Slavery in the Great Lakes Region of East Africa

Introduction

HENRI MÉDARD

While there is an extensive literature on the history and anthropology of the Great Lakes or interlacustrine region of East Africa (see Map 1), there are few works which focus on slavery. This constitutes a striking contrast with the historiography of the rest of the African continent in which slavery and the slave trade are some of the most prominent themes. Until recently authors working on the Great Lakes region considered that the institution was never a feature of Great Lakes societies or that it existed for only a very short time during the late nineteenth century and even then was of marginal significance, so they did not write about it.¹ In the late 1980s Michael Twaddle re-opened the debate by demonstrating the importance and longevity of slavery in Buganda, but little new work was done in response to this. Indeed, the idea for this book came when Michael Twaddle examined Henri Médard’s 2001 dissertation on the nineteenth-century history of Buganda and commented on how little attention had been paid to slavery in it.² This led in turn to a conference held in Paris on 16–17 December 2002, hosted by the Centre de Recherches Africaines (University of Paris I/Centre National de la Recherche Scientifique) at which most of the contributors to this book were present.

The starting point of the project, then, was the need to answer very basic questions. Did slavery exist in that region? How old was it? Was the institution marginal or central to these societies? The aim of this book is to demonstrate the diversity of slave institutions across the region and to open up the subject to future and more specific research.³

The debate originally concerned Buganda and Bunyoro, two kingdoms in present-day Uganda with intertwined histories. Why enlarge the subject to the Great Lakes region of East Africa? Why not a wider area, say Africa or East Africa? The primary reason is that the Great Lakes region has a distinctive cultural unity. It is a fairly coherent area, socially, religiously and politically (the region is dominated by the interlacustrine kingdoms). Many similar institutions can be found across the region, some of which should have been structured around slavery. Furthermore, by choosing the Great Lakes region of East Africa we have tried to avoid thinking about the area’s precolonial past within the framework of the colonial and postcolonial borders, and also to overcome the conceptual biases/distortions that result from the geographical organisation of colonial archives. This geographical separation is aggravated by contemporary language barriers that divide the region into Francophone Burundi, Rwanda and the Democratic Republic of the Congo, the former Belgian colonies, and Anglophone Uganda, Tanganyika, Zanzibar, Kenya and Sudan, the former British colonies. The picture is further complicated by the existence of the written sources relating to the pre-1918 colony of German East Africa (which consisted of Tanganyika, Burundi and Rwanda) and others relating to the Egyptian Sudan from the nineteenth and twentieth centuries written mostly in Turkish, Arabic, French and English.

In physical terms the Great Lakes region of East Africa extends from Lakes Albert and Kyoga in the north to Lake Victoria in the east, Lake Tanganyika in the south and Lakes Edward and Kivu in the west. In other words, with today’s borders, it includes southern and western Uganda, north-western Tanzania, Burundi, Rwanda and the eastern fringes of the Democratic Republic of the Congo. Its two or three subregions have long maintained intensive contact with each other (in the form of, for example, origin myths, migrations, trade and competition). Linguistically, languages spoken in this region belong to the Great Lakes Bantu family.

This collection of essays shows, firstly, that slavery was much more important in this region than has usually been assumed and, secondly, that the role of non-coastal Africans in the slave trade was more complex and important than has hitherto been recognised. These conclusions indicate that slavery and the slave trade in the Great Lakes region were not so different from the rest of Africa. In particular, these essays largely reinforce Edward Alpers’ emphasis on the importance of African initiative in the East African slave trade.⁴ But the inhabitants of the Great Lakes region were much more autonomous of the intercontinental market than Africans further south, whose trade was largely focused on Kilwa and the Indian Ocean slave network beyond. The role of local demand was even more important around Lake Victoria than in the Ruvuma and Zambezi valleys.

These concepts will be explored further in this Introduction. We have chosen to illustrate the arguments made in this introductory chapter mostly with references to Buganda. This is not to privilege Buganda over the rest of the region, but instead reflects the imbalance of sources and knowledge. Much more is known about slavery in Buganda than in any other part of the region.

