CC BY
62
Slavery and the Early State in Africa
Martin A. Klein
University of Toronto
ABSTRACT
In modern societies and in classical antiquity, slavery was largely a source of labor and slaves were at the bottom of the social structure. In earlier societies and in many many non-Western societies, most slaves were concubines, soldiers or servants of the state or of its ruling classes. They were thus often well off. The argument of this article is that slaves played a particularly important role in the emergence of the early state in Africa. The early state involved a struggle of strong men and their families to rise above their peers, to establish the right to rule others, and to impose themselves on neighbors. To do this in societies with no tradition of legitimate authority, they needed the ability to coerce. They also had to tax people who often resisted the emergence of the state. They thus depended on kin, clients, and slaves. Of these groups, slaves were the most dependable because they were totally powerless. They could not compete for power themselves and depended on the emerging rulers for their well-being. The article reviews the literature on African states to demonstrate the importance of slaves in their political structures.
The treatment of slavery by modem scholars is very much shaped by a concern with forms of slavery in the capitalist world, on its fringes and in classical antiquity, which until recently played a central role in European education. These were societies labeled slave societies by Moses Finley and slave modes of production by Finley's Marxist rivals (Finley 1976: 819-821)1. In them, most slaves were involved in productive labour, and a reliance on slave labour played a defining role in state and society. Slavery, however existed in the vast majority of complex pre-industrial societies.
Social Evolution & History, Vol. 8 No. 1, March 2009 168-194 © 2009 ‘Uchitel’ Publishing House
Its widespread existence created reservoirs of controlled persons that early modern European states could tap to meet the needs of their developing colonies. In these other societies, slaves were often used in very different ways. Productive labour was sometimes important as was labour on public works, but slaves were also soldiers, administrators, merchants, servants and concubines. In populous societies like China, slavery was not a particularly efficient way of meeting labour needs, but it was important in meeting very important felt needs in state and society, particularly the desire of the very wealthy and powerful for services (Wilbur 1943: 243).
In modern European slave systems, the state and the legal system played a major role in protecting and encouraging the use of slaves, but slavery was not significant in the structure and functioning of the state itself2. The situation was different in many earlier slave-using states. The slave trade into the Middle East, for example, involved relatively small numbers of people destined for productive labour. Slaves imported from both Eastern Europe and Africa were largely used as concubines, soldiers and servants. The largest user was the state itself. Large imperial harems were important centers of political life3. Slave soldiers were often the best trained and staffed the most important military units. Slaves, some of them eunuchs, played an important role in government. For the Ottomans, slavery was the way the court and the state reproduced itself (Toledano 1993, 1998; see also Erdem 1996). All princes were the sons of slave concubines and their careers were often managed, at least in early years, by their mothers. Slavery also provided the elite Janissaries, the corps of eunuchs, and young males, who through adoption into the homes of administrators, were able to rise to positions of wealth and power. If we broaden our inquiry and look elsewhere, we see that slaves in many precapitalist slave systems were more important as part of the state structure than as productive labour. I will be limiting my inquiry to Africa, South of the Sahara.
THE EARLY STATE
Recent thought about the evolution of the state tends to see an evolutionary process from decentralized stateless societies to chief-doms and then to early states. These transitions were not necessarily easy ones and probably in most cases worked themselves out over long periods of time. Elman Service defined a chiefdom as
a ‘socio-political organization with a centralized government, hereditary hierarchical status arrangements with an aristocratic ethos but no formal legal apparatus of forceful repression, without the capacity to prevent fission’ (1975: 16)4. The chiefdom is usually quite small. Much of political life is handled in face to face relations and fission is common, branches of the chief's descent group breaking away to live somewhere else. The early state imposes political and territorial links on kinship and has some formal structure of administration. It is larger, though not necessarily very large. Cohen (1978a: 35) goes further and argues that a major difference is that the chiefdom is prone to fission, which the early state controls through the creation of a bureaucracy, which he defines as ‘a set of officials, priests, nobles, slaves, eunuchs, and their suborganizations who manage the administrative machinery of the state hierarchy’. Cohen goes on to describe:
... the early state is a centralized and hierarchically organized political system in which the central authority has control over the greatest amount of coercive force in the society. Sub-units are tied into their hierarchy through their relations to officials appointed by and responsible to a ruler or a monarchical head of state (Ibid.: 36).
Controlling fission is not the only problem faced by the early state. The ruler must support a bureaucracy. In order to support what Service called ‘the legal apparatus of forced repression’, the monarch must tax or in some way transfer resources from his subjects to those who serve him. As states emerge, they also increasingly come into conflict with each other or simply seek to extend their resources by bringing hitherto stateless communities under the state's control. To do this ruler must have force at his command.
Several other things must be recognized. The first is that the emergence of the state is always linked to social stratification and to the transfer of resources from the mass of the population to a ruling elite (Claessen and Skalnik 1978). Second, it is important to recognize that political evolution can take different paths and involve different processes. It is not unidirectional and can take a variety of different forms. Third, the early state covers a variety of different kinds of society. The question that is poorly defined in the literature is when the early state becomes the mature state. Finally, the transition from chiefdom to early state was not necessarily
a sudden transition and that it was often strongly contested. People were not always willing to accept the domination of others.
A major problem the ruler faces is legitimacy. He often relies on members of his own family, in particular, his brothers and his sons. There may also be other noble families, who believe they have the right to rule. These are major sources of fission. Primary loyalties were more likely to a kinship group than to a state and that as the prize grows more valuable, those who want to contest its control become more determined. Furthermore, the number of potential contestants increases because one of the symbols of chiefly power is a harem, but the more wives a ruler has, the more sons he has and thus, more competitors for succession. In controlling the evolution of the state, the ruler has to gather around him men who cannot compete for power and who can only achieve wealth and power by supporting him. In more developed social structures, as in early modern Europe, kings could develop a cadre of officials of common middle-class origin. Such a group did not exist in chiefdoms and in most early states. The king could sometimes rely on priests, traditional or Muslim, and many did. In West African Muslim states, for example, there was often a committee of electors that came into operation when a king died and chose his successor. Muslim clerics were important on these committees. It is unlikely, however, that traditional priests were a major part of the administrative hierarchy of many states. The persons most likely to serve loyally were outsiders, persons who were totally dependent on the royal will and whose well-being depended on being his faithful instrument. Slaves were thus the quintessential royal servants, rarely the only ones, but often crucial.
