Ethnicity does not matter in the long-term perspective. Such was the conclusion formulated by a new generation of ‘Africanists’ in the 1970s and 1980s, and it was a kind of scientific revolution. Its defenders held that in sub-Saharan Africa, ethnicity had mainly been created through European colonial rule, and was, therefore, an entirely artificial concept.1 For a period that roughly coincides with the 15 years between 1975 and 1990, the attack against the well-established idea of primordial ethnic groups in Africa – which had dominated anthropological thought from the colonial period onwards – seemed to win the day.2 In spite of Eric Hobsbawm and Terence Ranger’s initiative to understand the ‘invention of tradition’ with a view to identifying the creation of group sentiment in a comparative and global approach, however, reflections of historians working on group identity in the African continent have rarely entered the debates on global history.3 While migration and connection – for example, over the Atlantic or the Indian Ocean – are essential themes in global historical studies, they do not yet interact with the analysis of ethnicity that
In the public debate about ethnicity, the new interpretations from social anthropology and historical research on sub-Saharan Africa have had very little impact from the outset.5 Even the ‘subjects of analysis’, including elites that would eventually read such studies, do not at all seem to feel that they live according to roles constructed by others. Among the local populations, we encounter a general feeling of certainty that ethnic criteria explain group affiliation and group hostilities.6 One might even argue that while ethnicity was deconstructed as a guiding principle by historians and anthropologists, the concept has become increasingly important for political and social relations in many parts of sub-Saharan Africa. Not only is it employed by political analysts and journalists from outside, who wish to simplify African topics for their readers or audience,7 but African populations seem to embrace it, without being manipulated to do so: categories of ethnicity appear to play an essential role in their life.
A good example of the reappearance of ethnic solidarity after periods of rupture is the effect of the 2007 elections in civil-war-torn Sierra Leone. In this small West African country, ethnic categories had been eclipsed in many areas during the 1990s, as a consequence of the Revolutionary United Front (ruf) rebellion.8 The civil war dramatically destabilised the existing patron-client networks based on ethnic labels.9 However, ethnic categories had not disappeared from national politics, as exemplified by the surprise win in the electoral
However, other examples from West Africa appear to show the opposite trend, at least at first glance. In Senegal, electoral behaviour and ethnicity do not seem to be at all linked: it has been held that ethnicity has lost its role and that the independent Republic of Senegal has been remarkably free from ethnic dispute as a consequence of successful social management.12 The obvious exception has been the separatist rebellion in Senegal’s southern province of Casamance, where the Mouvement des Forces Démocratiques de Casamance (mfdc) is undoubtedly dominated by Jola-speakers. Nevertheless, the movement’s leaders often describe their goals as ‘regionalist’ and not as ‘ethnic’ (while Wolof-speakers in the region indeed fear the ‘Jola’ as dangerous ‘autochthons’).13
In other Senegalese regions, it is far more difficult to find signs of tensions arising around ethnic labels. It would, however, be worthwhile investigating
In the present-day Trans-Volta Region of Ghana and in south-west Togo,16 ethnic allegiance is presented as irrelevant by central authorities. However, it has remained an important category of self-definition and has been crucial during moments of violent regime change. In particular, the Ewe-speakers of Ghana’s Trans-Volta Region claim to have been an underprivileged minority before the ascendancy to power in Accra of Flight-Lieutenant Jerry Rawlings in the 1980s and the electoral victories of the National Democratic Congress (ndc) after 2008. The Ewe case relates to the question of stability of national boundaries, as the Ewe language community can be found on both sides of the international border between Ghana and Togo. Cross-border claims of the Ewe-speaking community in the 1940s are, probably, the most outstanding West African case of political leaders demanding a revision of colonial borders and the creation of an ethnically homogenous territory.
All over West Africa – and elsewhere during the twentieth century, including in Western Europe as Bambi Ceuppens and Peter Geschiere have shown – ideas of autochthony and of ethnic claims have frequently been intertwined with questions of an ‘authentic’ or ‘traditional’ organisation of power.17 In the colonial period, this search for ‘authenticity’ was both a rural and an urban
In the border area between Togo and Ghana, chieftaincies are, to the present day, still fairly intact, at least on the surface of the institution. In current-day Ghana, chiefs appear to have considerable social prestige, while the decades of the Eyadéma dictatorship have made them a kind of ‘traditional bureaucrat’ in neighbouring Togo.22 By contrast, in the case of Senegal, chiefs lost their official role in district administration after 1959, although the institution continues to exist at village level.23 In Sierra Leone, the power of the chiefs waned only slowly under the independent state from 1961, as ‘traditional rulers’ continued to be of some importance during electoral events and in local
Ethnic claims are thus strongly connected to problems of the local and regional organisation of power, and therefore remain, in the contemporary period, an important category of political discussion.25 Although this does not yet prove the importance of the concept under pre-colonial and colonial conditions, it needs to be taken into account as a factor. In a second step, global historians need of course to ask if this trend only holds importance in sub-Saharan Africa. On the conceptual level, I will therefore discuss, later in this introduction, how the category is used in different parts of the world, and in the second chapter I will analyse the broader set of identifications on which individuals and groups relied in the nineteenth and twentieth centuries.