The Nature of the Sources

European sources relating to slavery in the Great Lakes region also tended to be distorted by the alternating imperatives of concealment or propaganda. In general, where slavery persisted, it was an embarrassment to colonial officials, though in the Great Lakes region this was more true for British and Belgian officials than for German officers. References to slavery are almost absent from colonial archival documents after 1900 in Uganda (at any rate those concerning Buganda, Nkore (Ankole) and Bunyoro). From the colonial records it would be impossible to guess that these societies had been structured around slavery only a few decades previously. The absence of post-conquest colonial references to slavery in Uganda may be indicative of a desire to avoid discussing the subject.⁵ Yet the ending of the slave trade had previously been used and over-used as a justification for the colonial takeover, not only by British, German and Belgian imperialists, but by most of the Christian missions as well (see David Northrup’s and Jean-Pierre Chrétien’s chapters). For example, the annihilation of Bunyoro in the 1890s was justified by the necessity of fighting the Arab slave trade, even though the Baganda and the Sudanese troops used by the imperial forces were as much involved in the slave trade as the Banyoro (see Shane Doyle’s chapter). The same can be said of the wars against the Baganda Muslims between 1888 and 1893 and of the Baganda rebellion and Sudanese mutiny of 1897–9.

In the absence of reliable colonial sources, Africanist historians have tended to turn to oral or written African sources instead. Documents that were written by Africans from the Great Lakes region in the late nineteenth and early twentieth centuries mention slavery, but only briefly. Most of the authors were high in the chiefly hierarchy and so were not preoccupied with such lowly people. Slavery was a dangerous subject for them, better to be avoided or belittled (the major reason for Baganda chiefs to fall out of favour with the British administration in the 1890s was the accusation of slavery). The Ugandan authors were Christians, and so did not wish to appear too knowledgeable about slavery, just as too much knowledge about women suggested sin (to make matters worse, most slaves were women).

Oral sources have seldom yielded much on the matter of slavery, for a very simple reason; early scholars, being unaware of or uninterested in the matter, did not ask about it. When they did, informants were not always eager to speak about such an embarrassing theme.⁶ However, E.M.K. Mulira in the early 1950s managed to collect for Audrey Richards a limited but precious amount of information on slavery using oral sources.⁷ Michele Wagner’s paper on Buha given orally at the conference and Eugénie Mujawimana’s undergraduate thesis on Rwanda show that the obstacles can be overcome. Jean-Pierre Chrétien, Shane Doyle and Edward Steinhart have also relied on interviews collected at different times during their fieldwork. It is our hope that this book will encourage future collections of oral data on this theme.

Early ethnographic work such as that of the missionary John Roscoe, who lived in Uganda off and on between the 1890s and the 1920s, mentions slavery, among many other things, but in the descriptive and unhistorical way of the time.⁸ The anthropologist Lucy Mair also offered rich insights, but she was interested in the Uganda of the 1930s, when slavery was no longer functioning.⁹ Quite logically many of the scholars in the 1950s and 1960s felt awkward about the exaggeration of slavery in colonial propaganda and literature.¹⁰ They strove to show newer, more positive and more contemporary aspects of the societies they were studying.¹¹

The 1970s did produce a handful of significant works, but these did not have the impact that might have been expected. In 1974 Roger Botte published the first article wholly devoted to unfree labour in the region, about the ‘abashumba’ in Burundi.¹² Their status was extremely close to slavery, and some of them could be and were called slaves (umuja), though others had kin and were free to change masters (see Jean-Pierre Chrétien’s chapter, as well as Edward Steinhart’s discussion of the related institution of ‘abashumba’ in Nkore and David Schoenbrun’s analysis of the root word-súmba). Though Botte’s article has been famous in some circles, this was more because of the contemporary Marxist preoccupations about social classes than the careful nature of his use of oral and written sources on unfree labour. It has been mostly unnoticed by scholars interested in slavery. A year later the same Marxist preoccupation about social classes led Walter Rusch to emphasise the role of slavery in Buganda.¹³ The Marxist approach has been very important in focusing attention on slavery, but its influence has declined as its theoretical aspect has fallen out of fashion. The best example of this phenomenon is in France where the formerly famous anthropologist, Claude Meillassoux, had in less than fifteen years fallen into total oblivion even among his former disciples and students, although his recent death revived interest in his work (he had, however, continued to be used in the English-speaking world). In the case of Walter Rusch, the language barrier (the book is in German) and his reliance on a limited range of sources (only published texts available in Europe) have made sure his book had a very limited impact.

