African American

African American

Slaue,ry in AFrica and khe Transahankic Slave Trade


After decades of neglect, slavery in Africa took the center stage in African historiography in the 1970s and 1980s. Sud- denly there was a lot of excitement about the subject among European and North American historians and anthropologists, although, some- what curiously, the then-vibrant schools of history on the African conti- nent – in particular the Ibadan school in Nigeria- showed little interest. The excitement generated an impressive amount of empirical research into African socioeconomic institutions, which the researchers labelled “slavery.” Historians studying slavery on other continents, espe- cially in the New World, where the slave population comprised people of African descent, were surprised to be told that “there were certainly more slaves in Africa in the nineteenth century than there were in the Americas at anytime,” and that

The scholarship of the last twenty years [1970 to IWO] has demon- strated that the variety and intensity of servile relationships and meth- ods of oppression that can be equated with slavery were probably more developed in Africa than anywhere else in the world at any pe- riod in history.’

Some of these researchers also claim that in 1897 the slave population of northern Nigeria alone was “certainly in excess of I million and perhaps more than 2.5 million people,” and that “Northern Nigeria’s slave sys- tem [was] one of the largest slave societies in modern history.J’2

But can we be certain that the institution studied by Atricanists ap- proximates that studied by historians of classical history, European his- tory, Asian history, and New World history? As Claude Meillassoux has stressed, “Slavery, rigorously defined, may have universal characteristics, but its definition . . . must be generally accepted if a real discussion is to take place.“3 Certainly, the use of terms such as slave and &VT invites comparison, and students of slavery in Africa have not been insensitive to the need to compare. In general, they have compared the slaves they discovered in Africa with those of the Americas. While conceding that there were some differences, it is argued that enough similarities existed for the people they identify as slaves in Africa to approximate to the universal social category of slave and slavery.* It has even been argued that there were extensive slave plantations in many African regions com- parable to New World slave plantations.5

A few scholars have expressed discomfort with this analytical thrust, which certainly characterizes the mainstream literature on the subject. Some have argued that the late nineteenth-century servile institutions in Africa did not approximate to slave or slavery, but, not finding what they consider an appropriate alternative term, they have reluctantly used both terms with quotation marks6 Others have employed terms such as captives or serfs without offering a convincing explanation for their use. Frederick Cooper has criticized this trend, pointing out that “the word ‘slave’ carries with it a bundle of connotations – all of them nasty. This has led some Africanists to use terms like ‘adopted dependents; ‘cap- tives,’ or ‘serfs’ for a person whom others would call a slave.” He adds:

The first euphemizes a process that was based on violence and coer- cion; the second distracts from the various possible fates that befell people and their descendants once the act of capture was completed; and the third misrepresents the nature of dependence and the slave’s relationship to the land.7

This intervention by Cooper seems to have silenced the few dissent- ers. Apart from one or two lonely voices, it has been business as usual in the study of slavery in Africa since the 1980s. Meillassoux has stated that the inclusion of non-slaves is responsible for the claim that slavery in Africa was mild or benign8 but apart from his very helpful effort to provide a universal definition, he has not shown which social categories


to include or eliminate. Similarly, Joseph E. Inikori has complained of terminological looseness in the study of slavery in Africa,9 yet has failed to show how to bring precision and discipline to the enterprise.

Such an exercise is attempted in this essay, which will present a com- parative examination of servile social categories in medieval Europe and precolonial Africa. I contend that slavery and serfdom under the socio- economic conditions of medieval Europe provide better comparative insights for a precise and disciplined study of servile institutions in pre- colonial Africa than the New World comparisons that have hitherto’ been conducted. The societies of medieval Europe were closer in all re- spects to those of precolonial Africa than were the New World slave societies, which were specifically organized for the large-scale produc- tion of commodities for an evolving capitalist world market. In addi- tion, slavery once coe,xisted with other dependent social categories in pre-capitalist Europe, and the coexistence of slavery and non-slave ser- vile categories has been carefully studied by students of European history.

Drawing on these studies, I will apply ,their methods of separating the various dependent social groups to the descriptive evidence on pre- colonial African servile categories, in order to determine which of them corresponds more closely to slavery or to serfdom or to some other category. For the exercise to be manageable, I have chosen two extreme geographical points in Europe: England to the extreme west and Russia to the extreme east. The exercise affords the opportunity to delve further into the old debate of the contribution of the transatlantic slave trade to the transformation of African servile institutions.

Slavery was first brought to England by the Romans, but their de- parture in 407 A.D. did not end slavery in that country. The chaos that followed the collapse of the Roman empire provoked slave raids and encouraged slavery in England and other parts of Europe. The Anglo- Saxons, from the Danish peninsula and the coast lands of northern Ger- many and Holland, raided England for slaves and eventually took over the country. lo Many of the indigenous Celts were enslaved. After their settlement, the Anglo-Saxon kingdoms continued to fight among them- selves and take captives. In consequence, slaves made up a large propor- tion of the population of Anglo-Saxon England.”

Slave raids and slave trading by various groups in the British Isles continued into the eleventh century, with Bristol as the main exporting port and Ireland a major export market. The Anglo-Saxon bishop of Worcester, who fought hard against the slave trade from England, la- mented, “You might well groan to see the long rows of young men and

maidens whose beauty and youth might move the pity of the savage, bound together with cords, and brought to market to be sold.“12 In the end, the conquest of the warring groups in Britain and Ireland by the Normans stopped the slave raids and the trade in captives. The West- minster Council of 1102 proclaimed that “no one is henceforth to pre- sume to carry on that shameful trading whereby heretofore men used in England to be sold like brute beasts.“13

But the Normans did not enact a law abolishing slavery in England. Slaves remained a statistically important part of the population. Based on the Domesday enumeration of 1086, it has been estimated that all categories of free peasants taken together constituted only 14 percent of the total population of rural England in 1086. Various categories of un- free peasants made up the remaining 86 per cent, of whom slaves were x0.5 per cent of the whole rural population and different groups of serfs accounted for 75.5 per cent. i* Thus, over a period of several hundred years, slavery coexisted with other servile institutions in England. The disappearance of slavery was a long process, in which slaves were gradu- ally converted to other dependent social categories.

This process of conversion has been examined in some detail by M. M. Postan in The Famuhs: The Estate Labourer in the XIIth and XlIIth Centuries (1954.). It started when Anglo-Saxon slaveholders, like their counterparts in the Frankish kingdoms, began to move their slaves, whether manumitted or not, into separate landholdings of their own. The new arrangement allowed them much freedom to cater for them- selves and maintain households of their own, while continuing to ren- der labor services to the lords in the cultivation of their estates. This new group of landholding dependent cultivators were distinguishable from their preexisting counterparts mainly in terms of the greater labor services they owed to the lords.

At the same time, however, the lords continued to hold other slaves directly in their estates. These were housed, fed, and clothed by the lords, and they spent most of their time working the latter’s estates. As a recognition of the significant difference between these groups of dependent cultivators, contemporary practice and modern medievalists limited the application of the term slave to those directly resident on the lords’ estates and completely maintained by them. The slaves enumer- ated in the Domesday Inquest of 1086 belonged to this category.15

The process of converting slaves to landholding dependent cultiva- tors in England was completed in the twelfth century, but by this time the most recently converted groups were still somewhere between being slaves and f&lly settled landholding dependent peasants. Their holdings


were small in size and had been carved out recently from the lords’ es- tates.16 They still spent the greater part of their time cultivating those estates, from which they earned much of their livelihood in food and money. In addition, like the fully settled dependent peasants, by law they were not free to move. I7 Postan refers to them as serf smallholders settled on servants’ holdings; medieval sources refer to them as bovatii or famuli.

What happened in England in the twelfth century, according to Postan, was a wholesale conversion of the remnants of the slave class to bovarii. As he put it,


a bzt-ius was a serf who possessed few rights or franchises denied to the late-Roman or German slaves, and whose greater independence resulted not so much from his superior status as from his separate holding and his life away from the .lord’s curia J&we]. It is &his_*“.* — physical separation that the key to th~zhange.must- be.sought. The smallholder was able to set up a family and have some inducement to re~loit tom _r-“–.– ..–.-

full the land of his holding.18 _ _

Postan insists that the smallholders settled on servants holdings, the bo- vu%, were not slaves:

Twelfth-century bovarii were not slaves; and neither were the bulk of their ‘Domesday namesakes. Twelfth-century bovarii were serfs, and the difference of status between slave and serf may have been very slight. Yet it is impossible wholly to identify medieval serfs with slaves with- out repudiating the accepted view of medieval serfdom as a condition intermediate (though of course far from equidistant) between those of slavery and freedom. Now and again the Domesday scribes, more especially those responsible for recording the Herefordshire entries, go out of their way to underline the difference between the boyarius and the slave.i9

Thus, by contemporary practice and by the method of separation and classification adopted by medieval historians, there were several cat- egories of serfs in England in the twelfth century. The common denomi- nator for all of them was the possession of land, which they cultivated for themselves. Also linking them was the labor service they owed their lords and their lack of freedom to move under the law. They were, on the other hand, separated by the amount of labor dues they had to ren- der, the amount of land they held, and the amount of time available to them to cultivate it. Serfs without slave origin, by local custom and tradition, had larger holdings and more time to cultivate them. For serfs