It is indubitable that in the post-colonial period, a number of major conflicts in the African continent have been caused or reinforced by perspectives of ethnic group antagonism.26 ‘Autochthons’ have been mobilised against ‘strangers’, and violence has been justified by rhetoric about different ‘tribes’ and the stealing of rightfully possessed land.27 In this context, the shock of the Rwanda, Burundi and Congo massacres had the most forceful impact on research paradigms and on the view of African group solidarity in a global comparison.
The Shock of the Great Lakes Massacres
With regard to violence defined as ‘ethnic’, post-colonial West Africa has rarely been on the centre-stage (with the exception of northern Ghana and Côte d’Ivoire in the 1990s and 2000s). On the contrary, the Central African experience of the 1990s has had a massive impact on public and scientific debates, and led to a revision of central premises in research on Africa. The Rwandan
The position held by Central Africa in the discussion about ethnicity as a historical factor brings us to questions as to whether such observations can be generalised for other parts of sub-Saharan Africa. An astonishing number of studies on African history and cultures do not even expound the problem that ‘Africa’ is only a geographically defined ensemble and, moreover, an artificial concept of thought, but not a cultural entity. It is very probable that the particular idea of a joint ‘African culture’ refers to a political dimension, which was developed during the colonial period and has been confirmed by the argumentations of leaders and intellectuals of emancipationist movements active from the interwar period. It is difficult to see why ‘African cultures’ should be part of a united culture, except if these cultures are described as ‘Black cultures’.
Ethnicity and Global History (and Historiography)
While ethnicity has become a factor in studies that are part of the move towards global history, such studies have mainly favoured the question of networks and shifts in identification during migration and diaspora situations.31 The move towards a global labour history is exemplary for showing the importance of processes in ‘non-western’ parts of the world.32 More often than not, however, it has not tackled concepts that are very much ‘reserved’ for one particular world
Some scholars argue that ‘ethnicity’ as a factor of cultural difference works similarly in entirely distinct geographical arenas of the world.36 Their argument is perhaps not completely groundless, but existing designs that compare, for example, conditions in the Balkans with ‘tribes’ in Central Asia and warring groups in Sierra Leone, lack historical grounding and reflection about the categories of group identification. In other words, before attempting a comparison that brings in examples from various continents, it would be preferable to obtain more reliable analytical results for the history of ethnic affiliation in any of the regions taken as exemplary. Also, the use of ‘ethnicity’ as a category needs to be more broadly questioned.
The incongruent employment of ethnic affiliation as a category is indeed very problematic for global historical studies, and an immense concern for any study that wishes to bring African history into broader, global debates.
For sub-Saharan Africa, ‘ethnicity’, as a trope, has somewhat replaced the colonial concept of ‘tribe’. In academic use, ‘tribe’ has become an unacceptable term – and rightly so, as it is based on a biased and negative image of static communities and of underdevelopment, an image that follows colonial traditions. The terminological change towards ‘ethnicity’ probably enhanced the accuracy of group descriptions and eradicated some biases. ‘Ethnic group’ now appears to be a widely accepted concept for the African continent, whether scholars regard those groups as ‘primordial’ or as ‘constructed’.
