The history of anthropological and historiographic studies of precolonial Africa between 1960 and about 1990 can be described as a long struggle against the weight of the great nineteenth-century myth of social evolution, as advanced by Herbert Spencer, Henry Maine, and Lewis Henry Morgan, and elaborated by later anthropology. Few would argue against the idea that, overall, social organization has progressed from simple to complex forms; problems arise as we attempt to fit African facts to the content of the evolutionary stages and the transitions between them. Jan Vansina's tour-de-force, Paths in the Rainforests (1990), in which he sketches the social development of Central Africa over two thousand years, provides an instructive example of progress in overcoming these problems, but also shows the continuing burden of nineteenth-century assumptions and unreliable twentieth-century anthropology. It is not my intention to evaluate the entire work but to focus on a critical section, outlining social evolution in the area around the lower Congo, which shows both the strengths and the weaknesses of Vansina's approach.Footnote 1
THE EVOLUTIONARY LEGACY
The evolutionary scheme begins with simple families that develop into unilineal descent groups (matrilineal, then patrilineal) and eventually into ‘modern’ states conforming to the European ideal. These three stages reappear in African Political Systems (1940), by Meyer Fortes and E. E. Evans-Pritchard, though without evolutionary implications and with an important modification to be discussed below. In Morgan's version from the nineteenth century, because monogamy did not exist at first, nobody knew who a child's parents were, but recognition of children's links to their mother gave rise to matrilineal descent. As the productivity of labor increased, fathers began to resent the ‘injustice’ of the rule that their own children could not inherit their property and therefore engineered a change to patrilineal descent. The continuity of the evolutionary tradition is illustrated by the fact that George Peter Murdock repeated Morgan's fantasy almost verbatim as his 1959 explanation of why the Nkundo Mongo group on the upper Congo River apparently adopted patrilineal descent while retaining kinship terms appropriate to matriliny.Footnote 2
Morgan was the first to discover that kinship terminologies systematically formed a limited number of patterns recognizable worldwide, and to suggest that these patterns corresponded to the social organization of those who used them. The different patterns are most easily distinguished by the terms used for cousins, although the terms applied to relatives in higher and lower generations may differ among examples of the same cousin pattern. Later, American anthropologists renamed and refined Morgan's patterns; a summary of those occurring in Africa is provided in Murdock's Africa.Footnote 3 Only two patterns need concern us here. First, Crow terminology, in which children of the mother's brother of the speaker (‘Ego’) of the same generation as the speaker (his matrilateral cross-cousins) are addressed as his or her ‘children’ (and thus are apparently ‘demoted’ one generation), whereas those of the speaker's father's sister (his patrilateral cross-cousins) are ‘promoted’ to ‘parents’. Second, Hawaiian terminology, in which all cousins, both cross and parallel, are equated with siblings. In decades of anthropological wrangling, the Crow pattern has been paired with another, called Omaha, which seems to be its reverse image. Identification of all these patterns depends on the ethnocentric and incorrect assumption that in any kinship system there is only one label for each relative, which can then be plotted on a genealogical diagram to reveal the pattern.