Joseph E. IniEzoti 43

of slave origin, the labor dues, the size of holdings, and the amount of time available to work them all tended to depend on the distance in time from their slave ancestors. On these differences, Postan observes:

The services of a smallholder regularly employed as a were, relative to his holding and rent, incomparably greater than the labour dues of an ordinary villein. . . . The latter, unlike the true manorial servant, was left enough time to work his own holding, and had enough substance to employ a servant himself if he needed one. There was no question of maintaining him and his family by the lord’s plough-team, food and money. In economic fact he was still a custom- ary tenant, discharging his labour obligations, while the fU-time famulus . . . was a labourer working another man’s land and deriving his livelihood wholly or in the main from his employer’s wageszO

ngland’s case in the gen- et-al context of medieval Europe, it is enough to refer to the summary statement by R. H. Hilton, who notes that the word &comes from the Latin word sms, which in classical times meant “slam”; while in the early Middle Ages the word-as derived from the ethnic term & By this time the descendants of the slaves of the classical period and most of the slaves of the Dark Ages had been settled on separate holdings of their own and were seti casati. Many of the slaves still resi- dent on their lords’ estates were Slavonic captives acquired by the Ger- mans in their eastward expansion and sold in the slave markets of Western Europe. Medieval historians, therefore, use both words to dis- tinguish between the conditions of the two broad categories of depen- dent cultivators. As Hilton put it,

Although some slaves in antiquity were by no means completely with- out property the distinction between slaves and serfs is based on the fact that, on the whole, slaves were the chattels of their master, em- ployed as instruments of production in agriculture or industry, receiv- ing food, clothing and shelter from the master and possessing nothing. . . . Some peasants were descended from the coluni of the late Empire, who . . . sank into ser6dom under the heavy weight of the obligations imposed on them by the estate owners and the state. . . . Other peasants were descendants of full slaves, some of their ancestors having been slaves under the Roman Empire, others having been en- slaved during the wars of the Dark Ages. What distinguished these serfs from their slave ancestors was, of course, the fact that they were now seti cmat& provided with their own holdings from the landown-


er’s estate. Other medieval peasant serfs were descended from free men who had entered into various forms of dependence under lordszl

Like England and the rest of Western Europe, slavery antedated serf- dom in Russia. But, while in England slavery died out in the twelfth century and serfdom in the fifteenth, in Russia slavery remained alive up to the early eighteenth century and serfdom began its development only in the mid-fifteenth, to be abolished by state law in the 1860s. The enserfment of the Russian peasants was associated with the construction and expansion of the Russian state. Muscovy had been one of several Russian principalities up to the fifteenth century. In 1462 a Russian em- pire ruled by the Muscovite state was established. The creation of the Russian empire was accompanied by the rise of a powerful nobility. The size of this aristocracy and its capacity to dominate the peasantry grew paripassu with the geographical expansion of the empire and the power of the Russian tsar. As all this happened, the previously free Russian peasants increasingly lost their freedom. The development followed an observable historical sequence. In the first instance the peasants’ general freedom was gradually curtailed in the second half of the fifteenth and the greater part of the sixteenth century. This was followed by a total . prohibition of their right to move in the late sixteenth and early seven- teenth centuries. Between the seventeenth century and the first half of the eighteenth, the changes were formally codified. Thereafter the con- ditions of the Russian serfs moved closer and closer to those of chattel slaves as the powerful nobles acted in total disregard of the law and tradition.22

Meanwhile the Russian ahulopy (slaves) were being transformed into serfs. From the mid-fifteenth to the early eighteenth century the en- serfed peasants coexisted with their older servile ‘ccousins,” the slaves. In the course of the seventeenth century, however, slaves and serfs grad- ually merged. The remnants of the slave class were moved into the rank of serfs by the tsar Peter I in 1723. 23 As to distinguishing the slaves from the serfs in the period during which they coexisted, Kolchin writes:

The main distinction between kholopy [slaves] and serfs was that whereas the latter were usually selfsupporting, growing their own food on al lot ted plots of land, the former were usually maintained by their owners; instead of living in a village with the peasants, they lived in or near their owners’ residences.24

However, the conditions of the Russian serfs deteriorated consider- ably over time. The practice of selling and buying serfs by noblemen,

Joseph E. Inikori 4 5

which began in the second half of the seventeenth century, became gen- eralized in the eighteenth. Kolchin states that in the course of the eigh- teenth century serfs were “bought and sold, traded, won and lost at cards.“25 Legally they remained serfs, however, and the Russian govern- ment made it clear that what was abolished by state law in the 1860s was serfdom, not slavery. In 18$3, on the eve of emancipation, there were a total of 11.3 mi.lhon male sex% (11,338,042). In 1795 the figure had been 9.8 million (9,787,802). 26 Assuming an equal proportion of female serfs, the totals come to 19.5 million in 1795 and 22.7 million in 1858. The litera- ture has regularly referred to all of them as se$.27

In the preceding examples the formula employed by students of Eu- ropean history to distinguish serfs and other dependent social categories from slaves incorporates unambiguous elements: first and foremost, the serfs or non-slave dependent people must possess a means of production (mainly land) large enough to provide an income to support a house- hold with unproductive children and old members; second, they must have enough free time to produce for themselves, in order to realize the potential income from the employment of their means of production; third, they must be allowed to retain for their own use as they pleased the income realized; and fourth, their residences must be physically sep- arated from those of their lords. On the other hand, for a dependent people to be classified as *slaves, they must spend virtually their entire wo&ing dry on their lords’ estates – they may be &tied some plots, but the size, and the time available to work them, would be so limited that they would have to be fed, cl+-& qnd housed by their lords; and their owners must be free, under the law and by tradition, to sell them to agy buyer. I propose to apply this formula to the dependent social categories of precolonial Africa.

Incidentally, the pioneers of modern African historiography con- ducted a somewhat similar exercise more than three decades ago when they debated the issue of whether the concept of feudalism could be applied to any of the political systems in precolonial Africa. Some as- pects of that debate are relevant for the present purpose. It is, therefore, appropriate to start with a summary of its major points.

Following ef6orts being made in the 1940s and 1950s to constitute a typology of political organizations in precolonial Africa, Jacques Ma- quet suggested in 1962 that “the feudal system deserves to be considered as an important type of political organization in traditional Afi-ica.“28 He outlined the defining features of feudal regimes that should be con- sidered in applying the feudal concept to Africa, noting that the concept was originally applied by historians of Europe to describe the main fea-


tures of the dominant political regimes that existed in Europe between the ninth and thirteenth centuries. The defining features were, there- fore, stated in accordance with the conceptions of feudalism employed by two broad groups of students of medieval European history: IMarx- ists and non-iMarxists. For non-Marxists the essentiai element of feudal regimes is the vassalage bond, while for Marxists, the disdguishingu element is the relation o extractian-defined by a particular form

* 29of laQ$$. Jack Goody criticized such attempts to apply the concept of feu-

dalism to precolonial African regimes. He began with the problem of transferring to African history a term for which there are conflicting conceptions, citing an Anglo-Soviet discussion on feudalism in which the English speaker dwelt on military fiefs and the Russian participant concentrated on class relations of surplus extraction between lords and peasants. Goody concluded that “the core institution of feudal society is seen as vassalage associated with the granting of a landed benefit (fief), usually in return for the performance of military duties.“30 He also examined specific cases of the application of the concept to African societies. His summary of Potekhin’s description of Asante is of particu- lar interest:

Pot&in writes that “Feudal land ownership constitutes the founda- tie; of feudal relations.” Land belongs to a restricted circle of big land- owners, while the peasant pays rent or performs services for the right to cultivate his land. In Ashanti, he finds “the exclusive concentration of land in the hands of the ruling upper strata,” together with the conditional land tenure and hierarchies of dependence “typical of feu- dal society.“31

Goody did not question the factual accuracy of the description, but thought that defining feudalism in terms of the ownership of land was not appropriate.

Of all the cases cited and summarized by Goody, only that of Basil Davidson specifically touches upon the question of slavery in Africa. Davidson’s main point was to show that African feudalism shared many common features with feudalism in medieval Europe. He pointed out that during the formative period of feudalism in the European Dark Ages slaves were gradually &nsformed into serf.., while in Africa strong

. . ; .-C_– ..I_.

states and empires destra tribal equahty and produced a mm- pendent people who were serfs and not slaves. Compawith England at the time of t&FNorman conquest, Davidson showed fi;lr- ther similarities:

Joseph E. Ipzikori 47

Thus the titles and the rights of great lords, the obligations of the common people, the customs of trade and tribute, the swearing of fealty, the manners of war-all these and a hundred other manifesta- tions have seemed to speak the same identical Ianguage of feudalism.32

Davidson seems to imply that slavery is incompatible with feudalism. For our present purpose, however, suffice it to say that we do not have to prove the existence of feudalism in order to establish the existence of serfs. Feudalism is a form of sociopolitical organization with several elements. At a given moment and place some elements may be present without others. Had Davidson been primarily concerned with the issue of slavery or serfdom in Africa, the right question to ask would have been whether the conditions of the people being considered approxi- mated more closely those of serfs or those of slaves in medieval Europe.