However, specialists of sub-Saharan Africa tend to forget that the picture is different in other geographical zones. In the Americas, the notion of ethnicity as a concept to describe group identifications is far more hybrid, and often contradictory. As for Amerindian groups as a subject of study, the expression ‘ethnic group’ is employed for smaller communities, and, ultimately, also for language groups (like Apache, Crew, etc).39 The same is true for the descendants of Maroon communities – communities of refugee slaves organised in
Concerning Europe, the use of classifications prompts similarly worrying reflections. The expression ‘ethnic’ appears to be mainly reserved for the Balkans region, and much less for the broader area of Eastern Europe.43 It is very unlikely that the category of ‘ethnic group’ would be employed for the Czechs or the Slovakians when discussed in the context of the Austro-Hungarian
In public debates, this choice might again reflect a certain kind of racist preconception. The inconsistent employment of the concept of ‘ethnicity’ in different zones of Europe is at least suspect: from this perspective, the Balkan wars as somewhat ‘archaic’ incidents where groups clashed over deep ‘ethnic’ roots and allegiances, appear as fundamentally different from conflict in ‘civilised’ parts of Europe. From such a point of view, the use of the expression ‘ethnic cleansing’, which became popular in representing the massacres of ‘Serbo-Bosnian’ troops and of the Yugoslav (‘Serbian’) army during the two war phases in the first and the second half of the 1990s, is not really surprising. The category of ‘ethnic group’ is also employed in order to describe the different sides during the conflict in former Yugoslavia.46 In some cases, the use of this expression seems simply absurd, and is, from an empirical point of view, contradicted by facts: thus, during the 1990s civil war in Bosnia-Herzegovina, any distinction through ethnic criteria is entirely inappropriate, while religious practice was partly crucial for mobilisation. Attempts to publicly brand the Muslim population of Bosnia-Herzegovina as ‘Turks’, show the artificial nature of ethnic concepts in this context. Individuals speaking the Shtokavian variant of the Serbo-Croatian language, and not being descendants of migrant populations were defined as ‘Turkish’ in war propaganda because of their Muslim religious practice: this demonstrates very well how flexible definitions of belonging can be
The Spanish state is the obvious Western European example to illustrate the terminological differences inside European history. Few of the observers would call ‘the Spanish’ an ethnic group (even in a study of migration); but, likewise few scholars would address ‘the Catalans’ or ‘the Basques’ or other regional identifications as ‘ethnic’ (although this position has recently changed through the manipulative effort of Catalan ‘nationalist’ politicians).49 On the contrary, in academic studies, those group solidarities would probably be called ‘regional identities’ and described as ‘nations’ in (mostly politically motivated) analyses that favour the autonomist positions of some regionally active political movements (in which ‘nation’ reflects an older sense, with a meaning of regional group in a neutral sense).50 However, it remains to be asked why Catalan (or, perhaps, Bavarian, or Auvergnat), should count as a ‘regional identity’ – but Wolof, Temne, or Ewe is characterised as an ‘ethnic group’.
More worrying and confusing from the global historical point of view, and very close indeed to similar problems to be found in North America regarding immigrant communities, it frequently occurs that individuals of (allegedly) visible African or Caribbean origins are described as members of an ‘ethnic group’, namely as ‘Black’ (ultimately vis-à-vis ‘Caucasian’).51 The latter complication in
Exemplary Experiences in West Africa
I will attempt in this book to come from one broader, West African region to results on global issues of identification and community-building. My comparative design will include the Wolof living today in the post-colonial countries of Senegal and the Gambia, the Temne of present-day Sierra Leone, and the Ewe of the contemporary countries of Ghana and Togo (Map 1). Much more than some weak linguistic similarities and resemblences with regard to patrilineal family structures, the three groups share a broader historical experience. I will in what follows enumerate relevant factors; the early interaction with European merchants, the spread of larger religions imported from outside, and the confrontation with measures of colonial control and activities of taxation, are quite similar for all cases.
All linguistic groups that are known about had been fairly stable in their group structures in West Africa’s coastal regions.53 They all had an early experience with European merchants coming to their respective coasts to trade, mostly in relation to the traffic of slaves.54 The Wolof-speakers in the states of Kajoor, Bawol, Siin, or Saluum in Senegal and the different smaller communities
Therefore, the Wolof, Temne and Ewe, and their neighbours, were influenced by their early interactions with Europeans.57 Wolof-speakers had been neighbours of French factories and fortresses since the later seventeenth century (and they had known other European merchants from the fifteenth century).58 Temne-speakers had to cope with the neighbourhood of the British colony of Sierra Leone only from 1787, but they had earlier contacts with Europeans.59 The Ewe were integrated in networks at the Keta Lagoon that linked them to the Danish in the region, and with British merchants at Little-Popo.60 The three regions therefore had a considerable number of middlemen as catalysts in a certain political and cultural homogenisation of the societies: such as the Signares in Gorée, Creole settlers and their Temne partners; and intermediaries in the Gold Coast’s southern coastal plains.61
In religious terms, all three groups were entangled, during the period between the mid-nineteenth century and the 1960s, in a process of religious conversion and homogenisation. This process has sometimes been discussed in the literature as a reaction to the shock of colonisation and of the
Both communities – Wolof and Temne – were less concerned by the effect of jihads, but most of their members turned to the Muslim faith through the long-term work of itinerant clerics (‘marabouts’). The process often advanced more notably under the conditions of colonial rule, when the ‘pacification’ of large regions had enhanced the mobility of preachers. The European colonisers in the end became the staunchest collaborators of the Sufi brotherhoods. They favoured the social role of those religious groups, and employed them as
The religious evolution of the Ewe is analogous to the advancement of Islam in Senegambia and Sierra Leone. In the region east of the Volta River, the new impact of the Christian religion followed similar patterns. Protestant missionary efforts protected by the German and British colonial rulers, later followed by Catholic colleagues, guaranteed that Ewe-speakers gradually converted.65 The link to official colonial policies of this missionary effort was even stronger than cooperation with Islamic leaders in other parts of West Africa: missionaries more directly took part in educational matters, and they held the keys to access to particular linguistic and professional skills that helped to start careers under colonial rule (something that Muslim brotherhoods could not offer). In all three cases, European officials regarded religious leaders as allies and collaborators in the control of the region, regardless of whether they were Christians or Muslims.66
Reactions to socio-economic change under European rule are also comparable. All three communities were inclined to resist the tax demands of the new colonial states. The Hut Tax War by the Temne-speakers in 1898 rebelling against British tax measures was the most impressive case. In the Volta River area, Ewe-speakers evaded tax demands using the colonial border, which they also profited from for contraband trade.67 In the coastal regions of Senegal south of Dakar and Rufisque, mass flights happened as well, although the refugees were Sereer-speakers rather than Wolof-speakers. In the Gambia, Wolof-speaking groups also fled from tax payments using their ‘borderland situation’.