One exception is worth noting, though. Rusch’s work helped John Iliffe in 1987 to write his brief but illuminating pages about the condition of slaves in Buganda. Iliffe’s focus on poverty and hunger gives an original view on slavery, illuminating other aspects of slave life than the usual labour, exploitation, violence and domination issues.¹⁴ Hartwig’s original work on Bukerebe (Ukerewe) has been widely read. However, the specialists of the more cattle-obsessed and land-based kingdoms further west detected few similarities between the societies they studied and this small island kingdom on Lake Victoria.¹⁵ Augustin Nzanze also mentions slavery very briefly in his work, but his concern is to imply that the exploitation of the Hutu by the Tutsi in the nineteenth and twentieth centuries was as harsh and shameful as slavery.¹⁶

The slave trade and its impact have received more attention than slavery as an institution, mainly thanks to the work of François Renault, Jacques Marissal and Eugénie Mujawimana.¹⁷ One important early work was John Milner Gray’s 1949 biography of a Muganda slave, Majwara. Gray showed how Majwara’s father, Namujulirwa, a very important chief of Kings Suuna and Muteesa, head of the large province of Buddu in southern Buganda, fell from grace and was executed. His family, including the young Majwara, were sold into slavery. Most of the article deals with the later life of this former slave after he was bought in 1871 from a Swahili trader in Tabora by the explorer, Henry Morton Stanley. Majwara accompanied Stanley, then the missionary David Livingstone, and then Stanley again across the whole of Africa.¹⁸ A number of other publications deal with the Indian Ocean or Sudanese trade, which taken together do contribute towards a global understanding of the slave trade. But these are essentially local studies, mainly focusing on other sub-regions outside the Great Lakes: Southern Sudan, the Nuba mountains, the Swahili coast, today’s border lands between Tanzania, Mozambique, and Zambia, or the Congo forest.¹⁹ Because of the organisation of the sources and because of the links the authors are trying to build between precolonial history and national history, the geographical focus can be slightly artificial.²⁰

These then are the sources most used by the anthropologists and historians working on the Great Lakes of East Africa. Three factors have drawn historians towards the study of slavery anew. The first is that teaching African history courses (which most of us do) compels historians specialised in this region to know a minimal amount about the general literature on slavery in Africa. This leads to new questions and new approaches to the region’s history. During the last thirty years knowledge on slavery in Africa has advanced so much that it has allowed regional specialists to look at their local sources anew. Pier Larson is the foremost example of this; he has used this huge general literature to enable him to construct a vivid image of slavery in Madagascar before the nineteenth century, despite the very scanty sources.²¹

The second reason is the systematic use by David Schoenbrun of historical linguistics as a source on slavery. The third reason is the use of neglected European sources, particularly the records of the French White Fathers mission. These sources provide especially rich insights into the history of slavery in the Great Lakes region of East Africa. Before the 1980s the written sources that had been most heavily used were those written in English (those of Protestant missionaries and colonial officers) concerning the period before the First World War, and those in both English and French (for Burundi and Rwanda) after that period. For obvious reasons, with the decline of slavery, sources related to our subject are mostly found in older documents dating from before 1914. The Catholic missionaries, many of whom were French, are particularly interesting because they actually bought and owned many slaves. The Church Missionary Society (CMS), the main Protestant mission (Anglophone and well-studied) in the area, rejected the policy of redeeming slaves, and so what they have to say on slavery is quite marginal.

It is not surprising that the first major breakthrough on slavery in the region was made in 1983 by Eugénie Mujawimana who was supervised by Father Roger Heremans, a White Father teaching history at the University of Rwanda. Unfortunately this dissertation is hard to come by, and has remained unpublished. It therefore has had little impact on the scholarly community. It seems, moreover, that Michael Twaddle’s pioneering work on slavery in Buganda²² came only after he had access to the White Fathers Archives, decades after he had first seen the CMS and colonial archives and conducted interviews in Uganda. But even with the Catholic White Fathers and Mill Hill Fathers, the amount of information collected on slavery is not large compared with other places. The White Fathers in Buganda bought only a couple of hundred slaves, whereas the White Fathers on Lake Tanganyika, just outside our zone of interest, during the same timespan bought ten times as many. This obviously had an impact on the frequency of references to slavery in the sources. Slavery in the White Fathers mission on Lake Tanganyika is much better documented. The same contrast can be found with other missions in Eastern Africa or Sudan. The Verona Fathers in Sudan, the London Missionary Society and Berlin Mission Society in south-west Tanzania, and the Holy Ghost Fathers on the Indian Ocean coast have deposited much more information on slavery than the missions in the area we are interested in here.²³

