A Politics Of Fear, A Religion Of Blame: A Comparative Study Of Kinga, Pangwa, And Nyakyusa Peoples In Southwestern Tanzania

A Politics of Fear, a Religion of Blame

A Comparative Study of Kinga, Pangwa, & Nyakyusa Peoples in Southwestern Tanzania

George Park © 2002

Communications: Dr. George Park: Prosit Anthropology 73 Lewis Point Rd, Charlottetown PE Canada C1E 1M6 Ph. (902) 368 2018 email [email protected]

TABLE OF CONTENTS Introduction 1 <1> The Matter of Authority Malatan politics 8 Protostates 15 Sanctions 17 Animism 18

<2> A Politics of Fear Concessions to fear 20 Malatan specialties 25 Four Kinga-Nyakyusa analogues 33 (1) The client-ruler strategy 34 In Kinga context 34 In Nyakyusa context 37 (2) A tempered aristocracy 43 In Kinga context 44 In Nyakyusa context 47 (3) Bachelor warriors—late marriage 51 In Kinga context 53 In Nyakyusa context 57 (4) Amity over kinship 59 In Kinga context 63 In Nyakyusa context 64

<3> Sowetan protostates in regional context Origins of the political archipelago 67 Acephalous communities in the regional culture 76 A premise of imparity 83

<4> A Religion of Blame Honour’s underworld 94 Uses of a bush religion: the social contract 100 Uses of a bush religion: the politics of fertility 104 Uses of a bush religion: ancestors & neighbours 113 Uses of a bush religion: embodied danger 126 Imaginary theatre 137 The politics of dishonour 147 Witch and state 159

i

<5> Priests and their Princes Histrionics of belief 168 Credos and credibility 177 Witchcraft, statecraft 180 Chiefly prince or princely chief? 187 Arena politics—’African despotism’ held in check? 200

<6> A Malatan Cosmos The mischievous ‘medium’: an extended case 210 Sifting premises 226 Konde/Kinga cosmology in perspective 236 Kinga/Konde uses of ritual 242 The viability of incongruous frames 258

<7> Taboo and Law The Everyman narrative 264 Picturing the Pangwa world 282 Performative and dramatic ritual 284 Intimacy and avoidance 286 Pangwa particulars: intimacy and avoidance 289 Konde particulars: intimacy and avoidance 298 Kinga particulars: intimacy and avoidance 306

<8> Given Terms of Existence About the deepstuff in history 319 Deep divergency: Kinga/Pangwa 322 Pangwa particulars: life cycle and life style 324 (a) Sharing 326 (b) Sexuality 336 (c) Gender-peer relations 344 (d) Sanctioning rights 350

<9> Kinga/Konde: Logics of Political Ascendancy What marriage implies 355 Logics of the age-village system 360 Logics of autonomy 367 A Profusion of Priests 376

ii

<10> Implications About regional cultures 380 How are taboos learnt? 381 Fear, anger, blame 388 Blame, taboo, and the problem of order 394 What grounds social identity? 400 Least polities 403 Logics of the Least Polity Rule 408 The management of intimacy 414

Source notes 431

References 451

Index [461]

iii

Introduction

A Politics of Amity Kinga use surnames but don’t usually suppose a surname will tell much about a stranger’s social identity. That is not because too many individuals bear the same patronymic (although in the case of the surname ‘Sanga’ they may seem to do) but because kinship by itself says little about where a person stands on the social map. Dialect and residential experience say more. But since loyalties are more dependent on amity than kinship claims, questions of social identity are easy to leave open-ended. In the colonial era (six decades ending with Independence in 1961) schooling, office-holding, and migrant labour brought in new career patterns. A man’s personal history was often written in his clothes, speech, and bearing. But in earlier times a man hoping to carry personal status credits with him into a new community would have been dependent on making or finding friends there. The important social boundaries marked political loyalties, not local descent groups. While surnames were passed on patrilineally by rule, they could sometimes be politically bestowed or self-bestowed for political reasons. Nomenclature distinguished ego’s kin by generation and gender, and cousin terms were behaviourally coded. But beyond the immediate family, kin ties scarcely organized Ego’s deepest loyalties. Exogamic rules were bilateral. If you discovered you shared a great-grandparent with your intended, your marriage plans were forfeit; and sexual dalliance followed the same rule, making a heterosexual incest taboo and exogamy one institution. In ‘kinship societies’ properly so-called, the scope of an exogamic taboo generally goes well beyond local definitions of incest. Where ‘incest’ is bilaterally extended, it goes well beyond the sphere of intimacy. A Kinga man in a new area may spend the better part of a morning trying to establish a putative kinship link with a stranger, but the satisfaction of success in that endeavour is not in knowing what mutual privileges they may enjoy but in the mutual good will 1

established in making the effort. The talk is for making friends. From four or five, after weaning, a child of either gender will be domiciled with age peers, apart from his mother’s house. The little cabins built and managed by the boys or girls of the hamlet were spatially set apart in whatever manner the self-builders chose. At herdboy age and wanting no further company but their goats, youths were only expected to build (for safety) within hailing distance. In old times and often enough more recently a father would spend only short periods in middle life sharing the house he built for his wife, since conjugality was for reproduction first and (quite often) last. The family system as a whole was not focused on close emotional ties, although dyadic bonding was stressed. An elder should love and shelter a (same sex) younger sibling into early adulthood, and in later childhood a girl will be as close to her mother as to her older sister or best friend—though certainly less open. Significantly, Kinga choose their own personal names and may add or switch to new as they please. I say ‘significantly’ because the custom, which I observed particularly in boys and young bachelor men, accords with a tendency to relegate family origins to background in the way reciprocal social perceptions are managed in community life. Unlike the girls, boys eat from many fires. Personal qualities, even though seldom sanctioned by open criticism, matter very much and continue to matter throughout a person’s moral career. I found boys and girls especially aware of that, both in themselves and in friends. Blind solidarity, whether with kinsmen or of the factional sort, was not in evidence. On the other hand, should a man stand in peril and witnesses dare not intervene they would know to call his next brother: not because kinship in itself would lend him strength but because no dyadic relationship was thought to be closer than the love of older and younger brothers. Where amity is the basis of most important social ties, it must be appropriately sanctioned. Both Kinga and their Nyakyusaspeaking neighbours managed this through new political institutions. They developed statelike systems of authority, establishing fighting forces to protect their power, and effective court procedures for handling internal trouble. To establish and fortify a position of authority in an individualistic society, no Sanga or Nyakyusa prince could afford to fraternize. This states the essence of their political systems, which could marry a culture of egalitarianism with a politics of fear. While his court must never fail to attract adventurous youth, and always be hospitable enough to keep them in service as long as they were fit, a major court had to be disciplined by the presence of revealed power—power dramatically revealed in life-giving ritual and in the uniquely sacral person of an anointed 2

prince. The moral architecture of the court had to satisfy the human equations entailed in a new pattern of careers for men, as well as all the new requirements of subsistence and social reproduction which the building of a protostate brought with it. Placed spatially between the Pangwa, who remained stateless, and Nyakyusa boasting a particularly vigorous and trickily balanced politics, the Kinga are structural cousins of both. The circumstance invites a regional study of social transformation. I know of no better case for exploring the uses of kinship and amity in the elementary game of building power from scratch. The special character of Sanga protostate politics is hard to grasp until one looks at the typical moral career of the barracksman—the long bachelor period in a big-men’s circle at court, eventual late marriage and re-rustication, the continuing salience of amity on a par with near-kinship in the domestic constitution. The Sanga recruitment system drew ambitious youths from the bush communities of a local domain, trained them as ‘king’s men’ for the whole of their young adulthood, then sent them back out, at last with a wife, to homestead some new place in the bush. In this way it became the task of a small sodality of court-educated settlers (now bearing the earned surname Sanga) to dwell, clear land, take charge of any bush hamlets in the area, and generally secure the place for the crown. A lay apostolate was in this way kept busy expanding Sanga rule; and it was this system of meritocratic circulation which required putting amity ahead of kinship as a grassroots basis for social alignment among men. But extended bachelorhood for both genders carried with it the implication of same-sex intimacy extending from late childhood to early middle age, since the female career so far as it was tied into court life was paced to be compatible with the male. The private sphere of life came so to be shaped to suit the special demands of the public sphere. The personal experience and outlook of Kinga men and women took on a distinctive caste only partially matched by their neighbours. The importance of amity for Kinga is only matched by the Nyakyusa-speakers in the Great Rift Valley (below the eastern escarpment from Kingaland) who segregate boys and young men off from a parental village in bachelor settlements of their own. But while the Nyakyusa ‘age village’ bears a clear family resemblance to the Kinga youth pattern, the eventual career patterns of Nyakyusa males and females are shaped to a different ethos in the public sphere. The lifetime labour of a Kinga woman from early girlhood is in cultivation and food preparation, whereas the Nyakyusa woman leans heavily on boys and men for work in the fields as well as herding. This Rift Valley division of labour suits a highly organized, centripetal 3

village system. There is no matching settlement in Kinga country. The mountain slopes are hard to tame, supporting rather a centrifugal pattern of scattered hamlets, and indeed a centrifugal daily life. Women gather in the dawn along the paths by appointment, heading to faraway gardens as if to devotions. Men in pursuit of their multifarious errands greet or join friends more casually—the whereabouts of a man are never easy to predict. In short, space is organized not by clustering so much as by networking. The concentrations of settlement are relatively small even at the great central courts. Still, looking beyond the Nyakyusa claim that all their princes originated in Kinga lines, what the two political systems do share is much: there is a common pattern in the structure of power, balanced as it is between secular-but-dangerous princes (the ‘chiefs’ of colonial times) and unheroic priests. These last are the people’s established intermediaries with those punishing divinities—throughout this region mythically conceived as belonging underground with the ancestors but as beings of extraordinary scope and sweeping powers over nature—and royal ancestors to whom political leaders and troubled commoner groups regularly or irregularly turn with grave offerings. Looking at this in the frame of a ‘regional culture of rules’ the difference between the two systems is at the level of corollaries not principles; but in a narrative frame the business day of a typical Nyakyusa prince is quite different to the one you might observe at a Sanga court. The show there was run by the trustee group of courtiers at each capital, acting as a team coordinated with respect to many particulars by an inner coterie of priests. By contrast, Nyakyusa priests were chronically divided against one another, and practical decisions of state were scarcely theirs to take. In the crosstalk during shared rituals at Lubaga, Kinga and Selya (Nyakyusa) priests were sensitive to differences of setting—Kinga could remark on a prevalence of witches in their hosts’ country—but found themselves deeply agreed on their cosmic problematic.

Family resemblances: Kinga-Konde Whereas the differences between Kinga and Pangwa can be dealt with by adding bits here and taking away there (such that Pangwa becomes Kinga sans Sanga, i.e. a ‘bush Kinga’ variant), KingaNyakyusa relatedness is deeply structural—to demonstrate it you must turn to the kind of transformational logic which can deal with ‘family resemblance’. One example is what I call the ‘antipolitan ethic’, the idea of conditional loyalty to the ruler and his (always personal) authority. For Kinga it is the stuff of folklore to unload blame on the ruler yet withhold any tolerance of mutiny—there is an unspoken 4

license to escape authority by moving on to a new place and new loyalties. For the Nyakyusa of Selya the initial situation of a malcontent is similar, and the outcome much the same. But contrary to Kinga practice, the process can take the form of banishment, as if it were unthinkable for an individual voluntarily to leave his fellow villagers. Officially, he goes only to escape the stigma of witchcraft. Implicitly (and only implicitly) that applies also to a Kinga malcontent, but that difference in sanctions is not superficial. Another family resemblance is to be found in the procedures followed by the two societies in recruiting and training young men for raiding and warfare. Kinga rely on wargames at court, Nyakyusa open the real political arena to warlike competition among young leadership candidates for cattle, wives, and followings—the celebrated ubusooka or ‘Coming Out’ process which must precede and eventually legitimate any princely succession. Every princely domain in Nyakyusaland calls the ubusooka in its own time, inaugurating a period of aggressive cattle raiding on its neighbours. Actual military schooling activities are more comparable to the Sanga schools, comprising initially far-flung raiding parties, later on the settling of scores through token battles staged as lively confrontations in full parade dress. It is the organizing ideas, the underlying plot not the warlike action itself, which is peculiar to each cultural community. When we turn to the character of a fully established prince the case for a clear difference is harder to make. Anecdotal evidence from the Western Sanga realm leaves no doubt a prince could be as darkly charismatic as any in the documented history of the Sowetan region. The two Kyelelos, father and son, who were both known to the Germans, seem both to have earned the surname mkali ‘the cruel’ or ‘fierce’. They were magnetic leaders, quite in the best Nyakyusa style, and believed to own the special supernatural gifts of a tyrant. Informants can’t always tell you which Kyelelo was the hero of which tale. But folklore in the same realm still recalled a princely court of a contrary style, where the incumbent was secluded from ordinary secular contacts, only partly for his safety from enemies. The peculiar dangers pertaining to his status made him dangerous in a godlike way to the mundane life around him. Lwembe, the divinity of Sanga blood propitiated by Kinga and Nyakyusa in common, was thought by Kinga to harbour just such powerful anger against his descendants. How far Nyakyusa princes, in finally undergoing the unnatural death of a ‘divine king’, shared in that quality and handed it on to a successor remains an open question. The death of the High Prince of the Kinga was of a distinctive kind, befitting the inherent danger of his carnal presence. For a month his realm lay in a leaderless, liminal state. But in both societies it was the priests 5

who conveyed powers proper to the throne from its last incumbent to its new. In both cases though by quite different dramatic means the political theatre of the priests reinformed their world of its one source of stability. These are the topics of this volume. It is hard to ignore evidence that the two protostates above and below the Great Rift escarpment developed interdependently over several centuries. Doubtless, the two bush cultures in place in a.d. 1600 were more distinct than the court cultures became, as they developed on convergent paths later on. But the importance of amity vis à vis kinship is deeply characteristic of both cultures and special to each. I find the political lessons of Sanga courts achieve a more credible time depth when they are linked to the Nyakyusa, and vice versa. What is shared and what isn’t shared of myth and cosmology needs to be discussed before much sense can be made of either political world. The main evidence for the beliefs must be the snatches we have of their ritual theatre, and those comprise rather less than a single and coherent system for either of these two court-centred peoples. Ultimately I think we must conclude that here, as so often elsewhere, the cultural heritage we tend to group under the heading of religion can be the most eclectic, imaginative, and prone to contradiction of all the departments of a culture. But where religion and politics (or read ‘social function’) do connive we have found an order of social reality we can know something about, and we owe ourselves the chance to explore it. At this remove in time the light I can hope to throw will not dispel all mist. But I like what I’ve found. It has more general significance than its local scale would suggest. Kinga occupied high mountain slopes in the Southern Highlands of Tanzania, tilling intensively and herding (mainly goats) in broken country looking down from a mile higher on the northern shores of Lake Malawi. As the particular boundaries drawn by outsiders on modern political maps offer slight clues to native history or demography, I find it convenient to draw ‘Kinga-centred’ lines of my own. I refer to the country they share historically with a handful of neighbour peoples in Tanzania (eastward in Iringa and westward in Mbeya districts) as the Sowetan region. It forms a good part of today’s Southwest Tanzania but, considered as a culture region, must spill over the modern nation-state boundary to include the Ngonde district of northern Malawi. This is because the people of that district belong ethnically with Nyakyusa-speakers of Tanzania, who share important religious observances with the Kinga, as well as comparable political institutions. The shape of this book, with its concentration on Nyakyusa and Pangwa comparisons, owes much to the excellence of the ethnographies available on those two peoples. 6

As the coherent part of the Sowetan region I mean to explore includes the three major ethnic groups hugging the north end of Lake Malawi, and since for the project in hand this part-region wants a name, I’ll call it the Malatan. Both these neological names properly reflect place. But where these bearings are ‘modern and political’ the rationale for this book is placing the Kinga where they belong in a much older history. Kinga are linked politically westward with the Nyakyusa-Ngonde communities, and historically eastward with Pangwa-Bena-Hehe communities. This agrees as well with Kinga oral testimony about migratory origins, which always posits a westward movement and claims to have fostered chiefly rule down in the valley and lakeshore by sending political emissaries that way to teach and colonize—and aggrandize themselves beyond all expectation. Everything about the traditions of a ‘westward movement’ takes its colour from heroic myths. But they hardly touch the real history of population movements which brought Bantu settlers into the high valleys of the Livingstone mountains in the first millennium of the present era, and by quite another movement brought Bantu into the Rift Valley. These two streams must have initiated the histories of, respectively, the Kinga and Konde. The name Konde is the broad tag Europeans first used to designate the whole ‘Nyakyusaspeaking’ people. It can be taken retrospectively to identify the acephalous population ancestral to today’s Ndali and Nyakyusa speakers. We know little in detail about their culture. But for the Kinga before the Sanga ascendancy I have the advantage of a detailed ethnography of a ‘tribe without rulers’, the Pangwa, in the highlands just bordering on the Kinga Eastern realm. In consideration of the detailed affinities between the two peoples, I dare let the Pangwa case stand-in for the earlier condition of the Kinga. Linguistically, Pangwa are to Kinga what Ndali are to Nyakyusa. The main exercise consists in ‘subtracting’ Sanga chiefly institutions from Kinga culture, then asking how well it compares with Pangwa. Something more may come eventually from archeology, concerning dates and artifact distributions. Today’s retrospect has to draw on the subtle but voluminous evidence of ethnography, padded with bits of early European documentation, and probabilistic reconstruction. As I must assume few readers will have access to the two monographs (in German) of Fr. Hans Stirnimann on the Pangwa, I have taken the liberty of translating key passages. This enables me to discuss (in outline, in Chapter Six) the putative ethnohistory of a proto-Kinga community which evidently lived by such rules as Fr. Stirnimann records, but which was politically transformed in a fashion which the known facts now allow us to grasp in principle.

7

FEAR & BLAME, CHAPTER ONE

The Matter of Authority Malatan politics In the Malatan region, comprising a crescent around the north shores of Lake Malawi, our best chance for a clear view of a precolonial world is in Nyakyusa country. This people occupied, under a hundred-odd ‘chiefs’, the floor of the Rift Valley. We have detailed knowledge specifically of the chiefdoms in and adjacent to the Selya area where Monica and Godfrey Wilson did their fieldwork in the thirties, and where the earliest missionaries left some lively records of contacts even before the German military government was installed on the lakeshore. The Wilsons worked on the classic supposition that there was one Nyakyusa-Ngonde cultural community which had lately been given certain local twists in response to special pressures. In effect, they set aside the possibility that the extended Nyakyusa language area could have developed through the mutual assimilation of various bush cultures in situ, such that cultural roots throughout the region might still feed on diverse traditions. I make a somewhat contrary judgement to the Wilsons’. In this book my premise is that protostate politics and assimilative social policies account for the spread of one language over the diverse natural settings of the Rift Valley country at the north end of Lake Malawi. I give weight to the linguistic kinship between Nyakyusa and the Bemba, a Western Bantu people of equivalent population occupying a high savannah forest region west of the Rift. Here is Audrey Richards on the people she studied in the 1930s: The Bemba occupy the high plateau land of [Zambia] stretching from Lake Bangweulu on the west to the [Malawi] border on the east, their empire formerly including the territory between the four big lakes, Tanganyika, Nyasa [Malawi], Bangweulu, and Mweru. The tribe numbers now roughly 150,000, sparsely distributed over an area roughly equivalent to the size of Scotland and Wales, giving a population of about 3.95 per square mile. (1951:164)

8

The corresponding density for Nyakyusa-speakers was over 50 persons per square mile at the time referred to. It is my premise that the linguistic connection dates to well before the storied founding of royal lines in uNgonde and uNkyakyusa by migrants from uKinga. Those traditions indicate an eastern rather than western migration source and place the beginnings of state-building at about a.d. 1600. But what is indicated prior to that was a long period of eastward migratory movements into or through the Corridor. If only a fraction of the evident Bemba-to-Nyakyusa rise in population density had occurred by 1600, we have to suppose that quite new forms of settlement and social organization would have been prompted by the overall fertility of the Valley floor, and the migratory traffic there. Monica Wilson properly paints the lands of the Nyakyusa and their Valley neighbours as an important northsouth Corridor.‡ Since my focus will be on longterm political processes in the region, and since the continuous development of political society in the Corridor after 1600 is mirrored in uKinga and the Sowetan political archipelago more generally, we are left with as much as four centuries of what was at least in its later stages armed history. It was not a history of conquest, though, but of sparring for advantage among small local polities, acting increasingly within a framework of alliance and mutual respect which allowed for the practice of intensive agriculture and peaceful community life. The Nyakyusa political system was idiosyncratic and obviously homegrown, under conditions of ‘recruitment expansion’ comparable to the well-known case of the Nuer in the wholly different ecological setting of the southern Sudan. Success in recruiting cattle, women, and fighting men to your ranks gives you an accelerating rate of apparent expansion in a given ecozone, at the (demographic) expense of your neighbours. From the viewpoint of anthropology, and specifically of ethnohistorical studies, the special significance of the Malatan region is the scope it allows for understanding political growth from egalitarian, acephalous political beginnings toward statelike structures. We have this advantage because our information begins while the regional story is in-process, allowing us close to the springs of change. For the Kinga, my fieldwork is supplemented in several pertinent ways by official colonial documents and German missionary reports from the first decade of occupation. Most of this is too superficial or particularistic to build a clear picture of precolonial politics, but it gave me eye-witness accounts which have helped a fieldworker’s mind to form pictures of places and events of a sort to stir recognition after half a century. In particular an 9

occasional remark about quarrels among local rulers in the Central and Western realms confirms my suspicion that political life was livelier than old men recalled—or than they were prepared to share. I was not so fortunate as to have any of this documentary information to work with in the field. Like most anthropologists of my generation I entered and left with virtually no valid historical information on the region beyond what I could get by ear. This volume attempts, as my earlier volumes could not, to deal not just with oral history but prehistory, reconstructing probable paths of change from localist to translocal politics in three historically related communities. What makes this project feasible is the wealth of documented evidence from uNyakyusa, most particularly the published work of Monica Wilson. There is less direct sociological work to consult on uNgonde, and although Owen Kalinga has dealt in detail with the oral history I have no keen, close-up sense for the details of stylistic and structural differences which may have arisen between the two major Nyakyusa-speaking communities. Still, that the constitutional difference is profound is clear and interesting in itself, particularly in view of its rather recent (if precolonial) development. The fact that something describable as a close-knit ‘theocracy’ arose from what elsewhere in the language area remained a turbulent rivalry amongst innumerable localized ‘chiefdoms’ forces us to recognize how near the whole region had come to owning the basic institutions of statehood. The Wilsons’ ethnography has been supplemented since my field years (1961-3) by some historical and critical work, most recently by Peter Weber’s fieldwork on their ritual life. This has helped clarify conditions in the Malawi-Tanganyika Corridor in the several centuries before the pax. We have skeletal reconstructions of the major migrations comprising the great Bantu expansion which began two millennia ago, re-peopling East, Central, and much of South Africa. It was a hugely successful human expansion on a continental scale entailing the sustained mobility and mixing of peoples, evoking deep currents of cultural creativity. It is apparent that continuous assimilation of diverse parties of Eastern (and some Western) Bantu herders and cultivators occurred on-scene in the heart of the Corridor region to form the proto-Nyakyusa community. The evidence suggests ages of largely peaceful mixing by small groups, some always on the move in search of space, wealth, and secure alliances in an unusually verdant, uncrowded region. All this would be taking place from Early Iron Age times, as the technics and practical knowledge associated with the main cultigens and livestock in the ecozone were being honed, and the land tamed to the new agriculture 10

of the Later Iron Age, after about a.d. 1400. By that date banana cultivars of appropriate type would have been well established in the Valley zones, inviting a growing population density, and out-migration from this Rift Valley community would have slowed in favour of more intensive agricultural practices. In all this, smithing would have played its part, first with iron tools for clearing and turning soil, and eventually with the iron weapons a developing chiefly system of authority could put in the field. Nyakyusa lacked direct access to iron but specialists in reforging could turn worn tools into weapons; Kinga were the major suppliers to some parts of uNyakyusa. They stood fairly well out of harm’s way in their high mountains but were scarcely isolated. Trade enjoyed important sanctions—we hear much of cattle rustling but little of the hi-jacking of a tradesman’s goods. It is likely the human and livestock numbers in the Rift Valley increased steadily before and particularly during the eighteenth and nineteenth centuries, toward the density levels of historic times. As this was happening I take it the distinctive institutions for which Nyakyusa are known took form, since these seem to comprise the developing structural basis for protostate politics, not the heritage of an ancient Valley or small inmigrant culture. I further suppose the unsettled, fissive, yet combinative nature of chiefly politics in the region is a fundamental, not an accidental or dispensable feature of the new politics. That is, rivalry is a prime virtue of the system. The reasons for positing an ‘assimilative’ history at the north end of Lake Malawi are several. There is the ‘corridor effect’ which Monica Wilson first pointed out to us. Terrains, microclimates, and major watercourses conspired to funnel migrations through this region: This wild belt of country, stretching from high and healthy uplands to the stifling lake-shore plains, therefore lies in the very heart of Africa. It has been a meeting point of invasions from East and West, and North and South. [1958: 1]

Probably the next-most important reason is the relative importance of amity over kinship as a building strategy in Nyakyusa social organization. This feature is shared with the Kinga and all their neighbours in the highlands lying eastward, and to a greater or lesser degree with the locally-organized peoples in highlands to the north and west. A much stronger emphasis on kinship would suggest a ‘power migration’ pattern instead of the migratory drift I assume, which has a fragmenting and recombinative effect. Associated with amity is a minimal bilateral exogamic rule. Second cousins will not 11

marry if the relationship is known. Beyond that point, kinship is no hindrance and bears no onus of mutual obligation. Only special patronymics indicate a putative common descent, stemming presumably from a small group of inmigrants uncounted generations back. The exchange of wealth in cattle for brides is not associated here with any definite pattern of exchange-alliance between local and kin groups. The stabilization of lineage groupings promoted by social contracts of such kind would have worked against the radically voluntaristic trend of the Nyakyusa social system, which convenes the scattered men of a lineage only for the funeral of a member and disposition of the estate. There are also softer indicators supporting an assimilative history. Delayed marriage for men is facilitated by an easy tolerance of homophilia and the instant friendships which tend to come with it. The traditional rite of passage for girls at the approach of marriage was employed to introduce a local group of boys and girls to an omnidirectional heterosexual life by secluding them, away from everyday contacts and duties, in the small isaka (the sleeping hut traditionally shared by a local group of girls) for an extended period of sacralized eroticism. Along with this celebration of not-for-keeps healthy youthful narcissism goes a love of individual display, evident in the theatrical war-pattern and eroticized dancing at funerals. These and the much-remarked Nyakyusa love of personal cleanliness and good housekeeping reflect in a similar way the omnidirectionality expected of youth. Strict kinship societies can’t afford such license.‡ Peter Weber has shown that a cult centre’s (quasi-territorial) sphere of influence served as a prime basis of political identity. That is, congregation conferred attributes of citizenship. The connection is that any inmigrant’s destiny, once he was settled in, was thought to depend on the success of local rituals. It was not expected to be his own ancestors but locally ascendant ghosts who might be putting demands on him. So, in effect, descent is by-passed, and assimilation involuntary. But that also means you won’t be long a stranger wherever else you may move. And it means that the heightened in-group ethic—Monica Wilson’s ‘good company’—of a peer village of men is a mute sign of the frangibility of the peer bonds it celebrates. The balance of amity and kinship is different for each of the three Malatan cultures but in turn separates them as a group from their neighbours. The license to move about in social space is a fundamental attribute of personal existence here and an ethical claim of voluntarism in a person’s political allegiance. This ethic is thus anti-political, and will be referred to as ‘the antipolitan ethic’. ‡

12

Where a special value is put on amity, individual moral careers are pursued within looser institutional limits and the limits which do apply may not be grandfathered on the young to the extent one might expect in a traditional African lifestyle. At the same time there is a gender dimension. Freedom of association may be far more tangible for males than females, particularly in uNyakyusa. The political life and active concern with the definition of rights through court procedure is, as so often elsewhere, mainly for men. Women can usually rescue a good measure of personal freedom by developing moral strategies which maximize the rewards of a ‘woman’s world’. Monica Wilson has shown how difficult this can be for a Nyakyusa matron, who must follow strict rules of submissiveness. I have found the Kinga woman’s world an effective refuge. It is made possible by an extended bachelorhood in the high style of youth, and after marriage a triadic devotion to food production, sisterly amity, and the extended nursing of infants (limiting pregnancies to four) followed after midlife by domestic privacy. But again as so often the world over, a discussion of political process can be tied up without more than a whisper about a woman’s role in it. In this respect, I admit Malatan politics is like any other. A man’s world joins a woman’s at evening meal, in the conjugal act, and in the ungovi. This last is the festive breaking of ground for new planting, in which a women’s team vies with a men’s, and all join in a lively beer party in the evening. These are the main episodes in which men and women interact on a common footing. For the rest, their spheres intersect without joining, and each man or woman must pursue the good in life independently. My analytic strategy is to use comparative gambits to achieve what straight ethnographic description can’t. Perhaps what this comes down to is that unless you have a clear notion as to what you didn’t find—but might credibly have been expecting to—you don’t know as much as you need about the activities you observed. One dimension of your ignorance will be narrative: you have discovered an interesting standard practice (x) but are unable to put it in a probabilistic time series (xt1....xtn) to indicate its likely origin or future. The experience is like a child’s who falls in love with a tune but can’t place it in a genre. Critics earn their keep by writing liner notes to help adults know what they have just heard. Another dimension will be morphological, putting (x) in a line-up of parallel institutions found in similar/dissimilar societies featuring what in our usual fuzzy logic we regard as ‘the same institution’—e.g. polyandry, age-sets, rank, law, sacrifice, and the like. Getting somewhere by working on a time series will lead predictably to getting along on pertinent morphological series as well, and vice versa. In short, one is able to work one’s way into a kind of historical depth and semiotic breadth of under13

standing one could not otherwise have reached. If this seems a mere form of words now, I trust it will take on substance as I proceed. What Fred Eggan called ‘controlled comparison within a region’ is key to studies of this kind. Where the comparative method is under discussion, ‘going global’ is often a disservice. Soup may have something to do with nuts but not much. Comparing peanuts to peas and garbanzos, or soups to juices and infusions brings you closer to the real world. The strategy is essentially qualitative, only quantitative when a qualitative gain is in sight. I prefer this distinction to the science/humanities dichotomy for quite a number of reasons. One anthropologist who has managed being ‘global’ without ceasing to be ‘qualitative’ is Claude Levi-Strauss, but I have no idea where to place his work on a ‘science—art’ continuum. A global structuralist of quite another stripe was George Peter Murdock. At once masterfully quantitative and singularly qualitative, he might be scorned in both camps, hard and soft. My sense for these two models is that, though they may too often have had to shorten Procrustes to fit a bed they’ve made, you will get a better sense for the real world of ethnological study by reading them both once than you can expect from reading either one of them twice. But if you want a sense for that world which will stay with you, you should turn from reading ‘global’ stuff to the particulars of a fairly coherent ethnographic region—as you turn in this case to the Malatan.‡ My special subject is the rise of political authority from a type of social system which can manage well enough without it. Unless you are willing to call a pecking order a political system, the field of the political is scarcely attributable to any non-human world. It appears gradually in the history of human kind, as ‘bossing around’ in small groups and eventually as systematic control of collective action in an effort to achieve ‘policy’. A large, well-controlled family is an effective political community in that some members are sacrificing private potentials to the benefit of ‘policy’—that is, the set of rules which makes the family ‘well-controlled’. The traditional Kinga family was hardly a political unit at all, though some dyads within it would qualify. ‘Authority’ is always part of the mix, but not necessarily a defining part: for instance, a mother and her young daughter may work together remarkably well without any ‘bossy’ episodes. Two boys may be enamoured and get on splendidly without evincing inequality in making decisions for set acts. ‘Authority’ develops where such inequality reigns openly, and ‘political authority’ begins where bossing is a transitive phenomenon—A controls C through B. Let us assume this is a nearly universal feature of groups of triadic scope and more. But let us simply say our interest here is in big and

14

obvious examples, such as can be identified without resort to splitting hairs. I’ll be using the terms ‘protostate’, ‘sanctions’, and ‘animism’ in special senses, discussed in the following sections.

Protostates A series of interesting models have been used by anthropologists to explore the formation and working of statelike African polities. The best models treat tyrannical systems as artifacts of contact with extra-territorial threat and draw attention to ‘pyramidal’ features in the way translocal structures may form in face of what we may call perennial or ‘resident’ threat, distinguishing that from traumatic encounter. Aidan Southall’s revised statement on the ‘segmentary state’ is offered ...to define the segmentary state as one in which the spheres of ritual suzerainty and political sovereignty do not coincide. The former extends widely toward a flexible, changing periphery. The latter is confined to the central, core domain (1988: 52).

The Malatan examples discussed here are (excepting perhaps the Ngonde) unfinished by this yardstick. Where their ‘segmentary’ structures are clear, the notion of ‘suzerainty’ suggests a hand-inglove relation between a central throne and a centralized priesthood. This might be confirmed in one version of the Sanga system in uKinga—that is, the four realms as seen from the Central court at Ukwama. But it would be rejected with feeling by priest or prince elsewhere. The makings of a ‘suzerainty’ are not there. Kinga priests were a brotherhood, and a jealous one, with quite variant ideas as to who might be primus inter pares. As for uNyakyusa, where princes abound but there’s none to claim headship among them, the same lack of fit to the ideal model is present in spades. As will appear, the Kinga have compromised their animism more radically than the Nyakyusa have, though the two societies share at least one established ritual centre in Lubaga, the seat of Monica Wilson’s ‘divine king’ Lwembe. As for central direction among priests, the Kinga could claim to have more of it than the Nyakyusa. The truth seems to be that while neither group found agreement easy, Kinga were prepared to work harder at it. The key to understanding either of these protostates is that they know only one level of boundary. There is no relationship of political incorporation. Each local domain (‘chiefdom’ of pedestrian scale), including what may be one of the acknowledged 15

prime domains, is politically self-conceived as an island to itself. In uKinga it has been convenient for me as an observer to use the word ‘realm’ to mean the inclusive alliance of domains centred on a princely ruler, as if it were thought of as a political unity. My reasoning is that it was such in practice, if not in the language of discourse, since it is unthinkable a satellite would fail to answer the call of a prince to fight real external threat. But in ordinary speech there is no such lumping. Each domain is known by its name, and terms like ‘satellite’ or even ‘central/ peripheral’ are not terms of native discourse. The ties which create a Sanga realm are radially drawn between one or more local rulers who agree to claim only the title and privileges accorded an untwa (my ‘local ruler’ or ‘lord’) and a pivotal domain whose ruler is accorded the title and privileges of the higher office, unkuludeva (my ‘prince’ or ‘high prince’). In the retrospect of the 1960s there were ‘always’ four avakuludeva and hence in my terms four realms, albeit one of them lately in a state of civil war. But Kinga elders had no need for separate terms other than the untwa/ unkuludeva distinction of rank. This was a difference that made a difference, and was jealously guarded by a ruling prince. Ambiguity never arose because the political universe was small and familiar to everyone. Every domain was called by a proper place name and every ruler called by the personal name he had chosen for himself. This applied even to Mwemutsi, though that name was hallowed by long tradition for the ruler at Ukwama. But I would note that Kipole, the woman ruling at Ukwama when the Germans came, was never to my knowledge called ‘Mwemutsi’. As for the other princes, the name most hallowed was Kyelelo. The eponymous figure was the regicide surnamed ‘the Cruel’ who reigned in the West only a generation before the pax. The name in that case was deliberately tied on account of its strength to the office rather than the man. Secular unity a realm does have, then, in plan, though representing the fact of unity directly would hurt the cherished sense of autonomy at the heart of Sanga politics at the level of domain. The fact of ritual primacy for one centring domain and prince is particularly clear for the Central realm, where the old structure persisted intact through colonial times, and the ritual claims of the Princely court were there to be observed. The planting of locally ancient grains was cooperatively controlled by priests for the realm as a whole. First fruits in due time were born to the centre as beer. But secular tribute was less regular. A few goats would be sent in from time to time, the number and timing negotiated by shadow-boxing— the necessary court-to-court communication was through the peripatetic priests, who were masters of the kind of indirection wanted for keeping the peace. Just as any local ruler must be jealous 16

of his warrior-following lest he lose too many to a rival, so must each realm keep jealous watch on a satellite which might harbour intentions either of challenge or secession. In uNyakyusa the same sort of contingent autonomy, with the rivalry and indirection implied, was openly sanctioned by constitutional principles. All known human societies have a place, broad or narrow, for amity as well as kinship. A few societies have reduced kinship to a bare minimum, others manage to expand familism virtually to the point of eliminating amity as a structural factor. What particularly characterizes the Malatan cultures is an easy balance between kith and kin relations, allowing something close to a maximum scope for voluntarism. The sexual angle would be hard to miss, at least for males, as homophilia is the normative initial orientation for both Kinga and Nyakyusa boys. Perhaps less obvious is the connection between easy friendships and high morale in youth—again, particularly where the bachelor life, sweetened by unproblematic sex relations, extends well into adulthood. Aren’t friendships everywhere easily made in youth? But friendship unconfirmed over time decays as kinship does not. Young men confronting new peers at a funeral in uNyakyusa were fractious not friendly. Older men in uKinga feared the envy of erstwhile friends. But to the extent that a highly structured kinship system makes for rigid loyalties it makes against a universalistic politics. Amity allows the more flexibility, and the Malatan protostates can be seen to optimize its advantages.