We can thus rely on a set of common experiences shared by the three larger population groups that are the focus of my study.
I will also show that another factor – that of the basic political organisation of communities – is crucial to the use of ethnic categories. For this reason, it makes sense to compare communities with important differences in their political structures, which were influential from 1850 or earlier onwards.68 The negative relationship between the formalisation and the strength of state structures on the one hand, and reliance on solidarities formulated through ethnic terms on the other hand, will be a principal observation in this analysis.
Political entities in Senegambia were mostly ‘strong pre-colonial states’ within a ‘relatively stable political system’.69 They were dependent upon larger pre-colonial administrative routines and institutions – including regular taxes, ruling dynasties, and a certain type of provincial government and officials. In the southward direction towards the Gambia River, Wolof-speakers still held political power, but the units in place were far smaller: in many cases, these Wolophone rulers were themselves subjects of Mandinka-speakers or Fulfulde-speakers. Therefore, for the Wolof-speaking members of such communities, political institutions were experienced as far less reliable.
Both the Temne and the Ewe had less elaborate political structures than those existent in Senegambia. Among Temne-speakers, political leaders never controlled larger territories. Some had at least a certain charisma as war leaders, but they did not command a provincial administration in their own right. Amongst the Ewe, the picture is mixed. Anlo on the Keta Lagoon, and Peki close to the Volta River, are two political units that could rightly be counted as pre-colonial states. After 1833 Peki had a certain influence, for decades, over a larger group of Ewe-speaking inhabitants of the region. Anlo made a name for itself, in the coastal areas, as a rather important political player. However, as I will discuss in detail, political leadership did not bring with it any more formalised, more durable position of power. I will shed light on the relationship between the pre-colonial political experience, and the emergence of identifications in the colonial period, through the long-term perspective used in this study.
Finally, we come to the effect of urbanisation as an experience under colonial rule, with beginnings that are even older. All three cases involved rural populations that fed one or more growing urban centres in the vicinity of their main regions of settlement – Accra and Lomé in the case of Ewe-speakers; Dakar and, to a much lesser degree, the agglomeration of Bathurst in the Gambia for Wolof-speakers; and Freetown and the Sierra Leone Peninsula for
Therefore, we find on the one hand considerable similarities in the historical trajectories of the three population groups used as case studies: their integration into the Atlantic slave trade, their early encounters with European merchants and company personnel, their similar experience of the European conquest, and the colonial contribution to the rise of a regional majority religion, which all follow comparable patterns. These common historical trajectories posed similar problems for the communities in question, and, in particular, for their political elites and rulers. On the other hand, the three groups analysed find themselves at very different degrees on the scale of pre-colonial state organisation. I will still discuss ‘state’ and ‘state institutions’ as one variable for forming identifications, in contrast to ethnic status, community affiliation, family group, or religion.
A comparison of three groups is challenging. This is even more the case if the results are supposed to be put into a global historical perspective. The approach necessitates a methodology that diverges from the standard repertoire of area studies, and is flexible to the needs of comparison and integration into larger trends.
The Methodological Panorama: A (Critical) Return to Colonial Sources and the Afro-European Encounter
Global history has long gone beyond the point where it had to receive input from specialists on sub-Saharan Africa to overcome its Eurocentric nature, as
Historians working on sub-Saharan Africa have, over the last five decades, established new standards in the employment of source material. They have become increasingly opposed to the Eurocentric and sometimes openly racist vision that characterised research on Africa until independence (and sometimes far beyond).74 This vision has been tackled by ambitious area studies, where historians have sought interdisciplinary approaches through the employment of field interviews and the quest for oral traditions in particular.75 The principal idea was to find alternative voices and to bypass the monopoly of documents written by the colonisers. This process has been successful in many respects, and particularly so where detailed regional studies on small communities are concerned.