Anthropologists have long sought sociological determinants for the occurrence of this or that pattern. The most common arguments have identified inheritance, corporate group membership, co-residence, or marriage rules and preferences as the practices with the best predictive value. For each case, the correlation is high, but there are always reported exceptions. For example, 84 per cent of examples of Crow usage occur in societies with matrilineal descent groups, and 90 per cent of Omaha usage occurs in conjunction with patrilineal descent.Footnote 4 Vansina says that in the Congo basin these correlations generally do not hold; his preference, as we will see, is to look to marriage practices for the origin of Crow terminology in particular.Footnote 5
Historians have understandably been turned off by such inconclusive debates framed in arcane language, but should take notice of the historiographic effect of Morgan's lingering legacy.Footnote 6 It left us with the impression that kinship and descent were fundamentally about domestic relations and therefore outside the purview of historians, who were concerned with the political and, therefore, primarily with the state. African Political Systems attempted to break with the nineteenth-century tradition, and the interpretation of Morgan's work associated with it, by insisting that social organization by corporate unilineal descent groups (Morgan's ‘gentile organization’) constituted a political system of sorts, an ‘ordered anarchy’ that might even be superior to the state because it was less likely to turn into despotism.Footnote 7 But Fortes and Evans-Pritchard retained the presumption that descent organization and the state were exclusive forms, and presented segmentary lineage systems primarily as structural constraints on political action.Footnote 8 Vansina transcends this opposition, which had been refuted by later research; he also intimates that descent systems could be the result of the agency of human beings rather than bodies of rules providentially established to contain such agency. He explains that the structure of a descent group, including its genealogy (the supposed record of the origins of its constituent segments), is a normative model and a means of defending competitive gains rather than a binding set of rules or a description of lived reality.Footnote 9
THE METHOD AND ITS APPLICATION
Boldly venturing where no historian had dared to tread, Vansina reconstructs the early history of the Congo basin, in the absence of documents and with little assistance from archaeology, using the method of ‘words and things’ to associate ethnographic and linguistic data with events.Footnote 10 The history of the form of words can reveal something of the history of their meaning. The meaning of reconstructed ancestral words can be established by comparing the meanings of their modern reflexes; thus the distribution of the modern forms in related languages can indicate the historical process of their diffusion and change. Such deductions always require interpretation and a sense, which Vansina amply demonstrates, of the ways in which one practice or concept fits together with another in the life of a given people, how the rise of a political system may have been made possible by the invention of new weapons, or how its diffusion was encouraged by the prestige of certain emblems of authority. Life in the rainforests, Vansina argues, was always more than the stories of ‘gyrations and migrations’.Footnote 11 As in all historiography, the narrative is partly a work of the imagination.
Vansina properly deplores the traditional ethnographic inventory of the customs of a ‘tribe’ supposed to have existed forever in its present form, without regard to the fact that all local communities had been drastically altered by colonial conquest. Contemporary ethnographic comparison of similar features, unlike linguistic analysis, cannot reveal the historical relations among them and the dynamics of change. It is impossible in a summary to indicate the sophistication of Vansina's procedure, but we must note that the method works best when the words designate, for example, plants and tools; less well when they refer to social and cultural values. A wordlist or dictionary inevitably reflects not only the limitations of the research on which it was based but the preconceptions of a writer who may take down the meaning that makes the most obvious sense before hastening on to the next word.
For the south-western forest area, almost entirely north of the Congo and inhabited by speakers of the Kongo language in its many dialects, Vansina sets out the evolutionary sequence from the basic ancestral ‘House’ (c. 1000 CE), as ‘chiefdoms emerged’, ‘principalities appeared’, and ‘kingdoms arose’ (fifteenth century). His intention is to reveal the unique legacy of African political institutions, and how, unlike so many others, they maintained local authority even while enlarging the scale of society. Population density at this time was very low, land readily available.
According to Vansina, the ‘House’ was a unit of food production headed by the founder of a family or his successor, and including among its ten to forty members not only his relatives by blood and marriage but clients and other associates attracted by the prestige of the head, who competed for followers with other big men.Footnote 12 It is traditional in anthropology, as here, to speak of the ‘fiction’ of common kinship of all members as a metaphorical extension to others of terms proper to the nuclear family.Footnote 13 So the big man was ‘the father’ of all the others. Competition between Houses for members required astute management of choices among various contracts of marriage, of which some transferred authority over all or some of a woman's children to her husband's House, others reserved such rights to her House of origin. Linguistic evidence suggest that some contracts included the payment of marriage compensation or ‘bridewealth’.
From this base, some Houses in the south-west grew wealthy, perhaps then to be destroyed by succession struggles. ‘The pressure grew on leaders to develop at least some principles of succession and legitimacy’, though the ideological model of the family was retained.Footnote 14 The first sign of centralization is the appearance of the term nkani, ‘leader’, related to the verb kanu (sic; read kana), ‘to judge’. Once a chiefdom appeared, perhaps about 900 CE, its neighbors would imitate it for their own security. Chieftaincy then ‘acted as a catalyst which transformed basic institutions over a vast area’.Footnote 15
The next step was the conquest of one chiefdom by another and the creation of principalities, signaled by the emergence of new titles of authority and such terms as mbanza, ‘central place, capital’, and nlunga, ‘copper ring of office’. This development originally occurred twice, among Kongo and Teke, in slightly different forms, well before 1200 CE. In principalities, distinctions grew between those in authority and mere plebeians, though the House retained its image of a bilateral family.Footnote 16
THE INVENTION OF MATRILINEARITY
Marriage payments
Meanwhile, as Vansina puts it, as the first Kongo principalities were taking shape, a dramatic development took place there as part of this process: the invention of ‘the principle of matrilinearity, which was later to affect a great portion of central Africa’.Footnote 17 The invention took place in three stages: an increase in the relative value of marriageable women led to the sense that wife-givers ranked high; this sense gave rise to Crow terminology, setting the wife-givers apart as a matrilineal bloc; and this nominal category became a perpetual corporate group when, as the political hierarchy grew, the bloc developed a long-term interest in a title to which its members could aspire. Vansina, aware of ethnographic challenges to the three-part set of African Political Systems since the book's publication, overrides the traditional separation of descent groups from ‘government’ and makes the political system a nest within which lineages grew. But the data on which his hypothesis is based are very shaky.