Sanctions A distinctive feature of human societies is an explicit set of rules for right behaviour. The rules are in any case supported by a combination of three sorts of sanction, but there will be uniqueness in the balance found. I call the three types transactional, mechanical, and authoritative. Durkheim long ago tried to express the contrast between the traditional microcultures ethnographers were beginning to study, as against moderns, by asking how rules are sanctioned in each case. I think he correctly saw that leaderless societies relied on ‘mechanical’ sanctions: control is diffuse (everyone sanctioning everyone else) and so in effect agentless. It is a mechanical system because it runs itself without ‘human intervention’. Malinowski, with the benefit of personal field experience, soon added what he called ‘reciprocity’: if you don’t cheat me I’ll deal with you again. Further fieldwork enlarged this idea. Many anthropologists encountered the occasional use of violence in the enforcement of rights—breaking the peace to restore the peace—in what ordinarily seemed a wellordered pattern of life. This is ‘transactional’ sanctioning: it is 17

neither ‘diffuse and mechanical’ nor top-down. The main agent is the party which has been victimized, and the sanctions a man and his family or friends may impose can look like revenge. It works for restoring the peace if and only if people generally regard it as justified action. That is, the new peace rests again on mechanical sanctioning. But this is the peace of a broadened circle of friends, a pedestrian community where everyone is known, if only at second hand, to everyone else and knows how to put in a claim for redress of a wrong. When outsiders to such a circle are involved in private conflict with members, and a resort to force spoils the peace between two communities, the eventual solution may be a resort to some sort of oracular or otherwise ritually sanctioned referee of claims. Anthropologists will think of the ‘leopardskin chiefs’ of the Nuer of the southern Sudan, who were certainly not quite ‘chiefs’, but also not quite priests. But there the variations are many, and as trouble cases accumulate there will be a strain toward reorganization. A new beginning then might lead to the constitution of political authorities and the pre-legitimation of unilateral sanctions.

Animism ‘Animism’ is one of those terms which eventually has to be wrenched free of the confusion bequeathed it by its inventors—or simply left alone. But the case against abandoning ‘animism’ is strong. Though it certainly has been doing some mischief, there is no polite way to extinguish a word. And for that matter, the original insights which gave the word currency were only subtly wrong. Animism is very nearly universal to folk societies, even where ritual theatre (ceremonialism) has begun to develop, and it is wise to tie the term to belief in a multiplicity of spirits intervening in the individual moral careers of the living. Like the word ‘family’, unless you know a lot about the variability of cultures you get in trouble using ‘animism’ as though you had resolved the question of how to gloss it. But the main reason for keeping the word is to revive the argument about how religion can work without worship. Animists propitiate, they don’t pray to, praise and glorify, or worship the invisible spirits they address. Like the religion of a Great Tradition, animism addresses the fear-inspiring experiences an individual meets in the pursuit of a good, extended, rewarding life. In the classic case a parent whose child is ill has turned to a diviner to discover the cause and thereby the cure. If a spirit being is addressed, its motive is resentment of the living, generally for neglect not of the child but the spirit. Where in a worship-centred religion the neglect would be in failing a religious duty (breaking a 18

hallowed rule), for the animist the failure is personal because the spirit addressed is apprehended particularistically, in a social relationship patterned on the dyadic relationships of family and closed community life. A step toward the exaltation of a spirit is a step over the ideal categorical boundary of animism, but it is likely to be a small step. Animists will always know how to accept new spirits without pushing out the old. The idea of witchcraft is at home in animistic worlds—as it could hardly be in a less particularistic religious world—and can be seen as another small step across a conceptual boundary. The same diviner who indicts an ancestral or other custodial spirit for causing misfortune in one case may indict a witch in another. Wherever animism has developed into a religion of blame, and mutual suspicion interferes with amity within a closed-community setting, this egoistic option of blaming a living soul can’t surprise. But crossing the further boundary, toward veneration of the spirits, goes against the essential thrust of animism, which knows no exclusions. Consider the plight of a diviner whose findings always pointed to one ‘God’. Unless he is in fact an established priest—as it were, hearing confessions—he won’t be worth his fee. What this points to is the salient structural feature which characterizes any system of animism. It is ‘cellular’ not ‘organic’, ‘granular’ not ‘sectional’, ‘popular’ not ‘ministered’. Life in each cosmic ‘cell’ is quite like life in another, though the names of important spirits may differ. You may not even have to bring your divinities with you when you move. Peter Weber has reminded us that Nyakyusa divinities work in parallel territories, each cell equipped with a full round of ritual services. An exception would be personal ancestral sites. A man may have traveled far before he is called to visit one, and may need to retrace his father’s path in time and space as well as his own. These considerations bear in turn on the question, what departures from the ideal type of animism can serve as clues to the level of politicization in a society. Further on, we stand to learn why religion and politics may seem closely linked in the early stages of protostate development—the subject of this volume.‡

19

FEAR & BLAME, CHAPTER TWO

A Politics of Fear Concessions to fear A protostate wants a light but effective superstructure. It lacks the grip on a people’s loyalty toward which it is seen to be steering. Assume the home-ground is not obviously more fertile or otherwise desirable than a lot of unsettled land nearby—a hard master will keep no men. The horizontal ties or personal bonds on which a community’s political coherence can depend must be a simple transform, a mere twist, on the normal extrafamilial ties of kinship and amity accepted in a pre-existing regional culture. Unless imposed from without, new social forms are the offspring of old. But the protostate’s power to flourish and grow will only derive from politics in the strict sense: some individuals must be conceded the right to command collective action on a panoramic basis. In most cases the spur to this concession of rank by ordinary people will be chronic insecurity associated with a new demographics. The folk are increasingly pressed by outsiders to their local pedestrian communities, putting claim on the woods for hunting or firewood, taking fish or roofing grass from the river bend, looking to court their women. The isolationist system of self-help sanctions isn’t serving well. The sense of sharing a common fate defines the referents of community everywhere, whatever the size of the group. For each type and each particular case, community has a signature style, because community is the product of human intelligence, whether we look at a dyad sole or a hamlet, at a ship’s crew or a widespread ethnic front. This distinctive subjective phenomenon lives behind the face of political growth of any kind, as it lives in the plans for a children’s drum dance or a daring night raid on a stranger’s cattle. Political rhetoric everywhere has to strike notes of common concern. I assume that in the formative stages of Kinga or Nyakyusa protostates the fear of violence from strangers had become real enough to generate a political response of defensive alliance among some clusters of intermarrying pedestrian communities, and I assume that eventually the core strength of these protostates did turn out to be military capacity. But if this suggests we should expect a pragmatic ethic to inform protostate politics, our premise 20

is wrong. The pragmatic or practical strain—the famous social contract—is bound to be there but, if it’s not quite hidden, it will at least be well camouflaged. A main causal string in the rise of statelike political systems is that the original concession to allow one person a command rank creates a system of power, and any enduring power must know the best way to watch its back is to grow. But a prime requirement for that outcome is disguising its arbitrary character as a negation of individual freedoms. So power, if it will be entrenched, will by direction and indirection present itself as necessary. Modern nation states have war machines and a stunning cult of monumentalism to make the point. Protostates, being prebureaucratic, are apt to depend for cover on a more oblique and colorful kind of ritual camouflage or inventive political theatre than fully developed states require. Statelike polities have arisen in many parts of Bantu Africa, and the process is often seen as evolutionary change from kinship to politics as a basis of social organization. What was special in the history of the Sowetan (SW Tanzania) political archipelago is that in at least three cases the organizing principle which set the stage for the rise of translocal politics was not kinship but amity. The initial mix of organizing principles for Bena, Kinga, and Nyakyusa peoples set amity at least on a par with kinship as a basis for lateral bonding among men. Bena men, while they merged with Hehe in war, are portrayed in the ethnographic literature as almost immune to the envy and distrust of peers which has been remarked for Hehe. Kinga and Nyakyusa, the two cases I propose to compare in detail, explicitly de-emphasize kinship ties in favour of amity. ‡‡ A separate consideration is this: to flourish, statelike polities require a broad code of civil peace which permits the extension of reciprocity beyond the limits of personal networks. This code is usually something which must grow slowly—it won’t be forced. Trade, intermarriage, pilgrimage, and an open pattern of migratory drift through the region will promote civility and help to limit the impulsive spread of quarrelling. Each of these institutions has deep roots in the Sowetan regional culture; it is likely they all antedate the archipelago’s politicizing movement. But the association of this civil pax with self-conscious ethnicity can be and presumably was accomplished politically. There are two important ways in which this would be done. One is habituating a population to relatively tolerable procedural interventions, marking and shaping the seasonal round or a lifecycle transition. Ritual regulation of the agricultural sector is found worldwide and everywhere wears the face of benevolent authority. The other, shorter road to ethnic solidarity is making regular use of political mobilization. Success in either case presupposes effective 21

sanctioning authority. The Sanga court’s imongo ‘marriage tax’ is an apt example of life-cycle intervention in a society patterning unusually late marriage; and the strict regulation of seasonal planting, though heavily cloaked in ritual, was sanctioned by force, banking on the everpresent fear of losing a crop. The court’s annual war-games employed the still more radically political tactic of mobilization. In a crisis, fear may be the motor and belligerence the vessel of political mobilization; the task of a political entrepreneur is then to attain a practical monopoly on the use of force in the public arena. But if the developing system is to be self-stabilizing, the course of choice will be to sublimate some of the crude modalities of war in favour of an outreach politics. For the Sangu and Hehe in the larger region of the Southern Highlands in Tanganyika, state-building in the nineteenth century was quick and brutal. So far as scholars have reached a consensus on the reasons, both mobilizations occurred in response to perceived threat and a newly perceived balance of opportunity in the use of military force. Coastal Arab ivory traders were accelerating their use of a long-distance route through the Ruaha valley and sparsely populated central Tanganyika plains during the first quarter of the nineteenth century, and had established a port and trade terminus at Ujiji on the lake by 1831. Ivory had become an increasingly competitive and bloody business in west-central Tanganyika, mobilizing hunters armed with guns and introducing an entrepreneurial dimension to chiefly politics. Few individuals living in the high mountains which look southward down the length of Lake Malawi had seen elephants, Arabs, or guns but all of them knew about such things, and most of the men had more than a passing acquaintance with the large armed forces Ngoni, Sangu, or Hehe could muster. Though none of this imported business had by the century’s end produced a profound direct effect on the course of life in the Sanga domains, there is reason to suppose that it would have done, given only a little more time. Ngonde, on the lake’s northwestern shores, met the assault from the south, and reconstituted its political institutions in the act of self-defense. In the Plains area particularly, and to an extent in Selya, the Tanganyika Konde were beset regularly from the south by pillaging Ngoni, and in the early 1870s from the north by the Sangu, but proved no quarry for either. A fair weight must be given to the effectiveness of the three protostates (Kinga and two Konde) in fending off intruders. Peoples to the north were devastated and scattered by the same Ngoni squads whose effect on the two Nyakyusa-speaking communities and the Kinga was transient. Arab traders were slow in making 22

inroads either from the south or north. Had penetration been easier the region’s history could have been as tumultuous as it was for its northern and southern neighbours. I think it important for our understanding of the early stages of state formation that this western end of the political archipelago did hold together through the nineteenth century. Had it not there would have been little left to show what intricately off-beat political systems had once been fashioned there. Militarily, though, the region’s experience in the later nineteenth century was clearly hardening. What we have to ask is what fate the Kinga and Konde escaped—What are the lessons for state-building of the Sangu and Hehe cases? Both of them exhibit the politics of fear in its immoderate posture. First the Sangu, then the Hehe were in the thick of the fighting for wealth and majesty, and a stake in the region’s perhaps overvalued future. An earlier understanding was that the critical stimulus to militarism had come from the south with the expatriated Ngoni bands, who were living by rape and pillage as they moved northward. They had reached the Lakes region, it was estimated, about 1840. But the Sangu insurgency was up and running, it now seems, at least five years before the Ngoni actually appeared in the Corridor, and Zwangendaba’s Ngoni band settled for a time in Ufipa, a fair distance northwest of lands the Sangu then occupied. The critical stimulus to war was temptation not alarm. Sangu were leading a scramble for custody of the Great Ruaha trading route into Tanzania’s interior. Coastal Arab merchants were taking out ivory in exchange for exotic goods like cloth, glass beads, and guns. The Sangu until 1830 would have been a relatively small but independent pastoral community in the lowland plains. But they were nicely suited to the task of squeezing the traders, and it’s likely they expanded very largely by recruitment from the several neighbour communities they overran in their scramble. Shorter (1972) calculates that they were the great power to be reckoned with in the Ruaha area by 1837. Ngoni weren’t there to challenge them until 1842. When they did arrive, having no use for trade or for guns and the hard-to-get powder needed to work them, and still disinclined at the time to provoke a powerful enemy, the Ngoni seem to have held off from seriously engaging the Sangu. ‡ Slaving had been an export business for the coastal Arabs from the first decades of the eighteenth century, and its increasing importance in the nineteenth, along with the now-ubiquitous ivory trade, accelerated exploitation of the interior populations. The ivory trade had become a stimulus to political entrepreneurship and the dissolution of borders. The slave trade as such never did put the burden on the southern highland area which it put on other regions. 23

But it is not to be thought that lurid accounts of this bloody business and the powers behind it failed to spread in all directions between the Zanzibar coast and Lake Tanganyika. Permanent enslavement of captives is reported, though hard to date, in traditions recorded by the British at Tukuyu, Iringa, Njombe, and of course Songea, with its generations-old Ngoni settlements. That the Sangu ecosystem favoured pastoralism and a mobile form of settlement must partly explain the impermanence of their ‘empire’. In the forty years between 1835-1875 the Sangu remained a power in the region which only the Ngoni were in position to attack or ignore. Sangu warriors doubtless learned tactics from these encounters, but the Sangu were pursuing a different game to the Ngoni. Did Sangu leaders establish some sort of ‘protostate’ in the forty years before their crucial defeat by the Hehe? Hardly more, I think, than the Ngoni had created a statelike polity during the same period: clear lines of authority and continuity in power were certainly there in both cases, but these were armies encamped where they found themselves, holding court in a fortress without a country to defend. Where they subdued a chiefdom in their pursuit of tactical gain the Sangu hegemony was an empty one. It quite disappeared at the end. It was as if Sangu knew only how to expand, not to prosper. The rise of the Hehe had a more solid base, but events moved quickly for them. Paradoxically, it is only after his defeat and death in a final battle with the Germans that Mkwawa’s statecraft came into its own. His story, like that of his father before him, is not one of overrunning, sacking, and dispersing ill-defended countrymen but of the pragmatic conquest and political assimilation of rival chiefdoms. Hehe were ruthless plunderers where they were facing an external enemy—Sangu or Ngoni—but their spoils were cattle and women, and a Hehe warrior made proud with his share. The two great Hehe leaders, father and son, had each his round of rivals to suborn, and in the process their statecraft grew with their military successes. What set apart the Hehe polity, as it finally was, from the rest of the archipelago was the fact that it had passed the point of no return by devolution into its original autonomous segments. The ethnic mix was irreversible—a state in the rough, perhaps, but it had by-passed the segmentary phase of development. How far had there ever been known in its territory the pyramidal links of a typical protostate, formed among some of its ‘distinct’ peoples? What we do know is that a round dozen of distinct ethnic groupings were twice incorporated into a single political system; that with the death of the first tyrant the system did devolve into its ethnic parts; but that the successor system which Mkwawa put together 24

and led against the Germans was taken over intact and remained so. That we can’t know much more reflects the tenuous, even evanescent personal-political nature of the structural links which make a tyranny work. Supported by frail roots in tradition, an overnight empire to exist at all can require a tangible future. Ironically, it is perhaps because the German conqueror, Colonel Nigmann, was such a fitting, in fact admiring, successor to Mkwawa that the Hehe moral identity survived the infliction of peace. ‡ Authority by its nature favours monopoly. What gave the politics of fear its radical civility in the Malatan communities was the antipolitan ethic limiting a ruler’s monopoly of power. To preserve the principle of amity, there had to be room within the larger league of chiefdoms for free movement of a self-armed man who found himself prejudicially treated by his ruler. But the politics of fear among Hehe (and presumptively among Sangu) was the monopoly of a single tyrant ruling over available civil space. It is a distinction crucial to the modern idea of statehood, but I think it required chronic warfare—cold or hot—to feed its legitimacy. Neither the Sangu nor the Hehe polity was organized primarily by kinship. Amity? good fellowship? civility? in the political circumstance of a tyranny these sentiments breed expectations which have to borrow trouble. The Malatan protostates appear to have been built on a more flexible, more politic politics of fear. We have to ask how it was done in each case, as the situations were quite different.

Malatan specialties Since our understanding of Konde culture in past time will always derive mainly from reviewing and rethinking Monica Wilson’s detailed ethnographies, my focus will often fall on the people and local customs of the Selya area. As it happens, this is the place Kinga visit in their main propitiatory rite, and the Wilsons’ records of those visits in the 1930s represent the earliest entry of Kinga onto the ethnographic stage: Godfrey Wilson was able to sit in while priests from the two communities haggled over protocol. Since a protostate process is generated through translocal dealings between separate, autonomous local domains, the degree of overall homogeneity is a function of history and local circumstance. Selya may seem in the literature to be the ‘heartland’ of the Rift Valley Konde settlement—it was certainly in no way ‘peripheral’. But the protostate process as we have been able to glimpse it was collectively evolving in circumstances of open rivalry and opportunism. There could have been no single blueprint for the hundred-odd chiefdoms the Germans found there. Early missionaries and later 25

British administrators, to whom we owe many important records, distinctly preferred an upland climate to the hotter plains, and accordingly took their local orientation from that subregion. With that in mind I have gone back to the broader (and now largely discarded) label ‘Konde’ for an inclusive reference to speakers of the Ngonde-Nyakyusa language in both Tanzania and Malawi. This allows me to say, for instance, that there were several, interlaced protostate processes going on among Konde chiefdoms in precolonial times. There wasn’t the strain toward a centre of the political field which, for all the belligerence of Prince toward Prince, the Sanga system cultivated. To the extent that Monica Wilson’s enchantment with the ‘divine king’ figure may have implied the living Lwembe at Lubaga provided a central political figure for the Konde world, I think we have been misled. It is crucial for this kind of study—dealing as it must with the origins of translocal politics—that the communities surveyed be at or still very close to ground zero in respect of occupational differentiation. The domestic institution should continue to provide a livelihood to everyone—the economy remains cellular. That means ruler, priest, and peasant families alike hew wood, draw water, and till the earth for their own consumption. This is the most essential distinction to be made between ‘protostate’ and statehood. But we have to deal with a transitional or ‘evolutionary’ process and can’t afford to discuss it in frozen categories. Do we call the Konde ruler a chief (Wilsons) or a prince (Charsley)? Are the Sanga courts in Kingaland three-tiered or two? Have we prince, lord, and lieutenant or chief, subchief, and headman? And shall we picture ‘priests’ or ‘ritualists’ or ‘councilors’ or ‘courtiers’ or ‘trustees’ or even ‘enforcers’ in action when the same man may be playing any of these different roles? Should we use the indigenous terms, peculiar to each of the three (and more) political cultures we have to consider? I have been eclectic. Especially in comparative analysis I have simply tried to use the set of terms which would be least likely to puzzle or mislead a reader. The network of specialized roles making the rise of political authority possible continues throughout the protostate process to overlay a cellular (peasant) socioeconomic structural plane. Consider the protostate even at its zenith remaining a society which can fall back to its politically acephalous groundplan without having to reinvent its world. We have to deal in the Kinga and Konde cases with two examples of entrepreneurial politics not with two entrenched political establishments. One stimulus to rethinking the Wilson’s work has been a number of historical studies, some in the field, to which a good many references can be expected. In consideration of the bulk of sometimes 26

conflicting matters of record with which a Kinga-Konde comparison must deal I shall also have resort to paragraph-length citations as the best way to make clear my construction of the facts as it is drawn from a particular documentary source. My intention from the beginning of this ethnographic project has been to work through a comparative analysis of the several statelike polities in the Sowetan region, always with the aim of learning how their institutions must have worked in late-historic times, and what the region has to teach us about the origins of large-scale political life. What I particularly suspect, but can’t in any way prove, about the Sanga system is that over the many generations (a dozen, at least) of its development the ultimate resort of political discontent among strongmen of a ruling Sanga lineage was flight to the Valley floor. Two features of Kinga-Nyakyusa relations suggest this possibility: the shared Lwembe myth of a magically gifted ‘younger brother of the prince’ who had to be chased down the escarpment to Selya; and the tradition there that (presumably renegade) princes from Kingaland had founded a number of the (hard to number) major Nyakyusa ruling dynasties. All this is implicit in the widely-held oral history of the ‘Nguluwe migrations’ from a sacred place in uBena to the Konde-speaking Valley floor. The willingness of a host society to accept such emigrants is not altogether mysterious. Just as a prophet is not honoured at home, so with an ordinary commoner who would be king. The readiness of the ‘donor’ society to eject such a person, along with armed retinue and some paraphernalia of rank, would spring from this: that any petty ruler displaced, any aspirant looked over in a succession to high office, or indeed any ambitious but thwarted ‘younger brother’ to a ruler might quite well know that setting up anew in the more open atmosphere of Selya was going to be easier than doing so at home. Colonizing political missions of this kind are finely documented for the Alur of the Interlacustrine region. If something like this hypothesis sounds realistic for the Malatan region, we are closer than before to understanding the way ‘amity politics’ could have come to coexist as it clearly did there with an unperturbed politics of fear. Another clue is this: an actual ruling prince unless active in war lives remote from the bulk of his people. In the Nyakyusa case it is partly because commoner villages greatly outnumber royal, and tend to be jealous about their autonomy. As I shall be at pains to argue later on, it is also because the most powerful Konde chiefs are those who have already devolved their work as hands-on rulers to their presumptive successors—‘chiefs’ of a younger, aspiring generation. An ‘old chief’ is in fact quite properly called a Prince, as I call his counterpart ruler in one of the four realms of the Kinga. In the Kinga 27

case a like remoteness is achieved for these Princes in two steps: within the high court domain, only the trustee families live in the immediate vicinity of (and have limited entrée to) the seclusive royal court; while the affiliated ruling court which centres a client domain typically rules less sanctimoniously and does not hide its own ruler from regular eye contact with his people. What I particularly see in the Selya-centred ‘age village’ ethnography of Monica Wilson is a set of rules for transforming Rift Valley cultures by translocal political action. Nyakyusa political life enjoyed and thrived on greater turbulence—if perhaps with no loss of very-long-run stability—than either Kinga or Ngonde. As unique and important as their age-village innovation was, two other features of their political system are needed to support a balanced view of Nyakyusa history: they seem to have lived in ‘villages’ so moveable as to bring the very use of that English word into question; and most appear to have lived in chiefless—effectively egalitarian—village communities clustered in political space around a central ‘aristocratic’ one charged with (inegalitarian) princely charisma. Though each of these clustered commoner villages had a spokesman-head, there are no records suggesting that this ‘Great Commoner’ could boast princely or chief-like authority. The rules precluded it. The likely result of such arrangements in a male-dominated society would be a fragmented pattern of loyalties and a substantial measure of volatility built into the typical male’s personal identity. If a ‘steady state’ is your standard for judging political development, Nyakyusa are only in early stages. But the reason for that is decidedly not a lack of political entrepreneurship or sophistication. Their political arena—the life-space allowed them by nature and their social surroundings—was vast and indeed politically dense compared either to Kinga or (I surmise) Ngonde, and not easily tamed. When I speak of political density in a pedestrian community I have in mind the sheer probability of meetings at crosspurpose among adult men, and the consequent likelihood of sidetaking episodes. To put the Ngonde case in a nutshell, this third protostate can be viewed as a Konde-speaking population separated by a geographic barrier from the rest, and thereby achieving some closure. They responded during the nineteenth century to external threat by pulling together into a ‘theocratic’ version of the less pyramidal Konde protostate model found north of the Songwe river. In theocracy the central office is not only remote but non-confrontational, allowing for a ‘soft’ form of central direction—Southall’s suzerainty—where a hard one might fall apart. The Sanga political 28

system had meanwhile created a ‘Kinga’ people united rather by the spreading acceptance of a ‘court culture’ which colonized (not uprooting or resettling) the ‘bush’ communities of their mountain slopes and hidden valleys, and maintained the pyramidal organization of ruling offices typical of a segmentary protostate through the fiction of fraternal amity among leaders and the regular passage of token tribute. Recall that Southall redefined his ideal type of the segmentary state as a blending of political authority and ritual theatre. ‡ But ‘ritual suzerainty’? Something approaching that is attained by the Sanga court, where in particular the High Prince is warranted to embody ritual danger of high order by his collateral agnatic ties to the apotheosized Lwembe, externally propitiated by representatives of all major Kinga courts. The propitiated being does not dwell, though, at Ukwama in the actual person of a ‘divine king’, nor can that actual person speak for the divinity. Indeed, the high prince at Ukwama exhibits a majestic mask of fear instead. Again, Lwembe in Nyakyusa society, where the priest-medium representing him claims no right of descent only of calling, strangely lacks both secular power and majesty. These two Malatan protostates exhibit such un-draconian formulae for ritual unity exactly because they are committed to limiting the monopoly of power which their constitutions, written into their ritual custom, assign to a seated prince/ chief. Crucially, the position of the priest in both cultures is that of a tertium quid, counselor to the throne but equally a councilor or tribune of the people. In each case priestly consensus is required to effect the installation of a successor to secular power, and in each case this fact is signalled by the priest’s control of the medicines of power-transfer containing body parings from (at least) the most recent in a series of incumbents to achieve majesty. ‡‡ I have felt the lack of a close account of Ngonde political culture, comparable to that we have for their Nyakyusa cousins, as that would have allowed case-based inspection not only of the Kinga-Nyakyusa systems as ‘transforms’ each of the other, but of Nyakyusa-Ngonde congruency as well. What particularly distinguished the Sanga system from many Eastern Bantu examples of the type was the importance given to a ritual theatre of solidarity— political, agricultural, and propitiatory. Kinga court culture didn’t share the ethos which made the (typically divisive) celebration of their kinship and communal rituals so important to Nyakyusa. Sanga war-games were set up between two sides randomly chosen but so biased that fast cliques were divided, and kinship seems to have carried no claim to exemption. The difference to Nyakyusa is clearly that there all intrasocial (within-village) competition was allowed to 29

play itself out, as it were, after dark. How far this preference for witchcraft might apply to Ngonde I don’t know. However sophisticated it may be, a political culture confined to a single pedestrian community—for Kinga my ‘domain’, for Nyakyusa the Wilsons’ ‘chiefdom’—doesn’t meet criteria for civility of the kind state-building requires. In effect, what is needed is the replication of a new political style in several (3 + n) domains, and the emergence of translocal linking institutions strong enough to foster a panoramic sense of common political identity. African ethnography presents two broad alternatives for linking a series of autonomous pedestrian polities while each continues to occupy its given patch of land. In local descent-group societies the links may be produced by simple rules of ‘clan’ exogamy which, in effect, bring it about that “we marry our enemies”—a situation of political self-help which serves to maintain (among the several domains linked in a marriage circle) a condition of parity and mutual understanding at the level of rules, but without sparking the institution of a political superstructure or a statelike peace-keeping orientation to territorial government. Using this and similar devices you can model most of the well-known African ‘tribes without rulers’. The alternative route is political. A ‘culture of rules’ as wide as the sphere of societal coexistence is wanted to nourish any statelike institutions, but one which can leave to each people the maintenance of its own characteristic ‘culture of sentiment’ (the deepstuff). A shortcut version of the political route was taken by the Sangu and Hehe in the formation of their state-like military régimes. But the basis of stability for these quick-and-dirty despotisms was in a continued income through the spoils of conquest. History in their part of the southern highlands political archipelago couldn’t wait for evolution but took a shortcut from a mosaic of rivalrous petty chiefdoms to a despotic territorial state in which a sense of common ethnicity precipitated from the common enterprise of predatory ‘nation building’. The Corridor peoples, though subjected to many or most of the same intrusive forces which seem to have triggered the Sangu (in the 1830s) and later Hehe mobilizations, proved less vulnerable. The Kinga and Nyakyusa political systems seem to have kept a less hurried pace; the Ngonde moved more boldly forward, still on track, toward a form of theocratic state. About differences at a deeper level between Ngonde and Nyakyusa cultures of sentiment we know too little. I regard these three ‘protostates’ as moving, each on its own zigzag path, toward a receding horizon of statehood. Each was caught at some point along its path by warlike intrusions and the subsequent colonial freeze.

30

Can I describe the paths? Can you reduce a history to narrative? There is no need in any event to go so far as to drown the tale in dogma. Inside all three communities, important intergroup linkages were effected politically and ritually. This entailed the invention of an aristocracy with (fictional) bloodline relationships linking the major rulers of the whole region as an hereditary class set apart, and the development of a priestly presence with strong professional ties across political boundaries. The sources of this structural change, like the sources of a river, were many and concurrent—none need be given temporal priority. The ancestral figure for prince and priest was the three-in-one ‘most respected elder’ of earliest-remembered times, who was vested with the oversight of political, ritual, and jural services within an autonomous face-to-face pedestrian community. As the settlement pattern within a given ecological niche matured, trans-local trouble cases became more frequent, and resort to arms called for a programme of policing the peace. With the gradual appearance of secular authority, the scope of ritual action came to include the explicit sanctioning of entrenched political power. Moots gave way to authoritative law courts for civil dispute settlement. In the mystical trade, independent ritual practitioners largely gave way to politically responsive priesthoods. But local factors affected the degree to which a pre-existing ethnic mosaic resisted assimilation. The social situation and character of prince and priest want separate consideration in each community. As might be expected in Eastern Bantu civilization, the founding fiction for an aristocracy played on the myth of a lone male immigrant, a ‘hero’ with implied special powers and gifts. The burden of the tale is that his descendants alone are inherently fit to rule. In regional prehistory the nearest analogue or archetypal experience of ethnic overlay is the coming of Bantu migrants with the arts of the iron-smith and a significant intensification of agriculture. But the provisional dating for this would be a period starting before the middle of the first millennium of the present era. There is no indication that the actual emergence of a ‘ruling class’ was so early. But linguistic-ethnic dualism may well have persisted in the region through the Iron Age and into the centuries of political expansion (the ‘African middle ages’ are here assumed to begin in the East African interior about AD 1400). The institutional feature matching the myth, pronounced in the Kinga ethnography but endemic to the region, is a device which initiates a court-bush interface. It entails the ‘sending out’ by an established court of Sanga avatsagila (client rulers) to colonize and extend the court’s dominion over peripheral settlements. This same ‘sending out’ becomes in Selya the famous

31

ubusooka or Coming Out, which culminates a chiefdom’s growth and instantly decentralizes it. More easily grasped than the development of a political role system is the quieter evolution of a priesthood. It begins with an earlier level of ritual specialization associated with ‘territorial shrines’ whose salient boast was an ability to control rain. They are ‘territorial’ in the same (ecological) way that rain is—that is, covering whole reaches of an ecozone and affecting a host of contiguous if politically autonomous communities. Local ‘chiefs’ in many parts of German Tanganyika were sometimes berated as ‘old rainmakers’ without enough secular authority to serve as effective tax collectors. In the Corridor region, rainmakers always had divinatory paraphernalia, and might be sought to by a private person on any sort of victim’s errand. They were wizards and occasionally would achieve charismatic fame through the whole ‘territory’ they served. Before state-building began they were local to one pedestrian community and active agents in supporting and regulating a religion of blame. This state of things still prevailed in some parts, generations after prompting a German district officer’s frustration with ‘old rainmakers’ posing as chiefs. Extraordinary circumstance may promote a mushrooming political superstructure, but the ordinary course of political growth is extra-ordinarily slow and its fruit not predictable. ‡ In the Sowetan political archipelago I find evidence that the politics of fear, which informs military action and the disciplinary procedures of the law courts, is generally balanced against a politics of blame. I refer to a politics of blame in order to focus on the manipulative aspect of a religion which begins to be centred at a court. I see the allocation of blame as a companion to the court’s commanding power to allay fear. This is a power which can’t survive the loss of its mystique. Quite simply, the responsibility for trouble can’t be allowed to stop at the top. Blame must be deflected. If we look at the matter in light of the secular/sacred distinction, so important to early anthropologists, the manipulation of fear belongs to the secular arm, that of blame to the sacred arm of government. The notion of a ‘court’ nicely embraces both: when discourse finds no conclusion, it is time for the dramatics. When danger is apprehended, to name a source of fear is to place blame. Should this lead to a witchcraft trial or ordeal, an accusation can be rendered lethal. In what we might view as a ‘civil case’ it can be enough to blame a witch; where the trouble is wider and vaguer you might want to blame a shade (divinity) considered somehow ancestral to your whole ‘people’. Either way, the political 32

theatre entailed reifies the perceived cause for fear, dramatizing it as incorporeal animus, but giving it the fanciful intensity and definition which allay skepticism, enlisting public concern. The Sanga court presumed to deal equally with pragmatic-legal and ritualmetaphysical determinations, and seems to have succeeded fairly well in controlling both spheres—although bush-Kinga patterns of self-help in civil cases were certainly not everywhere eradicated; and diviners, curers, and unattached claimants to mystical power were not always brought under the court’s umbrella. Witchcraft beliefs remained a bastion of ‘mystical democracy’, as did spirit possession among Nyakyusa. This is because anyone, of either gender, might have secret powers of the very kind claimed by the court. I discuss these matters, and assess the politics of belief, in Chapter Three, “A Religion of Blame”. But to find the place of beliefs in the two political cultures, I want finer-grained modeling of structural differences. There are four basic questions to ask.

Four Kinga-Nyakyusa analogues In the circumstance of a small-scale society the several major institutions have to fit nicely together. There isn’t room for many options at the deeper levels of fit. The four institutions I consider here are functionally alike in both societies, though formally distinct. Each of the four is necessary to the working of the others—they constitute a distinctive organizational complex. In that sense a good part of the core institutional structure of the two societies is approximately ‘the same’. This is not the sort of statement which would bother a sociologist or an analytically inclined historian. When you apply an easy concept (say, ‘the middle classes’) to a handful of different locations, and notice no distortion in doing it, you are making the same assumption of functional identity. You know it is a molar, not in any sense a molecular, identity. This play works reasonably well when there is a fairly close historical connection among the societies so compared. It works, for deepening understanding on both sides of the equation, for Kinga and Nyakyusa institutions. Formally the two systems are quite distinct, and for descriptive ethnography that remains the important part. The wonder lies in how near the two systems come to homology when you adopt an analytical mode and consider function. For comparative (analytical) ethnography that is the point, though the argument actually turns against ‘functionalism’. The premise is that many kinds of cat can have the same insides. It points to the importance of intercultural contacts on the level of social thought; and for 33

anthropologists of other-than-humanistic persuasion, to the importance of noticing ideas even when they are not floating on the surface of life.