The origin of slavery was probably the question of what to do with captives. Before slavery, captives could either be killed, tortured or incorporated. Women were certainly always incorporated. Gerda Lerner (1983, 1986) has argued that they were therefore the first slaves, but they were in most cases incorporated as wives5. First choice of female captives probably went to the chiefs, to war leaders, if they were different and to priests. These thus had more offspring, perhaps an advantage at an early stage of social evolution, but a problem as social differentiation increased. Once a community made a decision that male captives could be usefully exploited, the chiefs also probably got the largest numbers. These
slaves may have liberated the chief or his wives from agricultural labour, or at least, reduced the need, and they probably provided food and drink for the chief's hospitality. The acquisition of slaves in early societies was one way either big men or descent groups increased their numbers and thus, their power. Slaves in these societies lived within the family and had a vested interest in that family's well-being. Early slave systems often had rapid incorporation of the slave. Thus, when conflict arose, the slave was part of the royal or chiefly household and often became part of the royal entourage. With increasing differentiation and greater reliance on some form of taxation, that role may have become more important than his productive role, and if the male slave learned quickly, he could become an important advisor and emissary of the chief. My argument here is that small numbers of slaves in the chiefly entourage were probably often a major factor in the transition from chiefdom to early state.
In many parts of Africa states trace their origins to arrival of migrants with military skills, sometimes on horseback, and with a state model. The relationship between the new arrivals and earlier inhabitants is often preserved in social structure. The heads of indigenous families preserved control over the land and often held key offices. Traditions of the state's early growth often indicate a slow and tenuous process. The newcomers are often described as arriving with their dependents, their griots, and their slaves. This may be true, but traditions indicate a complex process. For example, in the Sereer kingdom of Siin (Senegal), the kingdom spread out from a base near the coast and over a period of more than a century, came to incorporate between 50,000 and 75,000 persons. Similarly in the neighbouring kingdom of Saalum, traditions of the third king (called the bur), himself a migrant in the 16th century, speak of a series of eight wars. Most of them were with ‘kingdoms’, which are now a short distance from Kahone, the capital. It is clear in Saalum that the transition involved conflict between a series of chiefdoms, rarely much larger than a village, that produced a state. In both cases, the transition in the form of the polity and the extension of its scale demanded a force, which consisted mostly of the young males, but within that structure, the early Burs needed an entourage they could depend on.
AFRICA BEFORE THE LONG DISTANCE SLAVE TRADE
The evidence on the evolution of slavery before the external slave trade became a factor is limited. Much of what we know about the early state, in Africa and elsewhere, comes from archeology. While the archeologist tells us about differences of wealth and the centralization, but he does not tell us how the king finds agents he can depend on. There is, however, some interesting data. In Paths in the Rainforest, Vansina uses linguistic data to reconstruct the evolution of human society in the Congo rain-forest. He sees as one of his base points the emergence of households headed by big men, organized within a kinship ideology, but based primarily on the ability of the big man to mobilize wealth and force. The key was the acquisition of persons. He acquired women through different marriage strategies and generally had numerous offspring, but he also used the gift of women to attract men. The acquisition of persons was clearly central to his power. Those persons included kin, clients, serfs (a term Vansina uses for hunter-gatherer dependents), dependents and captives. Free men seem to have been free enough to shift their allegiances. The big man could be most assured of those he controlled, his slaves (Vansina 1990: 74-76).
Big Men emerge at different times as central figures, sometimes as a kind of state system, at other times as the nuclei of centralized states. Vansina describes the process among the Mangbetu of the northeastern Congo, where improved weapons facilitated increased warfare in the 18th century. ‘The former term for “pawn”, nkoli, acquired the new meaning of “prisoner of war”, and with prisoners a form of slavery appeared’. Households were transformed into chiefdoms: ‘...clients had become more numerous, slaves were procured by trade or as prisoners of war, harems became larger, and Houses used specialized semi-mercenary warriors...’ (Ibid.: 173-177). As states emerged in the 19th century, slaves were important in different ways. Female slaves were a source of food production and a good to be distributed to clients. Though female slaves were generally preferred, male slaves were an important part of the royal army and could be among the king's most faithful advisors. The vast majority of male slaves belonged to the king. When an important king was killed by his sons because he was too reliant on his slaves, his most important slave commander avenged his death and set up a rival dynasty (Keim 1983: 146-147).
Many writers have suggested that slaves were important throughout African history because land was freely available and labour was not (Thornton 1992: 74)6. This idea was most fully developed by John Thornton in his provocative Africa and the Af-
ricas:
Slavery was widespread in Atlantic Africa because slaves were the only form of private, revenue-producing property recognized by law. By contrast, in European legal systems, land was the primary form of private, revenue-producing property and slavery was relatively minor... Because of this legal feature, slavery was in many ways the functional equivalent of the landlord-tenant relationship in Europe and was perhaps as widespread ... it was the absence of landed private property - or to be more precise, it was the corporate ownership land - that made slavery such a pervasive an aspect of African society7.
Slaves were used in Africa both for production and as instruments of political power. Thus early in the first millenium, there were palace slaves, who seem to have balanced the importance of near relatives of the king in the Axumite state (Kobishchanov 1978: 156).
With Kanem we have a clearer picture, but also the influence of the export slave trade and models of slave use from the Arab world. Kanem first appears in the historical record in the 8th and 9th century as a loosely organized and largely pastoral polity. Lacking other export goods, Kanem's links to North Africa were built on the slave trade. It is not clear how extensively slaves were used within Kanem in the early centuries, but there were clearly slave communities by the 11th century. Kanem went through a ‘time of troubles’ during the 13th and 14th centuries, but eventually reestablished itself in the late 15th century under Ali Gaji as Bornu. The nobility were assigned villages as fiefs to support their military and administrative roles, but were required to live in the capital, leaving representatives to manage their fiefs. Many of them also controlled villages of slave cultivators. Their influence was counter-balanced by palace officials, many of whom were slaves, some of whom were eunuchs. The state was divided into four units, two of them under the control of eunuchs. Ali Gaji also used the Muslim ulema to counterbalance the nobles (Barkindo 1985; Cohen 1978b).
Kanuri models influenced the Hausa states, which went through a period of Islamization and centralization in the 15th century. Mohammed Rumfa of Kano built a new palace, expanded the royal harem, constructed new city walls, and radically extended the role of royal slaves in government. Eunuchs were given positions that dealt with the royal household, but also with warfare, the city gates, the treasury, tax collection and the execution of criminals. Key positions also went to virile slaves (Stilwell 2004: 44-47; Hunwick 1985a). Similar developments seem to have taken place in other Hausa states and in subsequent years, were developed even further8. The creation of a professional corps of slave officials meant that there were slaves who married, raised families, had wealth and privilege.