It is true that in Belgian Congo, which incorporated the area in question, colonial rule drastically altered Kongo life; traditional trade, warfare, and chieftaincy were abolished, but not everything was equally affected. Social relations within and between lineages did not change much, nor did the vocabulary that describes them. Moreover, for at least part of the area we have a considerable body of information on precolonial (though not ancient) practices and usage, which suggests challenges to Vansina's narrative.Footnote 18
Vansina says that one source of the invention was the practice of paying matrimonial compensation, using goods that included iron and copper objects, and perhaps squares of raffia. The relative power of each House was a function of the size of its hoard of such things, which could also be used to pay fines and acquire dependents.
So vital were these goods to the health of the House that the Kongo and their neighbors in Gabon began to think that the goods exchanged for a woman in marriage were not really equivalent to a woman in value. The goods were worth more; therefore the group receiving the goods became the debtor of the group receiving the woman. This vision was reflected in a new type of kinship terminology, whereby the creditors downgraded the debtors by a generation and debtors upgraded creditors by one generation. This is the Crow type of terminology in which descendants of the mother's brother of the speaker of the same generation as the speaker are addressed as children of the speaker, and those of the father's sister of the speaker are parents'.Footnote 19
This folk explanation of the usage was offered to the world by the Nkundo Mongo of the upper Congo.Footnote 20
Throughout the south-west, there were many types of marriage, the choice among them guided by political considerations.Footnote 21 Though it is likely that in the early period some marriages required hoarded goods, it is clear that in nineteenth-century Kongo only marital alliances intended to be permanent called for such goods. In a highly competitive commercial environment, which increased in intensity from Malebo Pool west to Mayombe and the coast, one needed reliable partners. In the north, the Lemba association combined marriage, fertility, adjudication, and control of trade in a unique synthesis that became a regional system, an alliance network.Footnote 22 Initiation into the association came by way of a highly ritualized marriage in which valuable copper rings were exchanged. On the coast, another such association, Ngofo, used iron or ivory bracelets, and surely there were others. In Karl Laman's collection of indigenous texts from 1915, which refer mostly to central Kongo, permanent alliance was effected only by what he called longo marriage (from nlongo, ‘subject to ritual sanctions’). This required, in a report from eastern Mayombe (Lolo), a substantial payment in pigs to the heads of the bride's mother's and father's clans. At the time, pigs were also used in Lolo to pay fines; to buy titles, ritual treatments, and membership in other ritual associations; and in slavery transactions. For example, one pig each was paid to redeem children whose mother had become a slave. A longo marriage could be broken only by a return of the goods, even after the wife died. ‘Marriage’ with a woman who was already one's slave required no payment at all. After colonial rule had rendered such alliances pointless, all marriages were of the ordinary and frail type in which the parties to the contract were paid only ‘consumer goods’ (food, drink, cloth) whose function was simply to register the agreement.