(1) The client-ruler strategy. In Kinga context Kinga oral histories I got for several domains record the exile of a local wizard during early days of setting up the Sanga court, only to call him (which is to say his official descendent) back some generations later. These oral records refer to and clarify the transition from a pre-political system of governance to the Sanga court system. The business of a wizard is reading signs not commanding action, but wizards do deal in consensus, and that can be good enough when the problems of governance are of quite limited scale. When things were going wrong for a man in pre-political Kinga culture the ultimate move was quitting his community for another. That is the way things still worked for Ndendeuli when Gulliver was with them in the early 1950s: movement there was free of risk and impersonal constraint. It is a proverbial characteristic of the wizard that his stature and influence is magnified by social distance from his client, and that means in effect that close local groups don’t have their own wizards. Clients come to seers at need and scarce need heed them otherwise. Still, the kind of wizard we are dealing with is not a loner or an entrepreneurial professional of the sort we find at a later or more turbulent stage of social complexity. The simplest version of a sacred grove is a rain shrine, and the earliest wizards the Kinga know were rainmakers. This was a profession to which men gained entry through apprenticeship not by any formal succession to office. So a wizard with his postulants and lesser associates would always belong, in the classic manner of artisans everywhere, to a circle not a local settlement. What we learn from the oral record is first that a would-be Sanga ruler had to contest the power of such a circle and banish its main protagonists; and second that, once established in his own right, the Sanga ruler would find reason to bring back the local wizard, complete with his kitbag and support group, to merge them with the new court. This is the best explanation I have as to how we get the office of ‘priest’. A full-fledged court must have wizards of its own so the people will want to bring all their trouble cases to court. But to make this happen you have to reinvest the wizard as priest, conceding his

34

association with authority, and look to him as master of ritual—not, as before, the inspired stranger. The most relevant Kinga terms for office at the court are unkuludeva (prince), avanyivaha (the court’s trustees or strongmen), unteketsi (officiant at sacrifice), and umotsi (rainmaker). Where the first three roles are anchored in a royal court, the rainmaker is not. He is a self-employed wizard to whom a Sanga ruler would be seen to apply as client. The rainmaker’s trade and his kitbag antedate the court and can survive without it as long as the belief persists that his ministrations are needed to bring on the rains after a dry season. As for sacrifice, any elder is competent to propitiate an ancestor; the unteketsi deserves the name of priest because he is keeper of the medicines of the court, host of the sacred grove, and necessary as officiant at communal rituals. This puts him in the category of unyivaha, companion (trustee) of the court. The priests lived hard by the ruler’s enclosure in daily contact with him and other companions of the court (avanyivaha). They were engaged as a group in managing the ordinary business of government. Where the prince himself was no leader, a suitable man would emerge among the company of avanyivaha. The High Prince at Ukwama in historic times had come to play a reclusive, ‘sacerdotal’ role. But at a small court (that of an untwa or local ruler) the only senior ritualist might be the umotsi, master of the rainstones. As any sort of court would require a more-than-military presence he would doubtless be drawn into various official—priestly—functions as well. In discussing such a person’s status, no one Kinga term suffices; and the same applies to any court regulars. Their company was contingency-oriented. Within it role-differentiation had to be situational not categorically generated. In accepted social memory, and in popular conception, the highest court’s avanyivaha were not a class of oligarchs but were seen to be a company acting both for the crown and, as intercessors, for the people. Call them power-brokers. What most distinguishes the Sanga polity from the Nyakyusa is this gathering-in of influential individuals and their integration as a governing executive. As a group, this company stood between ruler and people, embracing the double duties of tribune and enforcer. The weight this gave to a well-established royal court allowed it to play the feudal game of fealty and tribute with a handful of satellite courts. How many of these were actually ‘founded’ by a Sanga client ‘sent out to rule’ we can’t know; but when it was convenient to say so, a long-dead ancestor could be named. Whenever a little band of goats was driven from satellite to centre as tribute, the paramythic untsagila, a lieutenant of long ago sent out from a mother 35

village to colonize a mere bush area of settlement, was being invoked at once by the client and the higher court, affirming the nature of their political ties. The affirmation was wanted because centrifugal forces always pulled at the political tether. A princely court was in no sense a command centre for the realm. By the time a client-ruler line could have matured to the point of setting up a court in its own right, able to boast a military, jural, and ritual presence on the Sanga model, its independent identity would have been made a matter of plain fact. But its symbolic clientship to the Prince, expressed in explicit acts of rank concession by untwa to unkuludeva, would have been equally well established. The social construction of the client-ruler figure, untsagila, grounded the untwa-unkuludeva relationship in a history of colonizing expansion. The father of all stories about banished wizards is the Lwembe myth—Kinga version—which pertains to the senior Sanga court, at Ukwama. The wizard Lwembe, banished to Selya in Nyakyusaland, can’t be allowed to return. As source of the greatest afflictions of nature the Kinga have known, he is too dangerous. His return would destroy the Sanga system of rule by disarming the secular power, whose tools are admittedly mundane. It is in this myth we see the key constitutional mandate of the Sanga court: secular power is its driving force, guided but never governed by ritual wisdom. It is a builder’s mandate, allowing for the recruitment of a trustee group on the basis of individual ability. The client-ruler strategy in its Kinga version allows for the gradual growth of a network of courts, each a replica of the original, founded by royal ‘brothers’ originally sent out as client-rulers with the mandate to extend Sanga rule over a broad country. In the ideal model which still governed Kinga political thinking in 1960 three of the client courts had grown to represent rival realms to Ukwama, and each of the four realms had in turn sent out client-rulers of its own to build up satellite courts. But even while formal battles as well as tactical raiding between princely realms went on, the client-ruler construction continued to be invoked. Tribute, though it was famously irregular, still was owed by the ‘younger brothers’ to their senior prince, and in a general crisis the priestly companies of the realms concerned would join to discover the ritual irregularity at fault in some failure to render tribute due. Oral historical evidence indicates that a client ruler could take with him (or send back for) a suitable member of the mother court to serve as ‘chief priest’ in the new court setting. In the one case for which I have name and dates the new priest came into a turbulent situation and was put straight to work as witch finder-executioner. 36

His further task (but for German intervention) would have been to build there a new branch of the priesthood, so assisting eventually in the establishment of a new autonomous local domain under the Sanga umbrella. In this latter-day example we can perhaps see an historical analogue to the mythical practice of bringing back the banished wizard whose local credibility was a precious asset. With the firm establishment of the Sanga protostate, it may be the prestige of a Sanga celebrant had become the only asset required. Alternatively, we may speculate that an aggressive witchfinderpriest had more than once been in the vanguard of a Sanga colonization programme. In several local domains in the 1960s I found the court ritual specialists bore local rather than Sanga surnames. But in the oldest princely courts the priests, though I supposed some of them would stem from pre-Sanga founders, all identified as Sanga. I took this at the time as evidence of a growing ascendancy of the political, perhaps linked to the climate of external colonial rule—so not “pure Kinga”—and the question remains unsettled for me. It is likely enough that the solidarity of the avanyivaha as a power group, and the principle of recruitment to position through apprenticeship and ability not heredity, would have combined in the bigger courts to turn over the (rhetorically ‘ethnic’) claims of a priest to be ‘tribune of the people’. If he was able at need to curb a ‘tyrant’ Sanga prince it was more as vizier than tribune. The head priest at Ukwama showed me his imatsi, the ivory-like shell badge of high authority. It was big enough for a prince.

In Nyakyusa context The client-ruler institution took a distinct form in Nyakyusa culture, where chiefdoms competed fiercely for followings, jockeying about for position both in spatial spheres of influence and in the military-reputational game, and depended more on charismatic than sacerdotal theatre for the retention of power. While important chiefs might be given a full divine-king mortuary treatment, there was no sacerdotal option—that was reserved for the perpetual wizard-king seat at Lubaga in the Selya region, serving various chiefdoms but with slight secular power. The interment spot of an important individual chief would in future be a place of propitiatory sacrifice but not a shrine of such danger and secrecy as was Lubaga’s unapproachable sacred grove, Lwembe’s seat from mythic times. The role of territory in Nyakyusa chiefly politics is puzzling as Monica Wilson presents the evidence from testimony only one long generation after the pax. But a careful reading and comparison with 37

German records (so far as known to me) shows that the basic Kinga system of bounded domains, aligned as contiguous territories in four bounded realms, had no parallel in Nyakyusa political thinking. This means that client-rulers could not be ‘sent out’ quite as (in the model realm of Kinga social memory) they were by a Sanga prince, with the mandate to rule a named territory and all who lived there. The initial task of the cadet Nyakyusa ‘chief’ on being sent out was rather to recruit than rule. That is, he must gather men, cattle, and women, pretty much in that order of priority. In fact, the Nyakyusa institution reads to me in the Wilsons’ presentation as something closer to an opening of the gates, a longawaited liberation of a warrior age-class, than to a pragmatic statebuilding move. The pragmatic version is, in my judgement, a textbookish confection so far as it is to be taken as a portrait of traditional times. It was presented first by Godfrey Wilson in an early summation of field data. Here is the fairy tale in brief: (i) An established chiefdom comprises (say) twelve senior villages each with three or four satellite junior age-villages. Two of the senior villages belong to the chief, who keeps wives in both. The satellites are dormitory villages for all bachelor male offspring of a senior village. The boy/youth herds, does the heavy gardening, and learns spear-work for his father. He is fed by his own and/or a companion’s mother with food he and his friends have helped produce from the paternal gardens. (ii) About thirty years after the chief’s Coming Out he will be ready to retire from active leadership. He and his whole generation of village leaders are feeling a ‘push’ from a frustrated younger generation of males aged 30-35. So the commoner villages, each in the person of its head, go together with the chiefly villages to stage a grand Coming Out. Two young chiefs and as many young heads from each senior commoner village are designated rulers of some (say, twenty-four) newly franchised cadet villages which the ceremonial organizers map out. The older generation bows out gracefully and passes power onward. In a grand ceremonial drama, a whole new political arena is opened up with fanfare and much excitement. By putting this custom on a par with the Kinga counterpart, and associating the institution with a ‘client ruler’ strategy, I am denying one crucial facet of the Coming Out which was dear to the heart of Monica Wilson. It was her notion that ‘the old chief would soon die’, and the stage was now set for a ‘divine kingship’ death by strangulation. I doubt this. I see it as fairy tale. I can’t conceive that the political entrepreneurs who managed to set up so marvelous and 38

so political a structure as a Nyakyusa chiefdom would just blow a whistle and say, ‘new game—start over’. The ‘graceful withdrawal’ (especially one to be accomplished through strangulation) reads like one of Malinowski’s ‘just so’ stories. ‡‡ Monica Wilson renews the pragmatic, orderly view in her tellings, so that the Coming Out (ubusooka) looks more like a communal rite of passage than the political tsunami I take it to have been. She definitely did not see the acquisitive elders’ age-villages as beneficiaries. The ubusooka rite is portrayed on the granters’ side as benefaction but can be otherwise seen by a skeptic. As tradition gives it, it seems an utterly reckless, unpolitical act. For my part, I see it as the culmination in power of a maturing chiefdom, which has produced a potentially powerful new generation and is ready to expand aggressively in political space. I see the Nyakyusa chiefdom as a loose confederacy of autonomous villages. The one close Kinga analogue to a Coming Out is the case of Kyelelo the First stepping back in favour of a son in order to continue consolidation of a Western Realm seated at Luwumbu. ‡ All the same, if Monica Wilson’s analytical emphasis on the theme of ‘rebirth and renewal’ be taken à la Durkheim—as invoking a collective representation of social solidarity in the very act of seeming political dissolution—we stand to learn from the exercise. She makes clear the breaking out of long-smouldering resentment toward the elders, of license for displaced violence against brothers as well as strangers, and the confounding of old and rigidly protected property rights in cattle. There is a paradoxical mix of peace with turmoil in all the descriptions we have of Nyakyusa political life in free-tradition years. We are told the client-ruler strategy marches to its own drum in each chiefdom, beginning with ritual seclusion of the new power-holders and finally a ceremonial send-off (the ubusooka ‘Coming Out’) from the ruling village. This enacts a stripping away of all the junior adult age-villages from each of the senior villages of that chiefdom. In the roster of these senior villages there are likely to be eight or more of commoner rank, each with a headman (‘Great Commoner’) in charge, in addition to the chief’s two. Ironically, the Coming Out celebration may typically have been the moment of greatest unity within a chiefdom. The ceremonial sendoff and (alleged) voiding of the old power pyramid is realized only once in a long generation (about thirty years) and always after much delay. Two or more young aspirants to chiefship—already aged thirty to thirty-five—are sent out in this way, each with a mandate to marry and command a new chiefdom comprising a full share of the young men stripped away from the old. 39

A good handful of spares, cadets ready to take over leadership at need, are put through the rite as well. Perhaps in the 1930s when the Wilsons’ fieldwork was being done, or in 1953 when Monica Wilson was on revisit, this affair could pass for an orderly ‘communal ritual’, a sort of biggish, long-delayed bar mitzva followed by youthful exploits. But before the colonial pax the sequelae would have been a fair mix of braggadocio and real violence. The competition among cadet leaders can’t have been quite unbloody, but should be imagined higher on death-defying than death as such. The new young chiefly contenders (client-rulers) depart each with their wonted male followings and a few women and cattle. Apparently these cadets have no mandate but to seek their own destiny—build a new village where they can, and there hustle themselves the new cattle, women, and gardens a real chiefly future requires. This account of the passing over of the torch of power to a new generation is the scenario their most reliable informants gave the Wilsons in the nineteen thirties. Assuming that the actual number of successful chiefdoms is either static or grows only gradually over time, about half the aspirants to glory must fail; and it follows that the other half must succeed by gaining a competitive advantage. This would be done directly by warfare or by otherwise exceeding them in recruiting men from opposition ranks and/or neighboring bush cultures, and in the regular business of taking or inheriting women and raiding for bridewealth. For ‘warfare’ a Kinga observer would use his word for ‘war games’: pre-arranged and ritually prepared contests between companies of men from each ‘side’ as decreed at the Coming Out, refereed battles ending with the first serious drawing of blood. We could all wish some scribe had been around only the one long generation earlier to observe an actual case and record how all this worked on the ground before the pax made the ubusooka a mere rite of passage. We have one picture of an older generation ceremonially sowing turmoil, and another of accomplished tropical paradise. But between the anthropology and early missionary accounts we do have an interesting body of evidence for reconstructing a credible model of the social system our ‘scribe in time’ would have found. I find the cadet generation a set of rivalrous client-rulers. I think the five key elements of the model would be these: (a) The physical life-expectancy of a thirty-year-old bachelor male would have cut back the likelihood of even one of the two chosen favorites living to become a major power. The system was open to ambition on the part of others in position to claim ‘aristocratic’ status. Only a few chiefdoms in any generation would achieve the 40

size and majesty to make their Coming Out a map-changing event. Splitting a weak chiefdom could only produce two weaker ones, and this must have been a common result. These assumptions allow for a realistic fit of the model to probable rates of demographic and political expansion. But by the same token, the sense of expansiveness would have been in the air. It was indeed a world renewal, as new villages and cattle byres were springing up and a fresh generation of men settling in to marriage and rivalrous begetting. (b) Patronage by a powerful chief of the older generation was important and put a premium on the client’s ability to subserve without losing credibility as a leader. A young political entrepreneur might play one patron against another. But patrons could do the same. I am presupposing a political game in which force was not the universal instrument but was one in which immoderate success was unlikely and moderate degrees of failure tolerable. (c) A fair proportion of the lesser client rulers could eventually decide for retreating to ‘frontier’ situations, at least on a temporary basis. With subsequent success in seeking their needed followings of women and men, a few such ‘colonizing’ leaders would be in position to move back at option into the mainstream politics of chiefly aggrandizement. The likely stage of ‘greatness’ for an individual chief would be but a decade or two; but by inheriting (with the usual priestly help) the name and office a skin-changing successor could extend the rule as long again. Thus a younger brother or son moving up to the office would continue its name with all the debts and credits attached thereto. This model stresses the entrepreneurial aspect of chiefly careers. It accepts the 30-odd year periodicity. When ‘things are going well’ the kind of change the ubusooka implies would be unwanted. But the unrest would have been shared by both generations of men, and the ceremony welcomed by both. ‡‡ (d) Warlike manners were put on for the occasion. Ever ready to psych themselves up for war, Nyakyusa men other than the chiefs were not deeply macho, not inner- but peer-directed. Peer relations being artificially dense and omnilateral, losing a comrade was not structurally traumatic. The obverse of the Nyakyusa emphasis on ‘good company’ within an age-village was the ease with which a deserter would be taken in by another such group. The model takes account of the antipolitan ethic so evident in Nyakyusa testimony: peer loyalty may even be negatively associated with loyalty to a chief. Having friends and most of one’s kinsmen in other villages leaves a man his own boss in the politics of commitment.

41

(e) What especially energized and maintained the prestige of the ‘Coming Out’ institution was the high-spirited style of youth and young adulthood, stemming particularly from the high morale generated in the age-village system for boys. Observers might be expected to assume that delayed marriage would produce frustration and dissension; the sexual equation in the Nyakyusa case appears to have been more like that of the Kinga, where homophilia is not thought in any way unnatural, than Monica Wilson’s informants intimated. Less like the Kinga experience was the opportunity every Nyakyusa youth seems to have had to form brief liaisons in bachelor years with (married) women of their own age. This risk enterprise was by all accounts a feature of their young lifestyle. It was ritually infused into the careers of boy-men and maidenwomen by their special days and nights of intimate seclusion before a girl’s nubility, friends with friends, learning about love-making by show and tell. This idyllic interlude—what Monica Wilson describes as their ‘initiation-marriage’ rite—was arranged for a self-selected group whenever one of the maidens in a circle was preparing to wed. Taken for granted: the groom (who was not invited) would ordinarily be a stranger. While Wilson’s informants stressed an educational rationale, it is hard to doubt two further implications: that the possible trauma of a wedding night was finessed, and that the radical exclusivity of marriage rights was mooted. Quite obvious is a third implication: a bridge was fashioned, whatever else the young people did in their private interlude, from homophilic to heterophilic sexual orientation. By the time of Wilson’s fieldwork the original African morality play was outdated: homophilia was being played down, the interlude itself finessed, and a maiden’s transition to adult sexuality was being eased by a few night visits with the husband-to-be, who was expected to have ‘external’ sex relations with her. That is, if the missionary mind has not unduly influenced our information on sexual matters. ‡ The colonial mind was set on taming the other specialty of youthful males, and in this was more certainly effective. The pax meant warfare ceased to be a viable alternative to (or extension of) litigation in the pursuit of justice or the reward of ambition. But in the real world of tradition the political consequences of the Coming Out in a major chiefdom—letting all cats out of their bags at once— was a triumphalist kind of world renewal. It appears to have worked by sowing trouble on a scale large enough to keep the political arena in business for another generation. Further discussion will be wanted in later chapters, where the Coming Out is discussed in relation to the militaristic mandates and situational logics of princes in these two protostates.

42

(2) A tempered aristocracy. Morton Fried gives us (in his Evolution of Political Society) a thoughtful and undogmatic discussion of the range of political complexity anthropologists have met in the field. As for interpretation, his main independent variable is social complexity. The range is from egalitarian to stratified societies. Between them, in the midrange of social complexities wanting political management, are Fried’s ‘rank societies’, and I expect it is at this stage of institutional evolution we might find the kind of protostate politics I’m concerned with. But Fried is no Africanist, and his mid-range examples are certainly not protostates. None of them shows the tendency toward central direction characteristic of the pyramidal polities of the Sowetan archipelago. What I do find relevant in Fried’s middle category is the appearance of rank differences without socioeconomic differentiation. This is a vital point if we are to avoid misunderstandings when ‘aristocracy’ is applied to Kinga or Konde social organization. Read no implication of class or caste. If Kinga rulers do collect more wives than commoners, near kinsmen of the rulers, having no court life to maintain, generally don’t. If Nyakyusa know an overclass it is rather a gerontocracy than an aristocracy. Elderly Nyakyusa commoners are as keen on polygyny and wealth in cattle as are their village peers whose patrician claims have not run out. On the other hand, it is doubtful that old men ever enjoyed a different or better diet or free time than the young. Social complexity? Perhaps it follows after political complexity. Even within Eastern Bantu civilization taken as a whole, the only cases of true stratification which come to mind are ethnic: in the best known of these dual societies, the Hutu-Tutsi (ethnically Bantu-Nilote) in Rwanda, it is arguable that prior to the foreign interventions of the later nineteenth century, civil patron-client contracts maintained this rank society in such good order as to preclude the development of protostate politics with its integrative moment—and in consequence opened the way to caste hierarchy in colonial times and a post-colonial government destined never to achieve legitimacy. Nothing comparable belongs in any critique one might draw up on colonialism (and most particularly the British period of indirect rule) in the Sowetan region. The Sangu and Hehe despotisms left no heritage either of stratification or (apart from the usual passage of chiefly office to a son) rank differences. Both the Sanga system and the perpetual turbulence of the Nyakyusa produce a significant slope of dominance: some men are more important, have more influence, get to initiate more action 43

than others. At any given time, only a very few men at a Sanga court have an expectation of joining the dominant group there. Most men will return with marriage to the peasant life at middle age. Most Nyakyusa live in commoner villages and have no expectation of joining a chiefly court. Yet these are not ‘bush’ villages. Even in a ruling village only a few men have more than a nominal connection to dominance, and since each commoner village has its own dominant group there is no instantly obvious difference in rank among the several villages of a chiefly domain. What is important to notice is that access by birth or merit to political ascendancy in either society would be seen by ordinary men to have the same conditional sort of permanence or preordination we associate with ‘aristocracies’. I qualify the local example as ‘tempered’ to reflect its subtle blend of the realized aristocracy constituted in a set of loosely hereditary offices and a nominal one constituted in a set of hereditary ranks. In either case, if many are called few are chosen to enjoy real personal influence, and in the competitive atmosphere of Nyakyusa politics the devil can expect to take the hindmost. It is to the point that the Malatan status differences directly of concern in politics are those among adult males, and limit but don’t determine access to office. Without an endogamous elite. and as a rank society can’t in a polygynous world proliferate ranks in pace with its proliferation of male offspring, claims to high birth could not easily flow along agnatic descent lines in the way nominally possible in a stratified society. Wealth in the form of cattle and wives is a key value in a masculine career among Nyakyusa. It is much less important for the moral strategies of Kinga men, but in neither society is the inheritance of wealth a source of great expectations. In various ways in the Malatan protostates, men concede rank to an office holder and with it concede him—by way of privilege—the capacity to hand on to offspring some of those moral qualities by which access to rank positions may be justified. But beyond that the people are jealous of according credit beyond its due. This is their main guarantee of quality in an officer who will hold his position for life, and life-tenure is in turn the key to stability of office itself. These polities have their strongmen, but they comprise a tempered aristocracy—something closer to a Jeffersonian standard, when seen in perspective, than the untempered situational power they nominally cede to the man in high office.

In Kinga context In the Kinga case, court-bush circulation was managed by cheapening the Sanga identity to the point that the name 44

warranted a man’s acceptance by and loyalty to the court and its élites, and little more. Consider the scene. When bush youths are recruited to the royal courts, they replace senior warriors recruited in the same way a generation ago. These seniors, according to the Kinga model, are each awarded a mature princess of the court, and with her the Sanga name to assure them status in the new offcentre settlement the group is to make its own. The best of the senior warriors, in particular, is so awarded and made group leader, charged with carrying the Sanga banner into frontier country. This man has now become untsagila (‘lieutenant’ perhaps) to the royal court. But internally—within the new frontier community—the role is better translated as ‘client-ruler’. In matters local, he is going to be a ruler responsible to his own people not the high court. As client to a princely patron he owes and is owed nothing until a specific, situational compact be made. Success in establishing authority will be slow work in the bush, but with it the client or his successor can aspire to taking on the rank and trappings of a local ruler, untwa. That will take building up a clientele of around a thousand households—not a simple undertaking. There must be a broad, unchallenged network of loyalties and an effective peace-keeping force. Maintaining men in barracks wants meat, gardens, festivities, beer; this in turn wants women to work the fields, and these must be the segregated royal wives and their full-grown bachelor daughters. There is also the duty of affirming loyalty to the royal court through nominal tribute, in order to ease a transition to near-parity in rank. Expansion of the Sanga protostate is not being done in ‘aristocratic’ mode, though it may be fated in retrospect to be so pictured. A new court is being built in and from the bush, its politicians yet unborn. It is also the case that any such pushy new political establishment needs in turn to develop its own staff, the avanyivaha, priests and magistrates and their chosen assistants. So a client has a long path to climb before he (his part now played by a lineal heir) could hope to see himself established as part of the Sanga ruling élite. But that is the way Kinga of the final precolonial generation thought it had been done—not how realms had been made, as they were somehow all there ‘from the beginning’, but how local domains would have been formed within the realms. It started with a small party, a handful of men with their new wives. Most of these parties, with average luck, would settle a new hamlet and survive there, or perhaps go to replace a less competent local establishment in already settled country. A very few leaders, cultivating old ties with others settled near, would eventually set up as petty magistrates,

45

settling small disputes, sending hard cases further to the court of the realm. The territorial size of the domain one local ruler could manage would depend on qualities of the land and landscape. The demographic size would be a function of that and the natural limits affecting a pedestrian community disposed to affiliate. Most of this, in any event, was already stuff of the misty past in the extant domains the Germans were coming upon in 1900. We are left to extrapolate farther backward toward Sanga ‘beginnings’ and the emerging idea of (four) realms in perpetual ‘fraternal’ league. System change was ordinarily slow, and social memory kept alive not the past but a code-like reconstruction of it. But there is ample evidence in the variant versions of an event I recorded at different courts, and in the German and British records of claim and counter-claim, that nothing political ever was so rock solid as public statements made it out to be. Only a popular ruler, for instance, could carry the role in the theatrically appropriate ‘African despotic’ style, and only the generous ruler remained popular. Kinga theory took no notice of what we see as role strain and ‘double bind’ or the aberrant behaviour we’d expect to result. We can say that the Sanga protostate was ruled by a small political class. We can’t say this class co-opted any but its ablest offspring. A few clear cases show that very able men from humble beginnings could achieve power and distinction within the political class. Priests managed their own fraternity. We can’t say how far in historic times they reached outside that fraternity for apprentices, but the rule favoured ingroup youth. Of the men actually born at a Sanga court, the great majority were royals, first-generation offspring of the ruling prince or his predecessor. These were the avapapwa, brothers and halfbrothers the prince could remember as infants carried on their mothers’ backs. They rated their own barrack house, and carried prestige above other Sanga, but could have no careers at court. Their futures were their own to achieve, as the prince wouldn’t have brothers near once their bachelor-warrior days were done. They would be moved off with friends. As with the Konde, the wisdom was that commoner stars were safer trusted with devolved power. The rules of marriage meant the avapapwa would have to find wives elsewhere. The royal heir (unkinga) was early chosen and succession was in the hands of the avanyivaha. There could be no contest for the throne. The only ‘aristocracy’ the Sanga courts were generating was that of the court culture itself, and this (as a disinterested observer must call it) was an open class—aggressively open, in fact, to recruits.

46

In Nyakyusa context Nyakyusa, like Kinga, honour the idea of aristocracy as a prop to the political establishment, but apply it hardly at all in social relations. When pushed to explain who would and who would not have claim to aristocratic standing, the Wilsons’ informants focused on first degree kinship to an hereditary ruler (prince or chief: omalafyale). But the practical test is more liberal. A man must have no recognized kinship of any kind to the ruling line, if he will qualify as candidate (at the Coming Out) for the headship of a commoner village. It is an office second only to the chief’s but is not hereditary. It is coveted but not dangerous like the office of chief, and not mystically guarded. The most economic way to model the phenomenon of aristocracy for Nyakyusa is to start with the myth of charismatic immigrants from Kingaland who founded the first chiefdoms. Then take the Coming Out ceremony as a re-enactment of this founding transformation of the political order. It is easy then to follow the logic of the Coming Out. It calls for a fresh beginning with a new set of chiefdoms, starting from scratch and with the improbable project of building a large and prosperous society just as was done in illo tempore. Thus among Kinga the royals, bearing a rapidly devalued Sanga name, quickly merge with the common folk and enjoy at court no more than a modest measure of ‘primacy among equals’, while among Nyakyusa the case is different only by reason of its formal coloration. High office must be seen in both countries to be inaccessible to all but a chosen few. To put this into the form of a practical rule you have to take office into account. Anyone in high office is an aristocrat in the telling sense that he is thereby presumed to be the heir of a solid line of men holding that office. Blood lines and primogeniture play a less cogent part in the continuity of this aristocratic thread than the ceremony of investiture itself, which is enacted in the fullest religious sense by the old prince-priest-chief. So predictable is it that the Wilsons would have been told succession to this highest office was by primogeniture, pure and simple, that I feel obliged to take it as a Just So story. The Kinga said the same about their succession but then made it clear on enquiry the heir unkinga must be deemed by the gathered avanyivaha to have met their every test of aptitude for the office. As in the Kinga case, in the Nyakyusa it is only in the very act of seizing the man and ritually investing him that the court elders announce a succession. Two chosen sons, each putatively by one of the old prince’s two chief wives, represent the 47

two ‘sides’ of the old chiefdom sectioned out before their births, and are fully proven adults at the time of the investiture. Since the whole chiefdom (both the elder community and the several new) must have faith in them, it is unrealistic to suppose unhealthy or uncharismatic individuals would make the cut. Younger brothers, in any event, inherit an office before sons if a chief dies in the decades between these political ceremonies of renewal; and a first-born who died young or otherwise failed to make his mark would never be remembered in a royal genealogy. When the whole idea of aristocracy is to strengthen the charismatic aura of high office, when the crucial time for this is the occasion of a coronation, and when the time and circumstances of the devolution of power on a chief’s heirs are most carefully controlled by a gathered host of elders, commoner rulers and priests, then politics can hardly fail to temper its rules with pragmatism. In short, primogeniture is the model rule: just so. Everyone knows if an ill-suited man is made chief the story has a dead end. Nyakyusa numbered in the 1930s about 150,000 and comprised about 100 chiefdoms ranging in size from about a hundred persons to three thousand. The many ‘minor chiefdoms’ were dead ends. They would hold no Coming Out, take no new cattle, and hold little land even while perhaps surviving for several generations. The Wilsons do not try to explain why chiefdoms still in the 1930s consisted in a group of age villages bunched close in space, with wide open (“empty”) country separating them from neighbours. But worldview is a basic condition: the saying, ‘a village consists in men, not in land’ (B: 45), is a prime clue to the way a chiefdom must be built. Nyakyusa forty years into the pax were only beginning to move into the open lands. A chiefdom in past time was the project of a single man and the achievement of a single generation, destined after its glory time to end in fission and new beginnings. The fledgling chiefdoms created would be led by two (or possibly more) young upstart chiefs competing for followings, cattle, wives, and the land which such prosperity can bring under cultivation. The Wilsons were told that at the Coming Out each new chief was told just where to settle his two wives and build his two stem villages. Doubtless this worked out in the 1930s under the pax. But before 1893 the Coming Out was the start of a new power game and a turbulent reorganization of space. The best plans of the elders could hardly amount to more than a starting drumroll. Mwaipopo’s chiefdom, which the Wilsons knew well, began as four, with the Coming Out of four chosen ‘brothers’. Mwaipopo conquered the other three and reduced them to dead-end subordinates of his own. 48

Had they not been seen as brothers it seems they would not have been treated so well. Throughout their accounts, the Wilsons make clear that defection was a constant concern of any chief; and the obverse of that concern would have been the countervailing ambition to recruit men by defeating their leaders. We know the anticipated prize for a Nyakyusa raiding party was cattle to be traded for wives; but a continuing deficit in cattle would mean a deficit in morale, and greater risk of defection. Nyakyusa men born in a stem village grew up from before puberty in (branch) age-villages—most of them consensually governed under a few older peers, their intimates—and were expected to work and fight through early adulthood for the primary benefit of the generation of their fathers clustered apart in the stem village. We know the active chiefs over a wide area were competing for a lion’s share of the same wealth. The masters of this game were the group of fathers. Their grown sons worked their fields and, where things went well, augmented their herds in order to bring them more young brides. This is the kind of game in which nothing succeeds like success and, the obverse, failure can be total. In the early stages of his political career a faltering chief would have to fear losing not just a few men but the loyalty of whole commoner villages he had failed to keep on close rein. Hence the spatial organization of the mature chiefdom was to surround its tree-shaded villageclusters with open country: as much a semiotic enclosure, I think, as a military defense. Hence also the importance of depending on the bachelor men to work their elders’ fields (where Kinga depend far more on women) and take food from the women of their paternal stem village. But a big, successful chiefdom at its height would sprawl far beyond the close-knit, concentric unit of its earliest stage. The sprawl is one of political influence and alignment not rule: no one village rules any other. But the key to understanding this picture is to have in mind that overall demographic expansion in a chiefdom, if there was any, was not what drove the Coming Out of client-rulers in the explosive manner it was managed. The reason for the explosion was structural. The institution of the age-village, with the prolonged bachelorhood of men it fostered, made this sort of explosion unavoidable. What the Coming Out did was make the explosion a controlled one: a ‘restructuring’. The Nyakyusa prince, like the Sanga, was bound to overproduce offspring. But this was not a technique for expanding his realm through colonization. A chiefdom as described in the ethnography comprised the ruler’s own pair of stem villages and (at the height of his administrative career) eight or a dozen satellite villages, all of these ruled by commoners. Commoner villages also would have by 49

that stage each its own group of younger age villages gathered around. In a chiefdom matching the model of the Wilsons’ informants, all the established villages would date back to the last Coming Out, and all would be on the same ‘careers’ track as the prince and his commoner village counterparts. Their strategy was to collect women and prosper, holding the next generation of males in thrall—extending as long as possible the social dependency of bachelorhood. With the next Coming Out ceremony, after thirty years and more of ascendancy, a now-reduced generation of male elders will grudgingly contract to ‘retire’ in favour of a more numerous and energetic generation, the greater proportion of these new men still unmarried. This is a deep generation, ranging in age from about ten to forty. They have to be re-sorted and their villages relocated, to make up two or more fledgling chiefdoms. This means stripping all the old villages, chiefly and commoner, of their satellite age villages. This pools all the men (say) between fifteen and forty into one ascending generation. What remains then of the old princely realm is an odd set of ‘retirement villages’ peopled by mainly old men, each with a spate of wives, young and old, and a host of children under the age of puberty. We are left to imagine how these old polygynous clusters, left now without the steady support of a young bachelor set, will break up. Odds-on, this part of the process will be erosive, not explosive. Most of the women with young children will be needing new and younger husbands prepared to do their heavy gardening and house building. The model seems to say the old husbands ought now to oblige by dying off quickly, but the model can’t be trusted on such a call. What can be trusted is the informants’ constant refrain: everything begins over again. Men of the right age and character, this suggests, will be ready for this once-in-a-lifetime opportunity. They won’t just stand around waiting. I should be quite surprised to learn either that the rate of adulterous elopements did not take off with a Coming Out of young men, or that an eloping couple in the new circumstance did not find safe haven in a chosen young man’s new village. Turbulence of this kind stands in strongest contrast to Kinga social life, where domesticity, though structurally minimal, is paradoxically stable. There is no evidence at all that the rampant polygyny of the Nyakyusa prince produced an aristocratic class or directly contributed (in the way princely polygyny at the Sanga court did) to an overall expansion of courtly political culture at the expense of ‘bush’ neighbours. Indeed, the court-bush circulation of the Sanga system is missing; the near-equivalent is a constitutional rule relating ‘royal’ politics to ‘commoner’: it is a rule of the ‘conflict-of50

interest’ type. It seems a born aristocrat cannot beget the future headman of a commoner village. That is, any young man with demonstrable aristocratic claim is ineligible to lead any but one of the two royal villages of his realm. Nyakyusa women like the Kinga spaced their births generously, and would have done so in monogamous unions quite as easily. The many male offspring of a wealthy and successful prince were in fact ruled ineligible for leadership office, as were their offspring in turn. But the prince’s polygyny did set a pattern for his peers. Where the magnified heterosexuality of a Kinga prince had no resonance in the moral strategies of ordinary men, the climax of the good life for ordinary Nyakyusa men was always polygyny, though it always started late and might continue so long as to become an obvious burden. The two decades of young manhood preceding his still-busy but settled life as an elder are the ‘political’ decades of a personal lifecycle, during which a new chief must establish his own sphere of influence. Nothing in the records from the missionary period or the ethnography suggests the morale of these young men was not consistently high. Married young women with their under-aged children will be moving by inheritance or otherwise from the old men to the young. Cattle will be passing in and out of private byres. When the Wilsons are told young men were kept in the bachelor state to keep them fit for fighting, it is a stretch to interpret the remark as a reference to sexual repression. It is more likely, at least, their prowess as fighters owed something to being unattached and unburdened with home duties, and something more to the well-known social rewards of manliness. High levels of testosterone mark the front-line fighter everywhere through history, as does sexual opportunism and a certain creative naivety about how power is won and held in the world. If there is an overclass in the Nyakyusa world it is, as we have seen, a compact local gerontocracy always specific to its village, not a nobility. The decisive evidence against a Konde aristocracy is the conspicuous autonomy of the ‘commoner’ villages and the opportunity they represent for meritocratic careers in politics. That village headmen are, in their own community, everything a chief is in his—war leader, judge, and ritual celebrant—makes every village its own player in the pyramidal polity of a Konde chiefdom.

(3) Bachelor warriors—late marriage Both Kinga and Nyakyusa rulers had at their disposal something very close to a standing army. In both societies most men remained bachelors, living hard by the chiefly courts, right through their fittest years. As the season or occasion demanded, 51

the court’s men cooperated in clearing or working fields and in the various tasks of maintaining court properties. In both cultural locales the rationale for late marriage was political: military and more. Whether the men were living (seasonally) in barracks or in segregated age-villages, there was a clear political advantage in having a compact, mobile workforce close to the court. The typical peasant division of labour gives structure to the household; among households the structure (even in the classic ‘kinship society’) is uneven and for political purposes may prove unmanageable. Kinga and Nyakyusa societies with their systems of delayed marriage added within the ‘peasant’ system a political division of labour centred on the village rather than the household. This allowed for de-politicization of kinship, and prolonged political socialization of the men to the advantage of central direction. Formally, the two societies used rather different institutions to achieve this result, and the sexual side of bachelorhood in the two social schemes differed accordingly. I know of no reliable rules for reconstructing the sex life of any human community on the basis of hearsay. Unless there are signs to the contrary, I assume in any case that unmarried adults will be accorded acceptable alternatives to going without sex. I expect such alternatives to be standardized, and the kind of inquiries which normally yield information on covert institutions to suffice for finding out in principle what these legitimate alternatives are. Beyond that, whatever one is apt to learn will usually have a speculative feel. Intimate sexual information, the matter of which tends to defy verbal formulation in any event, is generally guarded or misrepresented, leaving us with no reliable informants on the norm. But the quality of gender relations in public is normally observable. Unluckily, my gender was wrong for easy relations with Kinga women, and Monica Wilson’s wrong for chatting up Nyakyusa men. The area of uncertainty in both cases is the one least accessible to direct inquiry: for me, lesbian relations among Kinga maidens and married women; for Wilson, homophilic relations after the earlier years of a man’s bachelorhood. As for Kinga, intuition told me at the time that close attachments were as important for sleeping arrangements among young bachelor women as men. But for a married woman while nursing a child I found it credible enough that “a woman needs no more.” Many or most women marry to a new community away from old girlfriends of the isaka. Always when the four or five years of nursing a small child are done, and begetting a new one is on the domestic agenda, a Kinga man must prepare himself for the challenge. That is when a wife wants his best effort and will show her keen interest in the quantity as well as the quality of his performances.

52

I hesitate, of course, to borrow from the Kinga to fill the lacunae in their neighbours’ account. What is especially clear on the hearsay evidence when coupled with observable gender relations is that Nyakyusa men turn on to women more vigorously in young adulthood than Kinga men do. It seems likely this was especially so in precolonial times, when a young man’s strength and vigor gave him a stronger hand in wrenching cattle and women away from his father’s to his own byre and bed. But I see a special implication in the ‘initiation-marriage’ custom of secluding a friendly group of boys and girls in an isaka (“girls’ house”) all to themselves for a few days and nights of self-education. Both genders, it seems, want that infusion of heterosexual eroticism.