Data is limited for the earlier period of the Western Sudan. The earliest known area of complexity was the inner delta of the Niger, where Roderick and Susan McIntosh have traced urban civilization to the 1st millennium BCE. They have strong evidence of occupational specialization, but not of slavery or of a centralized state (McIntosh 1988). There is, however, strong archeological evidence of warfare in what is now southern Mauritania (Munson 1980). Presumably, this involved the taking of slaves, but we do not have a full picture of slaves in Ghana. The role of slaves in political systems increased as states became more complex. In 13th and 14th century Mali, however, slaves and former slaves played an important role at the Malian court, and one even briefly usurped power (Levtzion 1963: 345; see also Al-‘Umari 1981: 265; Ibn Ba-tuta 1981: 290-293). We have much more documentation for Songhai. Under both Sonni Ali (1464-1492) and the Askia Mohammed (1493-1528), Songhai armies brought large numbers of captives back to Gao. The palace hierarchy included both eunuchs and virile slaves. Eunuchs guarded the harem of the Askias and held a number of intimate household officers. They also held military commands and administrative offices. Askia had a standing army, which was probably a largely slave force and included a eunuch cavalry unit. There were also numerous slave villages, which probably fed the court and worked under slave officials (Hunwick 1985b).
The whole Sudanic region was marked by the development of states south of the desert that were able to prey on less well organ-
ized but more populated areas further south. Sennar developed on the White Nile in the 16th century. Both the monarch and the provincial lords maintained a professional army of slaves. The smaller courts commanded less than a hundred, but the larger ones had hundreds of slave soldiers, many of them with horses and chain armor. It was the largely servile cavalry who did most of the slaving. Bruce suggested that ‘Slavery in Sennar is the only true nobility’ (O'Fahey and Spaulding 1974). Many high officials in Sennar were also slaves. Dar Fur was similar. Between 1750 and 1850, Dar Fur was probably the most important source of slaves for the Egyptian market, but it was also a major user of slaves. Dar Fur also depended on slave horsemen to raid for slaves in areas further south. The palace involved a complex hierarchy of slave bureaucrats, soldiers, guards, servants and concubines, including both eunuchs and virile slaves. The most powerful figure, the controller of the palace, was an eunuch. Eunuchs used positions in the harem to influence policy and get important offices. The capital was surrounded by slave regiments. The court and the more powerful figures also exploited slave labour on agricultural estates (O'Fahey and Spaulding 1974, part II; O'Fahey 1985, 1973). The role of slaves gathered from areas further south was also important to the Central Sudanic states of Wadai and Bagirmi9.
Eunuchs were more dependent because they neither had nor could create family. The role that the eunuchs played generally flowed from their role in the harem, where they dealt with the ruler in an informal setting and were often close to royal favorites and to those of their offspring with a chance at succession. They often played a major role in the education of princes and remained advisors long after the prince left the harem. Both eunuchs and virile slaves were found in the palace hierarchy, but eunuchs tended to be more important in intimate roles. They were in positions of trust, could exercise great power if they had the support of their royal masters, or conversely could be sharply reduced. Stilwell tells the story of a slave official, Allah Bar Sarki, in colonial Kano who was promoted at British insistence to the post of waziri, normally restricted to nobles. Many free officials did not cooperate with him and when he started to use his position to amass followers and ignore the emir, the British resident agreed to demote him. He was then publicly humiliated by being stripped of his fine robes and
forced to wear only his slave loincloth in public (Stilwell 2004: 154-155). There almost certainly was some Islamic influence in the development of African royal slavery, as Levtzion suggests (1988), but it is well to note that we have the evidence of the use of slaves in government from other polities in both Africa and the Middle East.
For example, we find royal slaves in coastal areas of West Africa. Kea writes about the Gold Coast in the 17th century, a time when the area was an importer of slaves rather than an exporter. Every noble had in his household wives, concubines, slaves, pawns and clients (Kea 1982: 104-112). The states were quite small, but rulers maintained forces that sometimes exceeded 2,000 men, who served as bodyguards, protected the royal residence and accompanied the ruler to war. These royal entourages were ‘recruited from the slave population or they were hired professionals or mercenary soldiers’ (Ibid.: 134). They were a turbulent and disruptive group, who could be hired for the wars of others. In Benin male slaves were also important in the entourages of the more powerful figures (Ryder 1969).
THE ATLANTIC SLAVE TRADE AND THE USE OF SLAVES BY THE STATE
In Africa the sale of slaves ‘born in the house’ was generally prohibited as was often the sale of persons who had been integrated into the host society. The pool of slaves that was tapped by European slave traders was largely those in the limbo between enslavement and integration as well as criminals, witches, outcasts, and perhaps, enemies of the chief. The impact of the slave trade was probably slow, but its long-term impact was important. As the scale of the trade slowly mounted, peoples near the coast learned to more effectively defend themselves, sometimes by becoming slavers in order to better arm themselves. This meant that the slave trade pushed deeper and deeper into the interior. Whereas slaves exported in the 15 th century came from fairly close to the coast, by the 18th century, most were coming from the far interior, and merchants had become efficient at moving slaves on foot or in large trade canoes. Different mechanisms were used to extract slaves. Armies were sent forth from many kingdoms in annual expeditions. In other areas, slaving bands raided far and wide. Others
were condemned for debt or by the courts and kidnapping was a problem.
In the late 17th century, the development of sugar plantations in the West Indies led to a dramatic increase in the demand for slaves, which in turn, fueled warfare and the emergence of a series of large states oriented to the production of slaves and the transformation of others. Between the mid-17th and early 18th century, a series of states emerged, in which war was a central activity and slave warriors a central part of the slaving operation and in many cases of the structure of the state. Warfare became the most important economic activity of much of Western Africa because it rewarded the elites and because it was the way the state reproduced itself. Though slave exports increased steadily from under 100,000 during the first 50 years of the trade to over 6 million in the 18th century, the number of those enslaved who remained within Africa was probably larger than the number exported. While most of those kept were women, a certain number of young males were also taken in as apprentices in slave forces, starting out as grooms or weapon-carriers and eventually becoming full-fledged soldiers.
Oyo developed during the 17th century. Based on its cavalry, Oyo brought most of northern Yorubaland and important Middle Belt states like Nupe and Borgu under its control. Power in Oyo, as described by Robin Law, depended on control of wealth and supporters. This meant that powerful men invested in slaves and wives (Law 1977: 62-64). The most powerful, the Alafin probably had several thousand slaves living in his palace or adjacent compounds. They included eunuchs, who guarded the harem, but also held senior positions at court dealing with justice, religion and administration. The ilari were bodyguards, messengers and tax collectors. They were distinctive because their heads were shaved and scarred. Particularly important were the ajele, who represented the alafin in key towns, particularly along the strategic trade route to the coast. Recruited from among the palace slaves, their major function was to watch the activities of local officials and keep the Alafin informed (Ibid.: 67-70, 110-113). Most of the slaves came to the north and were Muslim, which increased their dependence. They were encouraged to remain distinct. Slaves also made up a good part of the army10.