However, these were not the issues queried by Goody whose main discomfort was with Davidson’s tailoring of African history too tightly to European history. At the then stage of African historiography, Goody thought that African institutions should be studied in depth in their own right without attaching European labels. While encouraging comparative studies, he admonished African historians to stay clear of terms such as tribalism, feudalism, ad+gitaljm, which invite crude– comparison.33

~ -I’~‘~^- -r-.– w-.- .–.. L- –

J. H. M. Beattie concurred; in his study of Bunyoro in modern Uganda, he found that the political regime there contained three of the five features listed by Marc Bloch as defining feudalism in Europe. The missing elements were a special military class and evidence of the disin- tegration of a pre-existing strong state. Beattie concluded:

Like Goody, I consider i t to be more useful and i l luminat ing to retain the term “feudalism” and i ts associated vocabulary for the complex European polities to which they were first applied (and perhaps to such other systems, l ike the Japanese one, which can be shown to re- semble i t in al l or most of i ts essential features) , and to describe the political Institutions of traditional Bunyoro and of other African king- doms as far as possible in their own terms.34

It is a pity that the debate did not proceed further and examine in depth the issue of slavery and serfdom raised by Basil Davidson. Appar- ently it was concluded that these terms are to be avoided in the writing of African history, but this conclusion may be misguided. As will be shown shortly, slaves and serfs, properly defined, can be encountered in


African societies in specific periods and regions. The problem is how to distinguish them from each other and from other dependent social categories, and trace their historical development.

Let us now attempt to resolve this conundrum by applying the for- mula employed by historians of precapitalist Europe as outlined above. The main difficulty is the scantiness of descriptive evidence on the de- pendent social categories in question. However, the research of more than three decades has produced a reasonable amount of such evidence to make the exercise feasible. What is particularly important, this re- search has been most intensive in the geographical areas where depen- dent populations in precolonial Africa were mostly concentrated:de Western Sudan.BGeCentral Sudan, and the East A f r i c a n –

– . – T h e s e constitute the main focus of the exercise, but

scattered evidence from other areas will be examined to provide the basis for a continental generalization. The late nineteenth century will provide a starting point, and then we will turn to the period before the transatlantic slave trade.

The dependent populations of the nineteenth-century Sokoto Ca- liphate of modern northern Nigeria were among the first groups of such populations in tropical Africa to be scientifically studied by modern scholars. In the late 1940s Michael Smith and his wife, Mary Smith, conducted an elaborate data collection on the socioeconomic conditions of the servile po of the b

ulations among the Hausa-FulA ria emlrate

F03kinthenorthernparts the Soti Calm ate.

S–z& *d–lve evl ence shows’s hierarchical sociopolitical or- ganization of the Zaria emirate. The population was distributed into two spatial locations: walled towns, which included the capital city of Z,?,; ,and agricultural villages where the peasants lived, groups of which were located near each walled town. Administratively, the ruler of the emirate was the emir of Zaria, referred to as a kinLby SM. Being part of the Sokoto Caliphate, he, was answerable to the overall ruler of the Caliphate, whose seat was in Smity. Assisting the emir in the administration of the emirate were fief-holding officials, all of whom resided in the capital city of Zaria. The walled-s and agricul- tural villages were held as personal fiefs by these officials, who appointed subordinate staff (jekadu) to oversee their administration. At the lowest level of authority, the agricultural communities were run by villagel, chiefs. Several of these agricultural villages, known locally as tinji (plfl ral, rumada), were made up of servile cultivators; Smith called theriu’- “slave-vill.ges.” They belonged mostly to the emir and his fief-holding

Joseph E. Inikov-i 49

officials residing in the capital city, although some also belonged to mer- chants and other non-office holders. Like the other villages, they had their own village chiefs.

The servile cultivators in the rumada lived in separate households, and possessed lands allotted to them by their lords. These were separate from the lords’ fields and the harvests were retained by the servile peas- ants for the support of their households. In the mornings these depen- dent cultivators worked with their families on their own farms from the early hours to 9:30, at which time they went to their lords’ fields. At midday the lords sent them food, and at 2:30 they were free to return to their own farms. Apart from the meal provided on the field, the servile cultivators were responsible for the maintenance of their households.

According to Smith, the dependent populations of Zaria were of two categories, the native-born dima (plural, dimajai) and those+” _-. brought from outside by capture or-bmhase. Those brought from outside could be sold, but the native-born could not be alienated. The latter “formed the main body of the mililia-y force, of the police, and of the administrative staff at a subordinate le.yel.“36

Polly Hill’s study of pmne of the largest emirates in the&&oto Caliphate (covering an area of over sixteen thousand square miles), adds f&ther descriptive evidence. Of particular interest is the servile village belonging to the emir. A member of the village of servile descent, who was born in about 1885, told Hill in 1972 how the village had functioned in precolonial days:

He said that in the early days there had been “about fifty” men and women slaves working on the land; that many were strangers who spoke poor Hausa; that they lived in separate houses they had built themselves, since there was no house corresponding to a tinji; that each slave was given a farm plot ~~WWCZ) for his own use, a plot that could be inherited by a son (though not by a daughter) or lent to ̂ _ someone else (MO), but could not be sold. He said that the usual crops (grains, cassava, groundnuts, etc.) were grown on the main estate ( J&&)~; that someslavesbeca&~rich enough to ransom themselves; that the slave households cooked or bought most of their own food, though during the farming season cooked meals were provided in the evening as well as on the farmland at midday; that the slaves were paid for any work (such as thatching) done for the Emir in the dry season;

that some free men worked alongside the slaves . . .37

In general, Hill’s description of the conditions of the servile popula- tions in Hausaland complements Smith’s evidence, although she is criti-

cal of some points. For example, she points out that the restriction on the sale of bonded people was not based on generational differences; rather, Hausa tradition and public opinion imposed severe limitations on the rights of masters to sell servile individuals living in a house- hold. She further states that most bonded people in Hausaland lived in households “with their spouses and children, richer male slaves being polygynous.“38 c3

Another elaborate study of wed liphate is the one on theyIda emirate

populations in the Sokoto Ca- by Michael Mason, who describes

how the Bida state, a umt of the Sokoto Caliphate, was established by a Fulani ar iP- Band exwm 1901. This expan- sio’n led to the subjugation of the Ns, an indigenous population in the area that became the B&.-emirate. Bida’s military incursions into neighboring territories also brought captives from other ethnic groups, * such as the Yoruba, Afenmai, Igbirra, and even the Hausa and Fulani. Both the subjugated Nupe populations and the captives brought from outside were settled in agricultural villages around Bida City, the capital- of the emiLate. These captive villages or settlements, known locally as tungazi (singular, tunga), numbered 55 by 1859; between 1859 and 1873, 694 new captive settlements were created; and by 1901 there were a total of 1,601 captive villages in the emirate. About two-thirds of them were l — – settlements composed of subjugated indigenous Nupe populations.., _~. Mason believes that the total population of these villages must have been aroundAoo,-ooo by 1901.~~

The dwellers in These captive villages did not work collectively on their lords’ fields, nor did they render any labor services. Rather, they produced entirely for themselves and paid tributes to their lords in cash (cawries) and in kind (farm products).4o Mason calls them slaves- * their settlements;slave plantations. However, he points out:

In Nupe, although slaves did not live together with the lineages of.7 their owners, we do not see that the social relations between slaves and their masters were fundamentally different from the relations be- tween the lords or egbazi (sing. egba) and the peasants who paid them i tribute. S.&es m-ants both paid tribu_te. Otherwise, the &yes. J: .~GJX their own- products.–

He adds that there was only one social relation of production in nineteenth-century Nupe, “the one which we have called ’ ributarv.’ This

k–lmode colored all relations between masters and slaves and eween lords and peasants.“41

Based on the descriptive evidence presented, how do we categorize

so SLAVERY IN AFRICA Joseph E. Inikoti 5 I

the servile populations of the Sokoto Caliphate in the nineteenth cen- tury? We need to stress that the studies presented do not cover all the emirates of the caliphate, and that the extent to which the emirates stud- ied are representative of the whole caliphate is not easy to say. These emirates do, however, adequately represent the ethnic composition of the caliphate: Kso Emirate (predominantly-); northern Zaria (servile settlements stumre held mostly–by-th!~ani) ; and-%& Emirate (predominantly non-~au~a&i&ni). Whether the geographical location and the ethnic composition of the emirates affected the condi- tions of the servile populations is unclear; what is clear is that they were all involved primarily in agricultural production. It should also be noted that in all the emirates of the caliphate, members of the Fulani aristoc- racy were large holders of servile populations.

Having said this, it is possible to agree that in parts of the caliphate some newly acquired captives, by purchase or through capture, and some of those who lived within the residences of their masters lived and worked under conditions that approximated those of-These would be the ones Jan @ogendorn refers to as unmarried-slaves who_._, _ —. .“- were fed by their masters throughout the year.42 (Certainly, unmarried members of servile households worked and ate with their parents.)