In Kinga context A Kinga person of either sex will have a ‘vagina’. Whereas in women they are found under the pubic arch, men have them in the backside. This is an important matter with implications for each major facet of Kinga culture. I have discussed the subject at length and in various connections in Twin Shadows and Four Realms, and refer the curious reader to those earlier volumes. Late marriage was a condition of the military stance of the Sanga courts, which depended on a standing crew of unmarried men living in barracks and comprising its all-purpose political workforce. While readily countenancing marriage in mainstream fashion as a universal institution sponsoring the legal and moral identity of human offspring, Kinga were distinctive in keeping adult men and women sexually apart through the whole high-libido curve of their physical life cycles. The rarity of such systems outside the urbanized world is not hard to explain. In the Kinga case, since ‘delayed marriage’ affects both genders, the puzzle for ethnography is the other way around: if they can do it (as indeed they did) without backlash, why don’t many other societies do the same? The best short answer might be that very special conditions are required for building up to such a system, and the same applies to maintaining it. I find it most fruitful (as argued in Twin Shadows) to consider the matter in the light of individual moral strategies. This is a question of feeling one’s way, now gropingly, now planfully, toward a rewarding life. The process begins with dawning adulthood and continues to the end. The governing conditions for Kinga are special, since neither gender is domiciled through childhood as part of a ‘family’. Well before adulthood begins a boy is at home sleeping skin to skin with peers of about his own age. The same applies to girls, and 53

individuals move smoothly into rewarding sexual relations within these same-sex peer groups. Sexual frustration as such is not a factor in these lives, though libidinal dynamics of course pervade the period of youth. When I deny that peer relationships were possessive I don’t mean they weren’t warm, only that the obvious shine of a new relationship was soon normalized as friendship—a personal involvement wanting good will, not demanding fidelity. For Kinga, it is important that most men exposed to the traditional court culture chose to accept its norms and had little interest in either early or plural marriage. The main inducement to marry was the need for a self-reliant existence after one’s youth was spent. The gender division of labour meant that a middle-aged man without a wife had no career as an independent householder. He couldn’t feed himself or brew beer for entertaining friends. He would have no hearth of his own. His stature and self respect could not be maintained. His former bachelor life with age-peers would feature in the lives of his now-married friends as so much fallen fruit. Before the pax there was no migrant labour to fall back on. The example of the polygynous prince or local ruler presented a reasonably positive prospect in marriage: Kinga wives are prodigious providers, and their husbands have lots of free time. For women in the traditional court culture the propriety of marriage was more important. Conceiving a child meant marriage. Bachelor women would have grown up under the tutelage of their older peers who, as married women, continued to monitor and school them in more formal ways on the specifics of a woman’s destiny. The moral lesson was unambiguously in favour of fertility, and marriage was the means. The Sanga ruler was interested in plural marriages and plentiful procreation. Sons would become the privileged men of his barracks, daughters the most productive segment of the daily workforce in his fields. Only by making his royal village obviously strong and productive could a prince or local ruler keep it attractive. Good political theatre can’t be produced on the cheap. Extending the bachelorhood of royals of either gender required extending it for both, and for both this entailed institutionalizing homophilia. This was the case for a royal village, whether on the scale of the Realm or the smaller Domain, and the centripetal attraction of young men away from outlying hamlets meant much the same economic logic would apply to the age of marriage for women in the peripheries. The men who courted them would most often be seeking a long betrothal, and I was assured by young men that this was readily accepted among the maidens. Everyone knew that with marriage the bachelor life was ended for both partners: a man must settle down to clearing, owning, and tending fields, keeping animals, and maintaining 54

a household for his wife. Domesticity itself was less complete for Kinga men than women, but marriage for both was a break—and for women a pretty clean one—from bachelor ties and bachelor ways, beginning a new phase of the life cycle. At the same time, delaying marriage was (at least, away from the court itself) a voluntary pattern not a prescription, and there is probably no knowing what the statistics would have been. The main thing to have in mind is that premarital pregnancies were systematically avoided, the first conjugal embrace legally enacting the marriage. Her maidenhood lost, a woman dared not—for she and her girl friends had always lived in the shadow of this taboo—return to sleep that night with her peers. Her commitment to a heterosexual life had a measure of the seriousness of trauma. Two queries arise about this picture when it is compared to other Bantu marriage systems. Did women really keep their virginity for a decade or so past the menarche? And supposing they did, how did they manage to keep up a sufficient level of fertility to satisfy the equations of population survival and growth? As to the first query, both the housing and working arrangements of women favoured peer control over their sexual contacts; and unless we contradict the uniform testimony of male and female informants we have to accept that pregnancies outside marriage were at least as rare in the past as they were in the communities I knew in the 1960s. There none of my informants could produce an actual case. In face of a weak cultural emphasis on the consummation of heterosexual desire, and quite general approval of non-exclusive, non-possessive homophile relations for both genders, I find it easy to drop my own folk skepticism on the matter of long-preserved virginities. So I find the second query the more interesting of the two. It has been a rule of thumb that anthropologists should expect an average spacing of births in the range from 18 to 30 months. Typically, a husband is expected to press for early resumption of intercourse after a birth. Natural conditions impinging on survival are quite often supposed to be severe—nature itself is thought to call for very high rates of reproduction to maintain effective fertility in a human community without ‘modern conveniences’. Yet there is the Malthusian view: if the fertility rate is moderate, the need for ‘positive checks’ such as warfare will be less—and so on. Kinga moral strategies meet Malthus half way. A Kinga woman aims to space her pregnancies at about 60 months. When she has produced four infants and brought them to full weaning she is reckoned to be finished with the conjugal relationship. Weaning waits until after the infant’s fourth anniversary. The rationale is not (as one might suppose) that a woman loses her fertility at this stage or risks her 55

health with a further pregnancy, but that her conjugal duty to her husband is finished. Heterosexual relations in marriage are for both partners mainly about reproduction. Erotic satisfaction is an important by-product. There are no reliable statistics to clarify questions about life-expectancy for men or women, or survival rates for children in traditional times, but I think it unlikely a woman would have passed child-bearing age before her mid-forties. Menarche in the nineteen sixties still came rather late (17 or 18 was a typical expectation, 15 was thought early) and women in their thirties commonly showed impressive youthful vitality. In the Sanga court culture I doubt that many women quit the bachelor life before their mid-twenties, though in the full bush culture (or in the Kinga-Pangwa overlap region) marriage would normally follow a few years after menarche. I think the point to take here is that the Sanga protostate created a healthy moral and social environment by extending the life situation of youth as it did, and limiting births. The general good humour and good heart for life of Kinga men and women, regularly attested by observers, reflects this. To this picture it must be added that Kinga women devote themselves to a nursing infant virtually to the exclusion of those already weaned. These older children no longer live with their mothers but are cared for by elder peers by day and (in a children’s sleeping house) by night. A visitor from Ulaya might suppose that parental neglect of a child, once weaned, would threaten its survival. But Kinga evidence supports a contrary view: the undivided attention of the mother during most or all of the precarious ‘first five years’ seems to make a greater (and positive) difference. It is also important to the fertility of the Kinga that daughters, a few years after weaning, have begun to spend more and more time, when they aren’t caring for small siblings, in helping their mothers with food preparation. In my observation, the mother-daughter relationship during a girl’s preadolescent years is the closest of all family ties among Kinga. I found the feeling-tone more sisterly than motherly. The next-close family tie is that between older and younger brother—certainly not that between father and son. A boy is expected to ‘grow wild’, fend for himself, and thrive on risk. Usually the near brother will be five or ten years ahead and inclined to be his brother’s guide to what comes next. The male role in mammalian procreation can require, in its minimal form, next to no distraction from other duties. In a human social order the problem is to make room for paternity in a man’s conception of those ‘other duties’. Kinga do this, as most other societies do, primarily through the institution of marriage. But the motivational context of Kinga marriage doesn’t include that 56

dramatic heightening of or conversion to heterosexual values in early adulthood, which anthropologists so often find elsewhere. Much world and time is wanted for conjugal arousal. The comfortable domicile a non-Christian man built for his marriage in 1960 was still regarded as hers, not theirs. For men wherever the court culture prevailed, the bachelor sexual adjustment was complemented not replaced by a monogamous marriage. A woman saw to spacing her children at four-to-five year intervals, and in that interval didn’t expect to sleep with her husband before she was ready for another pregnancy. The more Spartan quarters of a men’s dormitory (barracks) suited him, in any event, when there was a nursling about the private domicile. As for women, it was the man’s story that nursing mothers were satisfied without a sexual partner and only displayed their prodigious sexual appetites after weaning the child. Lesbianism after marriage, where it became explicit, was painted as a refuge for women afraid (from traumatic experience) of pregnancy. I find the deep bond of nursing was set up for a Kinga woman by absence of a deep-reaching, selfpossessing, sex-reinforced relationship of reciprocal dependency in the marriage bond.

In Nyakyusa context Where the age-village system was in operation among Nyakyusa, delayed marriage for men was a function of the value placed on polygyny and control of bridewealth cattle by a senior generation. Girls in the 1930s were allowed no extended youth but did in more-traditional areas enjoy at least an abbreviated ‘courtship’ phase. This was a time of closeted license which put a pleasant end to childhood. Before the pax young bachelor men had been at once protectors of their own and raiders of distant herds, so standing in a position to constrain their elders’ monopoly tactics in the bridewealth game. Though we know young men felt relatively powerless as the game was being played two generations later, we can’t quantify the difference pacification actually made in the typical marriage age for men. The Wilsons estimated men normally would be thirty to thirty-five before they were done working for their fathers and able to set up as married land-holders. Calculating on that basis, this would mean the older ruling generation was then twice that age, and the Wilsons’ evidence suggested this would have held for traditional times. So the change the colonial peace made was to take their leverage away from youth and make gerontocratic control of the women air-tight. We know Nyakyusa shared the Kinga institution of a girls’ house (isaka in both languages), and the Nyakyusa-speaking Ngonde retained the institution, delaying a 57

maiden’s marriage in more nearly Kinga fashion at the time of the Wilsons’ fieldwork. But there is no evidence from Selya that a girl’s bachelor phase was ever expected to last more than a few years into her full adulthood. In fact, the ample evidence of male dominance there in adulthood indicates a pattern distinct from the Kinga, as does the Nyakyusa man’s strategic career interest in plural marriage. But the arithmetic of the calculations about marriage age has to be challenged. The Selya ethnography seems to give us a nearly used-up generation of men in the stem villages, by the time they are willing to turn over power. But at the same time the age-depth of the new generation runs from (say) 15 to 35—twenty years. Applying this depth to the senior generation we get men in a range from 45 to 65. They are sectioned into age cohorts, of which only the eldest would in any sense be ready for retirement. Given the inheritance rules, a fair number of the men even in the eldest cohort would be younger brothers replacing a deceased founding member, and the process of replacement would be continuing. All this gives us a new arithmetic, and the gist of it is that the new generations setting out at a Coming Out have to be seen as planting satellite villages to grow up beside and among others still firmly established and still far from ready to ‘step back’. It seems we will have to be satisfied with our own mental sketches of the ground-plan of a chiefdom about five years after a Coming Out. All we really know is that the more senior of the bachelor age-villages of the chiefdom have been reallocated, and there will be a flow of energy from the new sites and a flow of cattle and women to them. We have a ‘world renewal’ by installments. Unless we assume that homophilia persists in an age village, though scorned as childish, right up to the point of its Coming Out, we should be asking how these young men achieve their high quotients of success in heterosexual enterprise, which is so evident later on in their lives. An obvious implication of the Nyakyusa arrangement would be that mature bachelor males were able to hustle themselves informal access to married women. It is difficult otherwise to understand their evident relegation of homosexuality to youthful insouciance. By all accounts, young Nyakyusa of both genders honoured public evening occasions in a euphoric and convincingly heterosexual style. By many accounts, aging polygynists had no personal magic for satisfying the sexual requirements of young wives, who normally spaced their children four or five years apart and had rich opportunities for dalliance. The facts as we know them give no confirmation for the notion that the young warriors of yesterday were compliant in being frozen out of the heterosexual game to the degree that, forty years later, even Christian young men still were. 58

Kinga women lived more fully within their own world than Nyakyusa women could in theirs, and Kinga women went about within their world with a far more autonomous air as individuals. Kinga men fully accepted their moral exclusion from that world on any terms short of marriage. Since attention to good manners is, but docility is not, characteristic of Nyakyusa masculinity, the point to take is that as long as bachelor men in their age-villages were seen to be compliant the system would work. It is a corollary that the peculiarly explicit moral emphasis on peer solidarity, without which the whole age-village system could not have worked, begins to make more sense when we see that the bachelor life was undeniably a good one, maximizing and prolonging the best rewards of youth. In this the two social systems of the Rift Valley floor and the Livingstone Mountain slopes are matched again. But you have to go far, to a region with developed age-class systems, to find a similar treatment of men’s youth and young adulthood. And that age-class system fits into a quite different political nexus.

(4) Amity over kinship The kinship-friendship balance in any human community is apt to be a delicate one, even where the slant of public rhetoric is onesided. I have discussed the importance of amity for Kinga and some neighbours in Twin Shadows, and the structural importance of the two kinds of personal tie in The Idea of Social Structure. Here I am concerned with the prominence of amity in Malatan political history. I believe the kinship ethic as such was a weak force in the region from early times. The main reason would have been continuous ethnic mixing under a slow pattern of migratory drift into and within the Rift Valley Corridor and associated Tanganyika highlands. The social fabric there is loosely woven. Individuals pursue their own moral strategies, their careers not fore-ordained or steered by groupthink. Neither men nor women live their lives out in the group they are born to. In any case, the unit of local solidarity and military power is— when you focus on actual membership—no proper ‘kin group’ at all. Students of African social organization will see the significance of this in relation to the rise of a Sowetan political archipelago. Throughout the broader region, the emergence of chiefly authority is associated with small local communities lacking clear-cut ethnic identities. What is foreign to Sowetan society is the phenomenon of clanship, where the clan is taken as a political building bloc deriving its strength from pyramiding, bred-in-the-bone lineage loyalties. 59

That means the classic problem of achieving a transition from the particularism of kinship to the pragmatism of political authority is finessed. There was no easy way the prepolitical egalitarian communities could have developed a translocal scale of organization through ‘filiation’ and ‘marriage rules’ when they had no solidary local or translocal unilineal kin groups. Clans, even of the decidedly impure sort we have learned to anticipate under the rhetorical surface of life in ‘tribal’ Africa, were not available in a population area settled by filtered-in, small-group colonization and imbued with a restless and individualistic antipolitan ethic. I don’t argue George Peter Murdock’s conception of Subsaharan Africa, with its social dynamics ever tied to rules of descent, inheritance, and marriage, is irrelevant here. Kinship is in the mix, though more important elsewhere in the region than among Kinga. But friendship is the prime mover and main bulwark of political loyalties. Consider the implications of an antipolitan ethic: the man without friends in the community will move away from trouble there, but where a man has many friends a ruler can’t afford to lose the one for fear of losing many. Rulers must be seen to be strong, but can’t afford it at the cost of being found unjust. There are reasons for my readiness to read backward from the cultural importance of amity ‘today’ to the regional culture of those misty times when the ‘African middle ages’ were new. A direct transition from kinship-organization to political is difficult to model. Starting from the kind of kinship-friendship balance we find in the regional bush cultures makes the modeling easier—the shift becomes more credible. Taking a very broad view, the main historical variable in accounting for differences between one Eastern Bantu region and another may be the settlement pattern. In some regions there is a history of ethnic consolidation on a large scale, often in response to organized external ethnic pressures. There have been fairly massive migrations. Then a pyramiding ‘clanship’ politics can be the mechanism for the rise of a ‘primitive state’. But in the Corridor region there was a continuous percolation of small parties into and through the Rift Valley, mainly but not exclusively seeking southward. Though the earliest movement of Bantu groups into and within the region are believed to have been eastward, oral traditions concerning later political colonization and the origins of local ethnic mosaics point to a subsequent drift of small groups westward (notably through what became Kinga country) toward the Rift valley. The comparative turbulence of Nyakyusaland’s politics, and the multiplex quality of its social and religious life, owe much to its geography. 60

If Eastern Bantu peoples in less than 500 years spread three thousand miles southward through sparsely settled lands from the equatorial interlacustrine region to the Transvaal, it was not in surges, waves, or mass movements. We aren’t dealing with conquest, invasion, population displacement. In one form or another, the original Bantu pattern of settlement in the Sowetan region would have been migratory drift: discontinuous, purposive movements by small groups toward proximate goals. The dominant motive for this kind of migration is scouting good living conditions, and this kind of pragmatic trial and error cumulates over many generations, given a thin population of indigenes spread over a vast area. This is the pattern of another well-known migratory settlement of continental scale, the peopling of the Americas: after a first, thin peopling of the continents by foragers, a gradual repopulation occurs through migratory drift by carriers of more effective technics. The extent and speed of movement is explained in the Bantu case not by finding empty country but more likely by colonizing already lightly-settled areas with the right tools and techniques to exploit them more productively. The result in the varied geography of a continent will be regions of migratory cross-currents, backwaters, and mainstreams of movement. The last may eventually see larger, concerted migrations in force; but for Eastern Bantu that would mean in their Late Middle Ages. For the Malatan region specifically, it is now pretty clear if only on linguistic evidence that there never was a north-to-south mainstream flow through the Corridor, and that pressure on the resources of nature was still moderate at the time of the pax. What seems to have happened in the Sowetan region after a period of light settlement is continuation of migratory drift to, across, and through the region of the Rift valley and its flanks, much of it favouring hill country. The period I have in mind now is an era of gradually developing availability of locally produced iron tools and experience with new cultigens, from about 500 a.d. to 1500. I take it to have been a long period of gradual and mainly peaceful change, while new tools and technics were making more intensive agriculture possible, specially adapted to the region’s ecological microzones. Cattle and small herd animals were supporting new prosperity wherever they could multiply; and this was true over the whole extent of Bantu civilization. In the Sowetan region, relatively isolated from major (Bantu and non-Bantu) ‘tribal’ movements, I take it political development after 1500 a.d. until the middle decades of the nineteenth century was evolutionary, not forced, but increasingly led by the spear. Much of the old freedom of movement continued in the Malatan subregion, and while transmuted by the 61

local political cultures was not replaced by ‘tribal’ boundaries, until the pax. The politicization of friendships supports a kind of individualism rare in kinship societies. Kinship systems, whether patri- or matrilineal, are important as rules for handling friend/foe relationships among men. Where kinship is weakly sanctioned (depoliticized) those relationships will hinge on other factors. One consequence will be that a man’s personal standing depends less on birth, more on proven character. This is the reason I must doubt primogeniture can be used as a reliable key to the actual pattern of succession to high office in Malatan communities. It is the very softening of birthright that allows Kinga to proliferate the Sanga surname to favour new lines of political loyalty without creating an unwieldy new dominance order, and the Nyakyusa in similar fashion to have an ‘aristocratic’ rank which fades away with disuse. Important members of the two élites, the (aristocratic/ exogenous) political and the (commoner/ indigenous) ritual élite, could each maintain some distance from lesser kinsmen, and so generally operate as independent agents unbeholden to factions within their community. Professional priesthoods emerged which were linked on a non-military professional basis and enjoyed a reach across political boundaries comparatively unencumbered by political jealousies. The political career of a Kinga prince could begin in peerdemocratic intimacy, moving through the collegial life of the barracks, to overnight ritual elevation to a status out of ordinary reach. In like manner, the Nyakyusa chief, after a life half spent in free intercourse with bosom peers, could be anointed successor to a mystically magnified office: henceforth, he will cultivate the respect due a man who is known to carry a python in his belly. Men will hesitate to cross him. The three cultural solutions (Kinga, Nyakyusa as of Selya, and Ngonde) worked out by the two élite professions are remarkably distinctive and, for that, the more instructive. In the Ngonde case, we find that traditionally all close kin-peers to the chosen successor to the Kyungu (theocratic) throne must be put to death by the priesthood. This could hardly be required only to protect the new ruler from hostile rivals, since the priests alone had the wherewithal of bestowing office. More likely, it was to cut the Kyungu away from his friends and his secular past—to execute an apotheosis. ‡ It is the difficult social engineering required of priests and princes in this formative period before the European pax which we must have in mind as context, if we’d understand how the politics of fear came to play out so distinctively in each cultural community. The 62

methodological problem is ‘archaeological’: reconstructing from fragmentary texts we chance to find still strewn like shards about the field of play. The idea is to reconstruct what is still knowable of their context, the once-living world from which they survive as scattered relics. Reconstruction is interpretive not demonstrative, and may be seen to rest on fuzzy logic, but without context none of the evidence can be assigned clear meaning.

In Kinga context The metaphor of ‘social engineering’ becomes appropriate (replacing the blander, autotrophic ‘social construction of reality’) as soon as political power is seen to have a centre. Amity is engineered in the Sanga court by eliminating competition among young men for females, substituting kudos; and by condoning non-possessive homophilia. Kinship loyalties are undermined in several ways but in no way disgraced. The system of recruiting young men to the court, even renaming them, reshuffles men (in a nominally patrilineal society) without regard to paternity. Land allocation is not by filiation but by a nominal contract of clientage wherein the linear representative of the original settler of a locality enjoys privileged status; but the term of his privilege is short, as land once granted does not revert to him, and the right of a patron is expressed rather in an economic than a political idiom. Ritual avoidance of intimacy is required between father and son, quite in the style of an add-on incest taboo; and there is no expectation that sons will settle by their fathers, though intimacy between brothers is specially privileged. In general, personal loyalties are to peers not parents. After weaning, a growing boy’s ‘mothers’ (women from whom he can expect nourishment) are so many he need hardly fear being orphaned, though the same does not apply to a girl. As the male life cycle turns around peer alliance but without featuring exclusive dyadic ties, so does the female. Kinga of either gender are expected to be open to new friendships yet faithful to old throughout their active years. Loneliness is deplored. Suspicion of betrayal by a friend is known especially to men and deeply disturbs them. I deem the royal court’s proliferation of ‘Sangas’ to have been a major strategy for damping the fires of kinship. There is a certain irony here, as the nominal function of bestowing one’s name on an ‘orphan boy’ would be to give him important rights in a kin group. But in this case the court in effect is making the person an orphan by the same move. He ceases to belong to his putative birth-lineage. Moreover, the massive production of male offspring at court, all of them calling one man their father and few receiving any tangible 63

career advantage over the talented ‘orphan boy’, dilutes the kinship value of the Sanga name in favour of a political meaning. The spatial dispersion of the settlement pattern for men still belonging to nonSanga surname categories, when combined with a de-emphasis of kinship ceremonies, further dilutes kinship to the point that chiefly politics has little to fear from kin sodalities which might oppose it. This is a point of radical difference between the Kinga and Nyakyusa protostates.

In Nyakyusa context Monica Wilson’s first account of Nyakyusa age villages, Good Company, takes its name from the insistence on amity among male peers who were expected to live in political community from boyhood to age. We are told that a man accused of veiled hostility might be exiled as a witch, though after prolonged residence elsewhere might return. How much circulation of this kind took place we can’t know, nor what mixes of motive may have been entailed in such roving. But taking an overall view I find two clear facts suggesting the internal movement of men and women had been persistent from the very start of the state-building centuries. A fair level of linguistic homogeneity characterizes the whole of the Ngonde-Nyakyusa territory; and the advanced political-ritual culture, which by the nineteenth century extended throughout, everywhere shows morphological continuity: the family resemblance of these institutions from chiefdom to chiefdom bespeaks their shaping by a translocal community of priests. Further, the wide political dispersion of agnatic kinsmen, persisting still in the 1930s, speaks for itself. Boys of a village in Selya worked together in their fathers’ fields and engaged together in raiding and warfare. But in the model life cycle there was (unlike the Kinga) little mixing about. Where individual Kinga would have each his different set of special friends, amity for Nyakyusa men was rather (in a quasi-political sense) comradely. Monica Wilson describes it as obligatory, enforced by ‘the breath of men’, which is thought to act (as a sort of floating mystical sanction) against secretive antisociality. All of this is pretty foreign to the Kinga way of thinking, which openly condones eccentricity and leaves grumblers to grumble alone. Peer solidarity in the age village does resemble kinship solidarity in its sanctioning of primary loyalties, and the father-son relationship is decidedly authoritarian among Nyakyusa though its local context is no ‘local kin group’. Kinship itself, however, is richly sanctioned by ritual—something quite missing from Kinga 64

experience, where a funeral is closer to a moot than a rite, and generates no festivities. Only a kin group with ‘central direction’ (e.g., an authoritarian family or minimal lineage) would possess the sanction of ostracism. For the most part, kinship societies depend on interlacing ‘loyalty networks’ not an authoritarian hierarchy, and the rights which come with membership are inalienable. Nyakyusa kin groups are of this sort: finite but unbounded. The age village, having no kinship nexus at all, only augmented amity, exiles but does it in a manner which tags it voluntary. A man begins to be seen as a witch. The ‘breath of men’ leads to confrontation. The man moves off to a far chiefdom. Years later he may return to his home village—he has not lost his claims there. This favours the antipolitan ethic of the region, which justifies a person moving on to escape from trouble. It also bears analogy to the tight sort of family system we know from the parable of the prodigal son. But as elsewhere, a witch can be formally accused when the allegation is hard meant. Then the accused will ask for the poison ordeal, which will be licensed but never initiated by the chief. To be exonerated a person must vomit. As the accuser also must take the poison, and as the dose is a harmless one, the innocent will know no fear. The chances then are better than even that one or both contestants will be found guilty and will be driven out. Losers must usually leave a cow or two behind, confiscated to feed and cool out the community injured by all this unfriendliness. In true exiles of this kind there is no voluntarism, and the exile is unlikely ever to return. Since blame has been fixed by a mystical truth-finding procedure, the bonds of amity can be terminated without leaving a legacy of kinlike ambivalence behind. ‡ The great lesson of this jural procedure concerns the structure of Nyakyusa society: exile is sanction enough (though men may say a witch ought to be speared) because the social world of the Chiefdom is the world in which amity must reign. Other chiefdoms of the country we now call Nyakyusaland (and think of as some sort of political unit) were always politically out of bounds. Though kinship crossed those lines, politically sanctioned amity didn’t. Games of intimacy and claims of right in the name of amity, as played among peers, belonged to the village itself, extending (as one may sense from behaviour at funerals) to other age-graded villages of a chiefdom under somewhat more formal rules. Brothers and halfbrothers were systematically scattered among all these villages, and a functionalist will easily find support in the ethnography for the thought that this diffusion of fraternal ties through the whole chiefdom would have been a major source of its unity.

65

Amity and kinship combine almost seamlessly in a society which gives primacy to the former. Parties of young men from other chiefdoms had access through kinship to important funerals, and would be expected to pick fights there. Agonistic encounters like this were probably as important as formal battle in strengthening political identity within a fully established chiefdom, but would have shown up the weakness of others. In times of turbulence, when the inner moorings of social identity and loyalty are not strengthened by stress they are weakened. As in kinship societies, where the puzzles of social history so often lie in explaining why some lineages thrive while others wane and disappear, so with Nyakyusa chiefdoms. The politics of amity is as dense a subject as kinship itself. It is only too easy to forget, in reading what Nyakyusa say about themselves, that every chiefdom was destined to split before its generational cycle ended, pitting brother against brother. This applies not just to the chosen leaders of each new ‘side’ of the old community but all the way down the line, for brothers generally grow up in separate age villages. All this is in spite of the fact that at death a man’s next full brother is the legal heir to his place in the world. Here as elsewhere kinship apparently survives betrayal better than mere amity. Analysis in these terms discloses a steep contrast between the moral careers of Kinga and Nyakyusa men. If Nyakyusa favoured letting a younger brother follow his elder brother into the same peer circle the result would be sexual intimacy between the two. This would entail deep moral dissonance in a relationship where inheritance rules allow an older son to move, on the father’s decease, into paternal authority over a younger son. To avoid this, Nyakyusa use the age village system to prevent such intimacy developing between brothers. The moral maze within which a Kinga boy pursues the good life features rather a father one must avoid—even avoid serving— and an elder brother one must love. So far as I am aware, this arrangement is unique to Kinga culture within the Eastern Bantu world, and fairly accounts for the uniquely transitive nature of a Kinga man’s privileged intimacy with a brother’s wife.

66

FEAR & BLAME, CHAPTER THREE

Sowetan protostates in regional context

Origins of the political archipelago The physical setting of the Sowetan political archipelago is the large inland sector of Tanzania still conveniently called the Southern Highlands. There is a relatively high rainfall area describing a broad crescent running south and west from the town of Iringa, curving finally further westward to sit over the northern end of Lake Malawi. For the most part the lands included in this sweep are high, ranging between 1000 and 2500 metres, but the lakeshore is lower at about 500 metres and the Rift Valley floor which Nyakyusa occupy rises as you move northwest from the lake to about 1500 metres. Apart from uNyakyusa, most of the developed political centres are high enough to be relatively cool and malaria-free. The highland vegetation as typed in the early twentieth century includes grasslands, especially wooded grasslands, woodlands, and forest. Rains range from moderate to heavy and the seasons from DecemberMarch to November-May. The better cattle areas are low-lying, the better gardening high. I speak of a ‘political archipelago’ rather than a ‘political belt’ or ‘crescent’ because protostate development appeared (and remained into documented times) in separate ethnolinguistic centres, island fashion, with buffer zones of independent bush cultures between. Historically, none of the five protostates I include in this regional study were ever cut off from one another. A simple proof of that is that Kinga smiths dealt with all of them. Communication from one end of the archipelago to the other was for the most part indirect. But information moved through plains as well as highlands; intermediaries were of various types; and the region as a whole was in touch with commercial centres on the far coast, and southward or westward with shore peoples on the two great lakes of which all these peoples were aware. I consider it certain that political thinking in the archipelago was influenced well before the nineteenth century by news of Arab and Portuguese commercial activities to south and north, and by tales inspired by coastal towns and great kingdoms as far away as Zimbabwe or the Interlacustrine region. The movement of political news from within 67

the archipelago would obviously have been of a more businesslike sort, but specifically political contacts were probably few, or of no great moment, before the nineteenth century. It is apparent that each case of political development calls for its own history and reflects both local peculiarities of place and culture. Events are precipitated interculturally by a dual process of social differentiation within and political expansion outward, which characterizes the chiefly organization of power. The same principles apply at higher steps of magnitude when you broaden your historical horizons, toward the ‘evolution’ of (say) Eastern Bantu civilization: the evolution of a cultural system is only history seen obscurely. A theoretical possibility still exists that we might acquire more cultural information bearing on the political past of each of the Malatan peoples, but the possible sources are rapidly narrowing to retrospective linguistics and archaeological studies. The more specific understanding we can hope to gain from ethnographically based comparative analysis is presumably going to be limited forever by what is already on record. While our ignorance is therefore, and will remain, abysmal, that does not excuse our applying what general knowledge we have of ‘social evolution’ (how things have gone elsewhere) to interpreting the texts we are given. Working on that premise, and working out perforce from my own fieldwork area, I have wanted to explore the political processes which have carried a series of bush cultures out of an egalitarian existence in separate pedestrian communities into what I call protostate patterns. These are systems of political organization from which a full turn back is unlikely. They lead on, or would if we can imagine them continuing so long uninterrupted, to the kind of largescaled human societies we know as statelike and ultimately as urban. There is something special to learn of human history (apart from the likelihood of interruptions) from such a case: we stand closer in time to the real circumstances of state formation with an Eastern Bantu people than with most. I offer no special suggestion as to why the archipelago arose where and when it did. It seems to me enough to have in mind that Eastern Bantu civilization some half-millennium ago had been reaching the stage where translocal trade and political development could be expected to occur in most suitable areas between the equator and Zimbabwe. I find it more appropriate than postulating general propositions about the evolution of all political culture to examine the hard evidence we have of how very different these nascent protostates can be, while arising contemporaneously and interactively within a single regional culture and in the larger context 68

of a remarkably homogeneous civilization. The Eastern and Western Bantu-speakers, taken together, comprise a recent overlay on diverse earlier African peoples, the product of a phenomenal migratory spread beginning hardly two millennia ago. The overlay process was one of assimilation and economic development not displacement. Within the continental latitudes we are considering, only a few islets of pre-Bantu settlement remained unassimilated when the ‘colonial freeze’ began. I found, to give an example, a little community of little people—somatically San (Khoisan), linguistically Kinga—continuing under Sanga rule. Archaeology will in its own good time tell us of many islands which had not held out so long, and perhaps of other such sweeping movements of rapidly increasing, ethnically linked, and technically advantaged peoples like the Bantu in Africa. What we can say for now is that there has rarely been such a long, wide spread of ‘stirring times’ and never so close in time to the rise of an empowered global curiosity able to investigate it. Considering the different paths this particular array of neighbour peoples pursued toward social complexity, I think the defining condition for the birth of a protostate process is not suzerainty or hegemony but authority of translocal scope: not a merging of one basic, pedestrian-sized community with another but their pyramidal alignment through indigenous rulers. The politics of suzerainty come later. The condition is neither allodium or feodum, since there is no comprehensive legal infrastructure against which the pertinent system of rights could be adjudicated or modified. The adjective ‘domanial’ does apply, since the ‘inward’ autonomy of each pedestrian community is primal, but applying the term doesn’t help explain the pyramidal alignment. In the protostates I describe there can be no presumption that the central domain a (ruled by a ‘high prince’ for Kinga, or a superior ‘chief’ or unrivaled ‘prince’ for Nyakyusa) can count on the support of domains b, c within the realm for help in putting down a rebellious lord d, who has refused to offer his wonted tribute. The boss is the boss because he can and will do the police work himself. Should he fail, the pyramid would persist only in modified form, even over time to budding off or accommodating to a new centre: it will not dissolve because it will continue to have recognized exterior rivals of a matching translocal order. In short, a set of translocal polities has evolved within the territory of an extended ethnic community; a founding premise is universal political rivalry among the pedestrian domains comprised, with higher-tier rivalry emerging among segments; and in each domain a legitimate, unrivaled ruling authority vested in a chosen person. The translocal authority is emergent, construing a constitutional past in light of 69

apprehended future needs for alliance, and only as strong as the practical expressions of such use, in the symbolism of tribute or inclusive tributary rites addressed to higher than mortal authority. ‘Divine kings’ if we have them at all among Kinga and Nyakyusa are not addressed as living gods. In the Nyakyusa case, the formative process will be rerun in the manner already discussed for each major political segment (realm, chiefdom) with the promulgation of a fresh set of local domains at about thirty-year intervals. This procedure is an accommodation to the absence in Konde constitutional thinking of a ‘demesne’, land attached to a domain, such that the polity can reproduce itself in the reproductive lives of descendent governments and their publics. Instead, failing the ‘natural’ route to corporate perdurance, new fledgling domains must be founded at appropriate intervals, centred in new cohorts of male descendants of the realm, whose careers will overlap, backward and forward, the old and the next subsequent political cohorts. In the Kinga case the situation becomes fluid (if much less regularly) whenever the dominance order within a realm is successfully challenged, and this inaugurates an expansive or reformative mode as new local rulers seek to recruit and accommodate new followings. In either case, the power-adjustments are inherently turbulent, but now in a warlike, political mode as the power to recruit becomes essential to establishing legitimate rule on the new plan. In either case, too, it may seem that turbulence must readily turn into chaos. It does not because the fighting is fundamentally intrasocial. Failing to understand that, with its premises and implications, has long been a stumbling block to our understanding of the role of force in political evolution. One reason is that a highly developed kinship politics with a rhetoric of its own has masked the control of turbulence under the guise of ‘tribalism’. That is, while ‘uncontrolled warfare’ could be seen to prevail at the margins of ‘tribes’, segmentary rivalry within any such community, being ‘intrasocial’, was limited in ferocity in accordance with the degree of ‘kinship’ entailed. What we have found in the Malatan cases under scrutiny unmasks the strategic nature of intrasocial temperance. In the case, at least, of Eastern Bantu civilization, an early expansive settlement process was effected by migratory drift and piecemeal colonizing stratagems, not by pushing forward an ethnic frontier or otherwise extending domination on a territorial basis. The Bantu ‘peoples’ involved in and resulting from this process were at any time loose alliances of intermarried clusters organized on a flexible or shifting network basis. The model which best fits the facts as we can know them would not presuppose ‘tribal’ or ‘band’ 70

loyalties, or stress a pyramidal structure of ‘unilineal descent groups’. The emphasis would rather be on a type of network ties making use of both kinship and amity, and creating clusters rather than bounded localities on the ground. Network ties of this kind are not cut off by common boundaries, since no two individuals (unless we think of newborn twins) are structurally placed quite alike. It is as local populations increase and shifting movements become less free that political units begin to form and stabilize as ‘local domains’ of pedestrian size with a recognized centre. But this first step toward the building of a protostate would happen, especially in broken terrain, only gradually and unevenly. The first appearance of translocal structuration would be in the familiar ‘tiling’ pattern (stretched and warped to fit a terrain) of settlements on the ground. The Sanga protostate was by most measures one of the stablest in the archipelago by the end of the nineteenth century; but German maps show its northern and eastern marches inhabited by peoples (Mawemba and others) too scattered and uncentralized to have earned administrative recognition under either Germans or British. Wherever there were not roads, missions, native courts, and colonial government stations, to the day the British gave up power in Tanganyika there remained a bush culture hardly touched by foreign contact. Within a day’s walk from the Sanga capital of Ukwama, scores of small hamlets had survived, hidden by the bush and terrain from all the amenities, indignities, and novel experiences of the twentieth century. All this is despite best estimates which put the origins of the Sanga courts and their unbush-like political system back four centuries. The survival in this mountainous country of a hidden bush culture is possibly the only real basis for admitting that Kingaland under the British was, as Boma officers at Njombe were prone to say, sheltered and ‘conservative’ by reason of its remote setting. Mountainous regions are famous for this sort of thing. Granted that these Livingstone mountains have no local ‘yeti’ myth, they are widely known for their flying python (dragon) divinities. It is true enough that the Sanga courts were spared the coming and going of traffic which so affected the main centres of Hehe and Bena life under the pax. But Kingaland was mightily disrupted by the 1914 war. Missions with their schools, government courts of record, and roads-in from Njombe town insured afterward that the main centres of Kinga population were far from fully sheltered. The peculiar sophistication of those centres, and the individualism fostered there, meant the decline of the old power system was matched by rising interest in the new. Young women, in particular, were drawn to the missions. Then with the growth in British times of tea and wattle 71