Obviously, the reliance upon oral information brings a number of problems with it regarding the verification of empirical data. Documents can be re-encountered, but for interviews the situation is mostly impossible to reproduce.76 A number of historians have discussed these problems in more detail, and have given practical advice for the use of oral information.77 In spite of these debates, some of the broader implications that a strong reliance of the historian upon field interviews causes would still need more discussion. I do not intend to offer such a discussion in my book, as my study can only employ these results from more regional and local approaches to question my own interpretation. My study will hopefully give a basis for new fieldwork. The research design employed here is not that of a local history; therefore, oral interviews do not have the same importance. Comparative analysis cannot be based on the isolated, village-level field interviews, and not even upon unsystematic interviews with political and social leaders, which are also typical for local histories. A good comparison of identifications over broader regions needs particular moments in which these identifications are mobilised; these moments are spread over a region and can often only be found with recourse to the unifying effect of the colonial encounter, and to colonial observations. Crucial moments of mobilisation are the situations of – administrative – encounters with agents of the colonial power, when spokeswomen or spokesmen of communities presented their identifications to these agents. This historical encounter can mainly be found in written sources, of which an important percentage has remained entirely uninterpreted.
For the broad and comparative research design, I will thus rely on an approach where the interaction between local populations and structures of colonial rule guides us towards the critical encounters in which identifications were presented and renegotiated. This approach allows me to profit from archival documentation in new, innovative ways (concentrating on ‘the African voice’ inside this documentation). Oral testimony may give additional illustration, and
Oral testimony still holds another role. Anthropological observations by European visitors were often integrated into local perceptions and group identification. In some cases, the results of interviews held in the colonial period were used in local strategies of argumentation. For example, the ‘traditional ruler’ of Kudje in Buem State in former British Togoland, as a local informant, referred explicitly to European scholarship in a 1964 complaint to post-colonial Ghanaian president Kwame Nkrumah. In this account, with which the ruler tried to foster his claims for an independent paramount chieftaincy, Akuamoa iv referred to the early correspondence of missionaries in the area. This correspondence, which Akuamoa had read in its published form, then became the cornerstone of the chief’s argumentation. His ingenious employment of available historical information, reinterpreted as part of his oral narrative, shows that informants were quite apt to integrate older oral information they had found in published form.79
Also – in the same region – the monumental account on the Ewe by Jakob Spieth was reread in its German version by locals who had obtained proficiency in the German language before 1914. They subsequently translated their impressions into demands for political privileges in the 1950s, for example with regard to the conflict between the communities of Ho-Dome and Ho-Bankoe disputing the paramount chieftaincy of Ho Division. The two communities in conflict based their claims on oral narratives that consistently built in references to the authority of the ‘Western scholar’!80 Similar
Migration claims are one of the typical elements of information that was orally given to colonial administrations in moments of interaction. Informants, who wished to sustain the privileges of their group, were often engaged in presenting long traditions of such migration.81 While the experience of long-distance movement has become a principal issue of global historical approaches, little has been achieved in view of understanding traditions of migration as claims for political primacy and demands for land. In a global history perspective, it would be an important future challenge to discuss the role of such traditions in the world and over time.82
It appears, therefore, to be reasonable to concentrate here on African voices as they appear in documentation. This appearance is, as I will show, very frequent. The key moments are characterised by the entanglements of the colonial encounter, in which informants had to ‘sell’ their role and their privileges, i.e. present them as ‘authentic’ to the colonial authorities. However, this approach naturally brings us to question, once again, the reliability of written material. If we want to use it to analyse the mobilisation of identifications and even claim that the voices of Africans can be filtered out of Eurocentric biases, we need to discuss its problems. Obviously, much of the documentation was produced by an administration of rulers who were relatively unfamiliar with local realities, whose impressions were skewed by cultural preconceptions and stereotypes, and who did not quite understand the strategic elements of using group labels – such as ‘Temne’, ‘Wolof’, or ‘Ewe’.
Colonial administrators were interested in a number of basic tasks, such as taxation and recruitment of labour. Self-perceptions of colonial subjects were not necessarily a part of their concerns. From the nineteenth century onwards, some European officials were nevertheless engaged in pseudo-academic classification, as ethnographic ‘amateurs’, although largely in an insensitive and unreflective manner. However, some of these filters have become better
Nowadays, the critical historian is able to cope with these challenges. Moreover, administrative reports carry an underestimated amount of voices of African informants – and the historian, with some precautions, is able to analyse them in order to explain group solidarities.86 Also, the colonial context needed some element of accurate reporting and of engaged analysis, and information coming from ‘native clerks’ normally had sufficient grounding in their societies of origin to bring in at least some of the respective narratives circulating in these societies.87 In other words, even if intermediaries attempted
With regard to political entities and their relationship to communities, European administrators were particularly interested in more strongly centralised regional structures, which they perceived as being in a ‘decadent’ or ‘primitive’ state, but as comparable to the results of nation-building in the European continent and elsewhere in the world.88 Where no identifiable structures were immediately found, it was often judged appropriate to create them, to in effect give the locals back ‘what had been theirs’. Classifying the customary behaviour of the surrounding ‘people’, or ‘peoples’, was thus regarded a necessary undertaking in order to govern those groups in ‘appropriate ways’. It is evident that such an enterprise again left much space for misunderstandings and mistakes, and opened the doors to processes of invention of traditions – even if Europeans understood some of the basic terminology of group labels in the region.