The explanation is simple. The Catholic missionaries in Buganda and surrounding areas had little need to buy converts because of their dramatic evangelical success among the free population. Most of the energy and the hopes of the missionaries were focused on these free converts, not on redeemed slaves. It is not surprising therefore that some of the richest documents containing life histories of Ganda slaves were collected not in Buganda, but on the southern side of Lake Victoria where the White Fathers were less successful among the free population and had more time to spare on their redeemed slaves.²⁴ In Buganda itself most information on slavery is to be found in documents dating from 1879 to 1883 before the number of free converts became overwhelming.²⁵ After 1883, only during the inquiry into the potential canonisation of the Uganda Martyrs of 1886, some of whom were former slaves, did missionaries pay particular attention to slavery.²⁶

After this point missionaries only tended to discuss Ganda slavery when it touched on a particular theological issue. Thus, for example, Michael Twaddle (at the White Fathers archives in Rome)²⁷ and Michael Tuck (at the Mill Hill Fathers archives in Jinja) (see his chapter) have studied documents dating from the early twentieth century concerning ‘Marriage Cases’. These were compiled when missionaries interviewed parishioners in order to make sure they did not baptise polygamists. A large part of these life histories refer to the period before slavery was abolished and before what Frederick Cooper has described as ‘the removal of its coercive and exploitative dimensions’. Such sources avoid the danger of reducing slavery to what was left, its residual ‘social dimension’.²⁸

The other important neglected sources on slavery are in German (see Jan-Georg Deutsch and Jean-Pierre Chrétien’s chapters). The standard colonial approach to slavery in Africa was that adopted by the British, French and Belgian governments, officially banning the institution but choosing to ignore continuing slaving practices. The behaviour of the German colonial authorities in Tanganyika varied significantly from this norm. The Germans kept slavery (but not the slave trade) legal and chose to control it openly (see Jan-Georg Deutsch’s chapter). For example, in Unyamwezi colonial law stated that slaves who had misbehaved could be caned but the actual execution of the sentence had to be carried out by representatives of the German authorities. The frankness of the German authorities on issues of slavery permits new insights into the nature of slavery both before and during the era of abolition.

The Impact of the Slave Trade

The use of these new sources in this book transforms our understanding of slavery in the Great Lakes region of East Africa. Many historians thought slavery and the slave trade appeared in the Great Lakes region only with the growth of the Swahili trade from the Indian Ocean coast during the nineteenth century. It is now quite obvious that this is not so. David Schoenbrun’s chapter using historical linguistics shows that slavery (or very similar institutions) was long established in this region, probably as old as the societies themselves. Michael Twaddle had an intuition of this point, acquired from more traditional sources, in his 1988 paper.²⁹ At first glance these papers go against the so-called ‘transformation theory’, which is best illustrated in Paul Lovejoy’s book Transformations in Slavery. This theory emphasises the leading role of the slave trade in the rise of slavery in Africa.³⁰ Instead, they would appear to support John Thornton’s theory that the slave trade grew rapidly in Africa because slavery was already well established.³¹ François Renault had made this point concerning the East African slave trade as early as 1971.³²

To say that slavery was ancient, however, does not mean that the institution never evolved. In other words slavery was subject to historical change. Gerald Hartwig showed this clearly in his work on Bukerebe in 1977. Slavery appeared (or new forms of the institution were established) in the seventeenth century, but was then modified in the nineteenth century through the impact of the Swahili trade.³³ Did the nineteenth-century slave trade transform these societies from societies in which slavery was incidental to their structure to societies where slavery was central? It is hard to detect a common pattern. Unyamwezi, according to Jan-Georg Deutsch, did follow this trajectory. But according to Holly Hanson’s chapter, slavery became a fundamental part of Buganda’s social structure in the eighteenth century if not before, prior to the kingdom’s connection to the Indian Ocean and Sudanese slave trade networks. In societies such as Burundi, meanwhile, it is still doubtful whether slavery had become central even as late as 1900. Most of the chapters in this book confirm that a major modification of slavery took place in reaction to the slave trade, but whether or not such change made slavery central is still a matter of debate.