, Based on the evidence presented, it is clear that this category of depen- dent cultivators did not constitute a large proportion of the total servile population in the Sokoto Caliphate in the late nineteenth century. The vast majority were householders settled on holdings of their own and were, therefore, Q&%-~~~ They possessed enough land and had enough&me and motivation to produce for themselves and maintain a____—- – household.43 Both in status and in economic independence, most of them were superior to the -of twelfth-century England. Many were in the category of the villein of medieval Europe, and, as Mason’s work shows, the dwellers of the captive villages in Bida had greater eco- nomic independence than even the villeins of medieval Europe. What is more, the dwellers of the servile villages in the Sokoto Caliphate were

-super& in economic independence and social status to the nineteenth-.v_..Im”x.- century Russian serfs. Applying the formula employed by European his-. “‘3. torians to distinguish between slaves and serfs in precapitalist Europe to the evidence from the Sokoto Caliphate therefore leads to the conclu- sion that the bulk of the servile populations-e c referred to as slaves by scholars were in fad&& cmatz call the servile villages in Bida and other parts of the C>phatPslave plantations,” we would have to accept that the medieval manor in

England and the serf village in nineteenth-century Russia were all slave plantations.

In fact, several of the authorities cited above are uncomfortable with the use of the terms slave or slavery to describe the servile populations they studied. For example, Mason explains:

I have used the term “slave” only because of its familiarity and its cur- rency in discussions of unfree labour in African societies. As I have suggested in the t i t le , “captive” is a more appropriate term, as it is principally the mode of recruitment which the West African “slave” had in common with his brother in Cuba or Brazil.&

While Hill writes:

Unlike genuine chattel slaves in ancient Greece and Rome, in the United States, and in Brazil and Haiti, who were always totally devoid of rights, farm-slaves in rural Hausaland normally enjoyed so many rights (including those of self-ransom) that it is reasonable to ask whether the term slave is, in fact, an appropriate translation of the

However, quite apart nom the imposs ib i l i ty of f inding English word, it is clear that present-day definitions . . .

are suff icient ly commodious to include the Hausa variant. 2

When a definition is too “commodious” it ceases to define anything, however, and Hill later complains: ‘So it seems that, after all, the use of the term ‘slavery’ does confuse certain essential issues in rural Hausa- land and that a substitute ought to be coined.“45


Michael Watts has also argued that the extent to which the upper classes in the Sokoto Caliphate depended on the surplus produced by slaves has been “implausibly inflated?/

In some of the northern and peripheral emirates, perhaps less than 10 ercent of the populace was servile, and while Kano and Sokoto may f-h ve been high-density systems, we are after all referring to a society const i tuted by perhaps 8 to I O mil l ion “free” peasants. . . . The rights of ownership and production that farm slaves possessed converged with the slaveholders’ ideology of the Is lamic patr iarch, assimilat ion, and the possibi l i ty of manumission. In short , the pol i t ical , economic, and ideological tendencies in the Caliphate were toward the production ofpeasants who, if not entirely free, could at least be taxed or retained in a quasi-cl ient status.*

– .

52 SLAVERY IN AFRICA Joseph E. Inikori 5 3

So, while some members of the servile population in the Sokoto Caliph- ate were approximately slaves, a more precise use of terms, in the man- ner of modern historians of precapitalist Europe, would compel scholars to describe the vast majority of them as- serfs. In other words, what the“-

i Bri&isGolonial administr- in Northern Nigeria in the early : twentieth century was more q than slavery:I

The evidence for the Western Su&n is similar in many ways to that of the Sokoto Caliphate. Martin A. Klein, who has written extensively on the servile populations of the region, describes how their labor was rationally exploited in the nineteenth century: “slaves were settled in separate villages or separate quarters and there was increasing control over labour, feeding and dues.“47 He points to two possible ways of grouping the populations: the native- born versus those acquired by pur- chase or by capture, and those who resided within their lords’ com- pounds versus those who settled in separate villages. The native-born could not be sold, while many of those residing within their masters’ compounds were newly acquired, along with some trusted retainers and concubines. The latter group were fed and clothed by their masters, and those settled in villages worked under chiefs of servile descent. As Klein reports, “ the larger numbers and the relative autonomy” of those living in separate villages “made possible the development of leader- ship.“48

The micro-studies by Marion Johnson and by William Derman pro- vide descriptive detail for the region. Johnson studied the state of Ma-

,sina, established by a Muslim teacher and his followers in the region generally referred to as the Niger I&&d (including T~Q&*-. and areas to the west). The state was c~63Xithe early nineteenth century and lasted until the 186os, when its capital was destroyed by Al-Hajj Umar The process of establishing the state and expanding its geograp–dc a r created subjugated populations, referred to as e (singular, di- madiu), some held by the state and others by private Fulani pastoralists. Many of these people were settled in villages and were given lands:-Gith one-sixth of the harvest to be paid as rent; a quantity of grains was also paid in dues called diawgal (settlements held by the state paid only di- amgal). Johnson–dismissed the suggestion that the settlement of the ser- vile populations & separate villages of their own was brought about by the founder of the Masina state, who transformed preexisting chattel slaves into serfs witli-d’efined rights:

The Masina system is so similar to the arrangements in adjacent terri- tories which never came under control of the LMasina theocracy, that

it seems more probable that Sheku Ahmadu [the founder of the Ma- sina state] formalized and regulated a pre-existing system.49

Derman’s study was conducted in the Fouta-Djallon area of the modern Republic of Guinea, where, in a Fulani-led j&ad from about 1727, the indigenous Diallonke, Susu, and Poulhs were conquered and-G reduced to servile stat% The ajugated mdlgenous population and captives brought from outside were settled in villages, where they pro- duced for themselves on lands rented from the Fulani overlords: “they were economically self-sufficient. They lived in their own villages, they cultivated their own fields and women’s gardens (although they did not own the land), owned property and had their own kin groups,“5o The rent for the land was ten percent of the harvest. The servile culGv.vtors often rented lands from th&.r-mti~, but sometimes from other Fulani landowners. Apart from the rent, they owed their masters lab&s, working “five days a week for their masters from early morning until early afternoon.” Although Derman contends that these people were reduced to poverty by the amount of labor demanded by their masters, the labor dues seem to be of about the same magnitude as those of Kano Emirate and Zaria in the Sokoto Caliphate. The fact that the servile population of the Fouta-Djallon grew naturally after French conquest ended the? wars and stopped the supply of new captives is an indication that the servile cultivators produced enough from their lands to support households with unproductive children and old members.51

The servile v-s also had a good amount of autonomy to look after their own afI%rs. For example, disputes in a village were settled by the elders, while political leadership was provided by a member whose master was a chief. This political leader was also the link between the servile village and the Fwverlords.

Taking all the evidence together, we can agree with Klein that many of the newly acquired captives and some others in the nineteenth-

servile villages were

in the region on the eve of Fyench conquest. We now come to servitude on the East Coast of Africa. Here, Fred-

erick Cooper has studied servitude in four important locations: two is- lands, Zanzibar and Pemba, where Omani Arabs produced cloves on a large scale in plantations; and two towns on the mainland, Malindi and

54 SLAVERY IN AFRICA P Joseph E. Inikori 5 5

Mombasa (both in modern Kenya), where Arabs, Swahili, and migrant Africans employed servile labor to produce grains for local and export markets. The evidence shows considerable differences in the socioeco- nomic conditions of the servile populations in plantations of Pemba and Zanzibar, they lived

these places. In the clove in huts ‘%cattered around

the plantations, dispersed among the clove trees.” The main tasks of the servile workers were harvesting, planting, and tending the young clove trees. There were two harvests each year, a large one in November or December and a smaller one from July to September. Planting, watering the young plants, and weeding were year-round tasks. During the har- vest season, these servile cultivators worked eight or nine hours a day, seven days a week; at other times they worked five or six days a week. They were allotted some plots where they produced some of their own food, mostly ew.52

In Malindi the servile cultivators worked in the grain fields of their masters in groups of five to twenty, under a headman of servile status. The work week was between fortv and fiftv hours. The cultivators were allotted small plots of between dred by fifi-y Y=h where they


two hundred by ten yards and two produced some of their own food.

hun- Dur-

ing the dry season, Thursdays and Fridays were set aside for them to work their own plots, while in the wet season (the main agricultural season) they had only Fr-dg off. According to Cooper, “Most i&or- mants claimed that masters provided slaves with a daily ration of food and that the produce of the slaves’ own plots was a supplement to this. Masters were also expected to provide their slaves with clothes.“53

In Mombasa, on the other hand, servile cultivators held by several masters lived in villages of their own, each village containing from fifty to three hundred people. Some of them worked on their own and paid their masters a monthlv or annual sum called iiara. Others cultivated the nearby fields of thiir masters without suG=n; a quantity of grain was paid as rent to their town-dwelling masters, &ho periodically went to the villages to collect it. The dwellers of these servile villages around Mombasa were left with much autonomy. Cooper reports, “In.._ the Mombasa area, slaves ‘often lived-inawcojng &ves_owned by various masters. They were governed by a slave headman elected by their own elders. He was expected to arrest any suspected criminals and send them to Mombasa for trial.“5.!