The Asante bureaucracy seems to have been less dominated by slave officials, though slaves were numerous in the army. Asante dealt with the problem of separatism by creating a bureaucracy, in which men rose to high positions on the basis of merit and service to the Asantehene (Wilks 1966; see also Chazan 1988). There were officials of slave origin and there was a significant number of slaves in the household of the Asantehene (Wilks 1975: 451; McCaskie 1995: 95-100; see also Rattray 1929: 92)11. Most of the lower ranks of the army were slaves, though McCaskie says that they could not own guns. They seem to have been issued guns when on campaign (Lewin 1978: 62; McCaskie 1995: 98). The role of royal slaves, then, as in so many African societies, was as the ruler's personal instruments and his confidential advisors. Male slaves were less important in Dahomey. The army was based more on conscription. There were eunuchs in Dahomey and Allada, but their role seems not to have been as important as elsewhere (Law 1991: 81-83). In Dahomey women, many of them female slaves, seem to have exercised some of the functions fulfilled by male slaves elsewhere (Bay 1998). In Whydah, there was a body of palace slaves. In the years before Dahomey's conquest of Whydah in 1727 King Huffon was trying to use these dependants to assert the power of the monarchy in a struggle with the more powerful chiefs (Law 1991: 103-104).
The major source of slaves exported from the Senegal and Gambia rivers was what is now western Mali. Here, Segou was founded by an age-set, the ton, which under the leadership of Biton Kulibali, expanded its raiding activities and liberating the ton from the control of their elders. In the process, they created a state, which became a major slave-raider. Some of the male slaves became tonjon, slaves of ton. The Segou state was built up by their slaving activities. They also at times controlled the succession. Along with a breakaway state, Kaarta, they were the leading source of slaves in what is now Western Mali. Slaves were the army and most of the administration, they sometimes interfered with royal succession and one slave chief even became fama (king) (Roberts 1987; Bazin 1975).
Among the Wolof and Sereer on the Senegal coast, slave warriors seem to have existed early. The first Portuguese visitors describe large entourages and refer to the powerful having numerous
slaves, but there are no references to royal slaves before the 17th century. Known as ceddo, they were attached to major lineages and did much of the fighting. Their leaders called farba played an important role in all local political systems. The title was originally of Mande origin. The role of the ceddo was expanded by Latsu-kaabe Faal, who united the states of Bawol and Kajoor in 1695. Latsukaabe increased his slave holdings, bought guns, the first to do so, and put them in the hands of his ceddo (Barry 1988: 127136; Searing 1993: 18-26). Other Senegambian states followed quickly, relying on slave administrators and on slave forces, which rose to about 500 men in the 19th century. The ceddo wore dreadlocks and bright clothes, were heavy drinkers and could be a disruptive force (Searing 1993; Klein 1977). The farba was not the senior chief, but in 19th century meeting with the French, it was often a farba who was sitting next to the ruler.
The picture we get of slave use in Central Africa is more ambiguous. There were well-developed institutions for the transfer of persons between lineages as either pawns or slaves (Miller 1981; 1976: 76-77). They were useful in getting rid of undesirables, but also probably cemented relationships of power and wealth, persons being transferred from the weaker to the stronger group. Male slaves were supporters and female slaves were the mothers of warriors and workers. Central African data was important to Miers and Kopytoff in building their controversial argument that African slavery involved essentially decreasing marginality and thus the rapid incorporation of slaves into host clans (Miers and Kopytoff 1977: 1-81). The rise of states in the region involved the effort to control rival lineages and competition from members of the ruler's own family and royal retinues, composed sometimes of slaves and strangers, were a variable. Miller writes:
The transfer of former lineage mates to dependent positions in other groups tended to occur more frequently in areas subject to several larger and more centralized political institutions, usually headed by kings ... These king's authority depended in no small part on their ability to hire or force people away from their primary loyalty to kin and to incorporate these slave-like lineageless dependents as royal retainers residing at court; there they became personal servants, craftsmen, mercenaries, and other sorts of agents of the king (Miller 1976: 77; see also Reefe 1983).
In spite of this, rulers constantly struggled to control competing claims to power. Miller tells the story of Afonso I of the Kongo, who in his 40 year rule, was for a while successful in using his alliance with the Portuguese to create resources and exclude collateral claims to power. The problem was that those seeking to increase slave exports were able to play off provincial authorities against the Kongo monarchy. By time Afonso died in 1543, the state had been much weakened12.
Further south, the most important use of slave soldiers was by the Portuguese prazeros. Holding grants called prazos in the Zambezi, these Portuguese officials took wives locally and created mini-states that used depended on their ability to use their relationships with both the Portuguese state and the various Shona authorities. They depended on chikunda, soldiers, traders and officials of the prazero state (Isaacman 1972; Isaacman and Isaacman 2004). Elsewhere, wherever Europeans and Africans intersected in the organization of the slave trade, whether under European sovereignty (St. Louis, Luanda, Zambezi valley) or under African sovereignty (Gold Coast, Lagos, Loango), slaves tended to be important not only as workers, but as traders, administrators and soldiers. Slavery was often the only way persons could be recruited.
THE NINETEENTH CENTURY
Beginning with the British abolition of the slave trade in 1807, there was a gradual closing of the export trade from Africa, first in the Atlantic and then in the Indian Ocean and Northeast Africa. The effect, however, was not to reduce the use of slaves in Africa, but to increase it. There were a number of reasons for this. First, states in much of Africa were organized around the production of slaves and depended on slaves for a wide range of purposes. The warfare engendered by the slave trade limited the possibility of any free recruitment of labour, though free labour migration did develop in southern Africa during the last third of the 19th century. Second, the abolition of the slave trade was linked to the industrial revolution, which sharply increased the demand for African products, most importantly in West Africa, palm oil and peanuts, but also ivory, gum, cloves, sesame, coffee, and late in the century, rubber. After British abolition of the trade, the price of slaves dropped sharply, but by the early 1830s, it had risen to the earlier
level (Lovejoy and Richardson 1995a, 1995b). It was to drop again in the later part of the century as a third factor came into play, repeating rifles that made it possible for small highly mobile forces to enslave more effectively than ever before.
