions of ethnic identity. Champions of procedural democratic processes would probably regard the Siltie case as a vindication of their argument for vote-centric conflict resolution processes.
Second, the role of the EPRDF vis-à-vis opposition parties is important. While the tone of the EPRDF's internal report on the matter seems genuinely contrite, it is clear that the party has tremendous influence in determining the outcome of such matters. It is also certain that the position it takes on the claims of any particular group will have a powerful role in determining the outcome. The party initially did not consider the Siltie people distinct, despite their unique language, discrete territorial boundaries and differential development levels. But it is not entirely clear that the EPRDF only changed its official position on this as a matter of political opportunism, since it could be argued that their change in stance was a reflection of democratic deliberation. Party members seem to have come to recognise the will of the Siltie people in this process, and shifted their party line in accord with it. And this despite resistance from powerful Gurage within the party, a fact only implicitly acknowledged in the EPRDF report, but which surely must have weighed heavily on the EPRDF in the process of revising its stand.
Both the Siltie and the Gurage people are represented in the House of Peoples' Representatives through the Southern Ethiopia People's Democratic Movement (SEPDM) of the EPRDF coalition.Footnote 40 There are several registered opposition groups, including the SPDUP, the Selti Nationality Democratic MovementFootnote 41 and the Gurage People's Democratic Front, as well as some minimal support for the larger 2005 opposition parties, the Union of Ethiopian Democratic Forces (UEDF) and the Coalition for Unity and Democracy (CUD). The opposition, which led the initial campaign for separation of the Siltie people, expressed resentment at not being able to capitalise on the gains of the referendum in terms of their own political standing. The EPRDF won in all four electoral constituencies in Siltie Zone by over 80%. The CUD ran candidates in only two of four, and the UEDF in only one. The SPDUP held a Parliamentary seat in 1995, but lost it in the 2000 elections and did not win it back in 2005. The position of the party is clear from a party statement dated 10 October 2002, which asserts unequivocally that ‘members of the party now ruling the Siltie zone were openly saying there was no nation called “Silte” and they were strong opponents of the movement towards identity recognition … Unlike their past utterances, now only for the sake of holding power they bear the name Siltie, and are roaming around to avenge their yesterday's loss’ (SPDUP, 2).
The SPDUP, pivotal in preparing the documentation of the case to the House of Federation, feels that the party was successful in pursuing its first objective, the ‘identity case’, but has yet to achieve its second, which is for the Siltie ‘people to fully govern themselves’ (interview #90). There is clear resentment towards the EPRDF for their about-face on the matter, which enabled them to maintain popular support in the Siltie zone in the 2000 and 2005 elections. But unlike the substantially more acrimonious debates between the ruling party and the opposition generally, the Siltie case is different in that the EPRDF took the unusual step of changing its position on an issue in large part because it realised how widespread and popular it was, and also because the basis for the continued inclusion of the Siltie within the Gurage was not based on the principles central to ethnic federalism, notably language identity.Footnote 42
Third, it is interesting that the position of the EPRDF was initially focused on the ‘unity of Ethiopian peoples’, and therefore considered the Siltie to be part of the Gurage. One respondent close to the issue said that, while unity could not be forced, the party has long been of the view that ‘the people do not want disintegration and they do not want their own mini-states. They have common interests and the same problems’ (interview #85). This is quite a different view from that claimed by the EPRDF's opponents, who cast the regime as interested either in eternal divisions as part of its divide-and-rule strategy, or in breaking up the country completely. By grounding its initial opposition to the Siltie separation in the complex question of unity versus separation, but still being willing to acquiesce when it was clear that popular opinion was not with them, it could be argued that the EPRDF bolstered democracy. I am not saying here that there may not be elements of a divide-and-rule strategy as claimed by opponents from minority or formerly oppressed ethnic groups. This may well be true and there is certainly evidence elsewhere to suggest it. However, the claim that the EPRDF is interested in the perpetual disintegration and perhaps dissolution of the country seems unfounded. In the early days of ethnic federalism, it struggled with the contradiction between upholding constitutionally protected rights to self-determination and maintaining the unity of the country at the most basic level. Sometimes it has come down in favour of ‘unity’ – particularly, it would seem, when ‘unity’ as a political principle coincided with retaining a firm hold on central power.Footnote 43
* * *
The Siltie referendum is an interesting test of some critical assumptions of procedural and institutional models of linguistic and ethnic justice. The question of ethnic identity was central, and was subjected to political contestation through a process of meetings, petitions to government entities, constitutional interpretation, and eventually, a referendum. Critically, it was a peaceful process, remarkable in light of the tensions surrounding ethnicity and language in the country.