estates (far away in warmer parts of the extended region) bachelor men of the court centres took to migrant labour as a substitute for the old barracks life, whose rationale had disappeared. By 1960 this meant the great majority of adult men were fluent in Swahili and well acquainted with town life. Ironically, this left the defense of native conservatism to the fully disarmed old male power élite, unconverted women, self-secluding smiths, and well dispersed bush villagers. The importance of Sanga sophistication to the Kinga experience of the colonial decades stands out when you consider the Pangwa, linguistically and culturally the nearest of Kinga neighbours. Fr. Stirnimann’s two-volume ethnography of the Pangwa, concentrating as it does on folk custom not ‘deep culture’, virtually supplies my need for understanding the principles which would have informed the economic adjustment, the organization of community life, and the ritual practice of pre-Sanga Kinga. I don’t suggest ‘Pangwa culture’ replicates ‘pre-Sanga culture’ in detail. To begin with, that would mean reifying the idea of culture as applied on either side: Kinga are not made to a template, nor are Pangwa. Families, hamlets, sections, domains, and realms have each their own peculiarities on the Kinga side; and Fr. Stirnimann reports just such internal linguistic and cultural differences on the Pangwa side as well. But the greater absurdity would be assuming four centuries of Pangwa history would have passed without change. What can be supposed to have continued, as any archaeologist will bear out, is a basic ecological and demographic plan, a pre-medieval level of adjustment associated with acephalous politics and pre-intensive or shifting agriculture, and with open patterns of settlement. I have described this elsewhere as ‘level one’ in a developmental scheme for the region. Many sorts of change can accrue in a ‘level one’ community without any escalation. ‡ All the same, Kinga and Pangwa as partners in a persisting regional culture were in close contact all through the medieval and in particular the early-colonial period. Pangwa usages, as recorded in the 1960s on the basis of living recall by aging male informants, do match the Kinga similarly recorded then, and match them remarkably well when Sanga-specific elements are set apart. The match in this pairing is system-wide but is much spottier for Kinga when paired with other acephalous neighbours: Mawemba, Wanji, Magoma, and Kisi. Three reasons may be adduced and are perhaps equally relevant: The regional culture at the dawn of the medieval period, before the rise of a political archipelago, may have shown a similar areal pattern of variation following ecological clines; the Livingstone Mountain area shared by Kinga and Pangwa may have 72

better protected these two peoples from disruptive external contacts through most of the period; and (turning to regionally internal disruptions) only the Pangwa among these Kinga neighbours escaped being drawn—always marginally, always to some limited extent—into the transformative state-building processes of the archipelago. One can point to low population densities and the slow rate of change prevailing in these highlands. It was the British boma (bringing to bear a mandated government systematized only when the colonial era was half done) that saw an arbitrary Kinga-Pangwa boundary laid down, leaving at least one ‘Kinga’ community (per earliest German maps) alienated to a Pangwa court of administration. Boma documents say nothing of this. The Kinga-biased history I got was that Sanga rule was in process of establishment across the Mgiwi river before the Maji fighting erupted. If so, this account would confirm the southward direction of Sanga expansion from the Eastern Kinga realm, which a concerted oral history claims. But Pangwa were not so fortunate as to survive to the eve of the twentieth century in a quiet backwater. Where the Kinga were able to make the Ngoni invaders convincingly unwelcome, the thoroughly decentralized Pangwa were caught militarily unprepared and suffered radical disruption and displacement over the two generations before the pax. The best way to understand the Kinga role in these forty years of invasive disruption is to say the Sanga princes were able to meet attacks by Ngoni and later Hehe forces well enough to feel themselves safely in control of their mountain slopes region. While their spokesmen didn’t say so, knowing and believing in their story in heroic terms, the Sanga courts must have been greatly strengthened by popular fear of external invasion from the 1840s onward. The Wanji (north from the Kinga but still at the high-plateau level) during the same times of troubles were drawn inadvertently into alliance with the Sangu (another full step northward and downhill) in wars with Hehe, though it seems the fighting never quite moved up the escarpment into Wanji country itself. The Wanji economy may simply not have been fat enough to divert a lootseeking enemy. The Wanji communities farthest from the turmoil of the 1870s, and removed from the missions and schools set up in German times, came through the pax itself largely self-sheltered by their native isolationism from acculturative contact. The Magoma as they came to be under the pax were a bush community marginal to Mahanzi and Kinga, but had been politically 73

colonized by a small party of refugee Sangu warriors shortly before the pax. The Kisi on the lakeshore were potmakers and fishers who traded primarily northward along the shore with Nyakyusa, but traded also with Kinga. These three Kinga neighbours, Wanji, Magoma, and Kisi, were small ethnic communities assigned for administrative convenience to the Kinga Native Administration and its courts when Indirect Rule was set up (1926) by the British. As for the fifth and final of the acephalous peoples encircling the Kinga, one group of the Mawemba were also assigned to the Sanga court at Ukwama but managed to get themselves realigned as Bena subjects on the ground that communications eastward were less difficult. The Mawemba people have never been the subject of a systematic count, were never treated as a distinct ethnic group by the British, and have never been systematically studied. Like the Pangwa, they were always thinly distributed over large areas. Unlike the Pangwa, they were marginal to two developing protostates, the Kinga and the Bena, in circumstances which allowed them to play a ‘boundary man’ game. Still in 1960 they identified as Bena to the Kinga and as Kinga to the Bena, so hoping to be left alone. My Kinga informants assured me this had always been practice, and what bits of harder evidence I found did confirm the possibility. But the clearest benefit of their boundary-man status emerges when we compare their story with that of the Pangwa during the long period of turmoil initiated by Ngoni invasions. Mawemba country compared to Bena, Kinga, and even Pangwa is ecologically marginal, lacking the occasional verdant valley lands which can support extended settlements and some intensive agriculture. Pangwa country, which was virtually undefended, offered such lands, and the stronger invaders soon found them. For Ngoni, settling in the ambiguous lands (i.e. Mawemba) between two armed protostates could have been no real alternative even if some bits of the country had looked especially attractive. Will the full history of Ngoni adventures in the southern half of Tanganyika, when it is as fully known as can be, bear out my supposition that the Sowetan political archipelago was quite well developed before 1840? My reasons for supposing so are both morphological and narrative. Narrative in that my account of Ngoni fortunes, from their first entering the Corridor region to their eventual rassemblement in what became the Songea District under the pax, presupposes their settling perforce in marginal country among acephalous peoples—failing, that is, to hold their own and hold together in competition for living space with Nyakyusa, Kinga, or Bena just as they are known to have failed in their contacts with Sangu and Hehe armies. Morphological evidence for the time depth 74

of the Sowetan archipelago is discussed in Four Realms and elsewhere. The Sanga political system is a close cousin to the Bena and to the Hehe system described by Ernst Nigmann. The Sangu, as nearly as one can tell, are not of the same tradition; and what the Kinga have in common with the Nyakyusa is somewhat less on the plane of political than social organization. This leaves the HeheBena-Kinga chain in the archipelago an historical entity linked by overlap to another, the Kinga-Nyakyusa-Ngonde chain. Hehe were first brought together as an amalgam of more than a dozen independent peoples of indefinitely related cultures spread about the eastern part of the Ruaha river basin in south central Tanganyika. This massive alliance was brought about by a programmatic series of intraregional wars of conquest waged during the two decades after 1855, and reflects the success of one Munyingumba, who parlayed a small hereditary chiefship in the south into a regional military alliance in which he played the tyrant’s role. To interpret the few bits of the puzzle we can be sure of, it is important to know that all this consolidation within the region was achieved before any sharp, warlike external pressure would have been felt. We have to deal with a response to news of danger and high adventure. Partly, the goal would have been a rassemblement in favour of caravan raiding or taxing, in the manner of the exploitative front being formed even then in the Western region under Mirambo (the ‘corpse-maker’) of the Wanyamwezi. Munyigumba did not famously engage the Sangu until 1875-7, and had a final standoff with the Ngoni in 1878-9 shortly before his death. Yet the Hehe by that time were a coherent regional force not a loose alliance of local commands, and were engaged in the accumulation of booty by slaughter of their African neighbours in all directions. They were even then preparing to compete with the Masai for pickings from the caravan trade in ivory and slaves passing to their north between Ujiji and the coast. ‡ A long struggle for the succession followed Munyigumba’s death but actually ended with the emergence of a second tyrant of instantly heroic stature in the person of a favoured son, Mkwawa. The short work he in turn made of welding the quarrelling parts of a kingdom into a single nation suggests that war and the business of building a lifestyle on booty had become the focus of the new generation he led. Hehe military enterprise was in a class with the Sangu and Ngoni. Although the Bena, living between Songea Ngoni (to their south) and the Hehe, were swept into the wars, it was not by wouldbe tyrants of their own, and the effect was far from unifying their several chiefdoms. Explanation may lie in the character of Bena political culture. There as with Kinga the politics of fear seems to have tended in style more to the symbolic than the savage. I have 75

elsewhere discussed the way political theatre was able to sublimate neighborly aggression between Kinga realms, and the Culwicks report Bena sentiments to match: war “was the king of sports and the road to wealth and honour.” Whereas eyewitness reports from the Malatan region and from Kinga folk memory make this credible, another chapter of the same folk memory concerns blood and gore with no relief. This was from external wars where the ethical infrastructure we call community or a common culture was set aside, and the sporting metaphor quite lost. ‡

Acephalous communities in the regional culture An adequate study of the ‘bush cultures’ of the whole region, had it been done in time—centuries ago—would likely have shown a more even pattern of family resemblances, institution for institution, throughout the region than we can posit on the information we have. That is one speculation. Others come to mind if I pose the interesting question as to how populous these boundary-straddling settlements may have been in the special conditions of a medieval Eastern Africa—how far did some of them form as refuges from (for example) Sanga politicking? Then there is the difficult business of taking into account a factor which goes by such names as ‘cultural intensity’, ‘sophistication’, or ‘complexity’. This factor has been basic to my conceptual modeling of Kinga culture as having the two poles, court and bush. The metaphor of a ‘political archipelago’ makes little sense if we do not assume that institutional complexity in the organization of power brings with it a like complexity in the other major spheres of culture, so transforming the condition of a people with respect to its future—a transformation we are apt to measure on a scale of ‘emergence’ or ‘evolution’. The Pangwa, stateless and untransformed into the 1920s, compare for social complexity rather with the ‘orphaned’ communities of the bush cultures peripheralized by the rise of the Sowetan political archipelago than with Kinga court culture. What is special about the Pangwa is that both ecologically and socially they constituted a securely established and self-confident population but never an integral people. They numbered (most probably, about 1840) between twenty and twenty-five thousand, politically decentralized and scattered often thinly over an area larger than the combined lands of uKinga and uNyakyusa, which bore a tenfold population. The Pangwa had not been marginalized by history, only left pretty well alone. But this ended with invasion at the midpoint of the nineteenth century.

76

The official position on Pangwa in 1960 at the British colonial Boma at Njombe was that the name applied to an extensive, underorganized, and underpopulated area rather than to a distinctive people. The missionized Pangwa studied by Fr. Stirnimann were descendants of a half-century of turmoil beginning about 1850, brought about by Ngoni scavenger bands who scattered settled populations throughout the country, appropriating stores, livestock, women, and boys. Ngoni here and there set themselves up as rulers, exacting tribute in food and beer. Only in the early decades of German contact does Ngoni dominance disappear from the cartographic record on the northern (Pangwa) side of the Ruhuhu river. By then the still-unassimilated Pangwa who had been carried off southward as slaves and warriors had begun their return and were reestablishing their communities. Two Ngoni groups had reached uPangwa through uKinga, where Sanga forces had out-fought or outbluffed them. Pangwa had never been disposed to defend their territories; like ‘bush Kinga’ and Mawemba their strategy of defense was flight. Kinga settlements south of the Mgiwi-Nyangara river, which became the official (British) border between the two peoples, appear on the early German maps, where Pangwa locations are also marked. It is certain that Pangwa were disrupted and displaced throughout the land by Ngoni; in some parts it is certain that there was much loss of life, and many boys were taken as recruits; but most Ngoni moved on southward to another frontier and then to full-scaled warfare (especially over cattle) with Hehe and Bena. Oral evidence makes it clear the Kinga in what came to be called Pangwa country were settlers (colonizers) before all these troubles began. Later on, early in the 1905-7 uprising against the Germans, virtually all spearcarrying men of the Sanga Eastern realm were massacred and the land devastated. Hundreds, or perhaps thousands, of refugees escaped across the rivers to Pangwa country where they were safe, and where some children grew to maturity before they dared return to reclaim the land. It is as well to review this history if only to reemphasize the extent to which the heritors of colonial thinking have overcommunicated the integral nature of the Sowetan ‘peoples’ by supposing they were ethnic groups with territorial identities and hard-edged borders. The main difference between the Kinga of the Eastern realm and the Pangwa with whom they had always easily intermixed was the importance of the Sanga courts to men on the Kinga side of a (seasonally traversible) river. But a history of troubles almost always misleads by dwarfing the importance of social and cultural continuity. Time and the pax have turned matters around, and Pangwa culture looks seamless in 77

Fr. Stirnimann’s study, just as that of Lupila ( just across the northern ‘tribal’ border of uPangwa) was for me. For all the past devastation on either side of the dividing river, the people who put things back together were or became mainly Pangwa on the one side and mainly Kinga on the other. The last time of crisis was already half a century back when our fieldwork was done, and the British administration had pretty well levelled the political scene for the region as a whole. Only our eldest informants had known the pre-pax cultures directly, and they mostly as children. Fr. Stirnimann’s Pangwa have reduced the Kinga settlements and incursions (about which we know from hard records) to a story in the very style Sanga tellers favour— there is a Kinga figure, a sort of demi-prince, who came to teach and lead, and whose line the British came to favour when they found the great Pangwa District they had laid out contained no chief. The whole ‘founder-hero’ tale is there, only lacking the hero’s gift of seed he had carried in his hair. How far should we suppose that earlier Kinga colonization may have affected the social organization of some Pangwa communities? Probably the modest Kinga settlement south of the Mgiwi river helped to broker some mutual influence; but that had always been a feature of the regional culture, since ethnic boundaries where tangible at all had always been porous. In the end, the main change to Pangwa circumstance was probably better access to the big, efficient Kinga hoe blades. There was no sign, in any event, that translocal politics had taken hold. The political summit was simply at the centre of any pedestrian community then; and only two or three of the larger Pangwa settlements would have been strong enough to regain a coherent identity after the Ngoni incursions were past. If I may judge by what I learned from Wanji and marginal Kinga, the strength of the ‘anarchic’ bush community was its ability to disperse before external threat. Everyone then lived close by some forests and knew how to lose an enemy and find safety there. Pangwa by 1960 were as profoundly affected by mission schools and the secular changes of the pax as Kinga were. But among the three mountain peoples—Wanji, Kinga, and Pangwa—I know from my own walkabouts that the first two harboured many unmapped ‘bush’ communities jealous of their immunity from these sources of modernizing change. I suppose that, given the size of the Pangwa area and the sparse pattern of settlement, and given the picture afforded by British tour reports, the same should apply there in spades. Should an observer-stranger pass by chance from such a community on the Kinga side of the District line into one on the Pangwa, would he or she find they spoke differently, told different stories, raised and ate different foods, and organized their 78

domestic lives categorically according to ‘Kinga’ and ‘Pangwa’ templates? I think not. When it comes to cultural distinctions of this order—what one finds in a few days’ visit—the best predictor of difference would be the distance by footpath between the one community and the other. Crossing an arbitrary ‘tribal boundary’ counted for little even as late as 1960. The surnames, the greetings, the temper of life wouldn’t change. In that respect, descending from Kinga to Nyakyusa country is a far more drastic step. The Sowetan peoples were living in a radically decentralized regional culture before the protostates began to develop, and many communities remained at the start of the pax more firmly based in that culture than in the culture of the nearest court. Yes, this is truer of the hilly lands (Bena, Kinga, Pangwa) than of the plains; but no, it isn’t impertinent to Hehe or Nyakyusa conditions. Looking at inter-political sensitivity, Kinga bordered on Nyakyusa, Bena, and Pangwa in that order. It was the Sanga chiefs who established the Kinga identity through a gradual process of political colonization and recruitment among Pangwa-like bush communities. As always in these cases, the pattern of development was to build outward first and foremost, building upward only as translocal ties grew firm. When time stops for the Sanga epoch, the ideological focus of their courts is to the west—the prosperous Nyakyusa-Ngonde subregion. At an earlier stage, the Sanga focus had been to the east. Their royal lines claimed the same mythical point of origin as the Bena lines did, the sacred grove of Nyumba Nitu. Nyakyusa and Ngonde royals made the same claim: their founders were said to have come directly from Kinga country and originally from Nyumba Nitu. That sacred site lies in Bena country at a point from which a hawk could reach any of the old Bena or Kinga capitals in about half an hour. A comparison of the ‘rituals of kingship’ among Bena and Kinga is enough to confirm beyond doubt that they belong to one tradition. As between Kinga and Nyakyusa the matter is not so clear. One of the small mysteries of the region, which has to be a clue to the cultural matrix of its politics, is the question why the myths of origin of these four chiefly lines remained anchored in Nyumba Nitu. It is a matter which ought to yield to further study of that cultural matrix. It is unlikely to be a recent invention diffused from one political elite to the next—which way would it have moved? If it is old, it is likely coeval with the movement of ‘chiefly politics’ through the archipelago, as the tradition claims, and suggests the regional protostate movement began in what we now call uBena. I can say nothing more final for a bigger mystery, the grand puzzle the Sowetan region presents to history. This is the matter of 79

learning what scholarship can educe about the human past from my ‘political archipelago’, the series of five Sowetan protostates, running from the Hehe to Ngonde. Morphologically, each is a transform of its neighbours in the series: a family resemblance runs right through the archipelago, bridging the many important differences. The problem is to get a clearer sense as to how these social systems are related and how and why they have diverged. My intention in this, the final of my three volumes, is to build on the descriptive and analytical base set out in the earlier monographs. As usual, it is the Kinga I hover over most, and my own field experience which supplies most of the regional verities. This is not because I propose my findings needn’t be questioned, only that I’ll have done what I can on that score. My main field intuitions on the Kinga stand radically rethought. I do confess I have wanted to make a (to me) coherent portrait of the people, the place, the past time. But to place the Sanga political achievement in its context in time and events I have to see it in its regional aspect—which just means rethinking all the other fieldwork as well. The only really probing comparison with the Sanga court system which recorded information allows is with the Nyakyusa, and that is why I turn so often to them. The Bena, had the ethnographic work been ampler and otherwise focused, might have served as well. But Bena rulers, in the descriptions we have, are certainly ‘chiefs’ not ‘princes’: the element of inherent majesty is missing, as it is with the Hehe. Bena chiefdoms are a set of equal and independent polities without the mystical bond of the royal Sanga or the hallowed aristocratic provenance of Nyakyusa princes. I am much taken with this unifying mystique. Hehe unity is of the Napoleonic type, based in a triumphalist past. Bena unity was always opportunistic. Both the Hehe at one end of the chain and the Ngonde at the other were exposed to rapid structural change in the generations leading into the colonial pax, which Kinga and Nyakyusa (and some Bena polities) reasonably well escaped. What fascinates me about the Nyakyusa is an intrepid commitment to turbulence in preference to stability, one which seems only to strengthen their political identity as a people. My comparative approach turns at this point away from matters of like and unlike usage to problems of reading the social mind and dominant patterns of motivation from a people’s institutions. The purpose of the Kinga-Nyakyusa study which follows is to illuminate their worlds as they were at the height of their politicall achievement. Afterward, to explore the historical depth of the world the Sanga régimes built, I turn to a different sort of comparison. The 80

Pangwa studies of Fr. Hans Stirnimann constitute one of two ethnographies which can serve as reference models for the world of the pre-Sanga communities in what came to be ‘Kingaland’. My other reference model is Gulliver’s searching work on the Ndendeuli, a seemingly still less ‘political’ people colonized (and, like the Pangwa, terrorized) by the Ngoni. Stirnimann’s Pangwa are presented under the dependable rubric of old-style ethnography. The folkways and mores, the organizational schemes and rituals of times past are described to us just as they were described by competent informants to the investigator. Gulliver’s work is a sociological study of the way Ndendeuli ran their daily lives and dealt with others. The fieldwork was done only a few years earlier than mine and Stirnimann’s. ‡‡ Of all the Sowetan cultures on record, Nurse found Pangwa is by careful linguistic measures the closest relative to Kinga; and comparison with my parallel work shows the linguistic measure of a two-thirds lexical overlap reflects well the degree of homology between the cultures more generally. If asked to predict how much lexical divergence would be found by applying the same techniques to dialects in the twelve local polities under the traditional Sanga courts, I would expect the least-similar pair to be on a par with the Kinga-Pangwa pairing of Nurse; at least, a finding of that much difference would not surprise me. But that least-similar pair would not include courts in the Sanga heartlands. It is worth reporting here that some informants at Lupila (Eastern realm, bordering Pangwa) held their Sanga ruler, who locally claimed the title of ‘prince’ unkuludeva, was in process of sending out client-rulers to the Pangwa when the Ngoni troubles erupted. This is in questionably perfect accordance with the picture of the past preferred in all the Sanga courts. The truth-value is unclear. But all oral history in that domain of the Eastern realm is only such as was able to survive the scorched earth of 1905. Their prince in death became a hero of mythic proportion, fated one day to return in full vigor even after half a century had passed. It is particularly to be noted that an attempt was made in the 1905 uprising to wipe out all Lupila males from the age of mischief upward. The Lupila I found in 1962 was reinvented yet deeply local. Linguistic evidence aside (and it was not available in convincing form until the mid-1980s), Kinga-Pangwa parallels in social organization and ritual were striking. The methodological difficulty I had had in dealing with my own field notes on ‘custom’ before Stirnimann’s work came out was resolved as I read his two volumes. It allowed me to sort out what belonged to the Sanga system and what to the pre-existing cultures on which the Sanga protostate 81

was integrally overlaid. I can turn to the Pangwa to explicate that and to confirm the degree of continuity between the culture the Pangwa recall and my own reconstruction of the developed Sanga court culture. What the student of any regional culture which is at peace must keep in mind is that ease of movement in that social space is at once a reflection and a condition of the quality of peace being kept. Breaches of the peace are sanctioned by self-help, keeping trouble localized. Variations in speech form and culture are granularly distributed, not gathered in discrete political packages. The preSanga population of Kingaland was not, in the political sense which has tended to dominate ethnographic studies, ‘one people’ with ‘one culture’. What the Sanga revolution did was to create the semblance of such a people and culture in the social imagination. The segmentary or ‘nesting’ structure of a protostate begins by giving a greater sense of unity to the local pedestrian community, and proceeds by raising the symbolism of unity to the more inclusive levels of structure. As everywhere in human history, the objectives correlative—the concrete acts and spectacles symbolic of unity— were in the Kinga case always ahead of the social facts they were meant to inform. My use of the Pangwa ethnography is analogical not comparative, in that their political system pre-pax, so far as it can be reconstructed, offers no significant parallels to that of the Sanga court. I can use Gulliver’s sociology of Ndendeuli analogically as well, though they are more distant—neighbours of neighbours—to the Kinga. Gulliver supplements the Pangwa material with observational notes on a regional group still living an acephalous political life in scattered hamlets. This window on a simpler, ‘anarchic’ social organization was still open to a sociologist in the 1950s because this people had been subjected and ruled over by the Ngoni. The British left the Ngoni in charge, and the Ngoni had no plans for bringing the Ndendeuli into more modern ways. Neighbours and Networks represents, for the larger Southern Highlands area of Tanganyika, a credible model for the freedom with which individuals and households might have moved about in quite early times on both sides of the Ruhuhu river which divides Ndendeuli country from Pangwa. I can presume but can’t know that this freedom of movement lies dormant in the ‘bush’ heritage of all the Sowetan protostates, and manifests itself in their antipolitan ethic: their readiness with censure—countervailing blame—for a ruler who fails to safeguard his people from hunger, sickness, and personal loss.

82

A premise of imparity Godfrey Wilson found in the chiefly priest Kasitile a matchless guide to the wisdom of his people. When the priest spoke of the two classes, commoner and chiefly, he was not thinking in secular mode: For us it’s taboo to accuse a man of witchcraft, but the commoners accuse one another... If a rich commoner dies and the heir buries him with only one cow, the shades are angry and the people also...People are angry...But the oracle says, ‘The shades are angry.’ We chiefs have bodies which are different, the commoners arrived first in the country, they ate raw food, we fear them saying, ‘They eat raw food.’ It is we who followed and brought fire. To us they confess (their witchcraft); to each other they say: ‘It is your father.’ Because if they confess to a fellow commoner saying, ‘It is we’, he says, ‘You are “eating” me, drink the ordeal’. The witch is driven out. But to us they admit saying, ‘It was we who were angry.’ They do not fear us. Commoners are the seniors in the country; they arrived first.... I fear for the ritual, I fear hunger, the hunger of the people. For I am the food of the country, I am the maker of food. ‡

Kasitile is saying the two kinds of Konde are two separate confessions with deeply different outlooks on the world. His is a premise of fundamental imparity. Chiefly persons are and must be few, yet without them the Konde people would be uncivilized. The distinction he is making is that between congregation and priest, ‘folk’ and ‘intelligentsia’, a people and its prophets. It is not easy to explain then that the commoners have their own priests, and that, in effect, they also have their own kind of chief-like heads who rule their villages and enjoy the supernatural power amanga allowing them to ward off aggressive witchcraft. These village heads, it appears, live on when their village has passed through its warlike stage to become the ‘commoner priests’ who feature so largely in Monica Wilson’s portrayal of priestly business. Commoners can’t ‘fear’ chiefly men, but not because of trusting them. The reason is, chiefs live in an orbit outside the close circle of envy, the village, where aggressive witchcraft incubates. Kasitile tells us that the two Konde confessions live side by side, each comprising a separate domain of responsibility. The key difference between them is in the definition of that sphere. Headmen have powers limited to a local sphere (the one pedestrian village), chiefs operate in a congruent manner trans-locally. There is 83

one pervading difference: appeals to the shades are impossible in the village locale. The appeals the Konde make to ancestors are translocal; and chiefly ancestors respond to trouble among any and all residents of the chiefdom. That is why when in the village it is said ‘the shades are angry’ (and not ‘your neighbours are angry’) the case must go to the chief, whose decision relieves any villagers of responsibility for the sanctions imposed. The buck is passed to the chief as if to a scapegoat. A reader will quickly see that the chief must not be tied (by culpability, by blame) as neighbours are tied to one another in a commoner village. That he has two villages, one on each more-or-less self-contained ‘side’ of the chiefdom, allows him leeway. Each of these villages is ruled, hands-on, by a ‘senior headman’ who with age and a successful career will take on a priestly bearing once he and his peers have become the arbiters of right. The political scope of a chief is larger. He must be free to judge those neighbours without prejudice. Specifically, when trouble arises among the households of a chiefly village, he cannot be put in the position of referring an appeal to his own court. Several of the special attributes and rules which apply to a chief appear to be designed to meet such embarrassments. There is the majesty of his wrath, which needs no argument to give it the force of truth. There is the supra-local ambience a chief begins to carry with him as he judges appeals from commoner hearings. Every chief had men in his own villages to handle local cases, doing what they could to settle them without appeal. There is the ordeal, which an angered chief will order when a man thought guilty by his peers will not confess and make amends—since the ordeal is most convincing when it only confirms the public’s sympathy, a man who feels ‘outnumbered’ is likely to confess before demanding it. And there are special hearings a powerful chief may set, to settle issues which apply beyond the sphere of responsibility of any single chiefdom. It is common practice for a man whose illness has been ascribed to witchcraft to sleep away from home. A night-flying witch is that easily foiled: This is a common precaution in case of illness, for witches are thought not to have power ordinarily outside their own village and never beyond the chiefdom. The fact that it is within the village that witchcraft is thought to operate is shown by the way in which one fearing witchcraft, or accused of practising it, moves to another village. Sometimes he goes to another chiefdom, but often he only moves to another village

84

in the same chiefdom—the chief himself may feel safe in one village of his country and not in another. ‡

It is the very vulnerability of a chief to witchcraft (or some other mystical power) which warns against supposing that a clean sort of secular authority has been achieved. In the Kinga case, a ruling prince was vulnerable to the extent that he might become immured in his stockaded compound. Kinga preferred to think him in danger of poison (sorcery) and kept a meticulous watch on his food and drink. But enemies could arise within the family, and powerful wizards consulted in uPangwa, whose occult powers could not readily be apprehended and foiled by the court priests. In the Nyakyusa case a chief, far from immuring himself, would only be careful where he slept, for witches’ habits were well known. But in either case we have to ask how well a Western idea of political authority matches the situation of a Malatan chief or ruling prince before the pax. The Wilsons were at pains to be clear about the nature of chiefly authority in precolonial Selya. Here is Godfrey Wilson summarizing a position which Monica Wilson subsequently echoed. For his ‘great commoner’ in each case, read her ‘village headman’—the less puzzling, though also less comprehensive, usage I have usually preferred: Within the whole chiefdom the political authority is vested in a chief (malafyale), whose position is inherited, and in the appointed great commoners, who have all originally been appointed not by the chief but by previous great commoners. The chieftainship has always been, as we should say, a constitutional one, since the chief can never make any important decision effective without securing the consent of his great commoners for it. The chief and great commoners together take all political decisions, try cases between members of different age-villages, and hear appeals from the decision of any particular great commoner. ‡‡

This picture is built around the idea of a chief with ‘constituted’ authority. But if Western terms must be used, I expect the ‘chief as scapegoat’ is a better fit to the situation. The chief in fact seems to have no business directly judging cases. A defendant is brought to him already judged by his peers. What is lacking is only his own acceptance of that judgement. The case must be rehearsed again. It is in being prepared to break the standoff that the chief assumes responsibility. If the matter is not witchcraft the facts should now be clear, and the chief can deal quick justice in form of compensation and a fine. For witchcraft or an another unprovable offense there is the ordeal. As I read this, contra the Wilsons’ position, the chief’s authority is not muted by the participation of village headmen in a 85

case, but stands clear of them as the sole source of judgement lying beyond popular blame. A good chief kept his own counsel and wore the mask of power in all his dealings. ‡‡ When Kasitile insists ‘they do not fear us’ he has in mind that members of the chiefly confession, since they cannot be envious of commoners, have no cause to use mystical powers against them. Commoners would not suspect envy in a person of the chiefly kind, or the witchcraft such envy spawns. A chief, in his night thoughts, is not biased by personal involvement in the quarrels of his people. But couple this with the consideration that strong-minded chiefs have been known to confiscate a commoner’s cow or his wife quite arbitrarily. So it is not that chiefs are blameless but that the blame of a commoner cannot touch them. This I take to be Kasitile’s theory of the two confessions. What is problematic in his world is that the unchiefly multitude should be insensitive of the moral burden carried by the few. But have hubris in mind: Kasitile has no admiration for the commoner priests he works with. He is talking about the structure of Nyakyusa society ‘as it was ordained’. If this is not quite the structure we find implicit in the way he and his colleagues actually do things, it offers a clue to what they are thinking as they do them. It pictures a society in which a chosen few are privileged with authority over the others. But it is well to consider this informant’s model as a formulaic one, not in itself descriptive. The problem it can help to resolve is the nature of political authority in a society where no man is likely to take orders from any other. Don’t suppose the Konde mind is not acquainted with authority as such. Women are elaborately obeisant to men, and privileged seniority either in kinship or political role-relationships is widely recognized among men. Monica Wilson said it plainly enough in the title she chose for her retrospective volume in 1977: For Men and Elders. What we must look for is the kind of authority which works across established networks—translocal, self-dramatizing, the kind born of taboo, spontaneously honoured. Kasitile is probably right in supposing that without men of his ‘chiefly’ confession there could have been no high politics for Konde. But he is demonstrably wrong in supposing the bearing of sufficient authority is a privilege of birth. The proof is in the great catalogue of cases Monica Wilson can supply. They give us chiefs in hiding and reticent, and commoner rulers who would set them an example. Here I offer an ‘extended case study’ based entirely on Monica Wilson’s work, of a senior headman acting in what appears to me a 86

sufficiently chiefly style. Mwambuputa was born 1880, is listed for 1938 as ‘the greatest in the country’, and later on in 1955 when he was in his seventies he had become ‘the senior priest in the country’ with 30 wives. His is the first of two senior villages of Chief Mwankuga. Though he is a commoner he knows how to act the chief. He took charge of a recent Coming Out (1930s) and there exhorted the young chief Mwanyilu, “Listen to the people: they are the real chiefs! Be hospitable! Greet people politely! Don’t beat your men!” ‡ We look back at him now in his fifties, wealthy, able, and forceful. He is having trouble with an assistant, Nsekela: Nsekela was an assistant headman in the village of Mwambuputa, appointed by Mwambuputa himself. Nsekela’s child cried by night and Mwambuputa was angry with him because he did not report the matter. “Why did you not tell me that your child was ill here? Why do you hide it? Indeed, you are bad, it is you who are making the child ill, get out of our country!” ...Nsekela denied the charge [of bewitching his own child] saying, “It is not I. No!” Then he went away to Selika about a case. While he was away his neighbours barricaded his doorway with thorns. On his return he found them, and he moved to Mwaihojo’s country (the adjoining chiefdom). ‡

Nsekela’s reaction on confronting the arrant use of authority is to go away on an errand and let the storm blow over. He knows more about this affair than we, but let us suppose this incident is just the culminating one of a series of incidents which Mwambuputa may have resented. Withdrawing is a culturally typical way of handling direct confrontation with authority—don’t slam the door, just let things cool. But when Nsekela comes back he finds his neighbours have sided with the chieflike Mwambuputa. They are letting this big man act in self-righteous anger, like a chief. Nsekela moves away. Still, eventually he may return in all innocence. What sort of game is being played? Mwambuputa himself had moved away from this same place all the way to uNgonde some years back. One story is that he moved into an inheritance there. More convincing is the story that he moved because two of his children had died—he moved in fear: Eating fine food alone is said to kill a man. ...It is said that Mwambuputa grudged his fellows certain food, so they were very angry and two of his children died at home. When he found that sickness was spreading at his homestead he fled to the lakeshore plain (MuNgonde), but there also he found that death pursued him, and he returned home again. It is said that sickness pursued him to the plain because he went off without performing the death ritual

87

for the children who died. So when he returned he found two bulls and killed them, saying: “Friends, I have come. I did wrong because I did not perform the ritual for your children with you.” ‡