Taking the evidence together, we agree entirely with Cooper that the servile cultivators of the clove plantations and in mainland Malindi can accurately be referred to as slave% Even though they had some small plots allotted to them, they did not have enough time to produce for-r

f :i


themselves and maintain a reasonable level of socioeconomic indepen- dence. The small size of the plots, less in area than a soccer field, is itself indicative. Although the demographic evidence is scanty, it clearly indicates that the servile populations in the clove plantations did not reproduce themselves socially, in spite of the favorable gender balance. The observation of contemporary visitors that the rate of reproduction was low and the death rate high is consistent with the hard data showing that mainland-born people were two-thirds of the servile population 1 in 1900-1901. Since large-scale slave imports from the mainland were concentrated in the first four decades of the nineteenth century, as the growth of clove exports suggests, a reasonable reproduction rate should have given rise to a much greater proportion of island-born people in the population by this time.55 This demographic evidence points further to the low lev$ of socioeconomic autonomy among the servile popula- tions of the clove plantatrons. ____ _ -._- “.–1 .–

-_^. — 3

mf those in the mainland area of Mombasa is a different- matter. There we are dealing with people whose socioeconomic auton- omy was greater than that of even the villiens in medieval Europe, not to mention that of the buvavii of twelfth-century England and the serfs of nineteenth-century Russia. It is an inexcusable terminological loose- ness to lump together the dependent cultivators of the Mombasa area+ and the servile populations of Malindi, Zanzibar, and Pemba under the same dependent social category. The former certainly resembled serfs more than slaves.

Having examined the regions with high densities of servile popula- tions in precolonial Africa, it is tempting to generalize that nothing very different existed in other regions. That may well be the case in places like Asante, Dahomey, and the Benin kingdom of south-western Nigeria. In such places, a significant population of dependent people whose socio- economic conditions approximated slavery existed side by side with others, greater in number, whose conditions were closer to those of serfs. There are regions, however, in which the people described as slaves by scholars had so much freedom that it would not be appropriate even to call them serfs, as in the case of the Puna and Kuni societies of modern Congo, where the so-called slaves labored in the same way as their free counterparts, could hold any &ice in “&&JQ&tt. and were- usually marrie$o&eeApouses from ~~e.‘lineages~~of-~i~-~st~rs.5~ There is~~?he interesting case of the servile warriors of Mozambique in the period from 1825 to 1920 who more resemble the knights of me-..s . . . 1.- ..,. .,, l., ..’ — .__._,, ~ _ . _ dieval EL&&~: ‘they lived in their own villages, scattered all over the land; they had their own leaders, appointed from among them by their

Joseph E. Inikori 57

masters; and they constituted the instruments with which the lordly class dominated the peasants, collecting tribute from the peasants and

$imposing discipline on the peasant communities. The most immt service they rendered to their masters, howeverqgw s military

1 AFhus, it is factually correct to say that a significant number of people

late nineteenth-century Africa south of the Sahara labored under socioeconomic conditions approximating slavery. However, it is also factually correct to say that the numbers of such people have been

of the people usually described as slaves at all. By the yardstick employed by histori-

s of precapitalist Europe to separate slaves from serfs, the &lk of *se some WerefrpeAeanle.

number of slaves existed in parts of r sub-Saharan Africa in the late nineteenth century, was this also true of

To answer this question we s with established centralized

state svstems in the Western. Sudan and in West-Central Africa. The -.


agrarian societies of the fifteenth and sixteenth centus-which subse- quently formed the nucleus of the Asante state, should also be exam- ined, for it is known that Asante economic and political entrepreneurs purchased captives from other African regions at this time.

The accounts of the first Europeans to come into contact with West -&can societies in the Western Sudan (the Senegambia area) do con-

tain references to dependent social categories that these Europeans clas- sified as slavery. Studies based on Arabic sources indicate that those sources also contain references to slaves in the societies of the Western Sudan. The combined authority of the Arabic and early European sources gave rise to the generally accepted view that the impact of the trans-Saharan trade occasioned the widespread use of slaves to produce commodities and to provision the aristocracy and state functionaries in the Western Sudan long before the first Europeans arrived on the West African coast. 58 As more detailed description of the material conditions of the populations referred to in these sources as slaves becomes avail- able, it appears that historians have been too uncritical of the sources in question.

From what we know of the historv of the Western Sudan from the beginning of the present millennium to the end of the sixteenth century, Songhay was the largest and most complex state system that ever existed – in the reglonbefore the coming of the first Europeans. It was heavily involved in the transSaharan,t~de and exhibited a highly developed classsystemJf there was an extensive use of slaves in the region, the


Songhay of the twelfth to the sixteenth century must be the place to find them in large numbers. S-M. Cissoko holds that indeed there were many slaves in Songhg. Based on the terminology employed in the Ara-..— – bit sources, he says that private and state officials of Songhay employed large numbers of slaves to cultivate their estates in the rural areas, but his descriptive evidence tells a different story. This can be best observed in Cissoko’s own words:

Large estates belonging to the princes and @z-Lz-~ of the great towns were worked by slaves, settled in farming villages. The askiya was him- self one of the great landowners. His fields, scattered throughout the valley, were cultivated by communities of slaves working under over- seers called fanfa. A sort of rent in kind was levied on the harvests and sent to Gao. The same happened with private slaves.j9_ _ _ _ . ._.

He adds further details:

The Ta’rikbs give us a few glimpses of country life. There is virtually no mention of peasants’ revolts. The rent demanded from the slaves by their masters was never crushing. . . . The peasants even sold part of their produce in the local markets, obtaining products like salt or

cm cloth and thus becoming involved in tradem60

no doubt that we are dealing with for themselves and paying

located in villages and Sphysically separated from their lords. The fact that the rent demanded “was never crushing” must mean that the dependent peasants could maintain households with unproductive children and old members and therefore were able to reproduce themselves socially. D. T. Niane con- cludes that these dependent peasants were serfs and not slaves:

Up to the present time [12th-16th cenG] in black Africa, before the

monetary economy developed, land was considered to be the indivis- ible property of the community. The kings and emperors had “human estates,” that is, Lands worked bv subjugated communities;-but closer examination shows that this was a system of serfdom rather than–..—r.——–..-cI_ -w-4.yl_f-e.xw,. b slavery.61

It is possible that some slaves did exist in the Western Sudan by the fieenth century, but newly available evidence indicates that the litera- ture on the subject stands in need of revision. The descriptive evidence in the sources should be carefully reexamined in order to separate true slaves, if any, from serfs and other dependent social categories.

JOSL@I E. Inikori 59

If it is hard to find true slaves in the societies affected by the trans- Saharan trade before the coming of the Europeans, it is even harder to

,, do so in the coastal societies of Western Africa and their hinterlands, y which were not directly affected by that trade. Somehow, Asante occu- . pied a borderland between the societies of the Western Sudan that were

directly involved in the trade and those farther down the Atlantic coast, which only received indirect effects. In fact, Asante economic and politi-

( cal entrepreneurs, who accumulated wealth from the sale of gold to the ) merchants of the Western Sudan, bought captives in then& and

sixteenth centuries to clear&-gin lands for agriculture. Some of the cap- tives came from the Western Sudan; others were imported from the Benin kingdom by the~~<~&~.&+ante entrepreneurs. This” situation ‘prc&pted some historians to argue that Asante already had a slavgclass by the fifteenth century. However, Ivor ,Wilks has now shown that the purchase of people by the Asante in the fifteenth and sixteenth centuries did not produce a slave class at that time. The wealthy Asante, who bought people to clear the forests, took pains to

a slave class. The matriclan institution was invented-..x”_.“m.– at this time for this particular purpose. Wilks expressed that:

Cl \

k$iQi ne of the major thrusts in Asante social “engineering” was towards

. i p revention of the cwolidation of a slave caste: those of unfree originse -.-__ were assimilated as rapidly as possible into the class of free Asante commoners and their acquired status afforded full protection of the


law. . . . But the precise way in which such task forces were incorpo- rated and assimilated into the open-textured matriclans must remain

.-a for the present a matter for speculation; ail that is sure is that incorpo-

% i ration and assimilation occurred, and that in consequence no slave.-

Therefore, to the extent that there were slaves in Asante in the nine- teenth century, their origin must be associated with conditions created after the coming of the Europeans.