Terminology
There is no evidence that wife-givers as such ranked high or that, in general, ‘women’ were reckoned as less valuable than the marriage goods; ranking is certainly not reflected in Kongo kinship terminology, which is not Crow at all. Kongo usage subsumes Radcliffe-Brown's two basic aggregative principles, ‘unity of the sibling group’ and ‘unity of the lineage’, into a much simpler and more general one in which any two persons who stand in the same relationship to another person or group are said to be ‘in the same place’, fulu kimosi, and are therefore treated as one.Footnote 23
By the ‘unity of the sibling group’, all of Ego's father's brothers and sisters are ‘father’ (se). Mother's siblings are likewise ‘mother’ (ngudi), and all cross-cousins, with whom Ego shares either a ‘father’ or a ‘mother’, are his ‘siblings’, mpangi. This fully bilateral system distinguishes relatives by generation and by the sex of the linking relative. Affines of the same generation are n'kwézi, from kwela, ’to marry’, or in alternating generations n'zitu, from zitisa, ‘to respect’. When plotted on a genealogical diagram the result of this kind of reckoning ‘individually’ (mu kimuntu) is read as Hawaiian. But Ego may also reckon mu dikanda (‘by clan’), seeing all the members of any clan or matrilineage as being ‘in one place’, and treating them as one person for the purpose of reckoning relationships.Footnote 24 So all of father's clan or lineage, of any generation, including all of Ego's patrilateral cross-cousins (his father's sister's children), can be called ‘father’ and all whose fathers belong to his own matriclan are his ‘children’. Ego therefore has a choice of terms to apply to his cross-cousins, depending on whether he chooses to reckon individually or by clans. Only if he chooses to reckon by clans will the terminology appear to be of the Crow type. If he reckons individually the result is called Hawaiian by anthropologists (unity of the sibling group), but only in the anthropological imagination are we dealing here with two different ‘systems’.
This usage, apparently both ‘Crow’ and ‘Hawaiian’, is not unique to Kongo. It is reported also from the Luapula valley, from the Plateau Tonga in Zambia, and for both the Ashanti and Fanti in Ghana.Footnote 25 Alternative terms for cross-cousins, implying a similar flexibility, are widely reported in Central and Southern Africa but actual usage has not received much attention and can only be inferred. The attribution of Crow terminology to the Kongo is the result of the hasty assumption that since they have matrilineal descent groups and may indeed call patrilateral cross-cousins ‘father’, there is no need to inquire further.Footnote 26 In fact, the ‘promotion’ or ‘demotion’ of cross-cousins simply recognizes relationship to a descent group and has nothing to do with the value placed on goods or women. It follows that what appears to be Crow terminology is a usage that came after and not before the emergence of matrilineal descent groups. Further discussion of the date when matrilinearity developed will appear below.
Moreover, in Kongo, terminological flexibility goes much beyond cousin terms. Ego can be both a member of his clan and one of its grandsons, reckoning relationships and using terms accordingly; if his father had been a slave of the same clan, Ego could also call himself its ‘child’, mwana. A man he calls ‘mother's brother’, ngwankàzi, is not necessarily a member of his matrilineage: if Ego's mother's father belonged to a given matriclan, his mother is ‘sibling’ to any man whose father also belonged to that clan, no matter how vague and distant the relationship; Ego may therefore refer to that man as his ‘mother's brother’, ngwankàzi.Footnote 27 Ego in fact has a choice of labels to apply to almost anybody, whether a ‘relative’ or not; as Kongo say, ‘it depends which route you are following’. His choice depends on the moral and political tone of the relationship it implies: respect, condescension, or friendly equality; exactly parallel usage in a ‘patrilineal’ society is reported by John Comaroff.Footnote 28
Jack Goody, discussing ‘cousin terms’, seems at ease with the possibility that any system may allow more than one term for a given individual, but makes nothing of it.Footnote 29 Most of the anthropology of kinship, on the other hand, has assumed, even insisted, that there can be only one term for a given relative.Footnote 30 Marvin Harris, for example, attempting to understand kinship in Moçambique, blamed his informants for what seemed to him to be confusion.Footnote 31 Anthropology could well abandon altogether the assumption that ‘kinship’ terms are primarily about family, with metaphorical or fictitious extensions to other relationships. Kinship terms, no doubt derived from the nuclear family, function as templates for conduct, to be implemented (even in the family) to the extent that individual motives and resources permit.Footnote 32 In many parts of Africa, the husband-wife pair, for example, is the template for certain relationships, regardless of gender. In that sense, all kinship terms should be regarded as ‘Legal Fictions’ rather than as records of biological links.Footnote 33 As Comaroff writes, ‘rules were constantly deployed to personal advantage’ rather than as prescriptions to be observed.Footnote 34
Matrilineages
The next phase of Vansina's argument is based on a serious linguistic error. He writes that having recognised the superiority of wife-givers and adopted Crow terminology, ‘the Kongo now called mother's brother ngwa nkazi, “my mother wife”, to stress the debt he owed’.Footnote 35 Relying on the Kongo dictionary of 1652, which fails to make the necessary discriminations, Vansina confuses nkàzi (class 9/10, low tone; a term for a person in authority, a lineage headman) with n'kázi (class 1/2, high tone), which Laman's dictionary translates as ‘wife’ but which can be used, like almost all kinship terms, indiscriminately for men and women: n'kázi ami, ‘my husband, my wife’.Footnote 36Ngwa nkàzi (ngudi a nkàzi) means ‘uncle’ or ‘authoritative mother’, where ngudi, ‘mother’, may refer to male as well as female ‘mothers’, as in ngudi za nsimba, ‘parents of twins’.Footnote 37 Presumably Vansina would not accept the translation ‘my child wife’ for the reciprocal of ngwa nkàzi, which is mwan'a nkàzi, ‘(male) child under authority’, or more simply, ‘nephew’, sister's son.Footnote 38 (Sister's daughter is busi, from buta, to engender.) Moreover, a man who had received a wife would refer to the wife-giver as n'zitu, not ngwa nkàzi. Nor did the term ngwa nkàzi, as Vansina says, replace ngudi a yakala, ‘male mother’, parallel to se dya n'kento, ‘female father’, which remains in use.