The Muslim jihads of the late 18 th and 19th century generally did not depend on slaves. In fact, some of the jihads began as efforts to restrict the impact of slaving on Muslim communities. Often, the jihad leader freed slaves willing to join his forces. Islam does not, however, oppose slavery anymore than Christianity does13. Most of the states formed in the jihad ended up slaving and exploiting slaves in a more systematic way than their predecessors. In Macina, in the middle delta of Niger (Mali), jihad leader Shayku Ahmadu had to create an army to protect his supporters, a bureaucracy to collect taxes to pay for the army and an education system to produce the bureaucrats and jurists the state needed. To do this, it was necessary to develop agriculture. Slave raiders provided the labour for this effort. Similarly, in a more prolonged struggle Al Hajj Umar Tal recruited men from his native Futa Toro to fight for Islam, but in later years men came from the Futa to join the forces of his sons in order to accumulate slaves and farm the land with them (Hanson 1996). Similarly, in Senegambia, many early ji-hadists used Islam to justify slaving activities that had no serious religious goal.
The jihadists found that they needed slaves in other ways. Usuman dan Fodio had attacked the system of royal slaves in northern Nigeria, but the royal slaves had developed knowledge and an esprit de corps that made them indispensable to a movement led by clerics who had never done anything but study and teach. The royal slaves, in turn, were like many civil servants infinitely flexible in what policies they carried out (Stilwell 2004, ch. 1, 2). Similarly in prolonged struggles the use of slave soldiers became indispensable. After returning from the pilgrimmage to Mecca, Umar visited many Muslim states and then settled in the foothills of Guinea's Futa Jallon, where he gradually accumulated disciples and slave retainers (Robinson 1985: 338). After the jihad began, some of these slave retainers received important military and civilian commands. In addition Umar's troops, which were often outnumbered, suffered heavy losses, but after major victories, many of the defeated soldiers swore allegiance to him and joined his army.
This included the once strongly pagan tonjon of the two Bambara states of Kaarta and Segou. With time, these sofa units of his army became more numerous than the Senegalese and Futa Jallon regiments he started with (Ibid.: 183, 267, 334-339).
At the same time, there was a penetration of the southern Sudan, southern Ethiopia and the interior of East Africa that lasted for the whole of the 19th century and consisted largely of slave military forces. The conquest of the Sudan by Egypt's Muhammed Ali was motivated in part by his desire for African slave soldiers, but it had the effect of incorporating the northern part of what is now the Sudan in the Egyptian state. From the late 1830s traders began penetrating the south in quest of ivory. They found an area with limited commercial development and few people interested in selling ivory to them. The traders early found that without a market, it was not easy either to purchase ivory or food and services they needed. They soon expanded their activities into slave-raiding and built zaribas, fortified posts manned by slave soldiers, fed by the labour of slaves, making their profits mostly from the sale of slaves. By the 1870s, the most powerful of the Arab traders Zubayr Pasha controlled most of the Bahr-al-Ghazal, a house of cards built on the labour of slaves and using slaves to man almost all positions in the state (Gray 1961; Beswick 2004).
The penetration of East Africa involved alternatively competition and collaboration between Arab and Swahili traders based in Zanzibar and African state builders, mostly Nyamwezi from western Tanzania and Yao from Malawi and Mozambique. The Arabs and Swahilis created cities like Tabora and Ujiji, where most of the labour was servile and most of the agricultural producers, who worked farms around the cities were slaves. Some of the traders linked themselves to expanding African rulers, but others created new polities (Sheriff 1987; Alpers 1975; Bennett 1971). The most remarkable was Tippu Tib, a Swahili trader from the coast, who built a state along the upper reaches of the Congo River in a rainforest area that was traditionally decentralized. It was based as much as the zariba of the Sudan on the labour of slaves as well as on administrators and military leaders of slave origin. Their African rivals operated in areas where polities were traditionally small and authority diffuse. To break out of the limitations on their authority, the more successful African state-builders used slaves
and what the Nyamwezi called ruga-ruga, rootless men who were left without family ties by the constant violence that ravaged much of East and Central Africa.
Their efforts were facilitated by the development of new weapons, first, breech-loading rifles that were loaded from behind, could be fired and re-loaded in prone position and were accurate at a much greater distance than earlier weapons, and then, repeating rifles, which loaded a magazine and therefore could keep up a stream of accurate bullets. Buying these weapons was as much a strain on African economies as modern weaponry are today, but they enabled small bands of men to dominate and destroy as never before. In West Africa, there were a series of these men, who created transient polities, often on the basis on slave warriors. The most effective was Samori, who first put himself into slavery to ransom his mother, but quickly became an effective military commander. He started his state-building efforts by using a levee en masse. After a clash with the French in 1883 he realized the inadequacy of his military weapons and shifted to reliance on a small body of highly mobile cavalry armed with late-model European weapons. Recruited from slaves, his sofa were feared and effective. Samori used Islam as an official ideology, but his wars were not jihads. Administrative positions within his rapidly expanding state were held mostly by the sofa and by members of his family (Person 1968-1975; Tymowski 1987; Legassick 1966).
There was a similar dynamic in Ethiopia. Though Ethiopia had one of the oldest state traditions in Africa, the 18th and early 19th century saw the Amharic state reduced to a virtually powerless shell as power gravitated to provincial rulers and to Galla invaders, who moved up from the south. From the 1860s, under three emperors, Tewodros, Johannes and Menelik, the power of the Ethiopian armies was increased as the empire, particularly under Menelik, extended its control over the southern part of Ethiopia. The slave trade into the Arabian Peninsula provided significant revenue and slaves were an important part of the operation. This was true not only of the Ethiopian state, but also of Jimma Abba Jifar, an important center of the slave trade in the South. The ruler, Abba Jifar II (1878) had as many as 10,000 slaves in his service. Many held high positions, where they were often preferred to the king's own relatives. The king also used wealthy men, Muslim
clerics and mercenaries, but the loyalty of his slaves was crucial to his virtually absolute power (Lewis 1978).
By the late 1880s, slave-raiding operations existed all along the southern limits of the Sudanic belt and involved not only older states like Dar Fur, Wadai and Bagirmi, but newer ones like Dar Sila, Dar Runga and Dar al-Kuti, all of them dependent on slave officials and a corps of slave soldiers. In 1875 Zubayr Pasha was arrested on a visit to Egypt. Three years later, one of his lieutenants, Rabih bin Fadlallah, began moving west with several hundred riflemen. He established himself in and around Dar al-Kuti in the area between the Congo and Chad basins and set up an operation based on the slave trade. Cordell (1982) describes the operations of Rabih and his rivals:
Starting with small, well-armed forces that accompanied them from the Sudan, they became warlords with their own armies created through the Turko-Egyptian practice of slave recruitment. Because they were operating in non-Muslim zones among small-scale societies, their arms and organizations gave them a tremendous military advantage. They raided for slaves, integrating boys and young men into their detachments, taking girls and young women into their followings, and giving the others to northern traders in exchange for additional arms, powder, and the standard goods of long-distance trade.