The other major region where recent research has produced helpful evidence on social structure before the transatlantic slave trade is West- Central Africa. This region is important, because it was not involved in the trans-Saharan trade, and it was one of the early major suppliers of slaves to the European traders. When Europeans arrived in the late fif- teenth century, the Kingdom of Kongo, the main centralized state sys- tem in the region, was contnnnng its territ&l expansion. Its political economy rested on the collection and redistribution of tribute products


from different ecological regions of the state. As the kingdom expanded geographically, the range of tribute products increased and the income& of the ruling elites came to depend on the tributes.63

This early tribute system apparently eliminated the need for produc- tion by slaves to support the rulers. There were political clients, servants, and prisoners of war, but no slave class; there was no trade in people of any type before the coming of the Portuguese. Hence, there was no word for slave or slavery among the pm-contact Kongo people. Anne Hilton states that as late as the seventeenth century ‘(the only term which referred to purchase [that is, purchased people], muntu LZ kusum- was a compound construction suggesting that the phrase had been d e v i s e d -condition.**

Jan Vansina’s elaborate linguistic study of West-Central Africa car- ’ roborates the conclusions derived primarily from archival sources. On the basis of linguistic evidence, Jan Vansina has stated that there was no trade in people in the region before the coming of the Europeans, and that the various words meaning “so in the region are “loanwords” that evolved as the EuroDean slave trade spread from the coast into the–__- – -I / –

articular, Vansina studied/ in1 detail the etymology of the which in Kongo meant ‘(se& before ISOO, but later

came to mean the traded “slave” in the major communities of Central Africa involved in the transatlantic slave trade.65

It is thus incorrect to say that the preexistence of widespread slavery in the coastal societies of Western Africa made possible the growth of the transatlantic slave trade. In a recent effort to revive this outdated explanation, Johns argues that African legal systems, which prevented private ownership of land, gave rise to the accumulation of slaves everywhere in Africa, and the transatlantic slave trade “was the outgrowth of this internal slavery.” Based on a misunderstanding of medieval European history, Thornton argues that “by contrast, in Euro- pean legal systems, land was the primary form of private, revenue-_-” – prc$cing property, and slavery was relatively minor.“(j6 In fact, a careful__ _. ..— examination ‘of ;he evidence points to a considerable similarity between the landtenu~stem i.n$e majo~s~of-~ed@a.l Europe and that in the major states ofEc&nialAfrica. As Douglass North and Roberte—-. -I. .– Thomas have pointed out, “Feudal law [in Europe] did not recognize the concept of land ownership,“67 and Goody makes the same point:

from the evidence concerning the inheritance of land at the village level [in medieval England] it would seem that here the idea that con- quest put all rights in the hands of the Norman conquerors was some-

Joseph E. Inikoti 61

thing of a fiction. Whatever the legal position on this abstract level, the medieval system in practice appears to display some similarities with African land tenure, especially in states like

m e 68

It was the same process which changed slaves and serfs to free peasants, as feudalism collapsed in Europe, that also gradually transformed the European feudal lord into a modern landlord.69 It should be noted that when the market conditions in Africa were right in the Nnth and twentieth “-@Af&an legal systems” could hold-k the de-‘ \ velopment of modern private property rights in land.

The evidence – clearly refutes the contention that the transatlantic / slave trade sprang from previously accumulated slaves in Africa. There

was no process in place that regularly generated people for sale in the coastal societies of Western Africa before the coming of the Europeans. On the contrary, it was the power of&rope-r captives to be exported–that-alegathering of captives for

ent is West-Central Africa, *and the best case in the region is the Kingdom of Kongo. This kingdom had never been involved in a trade in people, but it had the capacity to generate captives from outside the polity, and these captives could be sold if conditions made that the best alternative, economically and polit- ically, for the kingdom’s economic and political entrepreneurs. To start, it shoul be noted that the main initial Kongo trade with the Portuguese was in copper. It was about thirty years after their arrival that the Portu- guese shifted their emphasis from copper to slaves. Even by this time the King of Kongo had no immediately available “disposable people” to sell when the King of Portugal sent a trade mission to prepare the way for large-scale slave exports. As Hilton reports, while the Kongo ruler accepted the new trading arrangement and the gifts from the King of Portugal, he had no slaves to send even as return gifts:

There were in fact very few slaves available for purchase and scarcely ‘any had so far been exported from Kongo. In order to secure the re- turn gift Afonso [the reigning Ring of Kongo] had to raid the neigh- boring Mbundu . . . newly acquired captives being the only people who could, at this time, be legitimately sold.‘O

Thus, during the early years (up to the second decade of the six- teenth century), when the slave trade remained marginal and while cop-


per and cloth were dominant, not even the King of Kongo “had a large retinue of slaves,“71 let alone the title-holders and provincial governors under him. At that time the kingdom did not even have a standing army. In the course of the sixteenth century, however, the sociopolitical and military conditions associated with the transatlantic slave trade led to the accumulation of dependent populations in Kongo. We do not know enough about their conditions. The literature refers to all of them as slaves; we can only guess that many were approximately slaves and others serfs.

The state was eventually forced to establish a standing army made up predominantly of slaves. Aharo 11 (1587~1614) had sixteen to twenty thousand Tio slave guards. Officeholders and merchants in the capital city and other trading centers employed many servile people in agricul- ture to meet their subsistence needs. In the course of the seventeenth century the holding of servile cultivators in the central region of the kingdom became so widespread that almost every aggrieved person de- manded slaves as compensation.72

The Kongo evidence shows that transatlantic slave exports devel- oped without any assistance from a preexisting slave class. It also shows that political clients, servants, and other dependent peoples in precon- tact Kongo could not be sold bv law and tradition. What is more, thereI is no indication that the rulers had a pathological drive for the accumula- tion of people as an end in itself, as is often claimed. On the contrary, these rulers, like rulers in all precapitalist societies. saw people as noten- tial producers-and subie&hu

rs in the short run,

– -\ No one who knows the response of African political and economic entrepreneurs to European demand for agricultural and other products in the nineteenth and twentieth centuries will be surprised by this find- ing. Western Africans never produced ortradedcocoa and-rubberbefore–l^l- the comix.of , the.. Europeans.,-,.but when the latter dem.am5e.d those products and -were willing to pay prices that made their production re- warding, they were produced in rapidly growing quantities, although there was no previous experience to fall back on. Given the circum- stances of African states in the era

-..—__.- of?FZ transatlantic slave trade, it is

not hard to imagine that the political and economic value of European trade missions, like the one sent by the Portuguese to Monso I of Kongo in 1512, must have been easy to sell to the rulers of those states, since they could procure the captives from outside their polities at very little cost to themselves.

h I-

Joseph E. Inikoti 63

We owe a significant intellectual debt to the historians and anthro- pologists who have studied slavery in Africa for the past two decades or so. Their research has helped to call attention to social and economic issues in the study of African history after the initial emphasis on politi- cal questions. However, the sociology of knowledge that directed the research seems to have encouraged a lack of attention to terminological precision. As we know too well, slavery in Africa was a major theme constructed by the European slave traders to defend their business against the abolitionist onslaught in the eighteenth century. The same theme also became fashionable for the agents of European colonialism in the late nineteenth and early twentieth centuries, as the abolition of slavery was presented to the moral conscience of Europeans as a part of he EUTOPedC – ’ ’ * . ’ 3pa. The propaganda by the slave traders and by the agents of colonialism was so effective that it was given intellectual respectability in history textbooks. One part of the propaganda- the claim that the preexistence of widespread slavery in Africa gave rise to and helped to sustain the-transatlantic slave trade- was challenged ,~..&~,-mi&–~6-~s by Walter Rodney.73 It is fair to say——–___ _ that the subsequent encounter between Rodney and Fage set the stage for the research of the 1970s and 1980s on servile institutions in precolo-


nial Africa. Even Rodney, however, while questioning the existence of slavery in the coastal societies of Western Africa before the transatlantic slave trade, accepted uncritically the colonial propaganda that slavery w a s everywhere in late nineteenth-century Africa.

Given the sociology of knowledge that informed the study of slavery m Africa, it is understandable why very little attention has been paid to terminological precision, which characterizes the study of dependent social categories in &history of precapitalist Europe. I have attempted– to redress this weakness in the literature by applying to the Mrican evi- dence the formula employed by modern historians to se-es from_s& precapitalist Europe. The result shows that, while there were slaves in late nineteenth-century Africa, the bulk of the people hith- erto so described, resembled serfs. The phenomenon of what scholars refer to as inter-generational mobility among the slave populations in Africa’* – the tendency for the children of slaves to be- come free persons or nearly so – meant that the slave class in Africa could not reproduce itself, not only because its rate of reproduction was low, but largely because the children of slaves normally did not remain in&~; they either became freeor.&m<serfs. Apart from the clove plantations of East Africa, most of what-scholars have called “slave plan- tations? fact, serf villages. To accept them as slave-I- .


plantations, we must also accept the medieval manors in Europe and the serf villages of nineteenth-century Russia as slave plantations. In that case, Russia would have possessed the largest concentration of slaves in the nineteenth century, with over 22 million &..thc~n~~**–

The evidence also shows that there were no slaves in the coastal soci- eties of Western Africa in the fifteenth century to provide the spring- board for the transatlantic slave trade. It is, therefore, misleading to speak of cLtransformations in slavery” in Africa from the fifteenth cen- WV 75 What happened in the coastal societies of Western Africa and their hinterlands after the fifteenth century, following the socioeco- nomic and politico-military conditions associated with the transatlantic slave trade, was the transformation of free peasants, political clients, ser- vants,and ‘- – *allowing myxm-. plishrnents of the research of the last two decades, what is needed now is terminological precision. We need to find ways of separating slaves from other dependent social categories in precolonial Africa and reinter- pret the evidence.