While the peaceful outcome thus far seems to confirm the usefulness of a political contestation model for handling language decisions, this conclusion may be premature. The role of the EPRDF leadership in handling the Siltie case may have been the cause of the peace, not the referendum process itself. Interviews and data analysis suggest that the EPRDF made a calculated decision to support the Siltie position, perhaps to undermine the power of the Gurage, or perhaps in response to perceived threats to peace by the Siltie if they chose to do otherwise.Footnote 44 Either way, the resolution to the matter came only after the EPRDF supported the Siltie position, which was done using the administrative processes such as the Council of Constitutional Inquiry, an institution firmly under EPRDF control. Even more telling is the success with which the EPRDF has since undermined the power of the Siltie political opposition. Perhaps because the EPRDF took the supporting position, most members of the zonal administration are EPRDF members, not opposition, and were reluctant to discuss the role of the opposition parties in the referendum process. The opposition party confirmed its marginalisation in the wake of the referendum.
Significantly, the Siltie may be the only case of the peaceful use of formal political institutions to resolve competing ethnic group claims in Ethiopia. Consideration of the failures – ethnic conflicts which might qualify for institutional resolution, but either have not been pursued or have not been successful in resolving the conflict – provide an even more complex picture of how formal institutions are an incomplete solution to the conflicts of identity between ethnic groups.Footnote 45 In other work, I provide specific examples, such as the Sidama conflicts in 2002, the ongoing Oromiya–Somali border disputes, and the violence in Gambella in late 2003 (L. Smith Reference Smith2005).
For each of these three examples, there are myriad others. Rather than being just a few discrete conflicts based on ethnic identity, there are in fact quite a number of dissimilar and low-intensity conflicts across the country. While I have suggested elsewhere that there are considerable implications of not including the Gurage in the referendum process, the relatively successful resolution of the Siltie case to date reflects an exception to the rule, rather than the standard by which ethnic federalism can be judged. In part it is the rarity of such peaceful and institutionalised responses to ethnic disputes, no matter how great the level of political violence that precedes them, that makes the Siltie case important. However, the peaceful resolution does suggest the potential for institutional and legal processes to resolve disputes, rather than deepening and strengthening conflict. In Ethiopia, the House of Federation and the Council of Constitutional Inquiry are powerful democratic institutions as yet entirely underutilised.
Ethnic federalism is a unique effort to harness the democratic potential of dual citizenship identities in Ethiopia. It has been a mixed project, as we have seen. The institutions have in some cases peacefully and conclusively arbitrated and managed conflicts, particularly ethnic conflicts. At times, the institutions have demonstrated a surprising flexibility in resolving ethnic conflicts, as the Siltie case illustrated. Rather than hardening ethnic claims, the House of Federation and the Council of Constitutional Inquiry in particular have been pivotal tools of creative and democratic management of contested identity. Yet, in a number of other cases, these institutions have not been utilised at all, and in other cases, they have not been so successful. In part, these failures result because the institutions have not been merged with the process of social discourse over and about the nature and terms of democratic citizenship in the contemporary Ethiopian state.
Formal political institutions are limited, though vital, in resolving contested citizenship claims. They provide outlets for mediation and resolution, and they can structure the rules of political bargaining but, in many cases, they are bypassed or underutilised. Institutions are also limited in their effectiveness if they do not have mechanisms which recognise the commitments placed on African citizens by their civic-republican communities. In Ethiopia, there are a number of reasons for this. Certainly, party dominance is the primary one. Human and financial capacity variation between ethnic communities and the general lack of resources country-wide are additional factors. Critically, the lack of democratic dialogue in Ethiopia is a fundamental weakness of the present political system. While this may be facilitated by the party leadership, it pervades Ethiopian society. The responses of parents are typical: ‘We were told to teach our children in X language.’ ‘Yes, we were called to a meeting. At the meeting, we were told of the new language policy.’ This is not social discourse or democratic dialogue. It is the use of local political and administrative institutions for information dissemination and policy enforcement developed at higher levels. It does not reflect genuine citizenship participation or dialogue, and the results are just the kind of apathy and disempowerment of citizenry that one hears in discussions of politics in Ethiopia.
Important steps have been taken in recent decades to advance citizenship in Ethiopia. Formal, legal citizenship has been extended to all. Critical issues such as land ownership and ethnic equality were pushed into public discourse and onto the policy agenda. Though citizenship remains contested today, there is at least a public language whereby all Ethiopians can express their identities and help to re-frame the vision of the Ethiopian state. The 1995 Constitution and the institutions of ethnic federalism, however incomplete their application, provide a powerful foundation on which to build an increasingly inclusive and engaged citizenry. Though institutions are limited in arbitrating conflict, they can provide linguistic and procedural parameters for political competition regarding contentious issues such as ethnicity. Bolstering these formal institutions with informal (and formal) social discourse, through public democratic education in particular, as well as grassroots community development, will create avenues for Ethiopian citizens to activate democratic obligations from their communal identities into positive gains for the democratic polity.