It is obvious that African slavery cannot be studied independently from the slave trade. For example, in Rwanda one-third to one-fifth of the slaves were said by informants to be retained by the inhabitants of the major slave transit and trading points.³⁴ Nor can the international slave trade in East Africa and Sudan be separated from that of ivory. Even if one product seems to dominate, without the other the trade would have collapsed due to the combination of a lack of profit, transport or currency. Ivory and slaves were exchanged for a variety of imports from the wider world and facilitated the merging of the new long-distance trade with longstanding networks of African trade (in copper from Katanga or in less specialised products such as food, goats, salt or hoes.)

In the nineteenth century the Great Lakes region of East Africa found itself at the confluence of three trade networks of very unequal importance, stemming from the Indian Ocean, the Nile Valley and the Ethiopian Highlands. As Thomas Vernet pointed out at the conference in 2002,³⁵ the East African coast has a long history of slave trading dating from antiquity. This does not, of course, mean that this coastal trade never evolved. In the eighteenth century massive exportation towards the islands of La Réunion and Mauritius supplemented the older trade with the Arab world. The growth of the slave trade continued into the nineteenth century in spite of the progressive suppression of coastal exports. Overseas markets were replaced by demand from expanding plantation agriculture relying on slave labour in the Zanzibari Empire, the most famous product of this agriculture being cloves.

Before the nineteenth century most of the coastal settlements of present-day Kenya and Tanzania are thought to have not extended their contacts deep into the interior because of geographical obstacles. According to Andrew Roberts, the East African interior possessed a number of regional trade networks based on local production of, for example, salt or iron, or on ecological complementarities.³⁶ These networks were interconnected and goods from the coast, changing hands many times, would have meandered towards the Great Lakes region of East Africa (this is how François Renault understood the central slave trade in the nineteenth century, with a surplus of slaves flowing eastwards).³⁷ This explains the presence of glass beads, in very limited amounts, found in pre-eighteenth-century archaeological sites in Uganda and Rwanda. The interconnections may also account for the increase of epidemics linked to the growing contact between the coast and the interior which appears to date from the end of the eighteenth century. According to Sheriff, it was the constant rise of international ivory prices that enabled the unification of these regional trade systems into one single network during the nineteenth century.³⁸ But East Africa’s integration into the globalising economy was not solely driven by coastal initiatives. The presence at the coast of the inland groups that were most prominent in long-distance trade across the region, the Kamba (from eastern Kenya) and the Nyamwezi (from central Tanzania),³⁹ was mentioned for the first time in the seventeenth and eighteenth centuries respectively. It was only in 1825 that the first traders from the coast reached Unyamwezi. In c.1830 Lake Tanganyika was attained and about ten years later the lake was crossed. To the north, Buganda was reached during the 1840s via Karagwe.⁴⁰ Swahili were already present in Bunyoro when Speke reached the kingdom in the early 1860s.⁴¹

Map 2   The East African slave trade routes in the nineteenth century

The main route into the interior left the coast near Zanzibar from harbours like Bagamayo and proceeded through what is today central Tanzania to Unyamwezi where it branched off to either Lake Tanganyika or Lake Victoria (see map 2). Other routes, such as that through Maasailand are not as well known, though it was via this latter route that the earliest reports concerning Bukerebe reached the coast in around 1811.⁴² Insecurity and logistical problems seem to have caused this route to decline by the 1850s, before European interest in the interior of East Africa picked up. It linked the ports of Tanga, Pangani and Mombasa to Kilimanjaro and then went through Maasailand to Bukerebe and the east coast of Lake Victoria. It is unfortunate that we know so little about what is probably the oldest direct link between the Great Lakes region of East Africa and the Indian Ocean coast. We also know very little about the important Swahili community resident in what is today western Kenya, and their relations with Busoga and what is today eastern Uganda. At the end of the nineteenth century, though, the sources do indicate that the areas around Mount Elgon were hit hard by the slave and ivory trades coming from Buganda and Mumia.⁴³