It is a central goal of this book to understand whether the labelling of groups was only a top-down process fuelled by the colonial administration and by its closest collaborators. Several scholars have made this point over the years, but they have rarely analysed the concrete situations where communities or some of their representatives explained or ‘sold’ their group identifications to the agents of the colonial state. Therefore, I will attempt to understand under which conditions such concepts were crafted and received, adapted and utilised over time by the locals themselves. I will then interpret the preconditions of group conflict in the different West African regions.
Mediated by the colonial documents, it is possible to find out when local groups relied on particular arguments and labels to engage in a creative communication, in which those labels assisted them eventually to gain their point. During the colonial period, locals had no choice but to engage in discussion of those categories when approaching European administrators. Access to land, quests for political leadership, demands for administrative autonomy were all issues in which opinions about ethnic solidarity were important. Many of those formulations are documented, through petitions, letters, and process minutes. What they tell us about ethnic affiliation may not be the established truth over the centuries, nor need they reflect the effective cultural bonds during a circumscribed period. Nonetheless, such documents describe the limits of formulating solidarity, they correspond to how far local leaders could go in
This debate over ethnicity and solidarity is an aspect that in documents makes itself visible in important ways. For colonial documents whose authors had an established knowledge of the employment of local information, and with an awareness of the vested interests of informants appearing within the sources, profound analysis is nonetheless possible. Unsurprisingly, it is still more difficult to describe the situation at the end of the pre-colonial period through European documentation. However, even in that period, European visitors sometimes show good knowledge of local conditions and group relations, and in the more important cases it appears plausible that they must have repeated claims of reliable local informants (although this has to be pointed out on a case-by-case basis). Although such travellers’ accounts are, of course, ‘Eurocentric’ in the sense that we do not normally learn who exactly their informants were, there is no need to reject the information contained therein altogether, particularly if these accounts are compared with later, colonial sources.89 The changes of the conquest period (between 1840 and 1900 for our West African cases) lead in practice to an engaged discussion of group solidarities, and to the temporary appearance of extremely violent struggles. During these struggles, both the support of the European invaders, and that of imagined ethnic brethren, were frequently welcome. It is thus unsurprising that those circumstances offer us a large range of material, as local populations had to redefine their alliances and to reformulate their claims, which they did by interaction with the European conquerors.
The approach chosen in this study is, given the complexities outlined above, both modest and bold. I hope to interpret how, in their interactions with a frequently under-informed European administration, African communities (and, usually, their elites and spokesmen) in coastal West Africa presented themselves. Under which circumstances did they refer to ethnic labels? Can we make out a sort of consistent pattern that describes such recourse to ethnicity? Did the establishment of colonial rule after the turbulent period of conquest stimulate a more frequent employment of such ethnic labels – through their imposition on local societies by colonial classification, for example, as has frequently been argued in the literature? Or would the stabilisation of (alien) political structures mean that the recourse to ethnicity became less attractive? And, finally, how could such results usefully be put into a panorama of global patterns, concerning the use of ethnic claims?
Africans used their claims to belong to groups of a particular identity under very divergent circumstances. In the phase of colonial conquest, such claims were usually made to ‘inform’ a potential European partner about the legitimacy of warfare in a given region. The goal was to convince Europeans of the necessity of intervening in favour of the ‘traditionally legitimate’ party. In the post-conquest phase, with local warfare effectively banned, questions of local power inside the colonial structures, and access to land, became more important. The conquerors set up or employed ‘authentic’ native courts, where such questions would abundantly be discussed.90 European administrators became increasingly convinced until the end of the First World War, that an ‘authentic’ local administration was both useful and just. Officials in the field were thus eager to understand the group settings in the regions they ruled, and were more and more open to accepting information from the African side. African local authorities were the first to approach colonial agents, in order to obtain the latter’s support in land cases in which they referred to the ‘traditional rights’ of their respective group in a number of cases.91 However, while the brokering of ‘traditional rights’, and access to traditions of origin, were normally linked to local power, other individuals were not necessarily excluded from the process. In fact, particularly for the well-known case of the pre-colonial state of Asante coming under British control, it is frequently claimed that locals, including socially marginal groups such as unmarried young women, were very quickly informed about loopholes in the new set-up of administrative regulations that would allow them to obtain more individual rights.92 The same seems to have happened regarding access to land, and land claims in which individuals were remarkably capable in finding tactics to bypass the authority of local elders and chiefs were a common phenomenon in many zones under colonial rule.93 It did not take these individual applicants a long time to see the value of well-presented versions of group history. Therefore, they quickly began to submit their own ‘traditional’ claims before the colonial administration.