I . Paul E. Lovejoy, “Foreword,” in Claude Meillassoux, The Anthropalagy of Slavery: The Wmnb of Iron and Gold, translated by Alide Dasnois (Chi- CigO: UniVerSity of Chicago Press, 1991), p. 7. LOVejOy’S VkW iS totally contrary to Meillassoux’s argument in the book, which unambiguously states that tropical Africa was the last region in the world to develop a trade in slaves and the institution of slavery, many centuries after simi- lar developments in Europe and Asia and around the Mediterranean (pp. 20-21).

2 . Paul E. Lovejoy and Jan S. Hogendorn, Slow Death For Slavery: The Course ofAbolition in Northern Nt&tia, 1897-1936 (Cambridge: Cam- bridge University Press, 1993), pp. xiii, I.

3. IMeillassoux, The Antbropoiqy of Shmy, p. 22. 4. Martin A. Klein, “The Study of Slavery in Africa,” Journal ofA+an

N&q, 19, no. 4 (1978): 599-609; Frederick Cooper, “The Problem (3f Slavery in African Studies,” Journal ofAf;ican History, 20, no. I (1979) : 103-25; Paul E. Lovejoy, “The Characteristics of Plantations in the Nineteenth-Century Sokoto Caliphate (Islamic West Africa),” Ameri- can Historical Revim, 84, no. 5 (1979) : 1267-92.

5. Paul E. Lovejoy, “Plantations in the Economy of the Sokoto Caliph- ate,” Jownal ofAfican History, 19, no. 3 (1978): 34x-68; Lovejoy, “The Charaaeristics of Plantations.”

6. Igor Kopytoff and Suzanne Miers, “African ‘Slavery’ as an Institution of Marginality,” in Shvq in Afica: Historiial and Antbropolagical Per-

Joseph E. Inikoti 65

spectives, edited by Suzanne Miers and Igor Kopytoff (Madison: Uni- versity of Wisconsin Press, 1977). Cooper, “The Problem of Slavery,” p. 10s. Claude Meillassoux, “Female Slavery,” in Slavey in Af;ica, edited by Claire C. Robertson and Martin A. Klein (Madison: Univer- sity of Wisconsin Press, 1983). Meillassoux states that ‘<one approach to African slavery, which stresses its benevolent character by comparison to American or West Indian slavery, tends to play down the differences between slaves and other dependent or dominated social categories, such as pawns, serfs, or even married women” (p. 50). Joseph E. Inikori, The Chaining of a Continent: Export Demandfor Captives and the H&toy ofAjk’ca South of the Sahara, 1450-1870 (Mona, Jamaica: Institute of Social and Economic Research, 1992), p. 37. David Pelteret, ‘Slave Raiding and Slave Trading in Early England,” in Anglo-Saxon England, vol. 9, edited by Peter Clemoes (Cambridge: Cambridge University Press, 1981), pp. 99-114. Pelteret writes: “In con- trast to the Roman world, slavery seems not to have been an integral element in the social structure of the Germanic peoples living outside the Empire at the time when Tacitus was writing about them” (p. 100). But the Anglo-Saxons were involved in taking captives and ex- porting them to places in the Roman empire before they moved into England. That practice of taking and selling captives contin.ued for centuries after they settled in England. Pelteret, “Slave Raiding,” pp. 99, 102. Lzfk of St. WuZstan, &hop of Worcester, translated by J. H. F. Peile (Ox- ford, 1934), pp. 64-65, cited by Pelteret, “Slave Raiding,” p. 113, n. 107. Eadmer’s Histouy of Recent Events in England, translated by G. Bosan- quet (London: University of Wisconsin Press, 1964), p. 152, cited by Pelteret, “Slave Raiding,” p. 113, n. III. H. E. Hallam, “England before the Norman Conquest,” in TheAgrar- ian History ofEngland and Wales: Volume IA 1042~1350, edited by Hallam (Cambridge: Cresset Press, 1988), pp. IO-IL In The Famuhs: The Es- tate Labouver in the XIIth and XIIItb Centuries (London: Cambridge University Press, 1954), iM. M. Postan has suggested that the percent- age of slaves in the total population may have been less than the Domesday figures imply: “Corrected by the coefficients which histori- ans employ to translate the Domesday figures of households into num- bers of heads, the proportions of slaves to the total Domesday population would probably turn out to be smaller than the gross fig- ures in the Domesday might suggest. For we must assume that whereas a large proportion of the slaves were not in a position to estab- lish families, all the other social groups were counted in family units”

(PP. s-6). Postan holds that by 1086 the conversion process had not gone far


17. 18.

19. 20.




2 4 .



enough to engulf “that portion of the slave class whom lords were still employing in their c&a” (Postan, The Famulus, p. II). Postan, The Famulus, pp. 12-13. Ibid., p. 23. Ibid., p. 36. Earlier in the work, Postan had stated: “How fundamental the change-over from slave to bovarius was, is not a question to which a simple answer is possible. It may have signified nothing more than a modification of their status, and may not have been of very great economic importance. On the other hand it may have been one of habitation-the resident slave may have been transformed into a land- holding servant – in which case it was of great economic significance”

(P.11). Ibid., p. 9. Ibid., p. 25. This way of looking at slavery and serfdom would seem to be consistent with Meillassoux’s theoretical construct, which distin- guishes slavery from serfdom on the basis of the dependent producer’s capacity to retain a proportion of his surplus large enough to raise a family and maintain a household. Under Meillassoux’s framework, if the dependent producer possessed enough land and had enough time to work it and keep a surplus large enough to raise children and main- tain old members of the household, then he was a serf On the other hand, if the lord’s demand for surplus labor was so great that the de- pendent producer was unable to maintain a household, then he was a slave. Hence, the observation that in all slave systems the slave class is unable to reproduce itself socially, being born of the “womb of iron and gold” – capture by force of arms and purchase with money on the market. See Meillassoux, The Anthropology of Slavery. R. H. Hilton, The Decline of SeQdom in Medieval England (London: Macmillan, 1969), pp. 9-11. Peter Kolchin, Unfhee Labor: American Slavery and Russian Serfdum (Cambridge, Mass.: The Belknap Press of Harvard University Press, 1987), pp. 2-4, 41. The discussion of the Russian serfs is based largely on Kolchin’s work. Ibid., pp. 2-3. Ibid., pp. 37-38. Ibid., p. 43. By the mid-eighteenth century there were two broad groups of peasants, state peasants and serfs. The latter, owned almost exclusively by noblemen, constituted over fifty percent of the peasant population. The state peasants were made up of those who escaped enserfment and those recently freed (Ibid., p. 39). Ibid., p. 52. It is not clear whether male serfs refers to heads of house- hold or to all males, including children. If reference is to heads of household, then the total serf population was much greater than what is stated here.

66 SLAVERY IN AFRICA Jose@ E. Inikoti 67

2 7 .

2 8 .








3 6 .


Historians have debated the issue of whether the nineteenth-century Russian serfs were aCNdy slaves or serfs in the main, although their conditions came close to those of chattel slaves. This is an important distinction, contrary to the view of Kolchin, who thinks that the dis- pute is mere hairsplitting. The sale of the serfs certainly moved them very close to chattel slaves. But, as Kolchin himself acknowledged, most serfs continued to receive allotments of land of their own from their owners with which they supported themselves (Ibid., pp. 43-45). For this reason they were sewi and the demographic evidence indicates that they were able to maintain a household and reproduce themselves socially. Their sale was also regulated by state laws: Peter I decreed in 1721 that family members must not be separated by sale; and in the nineteenth century Nicholas I twice outlawed the selling of unmarried children away from their parents (Ibid., p. 117). It is fair to say that in the main these people were serfs, but over time their condi- tions came close to those of chattel slaves. Jacques J. Maquet, “A Research Definition of African Feudality,“Jmr- nal ofAfican History, 3, no. 2 (1962): 307. Maquet thought that the Marxist conception would be difficult to apply to precoloniaf Africa, where the notion of land ownership em- bodied in Roman law was non-existent. See Maquet, “African Feu- dality,” p. 309. Jack Goody, “Feudalism in Africa.,) “Journal ofAfi-ican History, 4, no. I (1963): 3. Goody, “Feudalism in Africa.,: ” p.10. He refers to I. I. Potekhin’s “On the Feudalism of the Ashanti,” a paper read to the Twenty-fifth Interna- tional Congress of Orientalists in Moscow (1960). Basil Davidson, Bluck Mother: The Years of the Afi-can Slave Trade (Bos- ton: Atlantic Monthly Press, 1961), pp. II-~. Goody, “Feudalism in Africa?,” pp. 10-13. Goody may as well have in- cluded words like shvey and plantations. J. H. M. Beattie, “Bunyoro: An African Feudality?,” Journal ofAfi-icn History, 5, no. I (1964): 26, 35. See Mary F. Smith, ed., Baba of &woo: A Wmnan of the Muslim Hausa, with introduction and notes by Michael G. Smith (London: Faber and Faber, 1954); Michael G. Smith, “A Study of Hausa Domestic Econ- omy in Northern Zaria,“A@ca, 22, no. 4 (1952): 333-47; M. G. Smith, “Slavery and Emancipation in Two Societies,” Social and Economic Stud- ies, 3, no. 3-4 (Dec., 1954): 239-90; and M. G. Smith, Government in Zazzau, r&o–r950 (London: Oxford University Press, 1960). Smith, “Slavery and Emancipation,” pp. 24+253,264-267; Smith, Guv- emment in Zazzau, pp. 86, 89-90; Smith, introduction to Baba of KY&o, p. 22. Polly Hill, “From Slavery to Freedom: The Case of Farm-Slavery in Ni-








44 .