The impact of long-distance trade through the Sudan, Congo and Abyssinia was much less significant than that of the Indian Ocean coast. Mark Leopold deals briefly with the Sudanese trade in his chapter. The chronology here is easier to establish. In 1841 the great flooded plains of the Nile (the Sudd) in what is today Southern Sudan were crossed for the first time by Egyptian boats. Before that no contact across the marshes can be safely assumed. The rare trade goods found among the Bari and the Pari came from Abyssinia and via Bunyoro.⁴⁴ The latter was reached by Sudanese traders in the 1860s, around the same time as the Swahili traders. Sudanese goods, however, were not competitive with items imported from the Indian Ocean. Instead the Sudanese interacted mostly through the use of force or political interference. Their relations with the indigenous rulers were characterised by conflict and violence. The successful resistance of Bunyoro and Buganda limited Sudanese influence to Bunyoro’s borderlands and the shores of Lake Albert. The highhanded behaviour of the Sudanese was very different from the submissive coastal Swahili behaviour in kingdoms like Karagwe, Buganda and Bunyoro. On the other hand, there was not much difference between Sudanese and Swahili behaviour in other areas, in the Congo forest, for example. It is telling that Swahili and Manyema who reached Burundi, Rwanda, Nkore (see Edward Steinhart’s chapter) or Bunyoro from the west behaved in a similarly brutal way to the Sudanese who came directly from the Nile. The Swahili and Manyema had adopted predatory habits in the Congo, because the longer distance their goods had to travel from the Indian Ocean meant that their prices could not compete with those of goods that had followed a more direct route. These raiders, however, were not as militarily organised as the formidable Sudanese, lacking the necessary physical might to overcome kingdoms like Nkore, Bunyoro, Burundi or Rwanda.⁴⁵ The antagonistic relations that resulted help explain the limited impact these Manyema slave raiders had on the Great Lakes region before the general free-for-all of the 1890s. As for Abyssinian traders, very little is known about them. They may have reached the area east of Bunyoro in the 1890s. At the end of the century ivory from Bunyoro may have been sold in Brava on the Somali coast.⁴⁶ The scarce sources found at the Uganda National Archives are full of potential lies and exaggerations, because the Abyssinian advance was used to impose restrictions on the movement of the local Ugandan population and to compel the British government to act against Abyssinian infiltration.

Coastal or Sudanese traders are easier to identify than local African traders in the contemporary sources because of their appearance (dress, language and religion) and the scale of their business. They have therefore attracted more attention from historians. It is nonetheless likely that traders from the interior, such as Banyamwezi,⁴⁷ Bahaya, Banyambo (from Karagwe), Basubi, Banyoro and Bakerebe did control much of the slave trade, though they usually operated on a smaller scale (one to ten slaves). These local traders were almost invisible to European commentators, and were such a normal sight that they were not mentioned by African sources. Despite these difficulties Eugénie Mujawimana has provided a detailed description of the different categories of Rwandese slave dealers between the end of the nineteenth and the beginning of the twentieth century. She describes large-scale and small-scale professional slave caravan leaders, and also occasional traders, some of them chiefs or bandits. She also mentions, but in much less detail, numerous traders from outside Rwanda, the two main groups being Basumbwa (who are sometimes considered a sub-group of the Banyamwezi) and Basubi (who are sometimes considered a sub-group of the Bazinza). Bahaya and in particular Baziba traders are also mentioned but to a lesser extent. Mujawimana unfortunately says little about the few Swahili traders. Comparing her analysis with the sources on which it is based, it is clear that she treated the Swahili more like a stereotype than an actual actor in the slave trade inside Rwanda. Rather than documenting the actual actions of the Swahili traders, she makes assumptions about them, just as her European sources assumed that their enemies, the ‘Arab slave traders’, were the key actors in the slave trade even if they were in fact absent or very few in number.⁴⁸

In 1971 François Renault wrote that in the nineteenth century ‘the [slave] trade was therefore mostly in the interior, finding its own market on the spot. What reached the coast was only a small portion.’⁴⁹ Just before, in 1970, Andrew Roberts had reached similar conclusions.⁵⁰ According to François Renault, two distinct but connected markets existed in East Africa, one around Unyamwezi and the Great Lakes, the other on the Indian Ocean coast.⁵¹ Jan-Georg Deutsch’s chapter reinforces this argument about the primacy of the internal trade, but the detail of his research requires a major change in the understanding of the East African slave trade. Deutsch tells us that the major slave market in this area of East Africa was Unyamwezi, not Zanzibar. This may also help explain why Kilwa harbour dominated the Indian Ocean slave trade to such an extent. The towns of the Mrima coast, opposite Zanzibar Island, faced too much competition for slaves in the interior. What is today central Tanzania constituted not only an important staging post for caravans travelling from the coast to the interior or vice versa. Unyamwezi was also a massive importer of slaves, with the demand coming not only from the export-oriented Swahili settlements at Tabora or Ujiji, but mostly from the Nyamwezi themselves for internal consumption. Part of the inflow of slaves into Unyamwezi came from Swahili traders who gave large numbers of them to local leaders, just as Hartwig has shown in Bukerebe.⁵² The study of the slave trade can therefore no longer be limited to the Swahili and Sudanese traders and markets. Other African buyers and traders must also be taken into account.