We have perhaps discovered why a crying child could so upset a man of Mwambuputa’s stature. He was driven from a position of importance as surely as if he had been a witch himself, for he dared not try to blame one of those he, with his ritually given amanga, was empowered to rule. But we need to move closer to the scene to judge what it says on the matter of authority and power. Three political generations are in play: (i) that of chief Mwaipopo and his younger brother Mwaipopo II who succeeded; (ii) that of chief Mwankuga, son of Mwaipopo I, and Mwankuga II who succeeds; and (iii) chief Mwanyilu. Mwambuputa belongs to the second of these political generations, and at the summit of his career it is he who will play the patron to the young chief Mwanyilu at his Coming Out. Years before, at the Coming Out of the young chief Mwankuga I, it had been Mwansambe, the most senior of chief Mwaipopo’s headmen, who had taken charge of arrangements for the new chiefdom, particularly selected Mwambuputa as a client. This implies a continuing relationship between the two. Judged a promising young commoner, Mwambuputa is installed as the primary senior headman of the new chief Mwankuga, and has an ulupando tree ritually planted in his name. The tree will mark the centre of a new political domain. The tree will grow with the new political generation as living confirmation of Mwambuputa’s important office—senior village headman, presiding over a new chief’s first village. But the site of that planting made for later complications. You may centre a new domain where you like at a Coming Out, but as there is no demarcated boundary people will suit themselves on that score. Mwansambe’s choice of a site was in Bujenga, where his nextsenior colleague Mwakalembo resided. These two men are of the Mwaipopo generation, which already in German times has begun to step back and cede space over to the next (Mwankuga’s) political generation. It is because Mwansambe was the senior headman for the chief Mwaipopo that he was able to overstep Mwakalembo, a person of his own generation, to favour one (Mwambuputa) of the next. In private, most people were agreed the land was Mwakalembo’s, even though he belonged (like Mwansambe) to a retiring political generation, and might have been expected to step back with good grace. Mwansambe eventually found, as if recognizing Mwakalembo’s right, a different place for Mwambuputa and his peer group. The corrected site was on Mwansambe’s ‘side’ (reckoning now in terms of the obsolescent geography of Mwaipopo’s retiring generation) but not this time in Mwakalembo’s back yard. Meanwhile the 88

ritually planted ulupando tree has grown tall at Bujenga, Mwambuputa has retreated to uNgonde, and the matured new generation (ii) of Mwankuga’s age-peers is taking centre stage. Seeing that and seizing the occasion of the death of Mwankuga I, Mwambuputa makes his move to return from abroad. ‡ It will help to review the double generational status of some principals. The Mwakalembo who saw himself diminished with the planting of that ulupando tree for a new, take-over generation (and who subsequently felt reprieved) was certainly of the senior political generation to Mwambuputa and senior as well by generation of birth. But the Mwakalembo who half a lifetime later appears to be threatened with losing he same place to a returning Mwambuputa was only politically the senior, having inherited name and status from his father. Even the Mwansambe who undertook the reinstatement of Mwambuputa was of matching age, having likewise inherited a father’s position and name. Seniority is an absolute of Konde political thought. It is the basis of Konde social organization as it is expressed in the age-village principle of peer solidarity, and it is the basis for the inheritance of position and wealth. But while ageof-birth seniority may be a given, it is only one of the ingredients of seniority as a social mechanism. An age-senior woman kowtows to a lad on the path because gender is a more powerful ingredient than age. The question we have to ask is what part such subordinated ingredients might play in a situation of emotional involvement. Suppose Mwansambe’s son in the scene we are tracing, the man who has succeeded to his father’s place, encounters the aging Mwakalembo, not yet replaced by his heir. The young man, to push his claim to precedence—political seniority—over this Mwakalembo, must push against a tangible preponderance of experience and personal alliance. But as soon as Mwakalembo himself is succeeded by a young heir, won’t the same move to bump him have more weight? Mwambuputa had been long gone, his perks naturally fallen to others, his political services supplanted. But his claim to reinstatement had the ritual sanction of taboo, in the living presence of the ulupando tree at Bujenga. Three men, age peers, have to agree on the project of reinstating Mwambuputa as a senior village headman under Mwankuga-II. We are looking now at the politics of flexibility: the system must adjust. In the event, Mwakalembo dug in his heels and made it apparent that, just as he (his father) had won this land by defeating a previous headman-owner in battle, so must there be a fight if he were dislodged in turn. The political priorities lined up to a stand-off. Mwakalembo had the inherited rights of a ‘father’ to Mwambuputa 89

and his peer group. Mwansambe could not count on support from his own chief Mwaipopo, who was keeping his distance. Mwakalembo-II (having built himself up as a man of Bujenga) would have had the popular support which neither Mwansambe (an absentee ruler) or Mwambuputa (with no current following) could expect. On the other hand, Mwambuputa by coming in at the funeral of Mwankuga-I could expect the patronage of the successor (a younger brother), chief Mankuga-II. The wrangling over claims and counterclaims was several times renewed without a decision, as Mwaipopo (with a full life behind him and now a salaried chief under the Rungwe Boma, and nothing if not circumspect) saw the risk of open dissension. ‡‡ Who will sort out the impasse? While the decision hung, Mwambuputa put up temporary houses in Mwaipopo’s along with the older generation. But Mwansambe had properly made the first move, and as a good patron like his father would have done, in time arranged Mwaipopo’s sanction for a major ceremony of reinstatement, attended by all the commoner priests and headmen—no chief, not even Mwankuga-II. It was commoner business. They took over the village (Igembe) where Mwankuga-I had come out, and with much feasting at Mwambuputa’s expense they installed him as senior headman of the new Mwankuga’s first village. All the commoner ‘politicians’ were there to witness an act of contrition. Mwambuputa (it appeared) should have sacrificed two bulls in the long-ago, to demonstrate atonement after the death of his two children. Now this had been done, and the people’s concern about the meaning of those long-ago events was cleared away. Here are some things that were said at the ceremony: [Of Mbuluko, who was being bumped:] “At first he refused, but they pacified him. He cannot refuse, he will move. They just considered where it would be best for the village headman to be.” [By Mwansambe to the senior men as they came together:] “Look after Mwambuputa in the war by night, look after him in dreams.” [About the legitimacy conferred on Mwambuputa:] “Before Mwambuputa killed the bulls we did not see that he had come back; now he has returned.” ‡

Can a headman have authority without power? That is being primus inter pares, managing not making decisions of importance. This is the situation the Wilsons construed as a ‘balance of power’. That is fair enough but not particularly revealing. There is a game of power playing out but it is not a contest of two sides. It is a game of placing blame where it will stay. It is a game with as many sides as men with stakes in the outcome. The scapegoat has come back, an 90

old case reopens. Now it appears it was the shades who were angry back then. Mwambuputa should have seen that, should not have thought witchcraft, should not have grudged his fellows the meat of sacrifice. We have a detail from Mwambuputa’s later career after he was re-established, which lets us see something more of a senior headman’s situation ‘on the ground’. The case concerns a certain M., who was angry with his neighbours. He felt wrongly accused of bringing trouble on a child. They drove him from his home by word of mouth. No one consulted Mwambuputa: Mwambuputa brought him back to a different part of his village... Then his neighbours grew angry asking why Mwambuputa, who himself had forbidden murmuring, had brought back one whom they had convicted of witchcraft! So they went to Mwambuputa’s place and expressed their anger to him. Then he agreed to go and expel him.... When we reached the house...the assistant headman of M.’s old section...stood out and said: “I have a word to say to my father Mwambuputa. I am astonished that you received back M. after I had driven him away because he is an evil man who does not like to have friendly dealings with us his neighbours.” Then Mwambuputa replied: “When I fetched him back I did not know you had convicted him of witchcraft, I thought he was simply moving.” “Yes,” replied the other, “that is why I have followed him, for your house is my home! If you love M. you too will be expelled!” And the others said: “Had you known about the case, we should have expelled you both!” “Let us tell him then,” said Mwambuputa... When he came Mwambuputa said: “Your neighbours have followed you saying that they have expelled you.” ...They said, “You knew yourself to be a witch, why else did you not show indignation and surprise and take your friends to Mwambuputa to be tried? Now you’ve given them power, and they have followed you here!”...Then Mwambuputa said: “No more words, leave the village, your fellows tell you to go!” ‡

For ‘Now you have given them power’ let us read, ‘See how you have turned our suspicions into certainty.’ The matter began with the illness of a child, which became a matter for alarm and an unfocused presumption of witchcraft. M. became the target when he showed signs of defensiveness, a suggestion of guilt. This increased to the point that no one would side with M. The child’s agonies flared up, and M. was driven out. For ‘He is an evil man’ let us read, ‘He is resentful and wishes harm to our child.’ His former friends are now saying he ‘knew himself to be a witch’ or he would not have been acting so defensively. From this I infer that it is possible for a man’s resentment to cause harm (because he is a witch) without his 91

intending it (because he is unaware of his magical powers). The people seem in their rough way pleased to have found the scapegoat they needed if they were to restore the peace broken by fear of death and a child’s mysterious seizures. If their culprit had chosen the route to Mwambuputa and the ordeal, they are saying, they would have believed its decision and thought better of him even if he failed the trial. In the other event, this tale would have no ending. Witch belief in this little drama appears as the safeguard of trust. In a peer village, the vertical dimension is absent which might have given peace-keeping authority to respected elders. Through the drama of scapegoating one member, the community achieves a ‘world renewal’ in microcosm. The excitement has been cathartic, blame has cancelled fear and must be held to. Nothing has been done, you might say, for the child. People knew of nothing that could be done but find the mystical cause of the illness and remove it. As it happened, the child didn’t die, but the drama was over in any event. The problem the community successfully addressed was not the problem of ‘our child’ but ‘us’. A credible psychological reading of the incident would be that M. set himself up for an outburst of hate through his own paranoid ‘murmuring’. Where amity is the nexus the breath of enmity won’t be welcome. The first instances we are given show M. refusing to share his friends’ food. Then he openly chides them for borrowing a bit of his firewood to heat their beer. Things had got to this point when the child’s screams alarmed the whole company. They instantly knew but couldn’t openly name a culprit—that would expose them to the charge of ‘murmuring’ themselves. They had to wait for a bold denial—“It is not I. No!”—for until that happened each of them was unsure of his own impunity. And how then to read “We should have expelled you both”? It clearly says “Don’t presume on our good will!” More than that it can be taken to say “We don’t need leaders” or even “We have better men for the job.” Mwambuputa has heard that before, when his own children were dying. He heard in the way one hears what is not spoken, and at that time he was unwilling to brush it aside. He has since laid down an impressive record as a settler of cases and local information manager. Interviewed about witchcraft he came back with an indisputably sophisticated judgement, adopting the view that openness about ill-feeling can always lead the way back to amity. “It is when he lets nothing out and stays silent, though angry in his heart, that he comes in dreams to throttle me.” As long as he keeps to his therapeutic perch above the fray, it seems a man like Mwambuputa can settle cases and grow rich. The key, I think, is 92

Kasitile’s insight: don’t be reached by their blaming and they will let you understand them. ‡ Shall we conclude our Mwambuputa has set out to govern the ungovernable? His is a society which cultivates amity at the localcommunity level as a condition of political health. The breach of amity is accordingly treated as psychic contamination, isolated, and purged. Recruited now to the ‘chiefly confession’, is Mwambuputa yet no more than a half-willing signatory to the dictates of mob rule? Men recruited to his confession learn to shed blame and keep their heads. An effective ruler, watching the power of consensus generate an impasse, is able to intercede, define the trouble, and pack it off on a (re-usable) scapegoat. Let chaos come out, and it is not quite as Mwambuputa would have it, that ‘therapy happens’. There must be an orgy of vilification abrogating the fiduciary status of at least one group member, who is in this way driven out. The ‘chiefly’ act is plausibly necessary to steer things toward this end, and so toward the redemption of amity. Can it be that all the chiefly man really need do is have no ‘real’ friends himself—none he can’t betray—and a good deal of meat on the hoof? I would rather argue from the evidence that politics among the Nyakyusa has become a profession surpassing the mere settlement of cases. Men who find themselves charged with authority quickly come to know they may as easily be discharged from it. It is an entrepreneurial game, in which the players put up the stakes and each defends his turf against rivals. May we suppose Mwambuputa might have admitted to favouring M. against the general will, and might have been expelled with him? The reason it didn’t happen was that Mwambuputa has learned to cover his back and play the political game, turning fear into blame and blame, perhaps, into kudos. In this view, it is the commoner village ‘chiefs’ who make a Konde community governable. The hereditary chiefs are involved in a game of their own, at a translocal level, with much the same rules. There is a balance kept. Show me a politics which doesn’t revolve about off-loading blame. Show me the chief in an antipolitan world who was never made scapegoat for local troubles in which he had taken no part. The strongest among the hereditary Nyakyusa chiefs stood high, as any reader of the literature will know. But they were many, not few. They hoed their own land, boasted many wives, and kept their own counsel. In a world where the deeper kinds of political involvement are local, the prime requisite of an appellate authority is keeping distance.

93

FEAR & BLAME, CHAPTER FOUR

A Religion of Blame Honour’s underworld In discussing the Malatan cultures it is best not to trouble with the distinction, so useful elsewhere, between ‘witch’ and ‘sorcerer’. Kinga don’t resort to autopsies to understand a death, Nyakyusa do. Kinga will blame a sudden death on poison, an undetectable objective device, not on witchcraft, which kills slowly, is known from dreams, and is subject to remedies. Nyakyusa may have obsessive fears of hereditary witchcraft; men, women, and their offspring are so branded. Kinga say a person can just reform, agree to get rid of his paraphernalia, and resume life without prejudice on parole. But when you ask further you find the paraphernalia of true ‘sorcery’ would never be found in the possession of an ordinary Kinga person. They are useless without the occult knowledge of a professional doctor umavi, who would have been consulted and employed but who would not be the ill-willed ‘witch’ within the victim’s personal circle. The suspicious stuff our Kinga ‘witch’ might have in his possession would presumably be charms for protection and good fortune. Most men are likely to have some, and with charms a good defense can easily be received as an offense. So witches may have to use sorcery to do their work, but that is because they are not inherently powerful. The difference between inherent powers and instruments of power which have to be consciously sought for—and may be forsworn—is a real difference, and distances the Kinga from the Nyakyusa mind. But Kinga don’t make the difference between ‘witchcraft’ and ‘sorcery’, and while Nyakyusa do they may use the terms interchangeably. As for enhancing political power with the mystical, Kinga and Nyakyusa agree on their python lore. If a prince has a python ‘in his belly’ it gives him witchlike powers, but there is no way you could prove anything about a python. Spirit pythons are no more tractable and no less prodigious than the giant serpents whose plain tracks and traces are said to litter the sacred groves.

94

In the literature, witchcraft transpires in three distinctly conceived frames. In today’s account of the Salem witch-hunt we have ‘the witch as victim’. In post-medieval Europe the witch was presented as a human embodiment of evil. The frame I’ll be exploring here may be called ‘the witch as creature of misfortune’. My approach will be interpretive. I take it that the high seriousness of the witch belief has to be understood dramaturgically, as it derives from a critical reversal of fortunes in the moral career of the person who will come to feature as a witch’s victim. There are three formal considerations which help to show how the invention and the persistent reinvention of witchcraft can be explained from the ordinary experience of misfortune in the context of a worldview common to Eastern Bantu peoples. The first point is the distinction between divination and prayer as a response to misfortune. Prayer is not available to the victim of accident, infection, or deep personal loss in the prepolitical, precontact Sowetan regional culture. When you go to a diviner with your trouble, you may discover you have offended an ancestor or incurred the envy of a witch. In either case, the matter of the enquiry is not to question your moral character or state of grace but to discover what small neglect of duty or show of pride has angered someone, living or dead, within your private circle. Divination is a prescribed procedure, and that very fact says personal misfortune is no one’s clear concern but your own. Consider the Biblical tale of Job and you will see that comparable cosmic concerns with justice are not invoked, theodicy and its ‘problem of pain’ have not been invented. The story of Job is nothing where no one is his brother’s keeper. This is my second point. My third point concerns the nature and uses of divination. The diviner’s client comes with his own agenda but without the needed degree of certainty as to a means for achieving it. The client is aware of risk. An ill-chosen path may jeopardize his future. He would not be going to the diviner if he did not sense danger in his situation. When the client is a victim of misfortune and senses that an act of unnatural hostility has caused it, he will be suspecting witchcraft. If by going to a diviner you always confirmed your suspicion and always continued straight on to confirming the identity of a witch of your choice, the formal procedure would scarcely have any real weight in the public eye. Your supposed enemy is Somebody’s brother, friend, or main support. Divination is the prime instrument of indirect action in most of the non-Western world. It is taken over gradually, with the evolution of jural institutions, by reliable courts. But divination, where it commands credibility, is a quicker, more private, and more efficient device. It gains its credibility by seeming to be objective, where obviously unbiased chance mechanisms are employed in the 95

‘choice’ between equally likely alternatives; or from the accepted disinterestedness of the seer, where the client’s anonymity is seen to be protected, as in some of the Greek temple oracles or a Sowetan witch ridder’s performance; or theatrically, when a godvoice seems to speak through the diviner in trance, or is ‘cast’ by the ventriloquist’s art. Systems of indirect action are virtually identical with what we call ‘civilization’ just as direct action, the quick resort to force and violence to gain one’s ends in a world devoid of polite procedures, is our prime criterion for ‘barbarism’. But diviners don’t deal easily with shades of doubt, and this leaves them best suited to cases which won’t come to proof against hard evidence: if there were no witches the diviners would have to invent some. Dramaturgical explanation ought not to depart from the path of common sense, but common sense observation can miss vital markers. In the study of witchcraft what can be missed is the special cast of animism, considered not as ‘belief system’ but worldview, an integral set of unexamined premises about the human condition. (So far as your ‘belief system’ is an ‘ideology’ it is a set of ‘examined premises’.) This distinction is important for understanding why a new ‘belief system’ like Christianity or Islam can fail to crowd out witchcraft. When a spirit world has no centre but projects instead the acephalous secular lifestyle of local traditions, it won’t be instantly centred by introducing a new and powerful spirit being. However attractive the new divinity, the old are not easily banished. The rhetoric of monotheism rests on a mystical cosmology which the traditional Malatan thinker can readily comprehend as an addition to his own, but would require profound conversion to realize as a substitute for it. Nor is the new rhetoric even a century on in history attuned to the specific nature of the philosophy it would replace. As if to accept this limitation from the start, Christian missionaries chose to bestow on their own chosen divinity the proper name of a locally important spirit, one among an indefinite number of haphazardly individualized divinities known by name throughout the Sowetan region. For Kinga the Christian divinity was cast as one of these (Nguluve), for Nyakyusa another (Kyala). Automatically, since the whole region is a seamless community for news, the newcomers’ premise of monotheism was annulled. For priests of either people, who were of course aware of an indefinite number of pre-existing (mythical) or heroic (legendary) divinities, it would not have seemed strange that the missionaries in the valley were representatives of one divinity while those up in the highlands spoke for another. Though the several identities of the handful of spirits the missionaries found in any local community must have 96

seemed ill-defined in the minds of their informants, the sheer existence and power of these beings had been massively established in the common knowledge of the people. Their (unrecorded) folk history surveyed a people’s timeless struggle to appease them. The typical Christian convert was focused well on a new belief system, but it was one framed within a well known cosmos. The same distinction—belief system vs. worldview—is important as an interpretive criterion for understanding the politicization of witchcraft in societies like those Kinga and Nyakyusa were building before the missionaries came. Mass ordeals and the execution of witches at the Sanga courts didn’t suddenly double the depth of the moral life for Kinga. But by taking some cases of witchcraft out of the sanctioning sphere of self-help and incorporating acts invisible by nature into their authoritative justice system as high crimes, the Sanga brought a fuller religion of blame into being. It didn’t greatly alter the ‘bush culture’ pattern of recourse to a diviner, which still called for the victim privately to prevail on the chosen ‘witch’ to revoke an awful curse. But witch-ridding became a standard part of the political theatre through which Sanga rule regularly refurbished its secular mandate. Had the system of the Sanga courts achieved a radically more substantial central direction in politics, I expect the spirit world as well would have taken on a steeper pyramidal structure. This shift is visible in the Sanga courts when they are compared to the Nyakyusa. But witchcraft as a belief system remains an ideology for ‘democrats’: if anyone, anywhere, can own the secret personal power to threaten the state, then tyrants beware! The antipolitan ethic lives on. The animist’s worldview is best known to students of theology for its inclusion of innumerable invisible spirits with mystifying powers. When you compare animism to Olympian doctrines like Hinduism or Pueblo or Kwakiutl ceremonialism, and then of course to monotheism, you can hardly escape the implication that the quantity of gods is quite apt as a criterion for sorting religions. But if you look more closely at the animist’s world you will notice that the only proper gods therein are sheerly mythical. They made the world and perhaps mankind but left them, once fashioned, to fend for themselves. They live within the ‘human origins’ frame. There are no shrines at which men and women can hope to communicate with these few ‘otiose gods’. Animists’ shrines are meant for less majestic, even hapless spirits who want—demand—attention and are continually bothering people for it. Their scope is always bounded, their style particularistic. Any of the standard tribulations of life, minor and major, may be read as signals of discontent from the spirit world. Animists live in a world watched over by small minds. These 97

may belong to a household, an extended kin-group, or a locality. Whatever the basis of their ties to the living, they are locally but few in number. Yet these spirit-minds are nonetheless many, as their spiritual clones are spread throughout the country in cellular fashion. Wherever you pass from one social group to another, if it is an easy passage, you find yourself in a parallel universe, familiar but not convergent with the one you left. Animism is the ground for all religions of blame, each being special to a regional culture. It is where blame can be shifted from autonomous spirits to hidden shades of the living that you have witchcraft. In its special fashion, it is an empowering system of belief. High religions, aiming at universality, sometimes succeed in displacing ancestor spirits because in kinship societies both the far and near ancestors, having particularistic identities, can only be maintained by the very closed groups whose powers a universal church will need to break and disperse. Animism is at its best and purest where ‘congregation’ would be a dead wrong word for the group of believers in Spirit X. Burmese ‘nats’ offer a well studied example. Eastern Bantu seem less inclined than some of the well known (precontact) West African states to politicize kinship in the form of clanship as a grandscaled basis for social control. Sowetan protostates, not based clanship, do not conflate a ruler’s ancestry with a people’s. Kinga legend goes so far as to bury a dissident divinity (Lwembe) in the ruler’s ancestral line as an irrepressible Trickster. So far as ordinary households are bothered by ancestral spirits, they are near kin, spirits of the recently dead remembered and ‘feared’ by only a few. The most general term commonly used by my informants in reference to other dangerous spirits is the Swahili shetani (from the Arabic and cognate with Satan). In comparative religious studies a usual term is from the Greek daimon—daemon or demon. For that matter, the english term ‘fairy’ didn’t always refer only to the diminutive and generally unfrightening sprites kept on the children’s shelf today. ‡ Shetani and ancestors are not often foremost in a Kinga person’s mind, though when they are there they are being taken seriously. Still, it is hard to say that even the Python class of spirits is dreaded in the Sowetan world. The salient attitude is fascination. Animism offers spirits of many kinds and magnitudes, but generally does not teach that invisible beings are morally superior to the visible—or indeed that they are morally impressive as embodiments of evil or of wisdom. In general the spirits are amoral in the measure most mortals are most of the time—busy with their own concerns and mindful of others only when crossed or neglected. By contrast, the Greeks’ Olympians were hugely magnified and understood in 98

passionate terms, found fickle with a favorite, ruthless tricksters otherwise. Not quite a matter of intimacy, any approach to divinity entails moral proximity and awareness of danger. Given, when witchcraft takes place in worlds informed by animism, that the accused witch won’t be seen in the lurid terms of premodern European demonology, still Kinga can’t be said to have trivialized witches even in 1960. First of all, the wonders of witchery are much admired by youths. And after hearing many tales of dire doings and narrow escapes I saw I was meant to take the witch scene seriously. But then, I confess, I was taken aback one evening during a fireside discussion of some legal court cases which had just been heard: I noticed a man in the circle with the telltale marks of Kyiganga, the Ngonde witch-ridder who had already been famous in the region for a decade or two. The seer’s signature razor-cuts leave three parallel scars right across the forehead. Looking for the signs, they are plain to see by firelight though in ordinary daytime contacts I’d missed them. I started looking at once and found fully half my little group of elders bore those marks. The main point to hold is that witches cause about the same kind of trouble ancestors do, and have to be dealt with in a similar, indirect way. I never heard of a victim who, once sure he had found his witch, attacked bodily. Every move in the game boasts of indirection. The first reason for refraining from violence is the obvious one: witches are dangerous. Witchcraft is a fine leveller, but only as a six-gun was in the old American West. If only one man is secretly armed, the playing field isn’t level. The problem is one of honour, squaring accounts at once in a public and a mystical frame. Compare theft. When a thief is caught in the act, public moral indignation takes hold. A thief threatens everyone. Procedural intervention is not wanted, justice should be quick. But witches can’t be ‘caught in the act’. An accused witch and his victim could travel comfortably enough together on their two-day journey to the favoured curer in uNgonde. We need to look over the social maze witchcraft has to work in. One of the topics is ‘honour’—it lurks in cultures which don’t put it forward as moral ideology. The special brand of honour which is pertinent here needs explanation. The Swahili proverb is: ‘Asiye na bahati habahatishi: When your luck is out you don’t push it’. A fatalistic sense of honour is the mainspring of much of the subsurface action in societies which manage the gap between intimacy and public order largely through a religion of blame. Honour is not a phenomenon of psyche but perceived biography, and so not of your continuous inner life but episodic—the memorable features of your exposure to others. Another person can observe that experience 99

but only in a dramaturgical frame. This takes following your moral career as ‘situational theatre’. Social interaction is seen as it appears in story: episodic and moved by circumstance, a rivalrous game with protagonists and antagonists, confrontations and dodges, stratagems and alliances. As you fare well your sense of honour builds—it is nothing to flaunt, only your ego’s nestegg. Kinga always hid their treasure, hating to be thought rich. But when suddenly your luck is out the moral capital you hoarded seems to be gone like so much stolen gold. The moral significance of honour and its loss is coded into the central mystery of the ‘high cult’ of a Sanga court. Kinga of court culture have been taught to place blame for crop-threatening weather and pestilence on Lwembe, the quasiancestral trickster figure whom they elaborately propitiate. I dealt with the Lwembe cult in the final chapter of Realms, and there is a small literature on this royal wizard, who is shared with the Nyakyusa. His storied place is near the beginning of the Sanga period—the beginning of Kinga political time. In the lore, Lwembe’s uncanny powers frightened his older brother the king (prince, chief), who banished him. But some time after his death in exile at Lubaga in Nyakyusaland Lwembe began to act like a typically neglected and resentful ancestor toward all the Kinga. In keeping with his royal lineage and the high drama of his banishment, his powers of harassment were great and godlike in scope. The fate of all crops and all people seemed to be at stake. An elaborate cult of propitiation developed as a ceremonial centerpiece to the Sanga court culture. At the sacred grove in Lubaga this godling’s living spiritual heir was handled as a sacred being in whose body Lwembe could present himself, venting his discontents and prescribing the conditions for assuaging his damaged honour. The myth projects onto a high political plane the ordinary person’s encounter with ajali, doom. What the cult builds upon the myth is mystery: Lwembe is at once the victim and the witch. As a simple proposition this may not be very impressive. But dramatically stated instead, it proposes a mystery of the kind which can draw out a religious mind. Though rooted in an animistic world, as a feature of the Sanga protostate the Lwembe cult is an institution which magnifies its world.

Uses of a bush religion: the social contract The site of interest for now is the Malatan region. I deal here with the way Kinga, especially, and Nyakyusa used elements of a preexisting bush religion in building their systems of political 100

authority. Of uNgonde, the third Malatan protostate, and its sacerdotal ruler the Kyungu we have a constitutional outline and a rich history but not enough of the all-important sociology one may glean from an ethnography based in direct and participant observation. I assume what I find most probable, that a traveler in the early 1600s would have found a Sowetan culture area and within it hundreds of autonomous local communities occupying dozens of distinguishable dialect zones. Some borders, especially where they coincided with natural barriers to easy movement, would have been marked by stylistic differences in certain artifacts. But trade in essentials like salt, unhafted iron tools, and pottery would have distributed through the region whatever portables were not in short supply, and significant functional differences in material culture would have reflected local ecological adaptations—not ethnic or otherwise political alignments. As with material culture, so in good measure with practical activities—the division of labour, the kinship-amity balance, marriage and domestic arrangements, and ‘bush religion’. The one major divide within this culture area would have been the Rift Valley escarpment between the lands now occupied by Kinga and those of uNyakyusa. Evidence suggests the cultural heritage of the Valley ecozone significantly differs from that of the mountain slopes eastward. Two items are particularly striking. In gardening, Nyakyusa men historically put in far more regular work than women— the reverse of the traditional Kinga pattern. This may reflect a different heritage from prepolitical times, but not one Nyakyusa shared with their Corridor region neighbours. In religion, the Valley peoples know spirit possession and inherent, heritable witchcraft. But everywhere stories about witchcraft float free. They are feather-light folklore of a sort which knows no boundaries but language itself. What deserves the name of religion is belief grounded in praxis, and the Kinga-Nyakyusa difference here is fundamental. It firmly sets the cultural heritage of the Valley zone apart, and lends special interest to high-level ritual cooperation between the two protostates, which bridges the deepest moral divide within the Sowetan regional culture. The evidence is clear that Kinga court culture evolved from, quite as much as it was imposed on, the prevailing culture of the bush. The objective part of the evidence relates mainly to the qualities of homology and variance among the five protostates of the Sowetan archipelago. Oral history has it otherwise, however. It is sometimes premised, on the strength of parallel local traditions, that a westward movement of chiefly politics (the ‘Nguluve movement’ of leaders descended from one Nguluve, an ‘Adam’ figure) started at a sacred grove (Nyumba Nitu) in uBena and had progres101

sively taken over egalitarian bush communities in uKinga and Nyakyusa-Ngonde country. The oral sources on which the premise is based even seem to imply that the original bush communities would have been overrun and absorbed in this movement. But a myth is often easier to tell than play out in the time and circumstance careful historical inference can allow. Are we to imagine a full-blown political system (unexplained and unknown to tradition) somehow lurking in that proto-Bena sacred grove at the outset? The party would of course move quickly, with the pace of mythical beings, sowing political seeds and moving on? Or do we posit an exotic alternative to this African magical account, featuring small parties of alien movers (Arabs) settling in with the gift of civilization (Sultanism)? You may find a dozen versions of this ‘aliens’ myth explaining Bantu accomplishments in (say) Great Zimbabwe at any bad library. But the likely source of African tellings about travelers with a mission to rule is the diffusion westward from uKinga of a stereotyped ‘royalty’ claim invented by early Sanga leaders to set their name and progeny apart. Sanga is a patronymic. The claim to a special and mystically privileged origin would have been important once heritable authority was established on a small scale and the idea of translocal politics was in the air. The claim of a sanctified origin for a brotherhood of rulers would nicely suit the needs of an ambitious chief wanting recognition as primus inter pares among his political neighbours. The Sanga origin of this stereotypical claim is ticked by the fact that only Kinga informants even in 1960 could pinpoint Nyumba Nitu, the grove from which in myth the leadership teams spread out, each to rule a country. The client-ruler pattern of ‘sending out’ a lieutenant (K. untsagila) is widespread and clearly older in the region than the Sanga system as such. It harks back, presumably, to the period of regular migratory drift, when a small migrant group would keep in touch with ‘home’ and provide a beachhead for others wanting to join them. But while in court culture the client-ruler role is securely secularized, the ‘Nguluve migration’ myth is a charter for a kind of sacerdotal authority centred in a princely figure, an heroic ruler, and not a reclusive god-king. So far as I can judge, missionaries were the first to take Nguluve for a deity. Kinga elders would chuckle about his name (“pig”), suggesting it might have to do with skin colour—Nguluve was perhaps an Arab who settled in. The magical living divinities (chosen successors to the names Lwembe and Kyungu) are peculiar to the Malatan subregion and belong to an historically later overlay. Why such further developments should have taken place constitutes a major project of this book.

102

A realistic view grounded in archaeology and ethnohistory would presuppose roughly concurrent developments in each of the five homelands under consideration, with stimulus diffusion accounting for some parallels, and local idiosyncrasies—whether of heritage or fresh invention—to explain deeper differences. Three centuries are, by a common reckoning, no more than ten political generations. I count ten full reigns no more than enough for the more-than-minimal chiefly institutions of these three protostates to have developed and, while showing so many subtle parallels and deep local divergencies, stabilized within the region. A great deal of what the protostates have in common with their neighbours can also be found among the less chiefly folk of the broader region, bearing all the marks of what is integral to tradition. I have elsewhere shown this in detail and noted certain constitutional provisions the several protostates share with less-political neighbours—provisions of the countervailing sort which no authority would have imposed. They are safeguards against the abuse of power, and would have been maintained by entrenched local groups jealous of their autonomy. You don’t impose a social contract. So far as it is real and so long as it endures, a constitutional contract is a bargain about the balance of power. ‡ Kinga myth does not claim otherwise—far from bringing civilization with him, the first Sanga had to trade his daughter for the secret of fire. The same applies broadly to Ngonde and Nyakyusa ruling institutions. Each of these three ‘chiefly’ or princely societies remained into British times political islands bounded by bush communities. Each in its own fashion had institutionalized an expansive procedure designed to colonize peripheral communities and incorporate them into one roughly pyramidal political system. Historically, what we have to consider in each case is the transformation of a relatively egalitarian and localistic type-society into one which has erected an authoritative establishment toward which the egalitarian ethic stands only as a countervailing (‘antipolitan’) principle. I suppose the main creative force behind such a transformation must have been conscious and political. I have in mind the open and consensual conversion of ‘leadership’ into ‘office’ through routinization and substructural innovation. The elaboration and differentiation of supporting roles seems to be best conceived as a semiotic process, and symbols do not pop up out of whole cloth. Nyumba Nitu emerges from myth and hadithi as the shrine of a vaguely remembered, apotheosized ‘Adam’ figure, Nguluve. To call him a hero is to confuse matters. There are no hero tales, and Nguluve is known by scores of peoples in the same way, as their own. But the principal source of the symbolic front of kingship is necessarily religious, and in this the Malatan cultures follow a general rule. Each, 103

however, has had to find its own way to a working system of translocal authority, and it is by understanding their differences we are likely to learn what we can of their separate political histories.

Uses of a bush religion: the politics of fertility I have to depend mainly on the two (out of three) Malatan chiefly societies for which there are ethnographic field studies. But Kinga and Nyakyusa differences are instructive and throw light on the Ngonde variant as well. Though united in sharing the territorial shrine of Lwembe at Lubaga, Kinga and Nyakyusa contrast in the way bush religion was incorporated into their protostate systems. There were three main departments of the religion on which an early Sanga leader, with his priestly accomplices, would have had to draw: garden and reproductive fertility ritual, the ancestor cult, and witchcraft. I take up fertility first. It is the classic focus of ‘agrarian religions’. If there is such a thing as an ‘agrarian mind’ it must brood most on fertility and growth. Sowetan peoples had markets but no ‘market mentality’. Theirs was also a focus which takes in the importance of human ecology. The conditions of survival and reproduction are necessarily different for Kinga in their high hills compared to Nyakyusa on the more populous, verdant, and ethnically varied valley floor. Intensive crop cultivation was coupled with cattle herding throughout most of the Sowetan region. But cattle, being hard to keep in the higher mountain slopes ecozone, played no important role in the rise of Sanga politics in uKinga. For Nyakyusa, cattle were the prime form of mobile wealth, and responsible as a predictable consequence for a good deal of political mischief. Without them, and the bridewealth system of marriage which made wealth of women as well, it is hard to conceive of the Konde social scene developing as it did. Modeling it as a sort of ‘factory system’ you have a chiefdom which at the high point of its developmental cycle comprises up to a dozen ‘human production centres’ each densely inhabited by a (non kinship) group of polygynists with their many wives and children. Peripheral to each of these production centres you have the ‘bachelor closes’ which are home to (non kinship) peer-groups of youths or young adult men comprising the workforce and fighting capability the production centres require. Scores of such chiefdoms comprise the country of the Konde—the linguistic community of Nyakyusa in Tanzania and Ngonde in Malawi. A certain degree of symbolic unity was affected in traditional times in the Selya area by virtue of the territorial shrine located at Lubaga. Local priests in 104

troubled times cooperated with Kinga in ritual address to Lwembe. His anger was thought by Kinga to be the cause of irregular, widespread (‘territorially’ inclusive) troubles. But where Kinga made a politically unifying thing of seasonal plantings, localism prevailed in the Nyakyusa approach to hydrological mysteries. Rainfall in the Valley does have visibly local patterns, and the regular seasonal round of garden crops was a local concern of each chiefdom and its rainmaker. Nyakyusa ‘rainmaking’ devices are the usual set for all of the Eastern Bantu civilization and certainly antedate chiefly politics. For the chiefdom we know best (Mwaipopo’s in British colonial times) the rainmaker was the hereditary priest Kasitile, who claimed collateral chiefly descent. A few samplings from the Wilsons’ fieldwork will elucidate some of the techniques and suggest the way this particular bit of lore fitted into the pattern of Konde social life. [Godfrey Wilson on the more centralized practice in uNgonde:] Kyungu M. had his rain-makers...Mugoma was a slave, Komanga a man of Ngonde. When the sky was hard he sent a message to them saying: “Put them into water.” If there was too much rain he said: “Take them out.” Each had two potsherds with two stones for each. To make rain they rubbed the stones with mutton fat and a medicine called ilifugo (fertility) and put them into the sherds and poured water over them. [Godfrey Wilson went with Kasitile to see his rainstones:] There was a large broken pot and three stones, two of them with holes an inch or so in diameter. There used to be five stones—two had been stolen by the Ngoni...When he wishes to make rain, water drawn from a waterfall is brought by Kasitile’s wife and left overnight in the grove. Then he pours it into the pot and puts in the stones... ‘the water bubbles; it is cold, but it seems to boil and the ifula go whirling and knocking against the sides of the pot’...The bubbling of the water is compared with rain. Kasitile did not know where his stones had come from. Mwakisambwe his great-grandfather had them first, then his grandfather, and father. [Kasitile is interviewed:] In theory the rain-maker always works under the direction of the political authority—the Kyungu or a chief...But when Kasitile himself was angry...the weather grew bad. ‘I thought, “It is not the custom with our forefathers for the tax collectors to come at all to men like me.” I grew angry and prayed...When I was in MuNgonde the country suffered from hunger, all their children were in trouble... Then I came up, and began to prepare my rain, and made the country good. Food started to be plentiful. Why should I be treated thus? I asked. And the shades and Kyala [a divinity] heard me, so it rained heavily (too much). Do I grow angry by my own power? I do.’ ‡

105

I suppose garden rituals since early Neolithic times have been co-opted everywhere to some sort of political use. Just here, the puzzle is what, in the historical situations of the Kinga and Konde, should be held to account for one particular cultural difference: the continuation of localist ‘bush religious’ praxis in respect to crops among Nyakyusa, as compared to the fuller cooptation of traditional rainmaking into a centralizing court culture among Kinga. It is useful to have Godfrey Wilson’s information on Ngonde praxis: here are rainmakers clearly subordinated to a centralizing monarchy with traits of ‘divine kingship’ absent in the Sanga ‘high princely’ courts. The association of ritual calendars with authoritarian control in the primordial agricultural civilizations of the Old and New worlds is well known. But in the absence of an élite theocratic class the weaving of ritual into the political order is sure to have been less than wholesale at the stage we have to consider for the Malatan region. Globally, garden ritual—grass roots religion—among decentralized cultivators is often associated with lineage or affinal politics. Where local descent groups prevail in combination with ancestor propitiation, the lineage elder typically plays the part of priest-officiant. In mixed-lineage communities such as we have in the Malatan region seniority in respect of local land rights plays a similar role. In the right circumstance, the intensification of a chosen person’s authority is predictable, and he becomes a pop-up spokesman or adjudicator whenever one is needed. As folk learn to defer to him, a conspicuous elder can find himself empowered to represent his community politically, whether in mending a broken peace or building an alliance where self-help (reciprocal use of force as a private, transactional system of sanctions) is not doing the job. In the Kinga case, one may guess, priest and warrior were from the start separate roles. But that surely would not have been the case with the Konde, where we see the two callings everywhere partly merged. Thus Kasitile calls himself chiefly and has hard things to say about the ‘commoner priests’ who are only aging village headmen. Konde chiefs also are felt to grow stronger spiritually as they grow weaker physically, and there are points in her monographs where Monica Wilson is clearly in favour of viewing secular chiefs as ‘divine kings’ in the manner of their death. I don’t find myself convinced, but I find the ring of truth in the notion that ‘still growing’ body parts (hair and nails) would have been taken from a powerful chief’s body in traditional times, to renew the chrism priests used in the inauguration of a successor. Why would not Konde priests, like Kinga, conceal the death of a powerful ruler until the stage was fully set for a successor? But ordinarily this would be unnecessary, as