The soldiers were slaves, his officials were slaves, and slave estates fed the court. Slaving was almost the only serious economic acti-vity and it was all done by slaves. In 1890 Rabih moved on, leaving a client in Dar al-Kuti. He defeated Bagirmi and conquered Bornu before falling to the French in 1900. Dar al-Kuti continued the same kind of operation until the sultan was killed by a French officer in 1911.
CONCLUSION
We can ask whether we have moved out of the question of the early state. Certainly, Samory and Rabih were using slave soldiers and administrators to establish the authority of new states. If their transient success inspires a sense of horror, it is a horror rooted in the convergence of centuries of slaving and a very modern military technology. Not surprisingly, colonial armies were like their African rivals made up largely of African slaves. The colonial regimes were using slaves to establish a fragile hegemony much as African
states had long done (Klein 1998). One effect of the slave trade is that it made captives available within Africa. For the slave trader moving slaves toward coastal entrepots, it was not important where he sold his slaves, only that he got his price. For would-be-purchasers, a person was often more valuable than the trade goods he could have received for selling that person. One result was that between the Zambezi and the Sahara, the most efficient way to recruit labour was to buy slaves.
The argument of this paper is not that slavery was indispensable to state formation. In many parts of Africa, slavery was not important to social evolution. Most striking is southern Africa, where the most important center of political complexity, Zimbabwe, is one that saw little development of servile institutions, even after it intersected with the Portuguese prazeros. The most dramatic transition from chiefdom to early state took place with the Zulu, but the Zulu success was based on the mobilization of all young men in disciplined age regiments. It was marked by a combination of autocracy and equality. Shaka during his short reign claimed the right of life or death over all his subjects, but his power depended upon battlefield results and those results depended on the cohesion of the regiment, which in turn was based on equality. There were no slave regiments. The most important derivative state was the Ndebele, which did impose on some conquered people a form of slavery, but servile personnel were not important in the army or the administration. Another derivative formation, the Ngoni, split into a number of sub-groups as it moved north. Always short of fighting males the Ngoni absorbed young captives wherever they went, but it is questionable whether these males or their female counterparts should be seen as slaves. The Ngoni preserved the racial egalitarianism of their Zulu progenitors. A similar convergence of autocracy and egalitarianism was found in Dahomey. Elsewhere, reliance on age-sets often served as a barrier to any kind of state formation. Walter Hawthorne describes the Balanta, who became slave raiders, but whose age sets were based on equality (Hawthorne 2003: 139-142).
I have dealt in this paper with three kinds of polities. I have not found much evidence on chiefdoms, though we can make an argument that the emergence of the early state was a result of the struggle of chiefs to control fission. On the early state, I have argued that the use of slaves was one way, perhaps the most effective, but only one way for a ruler to acquire loyal underlings. In the last part
of the paper, I dealt with aberrant political entities which were fragile because they were too reliant on slaves and on the hegemony of terror. I am not sure whether they should be seen as early states, but they certainly sought to control fission and conflict by relying on slaves. In between, there are some rather mature states like the Sokoto Caliphate, but it is striking that many of them found slaves to be useful instruments in trying to stabilize the exercise of power. None of them are states built on the participation of the governed, though many sought to find ways to ensure consent. Nevertheless, the premodern state was in general highly exploitative. The people it exploited often resented it, and for many, their primary loyalty was to local social entities. People governed or conquered by early states - and many conquered by more modern states - often resisted it. Rulers at any level needed the ability to coerce. They had a vital interest in finding men who would be loyal to the ruler rather than to other clans, descent groups or local communities. They could sometimes rely on members of their families, but usually only if those members were counterbalanced by men of lower ranks. The royal entourage was often a mixture of slaves, pawns, clients and strangers.
Of these groups, slaves had a certain advantage in that they had neither power or family. Everything a slave achieved was achieved because his only loyalty was to the king. For those who purchased slaves, their kinlessness was absolute essential. To be sure, royal slaves often achieved a corporate feeling, in a sense, family within their servile community. There are cases where royal slaves acted collectively in their own interest, for example in the control of succession. Generally, however, royal slaves were loyal, probably more loyal than brothers and sons. Few states relied on coercion alone. The traditional state combined coercion, rewards and ideology, and ideology was often developed to justify hegemony.
NOTES
1 See also Finley 1968: 307-313; 1980: 67, 79-80; Hopkins 1978: 99102. For a critical use of the concept of a slave mode of production see Meillassoux 1986.
2 Finley (1980) has written that there were only five slave societies, three in the Americas and two in Classical Antiquity. Many African slave systems had a higher percentage of their populations in slavery than Greece, Rome and the southern United States, and were more dependent on slave labour.
3 I have tried to explore the harem in ‘Sex, Power and Family Life in the Harem: A Comparative Study’ in Gwyn Campbell, Joseph Miller and Suzanne
Miers (eds.), Women in Slavery, forthcoming. There are a number of studies of the Ottoman harem, of which the best is Leslie Peirce (1993). For an African case study see Nast 2005. See also Mernissi 2001.
4 Elman Service quoted approvingly in Claessen and Skalnlk 1978. See also Claessen and Skalnlk 1981; Cohen and Service 1978; Earle 1997; Haas 1982.
5 Lerner is right in that these women were brought into the host society as captives.
6 The classic articulation of this idea was in Nieboer 1910. He also adds the proviso that technology was simple.
7 Jack Goody (1971) linked these low population densities to technological backwardness, which he linked to the mode of destruction being more important than the mode of production.
8 For Katsina see Yesufu Bala Usman (1981, chapter 1).
9 There is data on Wadai and Bagirmi in Humphrey Fisher and Allan Fisher (2001). For slavery and the state see, especially, chapter 8. Much of this book is based on Gustav Nachtigal (1971-1980).
10 Nupe also had a highly developed royal slavery. See Nadel 1942: 97, 106-107.
11 Coastal merchants, both African and European depended very heavily on slave labour for armed retainers, porters, and labourers though canoemen tended to be free. There was just not a large pool of free labour available and the conditions of the slave trade made movement difficult.