Poverty and the lack of food were major factors that led to the rise of slavery and the slave trade. Hartwig has shown how famines greatly increased the supply of slaves in Bukerebe. Using oral interviews and the local White Fathers archives, Mujawimana found similar processes occurring in Rwanda (see also Jean-Pierre Chrétien’s chapter). The slave trade became particularly important in Rwanda during famines, not only because of the breakdown of social relations but also because of the rise of insecurity and kidnapping. It is worth mentioning that enslavers in Rwanda and Burundi captured both Hutu and Tutsi indiscriminately, unconcerned by the distinctions between these identities.⁵³ Even in Buganda, with its reputation for precolonial food security, periods of hunger brought a surge of slaves to market.⁵⁴ In Burundi poverty was the main reason why people were compelled to become bashumba (servants or day labourers).⁵⁵ David Schoenbrun in his chapter notes, for example, that ‘the word -muhuuku (Nkore and Kiga) or -mupfûku (Hunde)’ meant ‘servant, bondsman, or slave’. The source for this noun appears to have been a verb, -kupuuka ‘to get thinner, lose weight; to fail’. Poverty and hunger influenced the origins as well as the experience of enslavement.⁵⁶

The antiquity of slavery in the Great Lakes region, which is emphasised throughout this book, has an impact on our understanding of the slave trade. Holly Hanson’s chapter, and also her recent book, (and Richard Reid’s chapter to a lesser extent), show a significant growth of slavery during the eighteenth century in Buganda.⁵⁷ This eighteenth-century crisis is fascinating. According to existing assumptions it seems too early to be linked to the Swahili or Sudanese trades. We have already noted that it is unlikely that a regular trade could have crossed the Nile marshes in the eighteenth century. It is more conceivable that the date at which significant contact was made between the interior and the Swahili coast may have been earlier than has commonly been assumed, as the technical and geographical obstacles here could have been overcome more easily than in the Nile Valley. David Schoenbrun’s chapter provides some linguistic evidence in this direction. It is interesting to note that Thomas Vernet’s recent work in the Portuguese archives has pushed the evidence of the presence of Kamba traders in Mombasa much further into the past (from 1829 to before 1689). As far as is known, the Kamba networks were not connected to the Great Lakes region, but this may be a sign that historians have underestimated the antiquity of relations between the coast and the interior.⁵⁸

The rise of slavery in the eighteenth century opens new perspectives not only on civil war and changes in the economic and political organisation of production in Buganda (see Holly Hanson’s chapter),⁵⁹ but also on the wider regional crisis associated with the breakdown of Bunyoro’s regional hegemony. During the eighteenth century Nyoro forays are recorded all over the region, in Rwanda, Buhaya, Buganda, Nkore and Busoga. The eighteenth-century explosion of Nyoro raiding was closely followed by the breakdown of Nyoro hegemony and the rise of a militarised Buganda, in a pattern which is similar to the evolutions of other regions across Africa in response to the slave trade. It is possible that the transformation of Bukerebe society may have been linked to this eighteenth-century crisis. Hartwig linked the emergence of slavery on this island to a series of invasions from Buhaya and the consequent establishment of institutions of kingship, which he believes occurred in the seventeenth century. It is likely, however, that a historian today, looking at his sources anew, might interpret the oral traditions differently. Developments in historiography would suggest that traditions which explain the establishment of kingship in terms of conquest should be treated with some scepticism, and that the antiquity of these new institutions may have been exaggerated in the oral sources.⁶⁰ Perhaps when Buhaya’s early history is finally written, Bukerebe’s history, and indeed that of the entire region, will become much clearer.