The (supposed) Crow system, Vansina believes, did not by itself produce matrilineages, but it set blocks of matrilineal relatives apart. Once the nominal category was registered in the collective imagination, it led to the creation of corporate matrilineal groups, but only after a need arose to safeguard rights to court titles, as principalities began to emerge. Vansina does not mention that matrilineal descent is necessarily exogamous and that therefore alliance patterns must have changed, but he writes that matrilinearity spread far and wide because women went to live with their husbands and took their matriclan names with them, continually generating incipient lineages in new villages.Footnote 39
Was this dispersal new? How far and wide? In practice, women marrying ‘out’ usually marry within walking distance of ‘home’, and in modern times the regulation of matrilineage affairs requires the constant participation of relatives from other lineages and villages.Footnote 40 Vansina leaves unclear the question of motive: why a group would adopt not only the kinship terms but the social organization of an in-marrying wife, as though it were a new dance. To belong to a matriclan (in practice, a localized matrilineal clan section) was not just a personal attribute; it meant being under the authority of a nkàzi. If he failed to live up to his responsibilities with respect to his dependents on the occasion of such family events as funerals, perhaps because they lived too far away, his authority lapsed; they became ‘slaves by default’ (sing. mwana fundu), or at the very least clients under the authority of the husband, not the foundation of a free branch of their original descent group.Footnote 41
Increasing dispersal, Vansina tells us, meant that lineages no longer knew just how they were related, but their ensemble took over an old name for House (kanda): ‘The creation of the position of head of the clan became inevitable, and it went to the head of the strongest known lineage.’ This development took place under the political shadow of the principality and represented ‘a defensive reflex for the autonomy of small groups against growing centralization’. Meanwhile, ‘chieftaincy now perforce rested on a claim of authority over territory alone, without any pretense of common descent’.Footnote 42 Just how these arrangements worked is obscure, as is the evidentiary basis for the story, though it is plausibly reminiscent of Maine's version of social evolution from kinship to territory.Footnote 43
KINGDOMS AROSE
Vansina continues on to explain the emergence of kingdoms. As a principality grew larger from success in war or trade, it might find itself unable to rule its lands successfully from a single central place. Instability ‘instigated the creation of yet another territorial level, the kingdom’, initially the Tio and Kongo kingdoms, in the fifteenth century.Footnote 44 This providential ‘instigation’, like the earlier ‘pressure to develop legitimate succession rules’ and the ‘inevitable’ creation of clan headship, is a sign of the residual functionalism that fills out key portions of the narrative. Vansina necessarily accepts what was once the received history of Kongo, conforming to the standard model of the state as territorial control derived from conquest, although it is well established that African kingdoms are characterized by a center rather than a periphery and control over people rather than territory. Vansina himself has pointed to the fashionable imitation and diffusion of prestigious emblems of power as a source of political influence, and acknowledges the role of ritual and metaphysical powers in chiefdoms and principalities, but seems to exclude such factors from his picture of kingdoms.Footnote 45 He believes that bilateral Houses served in the kingdom, as in the principalities, as political supports for the higher offices, though at the lower levels of social organization matrilineal descent was dominant: ‘It was only after the fall of the kingdom of Kongo that clan organization became dominant, as opposed to a bilateral House with rights to office.’Footnote 46