For a comparative historical study, using the European administrator as an involuntary intermediary, who collected, filtered, and interpreted claims of different forms of group allegiance, still bears the risk of a distortion of facts due to misinformation, misunderstandings, or simple lack of interest from the European side. Nevertheless, this method also has simple but unquestionable advantages. Accessibility and testability of the employed documentation is one part of this picture.94 Moreover, the engagement of the European administrators also allows the historian to formulate a broader perspective on processes. In a way, the colonial administration, by collecting material, already guides the scholar to locations where ethnic mobilisation was perceived to be strongest. By contrast, the fact that in some broader regions ethnicity does not appear at all as a particular topic in administrative reports, has – even in accepting the misunderstandings conveyed in some of the material – usually some significance. Obviously, in these cases, the historian is first obliged to inquire if the regional administration was particularly uninterested or inactive, or if the local situation was particularly complex to understand: however, long-term absence of any notions of ethnic classification, by a number of different individual administrators, given the growing interest of the metropoles in drafting ‘authentic native legislations’, is certainly a significant fact.95
In the context of colonial rule, it is equally necessary to keep in mind that the different European administrations bear some distinctions, although those distinctions were perhaps not as influential as scholars have long believed them to be.96 In their discourse, British administrators laid far more emphasis on the maintenance of ‘authentic’ local structures. One might thus expect that in the British-ruled territories appearing in my West African comparison – Sierra Leone, the Gold Coast Colony, Togoland under British Mandate from 1914, and to a certain extent the Gambia – the engagement of the administration in classifying individuals according to ethnic criteria would have been far greater. However, the French concept of ‘association’ led to the same effects. This concept became part of the French strategy of administration in the first two decades of the twentieth century, and more strongly so after the First World War. All in all, while theory and rhetoric emphasised a
Structure, Sources, and Limitations
Ethnic labels, whether long-standing or recent, constructed or ‘primordial’, are not only an African phenomenon but a problem that needs to be discussed in a global panorama. Moreover, such labels are never the only marker of identification. The nature of ethnicity is, quite frequently, defined on unstable grounds, and different scholars speak of different things when discussing questions of ‘identity’. The picture is complicated further by the obvious links of ethnic sentiment to other modes of self-definition, such as ‘religion’, ‘nation’, ‘state’, or, even, ‘race’ (the latter as a highly problematic, imagined category usually relying on perceptions of phenotype), plus, whenever we cross over into a more local arena, feelings of family or kin solidarity. Family and kinship are difficult to conceptualise in a broader matrix, as they are dominated by more localised patterns. The other different categories can, however, be deployed in a larger debate, in which they either combine with and ultimately reinforce the effects of ethnic group sentiment, or function as alternatives and counter-elements.98
Chapter 2 will engage in a conceptual discussion of ethnicity and of the various other categories of identification that can be set in relation to ethnic
As has been explained above, my comparative analysis of processes of ethnic mobilisation will rely on three principal case studies, each of which involves a larger community that is, today, considered to be an ethnic group (the Wolof, Temne, and Ewe). This analysis will naturally include the complex relations of the three communities with their direct neighbours. Relationships with neighbours are obviously important factors, as the concept of ‘otherness’ formulated on either side, often characterises the process of creating, maintaining, and employing internal group identifications.
Given the limitations of oral methods for this larger and comparative approach, which I have amply discussed above, my analysis will mostly rely on archive-based interpretation. For each territory, the interpretation of processes over a period of, roughly, 110 years (and some decades more in the case of Sierra Leone, where the process of pre-conquest diplomatic interaction becomes important by the first half of the nineteenth century), will be based on a larger territorial and one or two smaller, sub-regional narratives. The territorial narrative relies on the interpretation of ‘self-perceptions’ that local populations
The sub-regional setting describes, per case, one or two local regions, which appear particularly promising in relation to the larger territorial conditions. The more detailed cases can be found in regions where the group regarded for the comparison had a complex standing, normally because of conflicts with other groups, ups and downs in the assertion of political power, and shifts in the perspective on local identifications. The sub-regions – like the Petite Côte around Joal-Fadiouth in Senegal and Lower Saloum in the Gambia in the case of Wolof-speakers; the region of Port Loko for the Temne-speakers; the surroundings of Avatime in regard to the Ewe-speakers – are analysed in detail through data from an in-depth study of archival information. For the local view on group interactions I also use data from some selected local interviews (which with regard to the analytical approach taken, only guided as to principal patterns and questions). It has to be emphasised that I consider the regions selected as sub-regions as representative for the cases discussed, exactly because they tend to be encountered at the conflict-ridden margins of regions inhabited by the respective groups, where identifications often had an even more crucial function for group cohesion.