The existence of slavery in the eighteenth century raises new issues. Where did the captives go? Was there a trade? What were they used for? All these questions need further investigation on a regional scale. Wrigley and Vansina were the first to mention the desire for captives as an important element in warfare in pre-nineteenth-century Buganda and Rwanda.⁶¹ But they deal with this issue extremely briefly. Vansina seems to regard the importance of slave kidnapping during warfare as so obvious that he does not feel compelled to demonstrate its existence. He seems simply to rely mostly on the derivation of the word umuja (female slave), whose origins have been analysed in a recent study by David Schoenbrun.⁶² Wrigley has the same assumptions and tools as Vansina but he can at least rely on a wider range of sources, since written accounts indicate that Buganda was a provider of slaves for Bunyoro in a far away mythical past.⁶³ What made slavery suddenly attractive in the eighteenth century? Could it be that Unyamwezi created a regional slave trade a century earlier than has previously been assumed, or were there internal innovations in some of the kingdoms that suddenly made slavery more appealing or possible on a much larger scale? Michael Tuck’s and David Schoenbrun’s conclusions definitely suggest that the initiative came from an internal evolution within the Great Lakes region.

This issue of the internal evolution of slavery and the slave trade is extremely relevant to the transformation debate, which has hinged on the question of whether Africa played a major role in the international slave trade because slavery already existed there, or whether slavery became common in Africa as a result of the slave trade. In the Great Lakes, Bukerebe, Buganda, Bunyoro, Toro, Ussuwi, Buhaya and Busoga appear to have played the major regional role in the nineteenth-century slave trade because slavery within these societies was already considerable, whereas the early slave trade bypassed those areas such as Burundi and Nkore which had very few slaves. Certainly in the 1880s and 1890s, the role of domestic slavery in the development of the slave trade in Rwanda was very obvious.⁶⁴

The varying significance of domestic slavery helps explain the very unequal integration of Great Lakes societies into the nineteenth-century Indian Ocean trade. Areas in which slavery was already considerable participated willingly in the Nyamwezi and Swahili trade, whereas the others first refused contact, then either subsequently accepted it reluctantly and subject to restrictions or had it imposed on them by force. As a whole, therefore, the region was being dragged by external forces into new forms of slavery and the slave trade. The pre-existing pattern of slavery gave shape to the new slave trade, but once involved in the intercontinental slave trade, local slavery or dependence underwent noteworthy changes (see Richard Reid and Michael Tuck’s chapters). This is in concordance with Manning’s, Miers’s, and Willis’s arguments that the ‘expansion of slave holding’ tended to ‘corrupt and subvert and otherwise transform kinship systems’ and in this case other aspects of the ‘social order’ as well.⁶⁵ Thus the experience of the Great Lakes region does not fully support either side of the transformation debate (see David Schoenbrun’s chapter). Moderate proponents of the transformation theory, like Paul Lovejoy, who do not exclude autonomous inventions of slavery in Africa, are quite right to emphasise the impact of the intercontinental slave trade on local African slavery, but John Thornton is also correct in insisting on the key role of African slavery in the making of Africa as a major slave market.⁶⁶

The nineteenth-century slave trade had a strong impact on slavery. One must take into account that the long-distance slave trade in the Great Lakes region had a rather short lifespan compared with societies on Africa’s Atlantic coast. It is possible that the full impact of the new slave trade did not have time to develop fully in East Africa. If one considers Buzinza, Burundi, Rwanda and Nkore, it seems possible to detect an increasing involvement in slavery. For example, towards the end of the nineteenth century, Rujumbura, a small kingdom that was later incorporated into colonial Kigezi, organised itself for predation (see Edward Steinhart’s chapter). Nkore during the 1890s became more and more involved in the slave trade, emerging as a major market for slaves captured in Buganda civil wars, most notably between 1897 and 1899, and allowing Nyamwezi, Haya and Swahili traders to cross its territory regularly. In Rwanda at the end of the nineteenth century, during the reign of Rwabugiri, indirect relations with long-distance commercial networks linked up its indigenous war economy with the slave trade.⁶⁷

Hartwig’s pioneering work on Bukerebe showed that the growing trade in people in the nineteenth century resulted in the marketisation of social relations, bringing about new, harsher, commercial forms of slavery. Institutions akin to adoption, fostering and pawning were transformed into chattel slavery.⁶⁸ Jonathon Glassman has defined the commodification of social relations on the Mrima coast of the Indian Ocean as ‘[t]he uneven