106

ordinarily a chief would have spent his charisma and seen the identity of his successor assured long before death came. In the standing literature the presumptive course of development for a chiefly political system starts with some sort of social and economic leverage accruing to a central personage within a sizable pedestrian community. Sometimes there is a (seasonal) surplus of storable food, and the political leverage develops through the emergence of a ‘redistributional’ system centred in an authoritative office. Looking at such a system in simple secular terms there would have to be a flow of storable food to a centre, from which at need it would flow out again. This could make economic sense if it would hold down waste in ‘times of plenty’ by transferring supplies from private to ‘public’ hands, in favour of relief in a season of scarcity. But it only makes political sense—and proves workable—if this kind of skeletal redistribution is dramatized in the form of ceremonial feasting. This gives us a social-contract model for chiefly politics: moveable wealth, centrally stored for the whole pedestrian community, is defended there under a chief who maintains his power by feasting the people on a regular basis. The Sanga system can perhaps be fitted to this model, but the Nyakyusa chiefdom not—it is something else. ‡ What both Malatan political cultures do have of this model, and have in spades, is the ideal of the generous chief. It is his own wealth, though, in the Konde world, which he is asked to share. A Konde chief has two main wives and, with the rest of his peers, a continuing influx of ever-younger wives. He has sons galore to do his gardening and see to the expansion of his herds. His wealth is inherent in the system. But at the same time he must be ready to spend his capital wisely. Nominally, this may mean no more than placating his shades when they grow restless, but practically the bull must be dismembered and beef shared about to all the usual support groups. For the Sowetan region the advent of chiefly politics would have come in the Later Iron Age and propagated the lifestyle we associate with the ‘medieval’ period of Eastern Bantu civilization. New crops and techniques have filtered in. The presence of ample grains and pulses, easily desiccated for storage, opens the possibility of redistribution, but in most cases would not be enough. Where every household has hardworking cultivators, can store its own food, and regularly supplements by hunting or trading, even seasonal scarcity is rarely an emergency. Herd animals introduce a new set of concerns, inevitably ‘political’. Seasonal consumption gives way to longer term considerations. If a herd, with continued luck, will reproduce in geometric proportion to the years invested, so 107

does the risk of loss and the intensity of any future expectation of wealth. Herds are wealth as few crops in a subsistence economy are, and the presence of alienable (redistributable) wealth begets a new politics, quite likely armed. Without cattle, can we offer a set of equations which would account for the Konde age-village system? Cattle, sheep, and goats are important in the Malatan region. Hehe-Bena peoples into the nineteenth century were organized as scores of autonomous chiefdoms with more than passing interest in one another’s herds. Nyakyusa expansionism was traditionally focused on cattle raiding. Once the (deeply political) decisions had been made as to whose wealth a mixed group of youths had brought home, further decisions must be made as to whose new brides the new cattle might be traded for. Nyakyusa chiefdoms can be shown to be structured around the problem of keeping a large cohort of young men available in every village cluster, all prepared to work their fathers’ fields and defend their fathers’ cattle. They won’t do that when they have wives and kitchens of their own. For adventure, in the spirit of young men without women, they were groomed to enjoy undertaking dangerous raids on far off herds, to increase their fathers’ wealth in cattle and wives. Both kinds would eventually trickle down to them. But this was, as Monica Wilson reports at length, a formula meant to please men, and among men, elders. For our concerns here, it was a culture of affluence. Staple crops, particularly banana sorts, were abundant in all the favoured parts of the country and predictable enough to be of comparatively minor concern in Nyakyusa religion. In effect, ritual attention to crops was either left to the rainmaker—the same specialist the ‘founding chiefs’ would have found on their arrival in the country—or it was left to routine, being incorporated en passant into chief-centred rites. The first case is a little ambiguous. The Wilsons’ rainmaker, Kasitile, was, as one might expect in the region, a priest who had inherited his rainstones through agnatic descent and learnt his arts through apprenticeship. As such, his claims (though accepted by Monica Wilson) to the bona fide half-chiefly status of ‘younger brother’ to a ruling line are probably best considered fictive. The claim refers to a connection remote enough in time to lie beyond the reach of evidence. I read it here as I read similar claims from Kinga priests at Maliwa. The claim to a ‘royal’ connection had in colonial times become something you could scarcely deny in uKinga without seeming to deny a distinguished social position. At ordinary rates of reproduction, thousands of Nyakyusa and Kinga have grandfathered claims to royal blood. But the Wilsons’ rainmakers are wanted to act on a chief’s behalf, not their own. And though it is clear on evidence that a 108

Nyakyusa priest ‘of chiefly line’ was far from enjoying chiefly privilege in his community, it is just as clear that he was socially aligned with aristocratic interests to a point that justified a ‘chiefly’ or ‘princely’ demeanor. When must a lion tamer hide his pride? Sacrifices are made to chiefly Nyakyusa ancestors at the appropriate sacred grove. Commoner priests alone actually enter the grove, excluding the chief—he is inherently inadmissible in spite of his claims of descent. Then (whatever the occasion of the sacrifice) blood, bits of the dedicated meat, and beer will be spilt on the earth (and in the frame of religious mystery, right down through) at the shrine-stone in the wood, for the shades to feast on. Always there are “petitions for rain, fertility, and health”. The occasion for such a sacrifice might be an omen or a dream. Often the pressure would be fed by personal anxieties and sometimes by more than just a bit of bad weather; but always the bullock of sacrifice would have to be wheedled from a chief. These were the projects of priests, and any priest would be in position to touch more than one chief. Quite often, it was the quarreling of priests and chiefs which was blamed for a drought or too much rain, or for garden pests. That way a priest himself could become the occasion for the needed sacrifice, though he would obviously be blaming the chief for the trouble. Comparing the two cultures in this frame, the Nyakyusa look like a bunch of omnidirectional entrepreneurs. Kinga rely less on hustling, more on ritual theatre. Their priests, taken realm by realm, comprise compact groups. Whatever the internal differences, the group wants to move in one direction. It aspires with some success to solidarity with sister groups in other domains and realms. While Kinga youths were concerned as much as Nyakyusa with protecting their fathers’ herds, goats are a different animal from cattle, quick to scatter, safer by nature from predators, and marginally worth less as wealth. Herding was left to young boys. Easy come, easy go—goats had only nominal importance for the high purpose of marrying. And goats are very hard to steal. Older Kinga youths liked the adventure of a cattle raid, but (apart from any cattle which might be taken by chance in pitched battle with Kinga of a neighbour realm) preferred stealth and the magical arts to a bellicose style in their long-distance forays. As for gardening, ninetenths of the gardening work you would see done by young men in Nyakyusa country is done by women (of all ages) in Kingaland. On the other hand, in both cultures everyone eats alike, and Kinga dependence on fallible staple crops was always greater. The garden accordingly enjoys a loftier place in Kinga religion.

109

Though gardening for Kinga remained an individual enterprise, only cooperatively pursued through the customary ungovi work bee, the seasonal timing of a few traditional crops was calendrically regulated, as if to set apart these staples as the foods at risk of mystical punishment. When particular plantations were affected by pestilence the event would be treated as witchcraft or ancestor neglect, and as such would not be a political concern. But the kind of pestilence or drought which might strike the whole of a domain or the larger realm was the kind which could merit political blame. So the mechanisms of political-ritual theatre are called upon to exonerate the high court in advance. Any court claiming to possess a countervailing mystical power of defense against divine anger has exposed itself to blame when its powers are deployed in vain. The dramatic task of an elaborate Kinga planting ritual, buttressed by strict taboos on early planting, is accordingly to tag the danger to the advantage of priest and prince. If these taboos have not been broken, the buck passes upward. But blame must fall not on neglect by the Court. Some other, diabolical mischief must be discovered which only the Court with its priestly diviners could hope to identify and neutralize. Theodicy is the central contradiction in a religion of guilt. If a perfect God made us as we are, why did he intend us to know evil? A religion of blame derives from the great tradition of animism and does not have to account for inscrutable or in any way perfect gods, as all the evils of the mystical sphere are quite naturally inherent in the imperfect beings which command its forces. Kinga priests were the lawyers of a mystical arena in which these forces must be dealt with. Or we may think of them as lion-tamers. Python-tamers would be more exact. Nyakyusa say the snake in the belly of their chief is only the match to snakes in the bellies of witches he must fight in the community’s defense. The important ancestors whom the priests propitiate are those of their great chiefs. Hair and nails taken from these chiefs allegedly before death are key ingredients in the medicines a priest has hoarded away for use in the protection of the living. And the forces he must protect his people from are, more than likely, writhing in the bellies of just those chiefly ancestors in their underworld. The central contradiction in this religion of blame lies in the powerlessness of a living chief—a future godling—directly to intervene with those of his own blood who, having gone before him, endure only in ill-tempered envy of his own more robust living state. But it is a contradiction on which a priest must insist at any price. It is a special Sanga feature that, while the calendrical regulation of the gardening season (since it must be suited to local circumstance in each case) was carried out on the level of the local 110

domain, a fiction was developed that each sacred garden belonged not to the court of the local Lord [Untwa] where the garden lay but to the Prince [Unkuludeva] who ruled the inclusive realm. Considered as a stratagem, the effect was to pass the buck of blame along with the mystical burden of (dare I say?) ‘crop insurance’ to the higher court. It is not a trivial advantage of translocal authority as Sanga organized it, that blame is removed from the personally accessible to an inaccessible ruler. Ultimately, and it seems in traditional times pretty frequently, the great Sanga procession to Lubaga off-loaded the blame—accumulated at the courts of all the domains and realms—on the master Kinga troublemaker, Lwembe. High political theatre was the perk of a princely court. When the royal harvest was carried there in ceremonial procession from an outlier domain, the accepted meaning was not imongo (tax) or the rendering of a rent-like tribute but recognition of the mystical incorporation of this peripheral community as part-society within a greater realm. If that message was sometimes lost on an ambitious local ruler, the annual ceremonies which acted it out were vital in the popular mind and (if only in view of the human energy invested) unstoppable. At Maliwa the royal garden was devoted to sorghum (planted in January, harvested in August) because finger millet, the ‘original Kinga’ grain, did not thrive there. The procession would comprise some fifty persons. The Ukwama prince’s reciprocation was to present them with the makings of a true feast, a cow and sufficient beer, which they were to carry home. The meaning of this transaction is not at all what it would be if so large a contingent from Maliwa were to stay and feast at the Ukwama court. The two domains in such a hegemonic relation can’t be allowed to merge under a princely patron. This show is not about Sanga politics but about access to redemptive mystical intervention—access by the one (the lesser) autonomous polity, made through the more learned and powerful ministers of a sacred grove located in the other. But the people of Maliwa, no doubt, know how to read the subtext, which can only reinforce a hegemonic order in the realm. So far as the fertility of the royal gardens somehow contains, implies, and assures the fertility of all the others, hope can be seen to ride on the back of the prince. But in this drama the role of a secular ruler is by no account that of demiurge. Inspired ‘defender’ though he be, his special powers are those of an extraordinary, headstrong colleague whose existence endangers the cosmic order even as it lends its strength to it. Human fertility in the bush culture was not in itself a religious concern but was addressed by implication in rituals meant to assist in childbearing and childhood survival. For Kinga, as the ambisexual 111

ethos of the courts evolved, the tendency of the priests was to magnify the princely harem and press the incumbent prince to consider it his real workplace. Once installed, a new prince was abruptly removed from the homophilic world of ordinary manhood. The dangers of that world were not mystical: the problem was poison. As men grew older they learned to fear their friends. A ruler took no food or drink without seeing his official taster survive it; and a ruler slept only with women. In this way the heterosexual principle was given its champion, and the court was eventually magnified by the presence of his many comely daughters, the new set of princesses to whom men must defer. Picture these maidens going about always in groups, always light-heartedly tilling the family gardens in sisterly teams. Picture them singing as they worked, but don’t see Ophelia—perhaps rather Atalanta. When a princely court was well enough established, even a youthful ruler would not lead in warfare but let a local champion wear his kingly gear for the day. The prince should be dedicated, as his priests would assure him, to the work of the harem. It became the right of the throne to see that the avapapwa, the non-inheriting sons or ‘younger brothers’ of the royal line, and any of their barrack mates who would accompany them, were suitably married when the time came to send them away from life at court to settle and ‘rule’ disputes in some part of the periphery. In this way the work of food production, storage, and preparation at court was principally done by bachelors—maidens full-time, barracks men part-time. It was only with a man’s and his (royal) bride’s removal to the bush that their own delayed fertility began. There are parallels to this pattern in Nyakyusa tradition. There was delayed marriage for men (if hardly for women) and a massive ‘sending out’ of young men and women, a redeployment in space of the whole new generation, when it came time for the passing on of a right to rule from their elders. While Nyakyusa knew no ‘court and bush’ distinction, they lived in small, internally close-knit villages not spread out in hamlets. Each chiefly village would be surrounded by eight or a dozen commoner villages each with its own commoner head. So (as with Kinga) only a fraction of the population was politically close to a ruling prince and the ‘python in his belly’ which gave him sway. To be effective, political theatre requires this provision of aesthetic distance. The details of the passage of authority—how the Nyakyusa Coming Out ‘must have worked’—have been a favoured subject for term papers in anthropology since the system was first described by the Wilsons. For the present purpose it may be sufficient to have in mind two results: (1) Women, cattle, and fertility were concentrated in the polygynous families of men past their youth, all gathered in the stem villages of the chiefdom while their 112

sons, living in satellite villages of their own, did most of the work in the parental fields. (2) While the orderly royal genealogies made up for Monica Wilson belie it, every chiefdom really was painfully regenerated from a near-inchoate state every thirty-odd years with the formal abdication of governing authority by all the village officers. This was not a succession of new men to old offices but a managed secession of young men, moving from the chiefdom of their birth to one cloned from it, with a full structure of new offices. A Konde chiefdom unambiguously consisted in people, not the ground they happened at any particular time to occupy. It is important to see that it was not only the chief and his two villages of peers who passed over their powers and responsibilities relating to political reproduction to a new chiefdom, for as part of the same operation all the headmen of commoner villages comprising the body of his chiefdom did the same, stripping themselves of their young male workforce. This made for a combination of energy and social turbulence unequaled in Kinga country even in the two ‘frontier’ realms, East and West. Remarkably, as judged by the high morale and good life found there by Germans on their arrival in the 1890s, the Nyakyusa system of literally liberating youth at periodic intervals appears to have worked. The puzzle of the Coming Out is the Wilson’s great legacy to anthropology. Reconstructing it at closer range will be a project of later chapters, building on a fuller understanding of the Nyakyusa world.

Uses of a bush religion: ancestors & neighbours It is important to see some of what is implied in saying the Kinga protostate was not based on ‘kinship politics’—and why substituting an alternative ‘politics of amity’ only improves the focus slightly. A kinship constitution produces a segmentary society. If it has an apical office, its single occupant can only belong to one or another segment, and to hold the office he is virtually obliged to use tyrannical means. At his death, you must expect odd and problematic mechanisms of succession. At the death of the first Hehe tyrant Munyigumba, his military state fell instantly back into its original components. A ‘divine king’ can perhaps provide an escape from the particularism of a segmentary society but can’t give it secular unity. Kinship politics is also burdened by its inherent gerontocratic tendency, since the ‘natural’ leader of a kin group is its senior member of proper gender. Entrepreneurial politics of the kind we find both in the Sanga courts and the Nyakyusa chiefdoms would be stifled.

113

The Lwembe figure (Monica Wilson’s divine king) when seen from uKinga does serve to raise popular expectations above the level of a particular domain or realm. The great ‘pilgrimage’ of priests from all the Sanga courts to Lubaga in the Rift Valley evidently lumped the ethnically distinct Kinga and Mahanzi communities together, probably from the time in the nineteenth century when they were chronically fighting, and perhaps from before that. The same figure, seen from the Wilsons’ uNyakyusa or as a feature of the larger Konde community, being one amongst several rival divinities, can perhaps be regarded as a politically ‘neutral’ resource but hardly a unifying one. The formerly parallel figure in uNgonde, the Kyungu, did so evolve politically in the nineteenth century as a supra-segmental unifying force but had to do so, in good part, by shedding his traditional mystique—his implicit, not-visible powers. When a Nyakyusa man shifts from one chiefdom to another, as many do, he is turning to a new set of chiefly shades for his spiritual concerns. It is a move on the chessboard of personal destiny. Old men, it is true, are often found muttering to shades of their own. But a young man’s moral strategies are laid out on a wider gameboard. It is a lesson of the influential cults of chiefly ancestors that the welfare of a chief’s people is watched over by his shades, rather to the exclusion of their own. Thus Konde are supposed to go through chiefly ancestors to divine the reasons for troubles they share with their neighbours. Or to put it more succinctly, discontent can be blamed on the displeasure of dead chiefs. Kinga use of Lwembe as the butt of blame runs parallel to this but on a more inclusive scale and with a special twist. Lwembe is not a dead chief but his too-magical younger brother—a dead tyrant’s victim. The worldview which supports this kind of rhetoric about life and death and its meaning is older than the Sowetan political archipelago, and reflects exactly the transformation from an acephalous to a chiefly philosophic context. It is the differences between Kinga and Konde institutions resulting from that transformation which concern us here. My premise is that the differences between the two (bush) religions lying back of these two protostates were less before the political escalation began. That would be because the bush religion in the time of acephalous politics gave importance ‘democratically’ to Everyman’s dead, whereas in the new contexts we have to deal with a magnified ‘sacred ire’ ascribed to the shades of high office. While we may use such terms as ‘chief’ or ‘prince’ or ‘noble’ in reference either to Kinga or Konde officers, their political systems developed each in its own way.

114

No one studied the Kinga in the 1930s, or we should perhaps have had cases comparable to those the Wilsons offer, from which to judge the relative importance—frequency and severity—of witchcraft accusations in the two cultures. As it is I can only proceed, having uttered this caveat, on a basis of what I know. I find a greater prevalence of witches in Konde than in Kinga culture, and offer these cursory explanations: (i) The stresses of domesticity are distinctly less in Kingaland, for the reason that domesticity there is minimal. It is mystical (and mystifying) personal hostility which explains witchcraft, and this comes to the surface within close personal range. In particular, kinship ties are maximal-claim ties. You carry them around wherever you go, you can’t move away. Emotional involvements within this sort of tie are deep, complex, and ambivalent. By minimizing them, Kinga reduce the neurotic load they carry. (i) Where it comes to peer relationships, there is an important difference between the two cultural settings. Kinga amity is built very largely through a network structure. The older a Kinga male becomes, the freer (and usually wider) his pattern of association. The pervasive Konde amity structure on the other hand is collective: it is based in ‘compulsory voluntarism’—if you are not convincingly friendly you are asking to be named as the witch whenever untoward events break the brittle surface of village amity. (ii) Then there is the pervasive muttering of old men in the Nyakyusa setting. Each has pursued his moral career between kith and kin. Over time the volatility of ‘kith’ (village peers) impresses itself on a person who cannot himself confront the invisible masters of his destiny. He sees things falling apart. Perhaps he has had to admit he’s had witches for friends. We see him muttering where he sits facing the ground, not just to himself but to some other. ‡ Nyakyusa men’s moral strategies must rely on intense but fragmentary kinship loyalties. Local descent groups were rigidly disallowed, to the point that own brothers (for presenting the threat of dyadic combination or withdrawal within the new peer village) were never sent to the same close. The solidarity of the Nyakyusa village is that of propinquity not prescribed dependency. The ruling ethos is amity, rigorously maintained. Kinship does, all the same, pay off by giving ‘vertical’ structure to a chiefdom. If lineage had no credible value, wouldn’t even father-son filiation drop away? Then how would fathers, so busy collecting wives in their polygynous mother-village, manage their grown sons well? Nyakyusa had no translocal military arrangements. The sons (though not quite as among Kinga) live in their own small peer groups, not with brothers or cousins as such but friends. Yet they are all prepared to do the fighting to win cattle for their fathers’ new child-brides, and do most of the stoop work in all the parental kitchen gardens as well. 115

Nyakyusa couldn’t afford to let the mystique of agnatic inheritance pale. In going so far as to divert inheritance of wealth and wives down the full list of a deceased’s younger brothers, until finally reverting to his eldest son, they balance the structural weight of descent with that of sibling seniority. This offers a distinctive way of mapping a man’s social identity in light of his longterm expectations. The kinship map comprises an overlay on the collegial map of day-today life in a world where men grow up in peer-groups, not families. Mature men are important on the Kinga scene. They certainly run the courts of law and generally support each other. Around 1960 they were fast learning how to use bridewealths to get rich, and money management to get status. But they are not, like their Konde counterparts, a visibly wealthy class. Even as fathers in the longago, amid scenery now lost, their moral careers led into none of the domestic grandeur of a Nyakyusa polygynist. For a Kinga father it was good enough to say ‘a boy’s work is roughing it’. That is, after a boy is seven or eight a father loses control and, with it, responsibility. By the time a Kinga son is ready to marry, his father has quite finished with the work of marital relations. If old men still hold centre-stage in a Kinga community it is mainly in and through their minds. A Nyakyusa boy knows the cattle he needs are in his father’s keep. They will have to pass through the hands of all a father’s younger brothers (normally in the same chiefdom but not the same village), then through all the boy’s older brothers before coming to him. It may well seem to him his best hope of jumping queue is playing the dutiful son or nephew. Monica Wilson devotes a full volume of her ethnography to the rituals of kinship which are put on to teach that lesson. It is one Kinga know little of. Their animals aren’t milchers. The few cattle-keepers I knew kept a byre in the spirit of herding goats, doing it half as a gratifying hobby, half as cash business. Everyone said that before the pax cattle were rarely seen unless at a royal court. A goatherd has no great expectations. ‡‡ The Malatan peoples are decidedly not ‘tribes without rulers’, nor could their constitutions easily have descended from any nicely segmented kinship societies. ‘Tribe’ is in fact a misleading word to use anywhere in the Malatan or the broader Sowetan region. All three major peoples have convincing traditions of receiving and incorporating immigration without segregation on the basis of ancestry. As this has been the source of some uncertainties concerning the Nyakyusa, who do emphasize and celebrate lineage loyalties, and since comparison on this point will help to clarify reasons for the eclipse of the shades among Kinga, I’ll first approach the topic of 116

Kinga ‘ancestor religion’ as a variant of the Malatan pattern. It is not accidental that Kinga share their ‘high god’ Lwembe with the Nyakyusa, and can claim to share their line of ritually sanctioned rulers with both Nyakyusa and Ngonde. The Nyakyusa age-village constitution cuts across their lineage system, scattering cousins and siblings into different local groups. Though the highest offices are considered hereditary in the agnatic line, the system of accession at the chiefdom’s Coming Out ceremony is bizarre enough that some anthropologists won’t believe what Wilson has described. The institution proves worth circling back to from time to time, so different facets can be seen. The point just here is that the Coming Out anticipates the disengagement of the older generation of leaders from active politics, and does so by reproducing itself in a new generation. The new chiefdoms (for two at least are produced) are parented by the old, which retains its forms. Every chiefship is doubled, every headman below a chief must be chosen on merit from among commoners with no traceable connection to a chiefly (princely) line. Neither of the two new chiefs automatically inherits a father’s political weight. Charsley stresses in his review the absence of a ritual conferring office to the old prince’s sons, suggesting (in an “extreme formulation”) that the Coming Out only qualifies a new princeling to “seek to establish a ruling position for himself in competition with other princes of his own and of preceding generations.” It is because a Nyakyusa chief’s power is so personal, depends so on the appearance of ‘pythons in his tank’ or the mantle of charisma, that I suppose his magnetism would have tended to overreach the talent of a young successor. From that I take the view that, at least in many cases, the Nyakyusa Coming Out must have been the moral equivalent of the Sanga prince’s “sending out” of parties of braves with their new brides to establish new settlements affiliated to his own. The magnetism of a strong central court leads to demographic disproportion, political cross-currents. and a need to recirculate many members of the oncoming generation. It is probably true enough that most of the men a Sanga prince ‘sends out’ take a humble or at least a humdrum role where they settle— only a few become rulers or sire any. It must be true also that many Nyakyusa who Come Out with grandiose expectations are later found settling for the quiet life wherever they may find themselves. But probably a few do ‘colonize’ and so extend Nyakyusa politics a bit beyond its former boundaries. In each case slow growth would follow, though the political rhetoric may call for much more. As soon as we accept ‘entrepreneurial politics’ as the fittest model for the Nyakyusa case, these distinguishing consequences seem to follow. 117

Nyakyusa princes ranged from rich to poor in land and cattle holdings. The wealth of their local states varied radically because the rivalry of their realms was multilateral and intensely antipolitan— and all this despite the bias toward ascriptive agnatic alliances celebrated in their kinship rituals. Kinga by contrast lack any such histrionic occasions for the assembly of lineage members, except for funerals; and at these (as in fact at Nyakyusa funerals as well) the vast majority in attendance are not kinsmen but neighbours. Each society in its own way demonstrates the clash between politics and kinship. Neither culture can be said to have integrated the two. But it is the appropriation of the worldview and religious beliefs of the bush, turning its rhetoric and structural norms to political use, which allows the developing courts to give fresh form to the societies which cradled them. The business of politics is often conceived to be the allocation of public goods. But looking at politics along with kinship suggests the particular importance of allocating social identities. Kinship plays on the strings of family to net in uncles, cousins, ancestors, and descendants up to a given degree of relationship as the group to which ego ‘belongs’ and on which he or she must depend for social sponsorship of any important kind. More purely political systems of allocating identity must depend on some other basis for social alignment than kinship, and the tendency is to substitute kith for kin, propinquity for fraternity, locality for ancestry. This means we have to deal with two systems of dual identity, the one (Kinga) fuzzy and flexible, the other (the Nyakyusa with their lineal inheritance ladder and tightly bounded peer villages) quite clearly delineated. The main implication for our understanding of Malatan religion is that individuals in both these populations will have to deal with some added existential uncertainty in their dealings with the dead. With or without the imposing institutions of princely politics, most people most of their lives are deeply involved in efforts to control their individual destinies. Collective representations premised on political solidarity may have limited relevance to these concerns but can’t avoid stealing fire from the shades. Monica Wilson lists a series of offenses, all of which may be described as family quarrels, which are “directly punished by the shades.” But she concludes: ...The Nyakyusa speak of the danger of angering senior relatives, but they attribute most of their misfortunes to their neighbours, and they say that more misfortunes are due to neighbours and fewer to the shades than is openly admitted! ‡

118

Attributing misfortune to neighbours means witchcraft. In Nyakyusa culture, which is heavy with kinship ritual and ancestral lore, it is easy to see this as displacement. In fact, Nyakyusa shades are always thought to bring trouble on a younger person for disrespect of an elder. This man “mutters over the fire”—the shades hear, and are moved to punish with a disease or an accident. If there is a propitiatory sacrifice, it is not the shade but the living elder who must be reconciled. Such dense facework within a family is unthinkable for Kinga. Their bush culture, alive but often overshadowed in the neighborhood of the courts, seems to have struck the rough (if logically opaque) balance, typical for Eastern Bantu peoples, as between the ‘neighbours’ and ‘shades’ schools of thought in the placing of blame for misfortune. The initial appeal to a diviner calls for a choice between grievance and repentance. If the shades are angry the client should repent a neglect of duty and make amends. But if a neighbour has sent the trouble the injustice fixes automatically on the other. That is, a victim always earns the ill-treatment he gets from a ghost, but never deserves the mystical cut he may get from a wife, a brother, or a friend. The diviner’s call as between ‘neighbour’ and ‘shade’ decides the path a client will take in his quest—accepting or deflecting blame. If the client is inclined to accept the diagnosis, he will act on it, but need accept no responsibility for so doing. He has paid someone else to do that, and this was a disinterested person who acted as catalyst for an oracle. Under Sanga court culture the diviner was one of the prince’s men, avanyivaha, and made a similar choice, but in a new political and administrative context. Let a threatening illness befall the royal house itself, and the diviner must find either for hunting out a witch within the fold (which entails using the ordeal and can wrench the ties which bind a small élite together) or for propitiatory work in the sacred grove, which has broader implications than the ancestor cult of bush religion. It is just here that, for Kinga and perhaps for their Malatan neighbours, the exact character of the mythical shade Nguluve can be defined. He is the bridge between bush and court religions. When an ordinary Kinga man ties a goat at home in the morning and prepares a trip to the place where he can sacrifice it to an ancestor, that is usually where his own father was buried. At least for the second-generation ‘Westernized’ men I knew, there might have been little need for divining a specific ancestral complaint. When a subject is experiencing the spongy kind of reversal of fortune we associate with ‘depression’, he will have learned to blame neglect of filial duty. The trip to an ancestral site will bring him ‘home’ in a spiritual sense, and the feast he will host there will find him 119

among friends, old and new. In his supplication he will simply offer food, drink, and greeting to the shade(s) he addresses and ask that his gesture be shared with a few (named) ancestors of his mother’s agnatic line as well as his own, then to certain (named) friends he lost in boyhood “and all the other children,” and finally to Nguluve, who is understood as the apical ancestor of all the living persons present at the feast. They may be his own real or fictional kinsmen, or simply folk now living at this place. What the address to Nguluve implies is that all the shades a Kinga man knows comprise one face-to-face community, and this community includes as well all the shades these named persons would have known, and so on to the beginning, which is Nguluve. Kinga royal genealogies do not begin with Nguluve (as do the syncretic genealogies of Nyakyusa chiefly lines put together by Monica Wilson). The reason is that Nguluve is conceived as an ancestor far more remote than the first of the Sanga line. Antedating the division between court and bush cultures, he is no hero and was no ruler of men. As such he embraces, if not all humankind, at least all in the subject’s own world. He is certainly the ‘Adam’ figure not just for Kinga but for Bena, Hehe, Sangu, and other Corridor peoples going back to earlier Iron Age times before intensive agriculture with efficient earth-turning tools had begun. When a priest in the sacred grove at Ukwama deals with the shades of the royal line, he addresses the whole community of the shades in much the same way a commoner would at his own temporary, home-made shrine. Apart from some obviously Christian versions of the matter, I never heard a suggestion that Nguluve ‘rules over’ that community or possesses any unusual powers. Protostates do tend to recreate the invisible cosmos in the new image of the visible, but Nguluve belonging to all belongs to no one. The sole implication of the rituals we are given is that among shades Nguluve merits the universal respect he is shown by the living. He is cast of a different metal entirely to Lwembe. A man would not ask Lwembe to greet childhood friends. But beyond saying this, I have no idea how the priests at Ukwama would have characterized the godling figures for which we lack mythic support, or played them off against one another. The last man reliably reputed to be master of that sacred lore died very early in my residence in the field. What we will see is that Lwembe, though ‘younger brother’ to the Sanga princes of Ukwama, is indeed an exile who belongs to the Nyakyusa scene. It is Lwembe the Sanga court explicitly used, adapting a kinship fiction and a descent-based mode of religious thought so as to magnify the court’s importance and redefine its mystical potential. It can’t surprise us that both 120

Kinga and Nyakyusa rulers are supposed to have witchcraft in their bellies. Witches and shades share with the invisible agents of other animistic religions a reluctance to settle for what they have, an implacable sense of grievance. The grass, as it must seem to the uninitiated looking across to the world of the shades, is only browner on the other side. The shades are, in today’s parlance, ‘under stress’, and it makes them ‘demanding’. In their own terms, if not in ours, Kinga and Nyakyusa commoners recognize the syndrome also in their rulers. The argument that witchcraft was entrepreneurial magic in the bush culture and elevated to religious standing by the Sanga courts is plausible, perhaps, but not compelling. Witches are no more or less the objects of religious belief and ritual control than ancestors are, and neither has ever actually been seen. It is here that the witch/sorcerer conflation gives trouble in Malatan ethnography. The ‘witch’ you can pick out and bring before a court or doctor is quite generally described as an ordinary human being within or beside whom a sort of shadow or spirit-witch is alleged to hide. Any crimes and any witch-flights are done while the visible person sleeps, monitoring (it is supposed) in the same way one participates in the drama of dream. For Kinga, the usual evidence that Fulani ‘is’ or at least harbours a witch (and knows it) is his uttering of a ‘veiled threat’. This is not very different to the way most of us discover a friend is harbouring a grudge. Yet the witch accused and condemned in the traditional court procedure—assuming, as we probably must, that some were—becomes something more. Denial becomes defiance. When the highest drama of the court is reserved for the execution of a witch, know that he is possessed of a power to threaten the Prince himself. For a prince or other strong ruler in public drama to point an accusing finger is a sort of magisterial witch-threat: the recipient should not be surprised to fall sick or have an accident soon after. This offers a glimpse of the very engine of a religion of blame. It is tangible in just the episodic way high drama is. And there is high irony: a prevalence of witches has no appreciable effect on the level of amity prevailing either in a Kinga or a Nyakyusa community. Readers of Evans-Pritchard’s wonderful work on Azande practices will doubtless recall pondering how so much hereditary witchcraft could be proven without blowing the community apart. Understanding this in the Malatan context is a key to understanding the way order is kept there, and its relation to the culture of intimacy. Antipolitan values, placing independence before loyalty, are congenial to witch beliefs. Any worldview which gives a prominent 121

place to witches is biased, after all, against monopolies in the mystical trade. The Sanga courts adjusted to this feature of bush culture with the logic of a redistributional chiefdom. The court, in this scheme, trades on its magnetism as much as its discipline. But the give-away programme is balanced by another kind of theatre in the ordeal and the public execution of convicted witches. Nyakyusa, in their age-village programme, found a way to use discipline in the control of antipolitan tendencies, and retained kinship solidarity to the same end, though the two principles of solidarity were impossible to merge. As for witchcraft, the Nyakyusa chiefs took the lion’s share to themselves. The prototypical antipolitan is a sort of cowboy figure apt to pull up stakes when things don’t go his way. In the Nyakyusa peer village, banishment is always conceived as ridding witches, and we may suppose the cowboy scenario was more likely to fit the actual cases of (self-)exile than a heavy procedure like the ostracism of ancient Greece. Decamping in an antipolitan world will be more than one part voluntary. Kinga lads in the hills had an easier way with peer pressure. They were not assigned to a given village (as Nyakyusa were) but chose their own friends and built their own places with them. As a boy gained confidence and formed a wider amity network, it was likely he could untie himself from the family goats and take off to live with new friends. Amity all the while was being reinforced by intimacies. What this meant for many a young man was a gradual accommodation, a sort of ‘cowboy syndrome’, which might long prevent settling down to a life of routine. But easy friendships too easily fail, and older men might fall to suspecting an erstwhile friend of mischief when things began to go wrong. On this count the Nyakyusa may have it better. Their dislocations are fewer and less casual. The pressure to conform within a community of peers is comparatively rigid. First reports from the German missionaries who came among them a century ago were of a veritable paradise. This was a happy people. If this result had been achieved by a repressive politics, the politicians had done their jobs exceeding well. Still, three generations later Monica Wilson wrote that this was a society organized for the benefit of “men and elders,” cutting short youth for a girl before it could rightly begin, restricting the freedom of women all through their best years, and denying the perks of adulthood to young men by refusing them wives. The balance had steadily changed under the colonial pax. By contrast, the extended years of youthful independence for Kinga men continued as a high period of private pleasure. Morale was high, and fathers easily kept at a distance. Youth anywhere has to be filled with adventure, but youth may be short or long. 122