4s. 46 .

4 7 . 48.

gerian Hausaland,” Comparative Studies in Society and Histoy, 18, no. 3 (1976): 418. Ibid., pp. 402-404. Michael Mason, “Captive and Client Labour and the Economy of the Bida Emirate, 18y-I9oI,“Journa/ ofA&can History, 14, no. 3 (1973) : 459-60. Mason, ‘Captive and Client Labour,” pp. 465-68. Michael Mason, “Production, Penetration, and Political Formation: The Bida State, 1857-1901,” in Modes of Pr-oductiun in AJi*ca: The Precolo- nial Era, edited by Donald Crummey and C. C. Stewart (Beverly Hills: Sage Publications, 1981), pp. 214-15. Jan Hogendorn, “The Economics of Slave Use on Two ‘Plantations’ in the Zaria Emirate of the Sokoto Caliphate,” InternationalJournal ofAf- tian Historical Studies, (1977): 378. According to the more detailed work schedule presented by Hogen- dorn, the dependent cultivators in Zaria rose by 4:oo A.M. for the morning prayer, after which they went with their families to their own farms. At 9:oo A.M. they moved to their lords’ fields and worked until midday, when they rested and had their meals. By 2:oo P.M. work on the lords’ fields was over and they were free to return to their own farms. This schedule was followed during the farm season, which lasted (including the harvesting period) probably no more than six or seven months. During the dry season, from November to April, they worked mostly for themselves and were paid when they worked for their lords. See Hogendorn, “The Economics of Slave Use,” pp. 375-76; and Hill, “From Slavery to Freedom,” p. 418. Assuming that the morning prayers lasted an hour, the servile cultivators must have had at least six hours daily to devote to their own farms during the farming season (allowing for movement to and from the lords’ fields), while spending at most about four hours on their lords’ fields. Of course, there was a lot of time during the dry season to engage in non-agricultural activities for themselves: hunting, handicraft produc- tion, and so on. Mason, “Captive and Client Labour,” p. 453, n. 2. Hill, “From Slavery to Freedom,” pp. 397,413. Michael Watts, Silent Violence: Food, Famine and Peasantry in Northern Nz&ria (Berkeley: University of California Press, 1983), pp. 77-78. De- spite this telling criticism, Watts still talks of “several million slaves” in the Sokoto Caliphate in the late nineteenth century (p. 191). Klein, “The Study of Slavery,” p. 607. Martin A. Klein, “Slave Resistance and Slave Emancipation in Coastal Guinea,” in The End of Slavq in Afica, edited by Suzanne Miers and Richard Roberts (Madison: University of Wisconsin Press, 1988), pp. 208-209.

Joseph E. Inikori 69






54. 55.

56. 57.





6 2 .


6 4 .


Marion Johnson, “The Economic Foundation of an Islamic Theoc- racy-The Case of Masina,” ofAfi-icn History, 17, no. 4 (1976) : 488-89. William Derman, Se@, Peasants, and Socialists: A Former SerfVillage in the Republic of Guinea (Berkeley, California: University of California Press, 1973), p- 30. Derman, Se@, Peasants, and Socialist.., p. 34. Frederick Cooper, Plantation Slavery on the East Coast ofAfi”a (New Haven: Yale University Press, 1977), pp. xi, 156-70. Cooper, Plantation Slave?, p. 173. Cooper, Plantation Slavery, p. 173-76, 228. Cooper, Plantation Slavery, pp. 51, 52, 61, 131, 221-25. Klein, “The Study of Slavery,” p. 605. Allen Isaacman and Anton Rosenthal, ‘Slaves, Soldiers, and Police: Power and Dependency among the Chikunda of Mozambique, ca. 182j-3920,” in The End of Slave?, pp. 220-53. Walter Rodney, “African Slavery and Other Forms of Social Oppres- sion on the Upper Guinea Coast in the Context of the Atlantic Slave Trade,” Journal ofA@ ‘can Histoy, 7, no. 3 (1966): 431-43; John D. Fage, “Slaves and Society in Western Africa, c. I+45–I7oo,” Journal of Afi-ican Histoly, 21 (1980): 289-310; Claude Meillassoux, “The Role of Slavery in the Economic and Social History of Sahelo-Sudanic Africa,” in Forced Migration: The Impact of the Export Slave Trade on Afican Socie- ties, edited by Joseph E. Inikori (London and New York: Hutchinson and Africana, 1982), pp. 74-99; J. Devisse and S. Labib, “Africa in Inter-Continental Relat ions,” in UNESCO General History ofAfica: w Aficaj?om the Twelfth to the Sixteenth Century edited by D. T. Niane (Berkeley: Heinemann, University of California Press, UNESCO, 1984), p. 672. S. M. Cissoko, “The Songhay from the 12th to the 16th Century,” in UNESCO General History, pp. 202-203. Cissoko, “The Songhay,” p. 205. Niane, UNESCO General Histoly, p. 682. Ivor Wii, “Land, Labour, Capital and the Forest Kingdom of Asante: A Model of Early Change, ” in The Evolution oj’Socia1 Systems, edited by J. Friedman and M. J. Rowlands (London: Duckworth, 1977), pp. 523-24. Anne Hilton, The Kingdom ofI<onJo (Oxford: Clarendon, was), pp. 32-35. Hilton, %npium OffinJO, p. 233, n. 86. Jan Vansina, Paths in the Rainforests: Toward a History of Political Tradi- tion in EquatorialAfica (London: James Currey, IWO), p. 278; Van- sina, “Deep-Down Time: Political Tradition in Central Africa,” Histq




6 9 . 70.



inAfi-ica, 16 (1989): 352. The latter work shows the spread of the term from the slave trading coastal communities to the interior, along the main trade routes (p. 353, map 3).

66 . John Thornton, Afica and A&cans in the Making of the Atlantic World, 1400-1680 (Cambridge: Cambridge University Press, 1992), p. 74. This book focuses on a very important subject- the contribution of Africa and Africans to the history of the Atlantic basin; unfortunately, it is marred by conceptual weakness and factual inaccuracies. For example, the statement that “The average density in seventeenth-century Lower Guinea . . . was probably well over thirty people per square kilometer, or well over the average European density of the time” (p. 75) has no empirical foundation, both on the European and on the African side. The population of Western Europe was 61 million in 1200, 73 million in 1300, and 78 million in ISSO (D. C. North and R. P Thomas, The Rise oj’tbe Western World [Cambridge: Cambridge University Press, 19711, p. 71). With a land area of 898,804 square miles, the average den- sities come to 68 persons per square mile in 1200, 81 in 1300, and 87 in ISSO. Average densities in Western Africa did not reach these levels un- til the colonial period. It was these relatively high population densities of Western Europe, and the growth of trade and the commercializa- tion of socio-economic life to which they gave rise, that provoked the development of private property rights in land in Western Europe. The economics and the empirical basis of Thornton’s arguments are, therefore, dubious. Douglass C. North and Robert I? Thomas, The Rise of the Western World: A New Economic History (Cambridge: Cambridge Universi ty Press, 1973), p- 63. Goody, “Feudalism in Africa!,” p. 6. North and Thomas, The Rise of the Western World, p. 88. Hilton, Kingdam of Icongo, p. 57. Hilton, King-dam ofI<ongo, p. 78. Hilton, Kingdom of Konz.0, pp. 58, 78, 85, 122-23; see also John K. Thornton, The Kingdom of Kon.0: Civil War and Transition, 1641-1718 (IMadison: University of Wisconsin Press, I983), pp. 15-27, which de- scribes the urban setting for slave employment but does not trace the historical development of slavery in Kongo. Research in other regions shows a similar pattern. See Robert W. Harms, River of Wealth, River of Sowow: The Central Zaire Basin in the Era of the Slave and Ivory Trade, ISUO-1891 (New Haven: Yale University Press, 1981); Ralph Austen, “Slavery among Coastal Middlemen: The Duala of Cameroon,” in Slav- q in Afica; Joseph E. Inikori, The Chaining of a Continent: Export De- mand for Captives and the HistoT ofAfica South of the Sahara, 14~0-1870 (Mona, Jamaica: Institute of Social and Economic Research, 1992), pp.

Joseph E. Inikori 71

25-39; Patrick Manning, and Afican Life: occiafevttal, Oriental, andA@an Slave Trash (Cambridge: Cambridge UniversityPress, IWO); and Meillassoux, The Anthropdu~y of Slaveq, pp. 33,40, 43, 70, 239-40. Rodney, 431-33.

“African Slavery and Other Forms of Social Oppression,” pp.

Klein, “The Study of Slavery,” p. 605. Paul E. Lovejoy, Transjbvnations in Shve~: A Histaly of Slavery in A@ca I (Cambridge: Cambridge University Press, 1983).


Comments are closed.