12 Miller (1983) sums up the substantial literature on the Portuguese and Kongo.
13 And in both religions people opposed to slavery have turned to religion to justify their anti-slavery actions. I am working on a paper developing this argument.
REFERENCES
Alpers, E.
1975. Ivory and Slaves in East Central Africa. London: Heinemann.
Al-‘Umari
1981. Kingdom of Mali. In Levtzion, N., and Hopkins, J. F. P. (eds.),
Corpus of Early Arabic Sources for West African History (pp. 259-273). Cambridge: Cambridge University Press.
Barkindo, B.
1985. The Early States of the Central Sudan: Kanem, Borno and Some of Their Neighbours to c. 1500 A.D. In Ajayi, J. F. A., and Crowder, M. (eds.), History of West Africa. 3rd edition. Vol. 1 (pp. 225254). New York: Longmans.
Barry, B.
1988. La Senegambie du XVe au XIXe siecle: Traite negriere, Islam et conquete coloniale. Paris: Harmattan.
Batuta, I.
1981. Excerpt from Rihla (this is an excerpt from Tuhfat al-nuzzar fi ghara’ib al-amsar wa-‘aja’ib al-asfar, which is generally referred to as his Rihla). In Levtzion, N., and Hopkins, J. F. P. (eds.), Corpus of Early Arabic Sources for West African History (pp. 279-304). Cambridge: Cambridge. University Press.
Bay, E.
1998. Wives of the Leopard: Gender, Politics, and Culture in the Kingdom of Dahomey. Charlottesville: University of Virginia Press.
Bazin, J.
1975. Guerre et servitude a Segou. In Meillassoux, C. (ed.), L’esclavage en Afrique precoloniale (pp. 135-181). Paris: Francois Maspero.
Bennett, N.
1971. Mirambo of Tanzania, ca. 1840 to 1884. London: Oxford University Press.
Beswick, S.
2004. Sudan’s Blood Memory: The Legacy of War, Ethnicity and Slavery in South Sudan. London - Rochester: University of Rochester Press.
Chazan, N.
1988. The Early State in Africa: The Asante Case. In Eisenstadt, Abitbol, and Chazan 1988: 60-97.
Claessen, H., and Skalnlk, P. (eds.)
1978. The Early State. The Hague: Mouton.
1981. The Study of the State. The Hague: Mouton.
Cohen, R.
1978a. State Origins: A Reappraisal. In Claessen and Skalnlk 1978: 31-75.
1978b. State Foundation: A Controlled Comparison. In Cohen and Service 1978: 141-160.
Cohen, R., and Service, E. (eds.)
1978. The Origins of the State: The Anthropology of Political Evolution. Philadelphia: Institute for the Study of Human Issues.
Cordell, D.
1982. Dar al-Kuti and the Last Years of the Trans-Saharan Slave Trade. Madison: University of Wisconsin Press.
Earle, T.
1997. How Chiefs Come to Power: The Political Economy in Prehistory. Stanford: Stanford University Press.
Eisenstadt, N. S., Abibol, M., and Chasan, N. (eds.)
1988. The Early State in African Perspective: Culture, Power and Division of Labour. Leiden: Brill.
Erdem, Y. H.
1996. Slavery in the Ottoman Empire and its Demise, 1800-1909. London: St. Martin's Press.
Finley, M. I.
1968. ‘Slavery’. International Encyclopedia of the Social Sciences XIV: 307-313. New York: Free Press.
1976. A Peculiar Institution? Times Literary Supplement 2, July: 819-821.
1980. Ancient Slavery and Modern Ideology. New York: Viking Press.
Fisher, H. J., and Fisher, A. G. B.
2001. Slavery in the History of Muslim Black Africa. London: Christopher Hurst.
Goody, J.
1971. Technology, Tradition and the State in Africa. London: Oxford University Press.
Gray, R.
1961. History of the Southern Sudan, 1839-1889. London: Oxford University Press.
Haas, J.
1982. The Evolution of the Prehistoric State. New York: Columbia University Press.
Hanson, J.
1996. Migration, Jihad, and Muslim Authority in West Africa: The Futanke Colonies in Karta. Bloomington: Indiana University Press.
Hawthorne, W.
2003. Planting Rice and Harvesting Slaves: Transformations along the Guinea-Bissau Coast, 1400-1900. Portsmouth, NH: Heinemann.
Hopkins, K.
1978. Conquerors and Slaves. Cambridge: Cambridge University Press.
Hopkins, J. P. F., and Levtzion, N. (eds.)
1981. Corpus of Early Arabic Sources for West African History. Cambridge: Cambridge University Press.
Hunwick, J.
1985a. Songhay, Borno and the Hausa States, 1450-1600. In Ajayi, J. F. A., and Crowder, M. (eds.), History of West Africa. 3rd edition. Vol. 1, ch. 8 (pp. 323-371). New York: Longmans.
1985b. Notes on Slavery in the Songhay Empire. In Willis, J. R. (ed.),
The Servile Estate. In 2 vols. Vol. 2. Slaves and Slavery in Muslim Africa (pp. 16-32). London: Frank Cass.
Isaacman, A.
1972. Mozambique: The Africanization of a European Institution -the Zambezi Prazos, 1750-1902. Madison: University of Wisconsin Press.
Isaacman, A., and Isaacman, B.
2004. Slavery and Beyond: The Making of Men and Chikunda Ethnic Identities in the Unstable World of South-Central Africa, 1750-1920. Portsmouth, NH: Heinemann.
Kea, R.
1982. Settlements, Trade and Polities in the Seventeenth-century Gold Coast. Baltimore: Johns Hopkins University Press.
Keim, C. A.
1983. Women in Slavery among the Mangbetu c. 1800-1910. In Robertson, C., and Klein, M. (eds.), Women and Slavery in Africa (pp. 144-159). Madison: University of Wisconsin Press.
Klein, M.
1977. Servitude among the Wolof and Sereer of Senegambia. In Miers, S., and Kopytoff, I. (eds.), Slavery in Africa (pp. 335-366). Madison: University of Wisconsin Press.
1998. Slavery and Colonial Rule in French West Africa. Cambridge: Cambridge University Press.
Kobishchanov, Yu.
1978. Axum. In Claessen and Skalmk 1978: 151-168.
Law, R.
1977. The Oyo Empire c. 1600 - c. 1836: A West African Imperialism in the Era of the Atlantic Slave Trade. Oxford: Clarendon Press.
1991. The Slave Coast of West Africa, 1550-1750. Oxford: Clarendon Press.
Legassick, M.
1966. Firearms, Horses and Samorian Army Organization, 18701898. Journal of African History 7: 95-115.
Lerner, G.