Six West African and nine European archives have been consulted during the work on this comparative study. Among the West African archives, those of Ho in Ghana (the former administrative capital of British Togoland), those of Lomé in Togo, and those of Banjul in the Gambia have very rarely been used in studies on colonial history, and their interpretation can be regarded as highly innovative. The National Archives of Ghana in Accra contain a vast selection of
Each of the three cases discussed here is first presented in a general panorama, introducing the different (language) groups to be found in the respective territories, the more generalising results of sociological studies on those groups, the religious situation, and the specific experience of European conquest. Depending on the case, I also address a certain ‘state-of-the-art’ on ethnic interpretations, presenting the existing knowledge around 1850, and attempting to comment on the sources of those broader descriptions, and on their terminology and perspective. In subsequent parts of each chapter, I will concentrate on situations of intense regional debate about group rights, as they were defined in the interaction with colonial administrations. In this context, all the case studies have to follow a certain chronology that addresses key moments of community evolution in relation to the colonial encounter.
We find three typical situations for discussion about group labels led by the local populations themselves, belonging to particular phases. First, the actual situation of conquest, embedded in complex diplomatic exchange, is of great importance. It forced local populations to explain to the coloniser their predilections for certain forms of regional or local organisation, and it allowed ambitious individuals to counter established rules and narratives, and formerly subservient groups to elicit a reorganisation of local structures of power and dependency. The second key phase has to do with the reorganisation of administrative structures, during attempts at installing civilian rule, in the period of the First World War and the early 1920s. From that moment onwards, candidates with a claim to ‘authenticity’ had more chances to succeed. However, they had to know how to employ group labels and ‘traditional
For those, and for the other more singular periods, I will thus analyse cases of mobilisation and identification of locals, in their interaction with the colonial officials. This analysis will, in theory, privilege our principal sub-regional cases, but integrate them in a broader, global discussion of local identity politics and mobilisation strategies. In Chapters 3 (on Senegambia), 4 (on northern Sierra Leone), and 5 (on the Trans-Volta Region and Western Togo), I will point out how, over the above-mentioned period of approximately 110 years, local populations ‘sold’ their group identifications to colonial administrators, sometimes, if not frequently, with clear tactical motives. The explanations given for the different processes will be strongly linked to the role of political organisation and statehood; in other words, it will be shown that ethnicity had a role especially where the mechanisms of statehood failed or were non-existent.
This brings us back to the limitations of this study. First of all, given the conclusion of the methodological discussion, I do not regard it as practicable to carry out any statistically significant or larger qualitative interview series; my approach favours a broader perspective. Indeed, this should in the end be one of the particularly strong points of the present book. Second, the broader approach will be grounded in reliable local and sub-regional case studies, but it depends on the perspective of the representatives of the colonial power for the act of selecting those cases. Nevertheless – as argued above – while it might be correct to describe the colonial administration as racist, generalising, and often ill-informed, colonial officials had a natural interest in finding and defusing group conflicts, and they were the usual persons to be approached in conflicts
Third, the source material appears, at first glance, to sometimes be uneven. While the cases of the Wolof and the Ewe rely on documents that are mainly retrieved from African archives, and – at least for the twentieth century – on documentation that has so far scarcely been used or that is even unknown, the discussion of Temne interaction with colonial structures needs more input from European archives, which is due to conditions of access to materials. The different locations of documents can also lead to certain (although rather slight) differences concerning the time period studied. On the Wolof, the archival documentation is more or less balanced, with a strong input from archival documents stored in France and the United Kingdom for the period before 1914, and a near exclusiveness of the more detailed documentation from African archives for the period after the Second World War. The Temne case, on the contrary, has a certain bias in favour of the second half of the nineteenth century and the 1920s, two periods during which British officials were most interested in bundling documentation and depositing it in the Colonial Office. In the Ewe case, early documentation is particularly available from (sometimes printed) British files kept in the Colonial Office, and from the documents of the German administration of the new colony of Togo, while the British administration in Keta also left some valuable dossiers in the National Archives of Ghana in Accra. From 1915, this picture is completely reversed, with an increasing amount of detailed local descriptions, most of those to be found in Accra, Lomé, and Ho.
Those biases are no real obstacle to carrying out the comparison. On the contrary, the approach permits us to bridge certain lacunae in local material. Moreover, the existence of missionary societies in all three geographical arenas – most prominently the Pères du Saint-Esprit (Holy Ghost Fathers) in coastal Senegal, the missionaries of the Church Missionary Society in Northern Sierra Leone, and those of the Norddeutsche Missionsgesellschaft (Bremen Mission) in the area east of the Volta River – also contributes to making the individual regional situations more comparable. This refers, above all, to missionary documentation in the second half of the nineteenth century, before the actual conquest, when those missionaries had a more pronounced diplomatic role.103
In sum, even with the given limitations, the detailed and research-based comparative approach represents a study design that is as such entirely novel for historians of sub-Saharan Africa. The results can therefore be expected to lead to new interpretations, and will consequently be brought into a debate on group affiliation that links the perspectives of African and global history. Such a perspective constitutes important progress in debates on ethnicity.