Despite abundant scholarly work on the Kongo kingdom, the function, even the existence, of matrilineal clans at an early date remains obscure and contested. It was once accepted that the kingdom was founded by Nimi a Lukeni (the name is not well established), a prince from north of the river. Supposedly, the new kingdom was supported by representatives of original matriclans. Anne Hilton's sketch of the kingdom's component ‘kanda chiefdoms’ was based on reading too much modern ethnography into the sixteenth-century situation. Like others, she took it for granted that kanda denotes matriclan, although John K. Thornton found that the word could also refer to any governing group or even to choirs of angels.Footnote 47 While asserting the primacy of the kanda in the political system, Hilton also documented the importance of patrilateral ties. She then argued that during the sixteenth century, the abundance of slave labor freed title-holders from their dependence on the kanda, weakened the matrilineal structure, and introduced patrifilial inheritance and the formation of patrilineal social categories as the essential political units. In the coastal province of Soyo, this development went so far as to create corporate patrilineal descent groups. In the late eighteenth century, as all central authority declined, the elite revived ‘the kanda ideology’ and matrilinearity returned.Footnote 48 This confusing account has the merit of revealing the essential bilaterality of Kongo social organization and its openness to political engineering, which the anthropological preoccupation with descent has obscured; but the idea that matrilinearity came and went but later returned is not persuasive.
By a meticulous examination of documentary sources, Thornton has rewritten the story of origin, showing that it was based on a forced synthesis of seventeenth-century reports and oral traditions current in modern, matrilineally organized villages, projected into the past.Footnote 49 The supposed conquest belongs in the corpus of African myths about kings ‘from elsewhere’; it is a structural inversion of the Luba founding myth. The earliest tradition recoverable from the documents indicates that the kingdom was formed by an alliance of neighboring principalities around a central core, with subsequent accretions. Although part of the aggrandizement of Kongo was effected by conquest, territorial claims were always supplemented by marital and political alliances, but there is no clear evidence that matrilineal clans existed at all.Footnote 50
Although it would be very surprising if people did not calculate their kinship relations in making marriages and alliances, Thornton finds no evidence that they did so in the early period, and he concludes that the modern matrilineal clans may date only from about 1850. I have myself argued that matrilineal descent arose in much of Central Africa when the demographic imbalance brought about by the slave trade in the early eighteenth century made it possible to accumulate slave wives relatively cheaply, richesses en épouses gratuites, as one scholar called it.Footnote 51 This development would have begun on the lower Congo, but not for the reasons Vansina suggests.
CONCLUSION
All social scientists work within a cultural tradition as rich and as committed to certain values and a certain worldview as is that of the rainforests, which Vansina has done so much to expound.Footnote 52 Evolutionary theory, arranging social forms on a scale from the exotic to the familiar, created gentile society as an imagined historical object anterior to the state. This was also a progress from the irrational and disorderly to the ‘modern’ and rational. Functionalist anthropology later invented patrilineal and matrilineal societies as necessary types, reluctantly recognizing after a while that they did not exhaust the variety of acephalous societies and that everybody has relatives on both sides of the family. While these changes to the tradition were taking place, assumed realities often served to corrupt the data of ethnographic and linguistic ‘observation’ that subsequent scholars have to work with; just as, once upon a time, the assumption of ‘primitive cannibalism’ generated reports of the practice, so those looking for ‘the word for matriclan’ found kanda. Our contemporary representations, unbeknownst to us, shape the analysis, confer significance on some facts rather than others, and assure some stories of ready acceptance. Only in the mid-1960s did historians of Europe itself abandon the assumption that change was normally introduced by superior invaders, when in the aftermath of the Second World War, a generation of scholars grew up who did not share the myths of race and the mystique of empire.Footnote 53 In Paths, Vansina's account of local innovations in the rainforests is a major shift from ‘the myth of the jungle’ and the historiography of imagined migrations. He also does much to advance our understanding of African polities, small and large, as arenas of political competition, but even he fails to acknowledge the extent to which African kings depended on ritual and a reputation for occult powers rather than on conquest and control of territory. In the previous article to which this is a ‘footnote’, I drew attention to some new directions, a new willingness in current historiography to rethink the political.Footnote 54