1983. Women and Slavery. Slavery and Abolition 4 (3): 173-198.
1986. The Creation of Patriarchy. New York: Oxford University Press.
Levtzion, N.
1963. The Thirteenth and Fourteenth Century Kings of Mali. Journal of African History IV (3): 341-353.
1988. Islam and State Formation in West Africa. In Eisenstadt, Abit-bol, and Chazan 1988: 98-108.
Lewin, T.
1978. Asante before the British: The Prempean Years, 1875-1900. Lawrence: Regents Press of Kansas.
192 Social Evolution & History / March 2009 Lewis, H.
1978. The Galla State of Jimma Abba Jifar. In Claessen and Skalnlk 1978: 321-338.
Lovejoy, P., and Richardson, D.
1995a. British Abolition and Its Impact on Slave Prices at the Atlantic Coast of Africa, 1783-1850. Journal of Economic History 55: 98-119.
1995b. The Initial ‘Crisis of Adaptation’: The Impact of British Abolition on the Atlantic Slave Trade in West Africa, 1808-1920. In Law, R. (ed.), From Slave Trade to ‘Legitimate’ Commerce: The Commercial Transition in Nineteenth Century West Africa (pp. 32-56). Cambridge: Cambridge University Press.
McCaskie, T.
1995. State and Society in Pre-colonial Asante. Cambridge & New York. McIntosh, R. J.
1988. The Peoples of the Middle Niger: The Island of Gold. Oxford: Blackwell.
Meillassoux, C.
1986. Anthropologie de l’esclavage: le ventre de fer et d’argent. Paris: Presses Universitaires de la France.
Mernissi, F.
2001. Scheherazade Goes West: Different Cultures, Different Harems. New York: Washington Square Press.
Miers, S., and Kopytoff, I.
1977. African ‘Slavery’ as an Institution of Marginality. Introduction. In Miers, S., and Kopytoff, I. (eds.), Slavery in Africa: Historical and Anthropological Perspectives. (pp. 1-81). Madison: University of Wisconsin Press.
Miller, J. C.
1976. The Slave Trade in Congo and Angola. In Rotberg, R., and Kilson, M. (eds.), The African Diaspora: Interpretive Essays (pp. 75113). Cambridge: Harvard University Press.
1981. Lineages, Ideology and the History of Slavery in Western Central Africa. In Lovejoy, P. (ed.), The Ideology of Slavery in Africa (pp. 40-71). Beverly Hills - London: Sage.
1983. The Paradoxes of Impoverishment in the Atlantic Zone. In Birmingham, D., and Martin, P. M. (eds.), History of Central Africa. Vol. 1. London - New York: Longman.
Munson, P.
1980. Archeology and the Prehistoric Origins of the Ghana Empire.
Journal of African History XXI: 457-466.
Nachtigal, G.
1971-1980. Sahara and Sudan. In 4 vols. With an introduction and notes by A. G. B. Fisher, and H. J. Fisher. London: Christopher Hurst.
Nadel, S. F.
1942. Black Byzantium: The Kingdom of Nupe in Nigeria. London: Oxford University Press.
Nast, H.
2005. Concubines and Power: Five Hundred Years in a Northern Nigerian Palace. Minneapolis: University of Minnesota Press.
Nieboer, H. I.
1910. Slavery as an Industrial System. The Hague: M. Nijhoff.
O'Fahey, R. S.
1973. Slavery and the Slave Trade in Dar Fur. Journal of African History XIV: 29-43.
1985. Slavery and Society in Dar Fur. In Willis, J. R. (ed.), Slaves and Slavery in Muslim Africa. Vol. 2 (pp. 83-100). London: Frank Cass.
O'Fahey, R., and Spaulding, J.
1974. Kingdoms of the Sudan. London: Methuen.
Peirce, L.
1993. The Imperial Harem: Women and Sovereignty in the Ottoman Empire. Oxford: Oxford University Press.
Person, Y.
1968-1975. Samori: Une revolution dyula. 3 vols. Dakar: IFAN.
Rattray, R. S.
1929. Ashanti Law and Constitution. Oxford: Clarendon Press.
Reefe, Th. Q.
1983. The Societies of the Eastern Savanna. In Birmingham, D., and Martin, P. (eds.), History of Central Africa. Vol. 1 (pp. 160-94). New York: Longman.
Roberts, R.
1987. Slaves, Warriors and Merchants. Stanford: Stanford University Press.
Robinson, D.
1985. The Holy War of Umar Tal. Oxford: Clarendon Press.
Ryder, A.
1969. Benin and the Europeans 1485-1897. Longmans: London.
Searing, J.
1993. West African Slavery and Atlantic Commerce: The Senegal River Valley, 1700-1860. Cambridge: Cambridge University Press.
Service, E.
1975. The Origins of the State and Civilization. New York: Norton.
Sheriff, A
1987. Slaves, Spices and Ivory in Zanzibar: Integration of an East African Commercial Empire into the World Economy, 1770-1873. London: James Currey.
Stilwell, S.
2004. Paradoxes of Power: The Kano ‘Mamluks’ and Male Royal Slavery in the Sokoto Caliphate, 1804-1903. Portsmouth, NH: Heinemann.
Thornton, J.
1992. Africa and the Africans in the Making of the Atlantic World, 1400-1680. Cambridge: Cambridge University Press.
Toledano, E.
1993. Ottoman Concepts of Slavery in the Period of Reform (1830s-1880s). In Klein, M. (ed.), Breaking the Chains: Slavery, Bondage and Emancipation in Modern Africa and Asia (pp. 37-63). Madison: University of Wisconsin Press.
1998. Slavery and Abolition in the Ottoman Middle East. Seattle: University of Washington Press.
Tymowski, M.
1987. L’armee et la formation des Etats en Afrique occidentale au XIXe siecle - un essai de comparison: L ’Etat de Samori et le Kenedougou. Warsaw: Wydawnicta Uniwersytetu Warszawskiego.
Usman, Y. B.
1981. The Transformation of Katsina 1400-1883. Ch. 1. Zaria: Ahmadu Bello University Press.
Vansina, J.
1990. Paths in the Rainforest. Madison: University of Wisconsin Press.
Wilbur, C. M.
1943. Slavery in China during the Former Han Dynasty: 206B.C. -A.D. 25. Chicago: Field Museum of Natural History.
Wilks, I.
1966. Aspects of Bureaucratization in Ashanti in the Nineteenth Century. Journal of African History VII: 215-232.
1975. Asante in the Nineteenth Century. Cambridge: Cambridge University Press.