Compared to the condition of Nyakyusa youth, the kind of discipline—authoritative rather than, in Durkheim’s sense, mechanically or peer-enforced—which the Sanga had brought in had led Kinga youth, of both genders, another way. What Durkheim meant by ‘mechanical’ sanctioning is best seen as a feature of the everyday role network itself. The diffuse sanctioning of behaviour which constitutes such networks can produce a dense moral atmosphere which distributes self-realization unequally to men or women, young or old. Majesty had been a key to political power in uKinga, as it had not been in uNyakyusa. We are left with the interesting puzzle, how it may come about in human affairs that strong politics will coexist with high morale, grounded in felt autonomy, among ordinary people. The clue I see in this comparison is the sense that ‘strength’ in the case of a Nyakyusa prince/chief is not laid on but has to be achieved at the cost of stressful, aggressive action in a chronically turbulent social environment. Arena politics rules. The more monumental political theatre of the Sanga court as it existed before the pax was structurally firmer and allowed an easier transition to top-down colonial administration. The zigzag dialectic of change can be sharp or almost smooth. Like Nyakyusa, Kinga conceive witchcraft powers as egoenhancing. Men expect a ruler to be jealous of a witch in a way he can’t be of his priest. This is not because a priest is regarded as ‘on staff’ or because he is ‘one of the good wizards’. I think the reason is simply that priests at the Sanga courts were allowed to be professionals as distinct from being entrepreneurs. Priests have been able to uphold a high priority on esoteric knowledge and the arts of ritual theatre, in preference to competition for apparent mystical power. In a model-comparison with uNyakyusa, this difference is important to see. The priests in the more turbulent Valley communities are less effective in their management exercises. It is true when you look soberly at the evidence for a Kinga/ Nyakyusa contrast you find what is more nearly a continuum: between majesty and charisma, authority and belligerence, theatre and arena. Among Nyakyusa speakers we have been looking at Selya (not the more nearly theocratic uNgonde) while among Kinga we have been focused on Ukwama’s high court in the Central realm, not the long-drawn-out, always warlike insurgence of Kyelelo the Cruel in the West. Where there is chiefly belligerence there is struggle over followings. Where there is a single political and ritual focus there is the kind of stability which sublimates the spirit of violence. I think it likely that both political cultures knew oscillation on a structural continuum defined in this way. For Nyakyusa we have seen this built into the constitution of each autonomous major segment, each 123

chiefly realm, with its periodic ‘world renewal’, the Coming Out. For Kinga too much history has languished unrecovered, particularly for Eastern and Northern realms; and there is too much myth in the mix of history we have for Ukwama, to know as well where the balance lay. In this context the Kyelelo tale suggests a turbulent period began in the mid-1800s with the assassination of a weak or weakly supported ruling Prince, followed by succession wars which led in the direction of legitimating new boundaries within a common but expanded realm. The expansion was marked by the alliance of Vululile (cheated apparent heir to the usurped throne) with a Mahanzi local ruler. Prior to this military alliance against Kyelelo, the Mahanzi Mwakalukwa had apparently been linked with his neighbour Kinga communities by taking a key part in the frequently renewed ceremonial pageantry in appeasement of the territorial divinity Lwembe. Since it was just this divinity who was thought to bring plague, drought, and pestilence to valley and mountain slope alike, this procession bearing gift hoes must be seen as a grudging reenactment of a mythic drama with a territory-wide scope. From the beginning, Lwembe had demanded in recompense for his exile from the Sanga high court that a treasured set of consecrated hoes be returned to him. Each time his anger was seen to rise again, the sacred ransom was to be repaid. Tribute to a once-and-only king? That is certainly not what the Kinga would like the Nyakyusa to think of it. The meanings in the symbolism are many, and are discussed elsewhere, but just here the significance I would point to is the use of a religion of blame to iterate the terms of an old (of course unwritten) treaty of peaceful trade between distant partners. Kinga trade in iron goods long antedates the protostate development; and the Kinga procession to Lubaga presumably follows the web of footpaths used by traders before men carried spears. A prince in Nyakyusaland claims—must claim—the very powers he condemns in a commoner. It is that claim and those powers which create the princely calling there. The office takes its legitimacy from the qualities successfully cultivated by the incumbent, in whom these qualities are openly admired. A Kinga prince like Kyelelo is also terrible, and for it much admired. In both societies the priests are agreed that their propitiatory rites must be joined in by local rulers, but what seems required is hardly more than their presence at the sharing of the sacrificial meat. What I understand from this is that the priests in each case apprehend the power of secular command in the frame of natural history: there is a delicate balance between order and calamity in (as we might see it) nature and social history, and the priest owns the metaphysical art by which the cause of an obvious cosmic imbalance may be recognized and normality restored. 124

I resist any temptation to see this priestly intervention as oriented to the laying down of arms as between principalities. These priests are not doves of peace. At least in uKinga their role in the regulation of warfare entails the sublimation not elimination of violent undertaking. Their counsel to the prince is ever to mind the murmuring of his people lest he lose his braves to a rival. Yet for all that, we should understand that they always work by indirection. Kinga priests like the Nyakyusa are in their own way menders of the peace. What else could be the point of their insistence on doing the intricate work of a ‘ritualist’ and their obvious pride in it? But the situation of the priest in the two cultural settings is distinct. Kinga priests counsel, protect, and even discipline their princes. The exception which has to be made for the case of the preContact Kyelelos, the rebel father and son of the Western realm, is such as to prove the point. But Nyakyusa priests, as we know from the Wilsons’ careful witness, though they might badger to death the poor unwilling colleague they had crowned and called Lwembe the living god, king and founder of all princely lines, they might not get the time of day from a truculent prince. The counseling, cautioning, and doctoring agency in a Nyakyusa realm was not a concerted agency within a royal court but articulated around the sacred groves now grown about the graves of past hero-princes—so scattered about among the several components of a loose temporal community. The metaphysical agency was lodged in a band of ‘aristocratic’ priests forever at odds with another band of ‘commoner’ priests; and, quite apart from this field of influence, there was the staying power of a tight array of commoner village headmen, ‘great commoners’ whose cooperation in military matters was all-important to their prince and his luminance. I can only think the two-way stretch of the Kinga priest must have been easier to cope with than the four-way stretch of the Nyakyusa. Monica Wilson had her ‘inside informant’ Kasitile among the priests articulated around the Lubaga shrine of her ‘divine king’ Lwembe. Kasitile complained constantly about collegial infighting. It may have been easier in European times for a secular ruler to ignore the little cluster of priests around Lwembe at the Lubaga shrine, even seeming thus to thumb their noses at this high religion. Certainly the predictable decline in popular belief in the efficacy of this divinity in causing, then curing environmental crises was already evident in the 1930s. But it was before Contact that Prince Mwaijonga had made war on the then Lwembe, killing some of his men and rustling off some cattle. So we have a genuine divine king, complete with the requisite ritual, who can be viewed by a warlord in a secular frame and look no grander than another petty prince 125

competing with his truculent neighbours for the security of life and property, and for honour. In the better-contained and stabler realms of the Kinga—not including, for certain, Kyelelo’s before the pax—the prince was more symbol than leader. But in all cases, as with every Nyakyusa prince, the mystical persona of a sanctified leader could be deeply hurt by an admission of weakness or fear. A key to the difference is to be found in the part ancestors played in the two protostate systems. For Kinga the bush religion prevailed. Ancestors featured in the private moral career of an individual; they were always addressed by men through a nearest of kin and pictured as an unnumbered community of shades dwelling as if within the acephalous frame of the old bush culture. For Nyakyusa the shades, as they may be pictured by a consultant diviner, are deeply involved in monitoring the moral careers of men and women, punishing transgression of meticulous rules for social behaviour within the home, with affines, and the larger agnatic networks of men and women. It is in a real sense the shades who keep the threads of kinship taut in a society otherwise ruled by a brittle politics of amity. The political religion of Nyakyusa articulates around princely graves, where sacred groves are allowed to grow. In special cases, presumably in keeping with a reigning successor prince’s own stature, these sacred groves might gather enough of a mystique to merit the full paraphernalia (pythons, rainstones, and much else) of a major grove. The Lwembe grove at Lubaga had perhaps been one such. But in historic times it served a wider pedestrian community than any political realm, and fully deserves to be conceived as a territorial cult centre. With documented early observation by missionaries and later ethnographic study concentrated in the general area of Selya—in effect, at the more temperate altitudes—we simply don’t know how many somewhat comparable groves may have existed in 1890. We shall certainly never know how many, once planted, had by that date already ceased to exist.

Uses of a bush religion: embodied danger In ‘majesty’ we have the off loading—to the throne itself and all the other props and effects of political theatre—of all the charisma of founding heroes, whether taken from life or legend. Majesty also comes in an implicit form: Sangilino of Lupila (Kinga Eastern Realm) needed no regalia, but his was a quality rarely found and not transferable. Malatan charisma lived on well enough in its institutional form into the 1930s, embodied in the surviving old-time chiefs, that we can know its qualities from the Wilsons’ fieldwork. It was to disappear, of course, with the new generation of educated public 126

servants. The good political entrepreneur of tradition was big, decisive, and full of himself. The public was his claque, everything he would do must be made a quantum larger than life. The driving force for all this is explained as ubulosi (K: uvuhavi) witchcraft in one of its several manifestations. The poetics of embodied danger lends drama to the ‘constitutional’ stand-off between ‘evil witch’ and ‘charismatic chief’. Monica Wilson wanted to see it as a mystical ‘balance of power’: Wilson: The constitution was a balance between chiefs and commoners…Chiefs maintain their position by feasting their men on beef, and this…prevents witches from attacking; sated with meat their pythons ‘lie quiet’. ‡‡

Where the Nyakyusa prince was in his person a battleground between the forces of vitality and corruption, except for the living ‘divine king’ Lwembe the Nyakyusa rulers were not set apart on that account as sacred personae dangerous to and endangered by the hustle and bustle of ordinary life. Their freedom of movement and self-expression, their readiness for personal confrontation and their boldness of public demeanor were not impaired or restricted by taboo. It is especially interesting that the Kinga rebel Kyelelo the Cruel was, by all accounts, just that kind of ruler to his people. The reason is certainly not that he was un-Kinga but that what came to the surface of Kinga culture in one circumstance need not be replicated in another. Kyelelo’s more orthodox counterpart and enemy in the Western realm was Vululile. Informants at his capital gave me a Sanga version of their royal court and the special obsequies their ruler must expect. Details were not as impressive as those I had got at and for Ukwama, but the scenario was the same. The Kyelelos, on the other hand, were political entrepreneurs. Sacerdotal kingship was not their thing. One must imagine both the rebel father and his successor son knew and approved of Mwaijonga’s tilt with the puppet Lwembe. With pythons in your belly you can tell a priest to watch his tongue. When the elder Kyelelo finally wearied of battle he set up his son in a capital village of his own and handed on the hero name. This is not the way of the legitimate Sanga prince whose power comes out of the medicine horn of a priest. A High Prince is rather ‘divine’ than ‘charismatic’, having in common with his ‘younger brother’ Lwembe that he will step down only in death. But Kyelelo’s retirement in his ‘golden years’ would not seem incongruous to a Nyakyusa observer, as it is the young bachelor men (with their future chiefs among them) whom he will expect to do the fighting while the elders settle down to grow rich in cattle and wives. Ordinary men in both cultures 127

marry late, at least a decade after they are fully fledged as warriors. The Nyakyusa prince/chief is no exception to the rule of the place, as he only takes his rank with the Coming Out of his agemates and only gives it over to the next generation when the younger men have proved themselves ready to ‘come out’. This means a prince stands to reign some 25 to 30 years as a hands-on ruler. His preparation will have been an extended decade of leadership in defending and augmenting the family herds. At the climax of a successful reign there will be a surfeit of wealth and wives to keep him busy. I think I am right in doubting Monica Wilson on the brevity of his future after the Coming Out of a new political generation. To me, the decisive evidence is the known career of his village headmen—his age peers and colleagues in government throughout his reign—at the same point. All of these ‘retiring’ headmen take up now the final stage of their careers, as ‘commoner priests’. If the chief has to settle down at last as a diminished ruler, I don’t think his realm is diminished, only the active, hands-on role he has hitherto had to play in governing it. Do his pythons desert him? We hear of no such belief. As to reclusion, I take that to have been a matter of degree and taken gradually. For convenience at this point, let us picture a chief who has just turned over the ‘entrepreneurial’ side of his calling to his chosen sons, and let us suppose his transformation now can be characterized as moving from ‘chief’ to ‘prince’. Just how ‘sacerdotal’ will such a prince aim to be? The Lwembe figure has been described by Monica Wilson as a divine king with hardly any secular power, living in partial isolation. The evidence for ‘divinity’ is that living parts of his body are taken by the local priests when he is near death; then he is smothered; and medicine made from his body relics is used to pass on his identity to a successor priest. But similar procedures are evidently employed at the passing of a great chief; and since he could have been no recluse there could be no serious case made for ‘divine kingship’ in any full sense of the term. The obsequies of the high prince at Ukwama show that the dangers associated with his office were carnal and inherent, and must be prevented from passing to the shades’ world. Whether the prevailing sentiment was more ‘We must not lose these powers!’ than ‘We must not let them loose!’ isn’t decided by the evidence we have. But forcing the corpse to swell up and burst (as was done at Ukwama) is directly contrary to rites of divine kingship everywhere they are practised. And in life, the Sanga unkuludeva was forbidden entry to any sacred grove: never abstaining from relations with his wives he would defile anything sacred. Neither of these two figures can be reckoned a priestly head of state: a Sanga prince was never priestly, nor a reincarnated Lwembe princely. But we are free to 128

suppose there was logic in the way the Prince’s body was handled by his priests; and that logic would almost have to begin with the need to prevent a meeting of Lwembe with his perpetual bane, Mwemutsi, who had banished him in heroic times. The model of ruler as exemplary masculine leader is shared with many Eastern Bantu peoples. The Malatan pattern includes popular concession of the ruler’s hereditary right to command in matters (as we would put it) “of state”. The variant practice at the Sanga courts was to relegate the practical war-leader role to a commoner unyivaha. Monica Wilson would recognize this person as a transform of her ‘Great Commoner’, here attached directly to the ruler’s court. This is the man who would, in effect, play ‘Chief’ as distinct from the ‘Prince’. Like the Nyakyusa ‘headman’ or ‘great commoner’, he could stand close to the Prince and even resist without threatening him, being ineligible to succeed him in office. After installing a new Prince, Kinga priests would soon busy themselves seeing him withdrawn from the public drama of the law court and other secular matters. He must accustom himself to a theatrically sacerdotal role, spending his personal energies productively within a quickly growing harem, and turning his new mystical powers under priestly direction to ritual use. The signal difference in Kinga political religion, which is also found in Ngonde, is that the chiefly presence of a prince was not trumpeted by those mystical powers ‘in the belly’ which so reinforced the autonomous power of a Nyakyusa prince/chief. We are left then to see the case of the Nyakyusa as the special one among the Malatan three. As to reasons for this, I look to the inherent instability of a political system which starts anew with each generation. The number of princely realms and their locations fluctuated over time. Alternative shrines abounded, several of which operated as territorial shrines affecting the common weal—in their own territories quite as important as Lubaga for Selya. Given that no pyramidal political structure had evolved there, it is hardly surprising that Lwembe was never as central to Nyakyusa thinking as to Kinga. Still, whatever chapters in regional history have been lost or stripped of the necessary detail, we can regard the ‘divine king’ establishment at Lubaga in the light of political theatre as a major constituent force for continuity in a country thriving on rivalries. As such, that territorial shrine was probably stretched to its limits with the turmoils of the later nineteenth century: Ngoni invasions, slavers on the lake, and attempted conquests by the growing military states northward. The importance of such troubles in spurring hierarchization of the Ngonde state south of the Songwe river are well documented. What is important for us to see is that the politics of these protostates 129

depended on the sacralization of authority, and the vessel for this could be either more charisma (python-deity in the belly) or a more developed political theatre (ranging from ‘majesty’ ritual to heavy heroics in well-staged warfare) or any combination. Reviewing the variety within the span of cases on which I could get details in my fieldwork period, in the 1960s the main obstacle to generalizing is the destruction of the Eastern realm by (largely local) forces mobilized by the Germans, so that only folklore and a fragmentary oral history remains to represent that emerging realm; and some cooking of the data recorded in Colonial records in the West. During early German times a now-lawyerly ‘Kyelelo II’, rising from rebel to officially ensconced Prince of the West, was able to mask the extent of destabilization in that realm. This victory was achieved by diplomacy (only partly through missionaries) where war had consistently failed—a twist on the usual take. In the West it is apparent that Vululile, the orthodox claimant to the power usurped by Kyelelo, having been forced to depend on Mahanzi support in a new, half-foreign capital, could only turn to charisma for lack of majesty. In the East it is clear enough that the prince on-the-make Uliluvilo had been waging an expansive war on his southeastern reaches, but he had not (or had not securely) established Igumbilo as the main sacerdotal court for the East before the Maji Maji devastation. Uliluvilo had been ‘sent out’ by his father Sihudika from Igolwa. Was he, like Kyelelo II in the West, a favoured son and untsagila sent out by his father to continue the expansion wars of this realm? Or in what sense and in what degree could he have been the rival of his ‘father’? And how many biological generations do ‘Uliluvilo and his father Sihudika’ actually represent? I remain unsure but favour correcting for a likely foreshortening in that half-buried chapter of Kinga history. As for records, the Germans having laid waste the East, their cartographic information is slight on political information and hard to interpret. Turning to the Northern realms, since we do know that Ndwanga of Ilevelo well into German and even British times would be challenging the sway of Uhugilo as the seat of that realm, there is reason to think the sole still-point in the Sanga system was the court of Mwemutsi at Ukwama—the preeminently sacerdotal seat, acknowledged Alpha of the four, yet one whose precedence was continually challenged on secular ground by the early Kyelelos. The working conclusion I have come to is that the Sanga system long had depended for its anchorage on the Lwembe cult; that relations among the four realms (or however many of them a Sanga leadership would have been ready in a given year to recognize) were always informed with rivalry; and that the movement of tribute, though regularly toward Ukwama, never was so regular or so weighty 130

a prestation as to confer a secure sense of permanence on the pyramidal arrangement. A ruler’s real guarantee of security in tenure lay in the strength and loyalty of his avapapwa, the inner circle of Sanga ‘younger brother’ bachelor warriors living at his court, and the effectiveness of his management corps among the avanyivaha. The rest was up to priestly theatricals, and I suppose we should be looking for the ‘spin doctors’ minding that side. Right through Kyelelo’s rebellion the world to the north and east of the Malatan region was teeming with movements and countermoves of quasi-Napoleonic style, in response to the proverbial winds of change blowing from South Africa and the coastal centres of trade in gold, ‘white gold’, and black slaves. If the response of the Malatan peoples was not to look for charismatic tyrants able to build big coalitions, the reason was not strictly geographic. Kinga, Nyakyusa, and Ngonde made less political use of kinship rhetoric (with its particularistic daimons, the ancestral shades) than their neighbours. Their deepest principle of cohesion was amity, and the political face of amity is the antipolitan ethic. Malatan man in this enjoyed a typological isolation within the larger Southern Highlands region. It served as a sort of vaccine against charismatic infection. The claims of friendship are easily broken by acts which a tie of kinship will handily survive. The ruler-subject tie in an antipolitan society lacks the sanction of the maximal-claim or ‘kinship’ tie: the subject will not necessarily lay down his own life for that of his leader. Even the great, charismatic Kyelelo, finding himself wounded and alone behind enemy lines, could not expect a rescue attempt by his men. As with friendship, an ordinary Malatan man wants his claims on a ruler affirmed, never disaffirmed, on each encounter. The underlying rubric is amity (patronage) not kinship (paternalism). The disestablishment of family domesticity, and of the solidary thinking about kinship which it fosters, was built into the early life-cycle, for Nyakyusa in their age-village system, for Kinga in the segregated lives of children in their own-gender homes. One obvious consequence for religion is the eclipse of the ancestral shades, but a deeper implication is bred-in unease with authority, which can give a brittle character to political loyalty. PreContact Ngonde decentralization (except in ritual hierarchy and the control of the central court over external trade) puzzled Godfrey Wilson. Perhaps in the context of Malatan regional culture it should not have. The antipolitan persuasion necessarily produces a dialectic as between ideals of loyalty and habits of freedom, myths of belonging and expectations of self-reliance. As a source of creativity and entrepreneurial energy this dialectical tension should not be confused with disorder. Its mark is to be seen in the elaborately 131

balanced political structures of each of these three societies. But there is also a darker side in the religious expression of mortal fear and discontent. High office embodies in Ngonde, Nyakyusa, or Kinga cultures a popular ambivalence toward power. Ngonde nobles used to kill all but two sons of the Kyungu in their infancy. At every level of rule the key office was guarded by a council of nobles or Great Commoners whose combined powers of witchcraft could check-mate their ruler’s. The same show of countervailing power, here embodied in python-familiars, held the Nyakyusa prince to responsible rule. As for the Kinga prince who must choose seclusion, the motive was unabashed fear for the resentment of others toward his power. Everything he ate or drank must be tasted beforehand for poison. The schoolboys who scoffed at Mwemutsi Suruali in the sixties for spending his truly enviable salary as Paramount Chief on bottled beer could hardly have understood his reasons. Before the decay of his royal court, everything the ruling prince ate or drank had been tasted beforehand for poison. His daily existence had been sacralized— though he had not the same ground of fear as the early, reclusive Kyungus in the Ngonde court. It was said they must be throttled if they lost only a drop of blood by accident or fell into light fever. The institutionalized paranoia of a reclusive king is not found among Nyakyusa, as the ‘divine king’ Lwembe was not expected to play defender of the realm, but the burden of danger in that office was still enough to keep the kingship at Lubaga empty from early in this century. The reason for the dangers in taking on the mantle of majesty is not simply in the classic (Arthurian?) burden of kingship in a stillegalitarian society, or in some contradiction between the practical demands of good order and the intuition of communitas. We have to look further and specifically to the religious culture and its images of mankind. The politicization of the Kinga priesthood distinguishes that constitution from examples wherein the association of religion and politics seems only implicit, or embedded in kinship. Such are lineage and clan based societies, including those with heavy affinal emphases in the political arena. Kinga priests enjoyed a remarkable ambience, crossing borders with impunity where a prince dare not go. The priest was kingmaker and as such able to operate well beyond his own domain—even in another realm, as the nefarious history of one important line, the Sanga Nyengwa, serves to witness. ‡ A Kinga priest did not carry a specifically religious burden in the full sense my rendering of his title in English may suggest. He was a ritual specialist attached to the court as one of an élite group with privileged access to its sacred grove. He was not a healer or spirit 132

medium. As unteketsi he made offerings and executed sanctifying procedures on behalf of court and country. But as a courtier unyivaha he might be found leading a roving party of avasangutsi revenue collectors out after imongo from a bush hamlet. This was of course low, not high political theatre, and the scene was blurred in the social memory of the nineteen sixties. Whereas the Mahanzi rainmaker before Mwangawa was a private person with a religious calling, and he may be taken as stand-in for the Kinga priest before Sanga times, later court culture created an independent, mystically sanctioned authority largely immune to interference by secular rulers and even in position at times to patronize them. The children of the courtly priest slept with those of the Prince. On his breast the priest wore the regional symbol of authority, the imatsi conus-shell emblem, a perfect spiral disk traded inland from the coast through many hands, and elsewhere restricted to chiefs. Religion and politics were largely emancipated from ‘tribal’ institutions, having come to be, in an emergent polity of its own sort, instruments of state. ‡‡ So far as religion can be thought to reflect and sanction features of the social structure, an important dimension of this constitutional transformation has to be the shift away from ancestor cults with their particularism toward an inter-ethnic ceremonialism focused in the annual oracular drama at Lubaga. As the magically dangerous younger brother of the High Prince, Lwembe carried out of Kinga country the image of the ‘natural leader’, strong both in combat and magical-mystical powers. The type of the Sanga ruler left behind is, where the man himself is strong, of secular bent, and where otherwise, so withdrawn as to be no personal hero at all. Lwembe embodies the very forces of nature which the old rainmakers were charged to monitor and control with their wonderful rainstones. By placing a collateral ancestor of the Sanga ruling line in a position of responsibility for all the sweeping forms of devastation nature can visit on all the realms of the Sanga and central Nyakyusa it seems our ever-conniving priests have wanted to sweep away localism in both its religious and political forms. How was the great transformation engineered? It tells us little to say the Sanga rulerships ‘evolved’ along paths much traveled on the continent. The part of the common people in the change may have been mainly passive, but that of the sacred and secular rulers was not, and both would have had to keep in effective contact with individuals they were recruiting to new roles. Since it seems too much to suppose that the early Sanga rulers, however clever, would have realized a religious ‘revolution’ was wanted to secure their power, it is reasonable to suppose that the driving energy for the whole transformation would have come from the competition for 133

dominance between the two principal agencies. Each would have sought to extend the scope of his own kind and sphere of influence. Confrontation on certain issues would have forced at least the kind of innovation which comes of sidestepping—the ‘workaround’. The impression I had in 1960, when the whole native cosmology was in limbo, was that no one conceived of it as a distinctively Kinga way of seeing the world. It was not a human achievement to have pride in or claim the right to preserve as history—any special feature I might mention was taken as detail. The whole fabric was classed as general knowledge about which only some few details might be debatable. I came to see that in all probability the same would have been said to a traveler who had come to them in 1860. In short, as there can be no historical consciousness where there are no alternatives to ponder, all of this came down to the present in a package, fait accompli, from very early times, as explained by the mythical arrival of the first Sanga, his reconstruction of bush culture, his confrontation with a younger brother, and Lwembe’s escape to Lubaga. Any talk about the gradual politicization of the priesthood and its consolidation as a governing institution is thus my talk, not theirs. Was it witch-ridding which would have given the priests their first specifically political role in an expanding court culture? There was no one to confirm or deny. If I now suppose that the peculiar powers of the priesthood had to have been won at the expense, in final analysis, of the princes? I’ll have to make that my own text, as it isn’t theirs. The once-and-only Queen Kipole succeeded her father Mwemutsi Nyanzululu in about 1897 after allegedly poisoning him— she was the trustee who brought him his food. When her younger brother Dembademba came of age and took over the throne as Mwemutsi Mlambakyuma he was plagued by deaths among his wives and children, all ascribed to the witchcraft of yet another sibling, Mwakanema. Then with priestly efficiency all of this witchcraft was turned back on Mwakanema’s own wives and his issue by protective medicines the priests were able to obtain from afar, through their wide-spreading network of professional associates. No Kinga would argue, from this text, that the priesthood was not an indispensable ally of the crown. But the text itself (lest we forget) is from start to finish the telling of the priests about their own powers. Parallel texts one might take from the priests of the Kyungu, divine king of Ngonde, are not quite the same. His seclusion is rationalized not in terms of plots but another sort of paranoia. His bane is his extreme fragility, an uncanny vulnerability to accident. It is a personal fragility which apparently comes with the office. Pondering it we may want to turn with more understanding to Kyelelo the Cruel.

134

Is he not perhaps some sort of Achilles figure, a hero unscathable and unafraid to boast of the fire, the python power, in his belly? Or consider once more the living Lwembe in Nyakyusaland. Setting aside the odd-yearly Kinga procession, he seems to have enjoyed far less in the way of chiefly income than either the Kyungu or the prince of a Kinga realm, yet like them he has surrendered any clout he may once have enjoyed. He is no mediator between ordinary men (whom he seems not to meet) and gods but a recluse who deals only with priests. Though his great power is anger, it is not the kind of anger a secular prince, claiming descent from a hero, is allowed to have in his belly. This anger is initially not seen but inferred by the priests from bad weather and deaths of children, and attributed by them to the restless spirit who speaks through the lonely figure they will confront at Lubaga. The first Lwembe, as some say, came down from the Kinga mountains to find a scattering of poor folk without fire or any other civilized advantage. He brought them new crops, cattle, civilization. Today all the Nyakyusa chiefs who claim descent from him speak of themselves as having in this way brought civilization to a savage land. But the living Lwembe at Lubaga betrays no claim to a hero’s title. If there is power in him it is only the power to speak with the voice, and stamp about with the feet, of a god. Nyakyusa seek through him to appease not a copy but a true god whose anger has festered from neglect (mirrored in the figure who inhabits his shrine) and will burst upon the world bringing deaths or drought. This ‘divine king’ is perhaps a disappointment to the anthropological romantic but a compelling reminder of the inescapable theatricality of any effort to represent a religious idea in dramatic form. In Godfrey Wilson’s terms, the succession of living Lwembes “never achieved any secular paramountcy at all...because they were poor.” They remained “religious paramounts. Chiefs from all over the district sent gifts to them to ask for health and rain”. In fact they were allowed neither majesty nor charisma by the restrictive taboos under which they must live. As a mortal man Wilson’s Lwembe lived, to whatever extent he could not personally rise above it, in fear of illness and unnatural death. The last Lwembe under the traditional dispensation was gone by the time of the Wilsons’ fieldwork in the 1930s. There were candidates for the office then but none who chose to be chosen for the dreadful honour. Peter Weber reports a new living Lwembe in the 1990s, operating in quite a different style but keeping the brand-name alive. Reviewing the three Malatan cultures at the coming of the pax we find three kinds of political system. Ngonde is a self-contained, 135

monolithic theocracy. The four realms of the Kinga comprise a radically segmentary protostate whose major segments are weakly linked by the perpetual kinship of the four rulers but rather more securely by priestly cooperation. Nyakyusa chiefdoms are autonomous, locked in a turbulent competition for political-economic space, for wealth in cattle and women, and for manpower; they are linked by multiplex ties of kinship which intermittently effect the realignment of local alliances; and they maintain the semblance of constitutional stability in the irregular propitiation of hero-gods most prominently at the Lubaga shrine of Lwembe. In the three Malatan constitutions one constant factor stands out, the mandate and professional character of the priesthood. Is there nothing but wind in the belly of the Mwemutsi, High Prince and King-in-Seclusion, foremost among the four royal rulers of all Kinga? Boys in the sixties used to joke about the belly of their Paramount. Still after sixty years of acculturation to colonial conditions (and in the case of Mwemutsi Suluali, who was to be last of the long line, a good but wholly wasted European-style school education) the Prince did not know how to run his own court, hear cases, talk policy, or mediate effectively with the Boma government. Big and overweight, he was hardly ever seen sober and rarely could summon up the full dignity of his person. He was surrounded by unruly kinsmen, always there to carp and mock. He told me things were falling apart because Lwembe had not been appeased, almost in all living memory, by his people. Yet this most powerful scion was powerless to move his people. The old priests, who knew the requirements of ritual, were dying away. For all their political acumen, they never got a purchase on the system the pax brought in. Yet as I gradually came to know, the Prince never had in man’s memory played shaker and mover at Ukwama. No Mwemutsi was the one to intercede for a troubled people with his angry ‘younger brother’ Lwembe; he never took the mover’s part in calendrical garden rituals. It was never the prince and his men, always the priest and his avanyivaha who must do the work of smoking out a witch and subjecting his soul to trial. In all respects but one the religious burden of the Sanga prince was nominal, easy as firmly pounding a rubber stamp. The exception, of course, was sexual fertility. The royal belly was and must ever be heavy with a burden of heterosexual lust his countrymen found bordering on the unnatural. It is probably sound to attribute the special character of the Sanga kingship to the special fact of the royal isivaga, the teeming bachelor residence maintained at a royal court. Nyakyusa-Ngonde culture was not built around a court/bush dichotomy and didn’t in the same degree practise the temporary elevation of younger men 136

to a privileged standing premised, in effect, on sexual inversion. For Nyakyusa youth homophilia is a temporary, imperfect alternative to sexual abstention during a frustratingly prolonged bachelorhood. The formula for the successful life is plainly visible in the age-village of his elders: the heterosexual life brings wealth in cattle and wives, and the authority of senior standing. Nyakyusa say homophilia is a resort of youth, thankfully forgotten with maturity. In Sanga court culture the elevation of heterosexuality—in the big but reclusive persona of the Prince/King—to a transcendent plane can be seen to have provided a symbolic bulwark for marriage in a masculine community whose commitment to that institution might otherwise flounder. But to show that religious phenomena are psychologically appropriate and functionally justifiable is not to explain their special qualities at the level of specificity on which a believer experiences them. The Sanga prince was a fearsome and fearful figure even if veiled. By the magnificent style of his court he had to be known as a great power, yet he was privately slave to his sexual duties. The person of wealth and majesty attended by a lavish court was at the same time a poor man cut off from his friends by the failure of trust. Like the living Lwembe at Lubaga, the living Mwemutsi was a kept symbol, a puppet in the robust image of the ancient Hero said to have begotten the line. Was there genuine fear that the High Prince might fail in his fertility and all Kinga in consequence fail in theirs? If so, it would certainly be fear of the sort we call ‘religious’—congruent with a state of general confidence in Sanga institutions, and reflecting the competence of Kinga priests to manage such a situation handily, should it arise.

Imaginary theatre Witchcraft may be everywhere the dark and hidden face of an (elusive) ideal amity, but in societies whose first structural principle is friendship, this is no mere truism. Like friendship, harm in Malatan belief springs from human will not personal or cosmic destiny. Intent and intender are not so manifestly one that it always would be logically necessary, to kill the intent, to kill its author. That is the meaning of the fact that, while witching powers are certainly thought to be controlled by a living person, they are as a rule for Kinga lodged in a separate, objectified vehicle which a skilled doctor might find and destroy. The Nyakyusa construct is a little different, but offers the same escape from ‘witch burning’. In Konde thinking, a 137

witch need only be chased beyond the narrow horizon of his chiefdom, to lose all occult power. When after some years he is allowed to return it is expected he will offer an act of atonement. This is not the witch of folklore, unfixed in time or place, but the witch in local experience. The power of wish-gliding is something else—uncanny and very likely to have been negotiated with a python. A witch supposed to have won such powers would be beyond pursuing mere personal grievances. When people see a fireball gliding over treetops they say it is one witch attacking another. Talk about witches makes them a society of their own, paralleling the society of shades but inverting it in one respect: their greed is insatiable. The fairy tale witch is unstoppable by any means short of death. There are covens, we learn, wallowing in necrophagia. These witches can’t wait for a normal death but press each member to supply them from among his/her family and friends. This witch of the fireside tale is of a different kind to one who might be named in an actual accusation. The root difference is that an accusation will be based on the intuition of motive to harm, where the phantasy witch is a monster quite beyond ordinary human motivation. The objective correlative is certainly the African vulture. Vampires pale by comparison. But it seems daytime imagination most often takes off from the ground not mid-air. ‡‡ Still, mid-air thinking can come down to the ground. Here is an informant who lived much of his adult life as a chauffeur in the coastal regions of Tanganyika. He reflects general opinion about the prevalence of witches, always worse than ever now: Lukasi: Witchcraft has grown beyond all bounds in the years since, in Old Kingaland, the witch was killed off before he could spread his evil far and wide. Now you approach a witch very slyly, making friends with him over the beerpot. “How can you like Fulani, he is up to no good and needs to be done away with, eh? If I had the power I’d have got him before this trouble began.” The other fellow hears and agrees he might help. “I’d even give a cow,” you say. “A cow in calf?” says he. “A cow in calf.” The contract has been made. “There are many witches about from whom the medicine can be had,” says your fellow. He would never admit to being the witch himself.

It would not surprise, if such a truth could be known, to discover that all the wizardry of Old Kingaland was negotiated in just such a sly and inconclusive manner. The first man begins with a grievance, the second with a sympathetic ear, and they end up with improbable commitments. But supposing Chance not long after such an alcoholic conversation were to do the wizard’s work for him. Then he would be in a strong position to collect the payment, making the 138

thing a done deed and witchcraft real. People say of such a man, “Ali n’inumbula ja ‘gukanu, He has a wild beast’s heart.” But this refers to nothing secret, as the words only describe in metaphor human conduct everyone has seen. The technics of witchcraft are hardly more mysterious to Kinga than flight and pharmacology to most of us. The fascination is with the mind of the witch, its presocial ethic— its wildness. These are the mass murderers of common discourse in a technically simpler world. Witchcraft is not an aberrant epiphenomenon but a firm facet of Malatan belief. It parallels the ancestor cult and is at least as important in all three cultures. The cause of your illness may equally well have been the slighting of a friend or a god. Usually both the divination of an illness and the cure are accomplished by ‘objective’ means. For his coughing, the Nyakyusa priest Kasitile consulted in the order given: a diviner who struck a spear, one who used a stick in the ground (two visits), a diviner using a ball of stiff porridge, and finally one who used an axe, coppers, and a pot of ashes. The cures applied or prescribed were not confessional or other self-transforming ritual but objective—medicines or sacrificial offerings. Kasitile’s religion is one of external conditions not inner states. Among Ngonde baneful ghosts “come constantly to ‘brood’ on man and his works and have constantly to be satisfied and driven away by ritual”. Kyala, Kasitile’s spongiest divinity, was said to be angry and brooding over him. Up in the high hills Lwembe would return to Ukwama, unseen but sensed at treetop level, so that men becoming aware of him would know a solemn act of propitiation was in order, even before he had clearly shown his anger. That brooding mind is the core mystery. The procedures for responding to its presence, though mostly esoteric, are technical. ‡ Witchcraft in its usual form in uKinga can be termed ‘objective’ in this regional culture since the power to harm is usually vested in medicines. When they are destroyed, a witch’s power, at least for the time, is gone. The python of the Ngonde-Nyakyusa witch is another such ‘objective’ creature of fantasy. The special quality of charismatic charm a follower might perceive in a dramatically competent and ‘chiefly’ prince in command mode is hypostatized as python-ness. Saying that a man ‘has a python in his belly’ is therefore not quite saying these powers are immanent and so inseparable from the man himself. Charisma like friendship is a quality which must be sustained at every opportunity. Paradoxically, that is why the Sanga unkuludeva, whose charisma is staged not selfdramatically earned, can’t revert to a normal, uncharismatic existence and why a ruling Nyakyusa prince can seem to pass his python along to a successor when his vital charm can no longer be 139

maintained. To say this I have to bracket some of Monica Wilson’s findings about her ‘divine king’. Since an important chief’s body could be sacked (prescriptively before the last breath) for its mystical parings, it compared in death with the Lwembe’s body. But in life the difference was stark. A princely kingdom was won by the spear; only after death was a great prince apt to appear in the frame of divinity. Lwembe’s kingdom was ‘not of this world’. The shared rites of final passage have their meaning in the special need to solemnize the particular kind of succession which is not given by nature but must be conferred by rite. The after-living hair and nails of a charismatic man are known ingredients of the chrism by which this conferring of identity can be achieved. When you catch sight of a python near its watering hole you will behold a being with uncanny powers, but you are not seeing a shetani (daimon) on the way to possessing a mortal. Nyakyusa do, far more than Kinga, credit ‘subjective’ divination even to the point of accepting a ventriloquist—that is, not a consultant diviner but a self-proclaiming and self-serving divinity. Even the early missionaries found this level of credulity surprising, but in context and retrospect it fits well into the picture we have gained of allegations and counter-allegations among priests there. This was a society in which the gods particularly helped those who helped themselves. A python is most often visualized in uKinga not in a chiefly or wizardly body cavity but wish-gliding through the air or under the earth. Even the flight metaphor would hardly surprise a native who had seen a python high in the trees. It hardly needs be argued that the Lwembe who visits in that guise a sacred grove in uKinga is not the living shrine-keeper of Lubaga but the trickster’s shade himself. I never heard that a living witch can become a python or a python take on an individual’s human form, in spite of the currency of pan-Bantu fantasies about skin-changing. Even in the accounts of leopard trouble I gathered in the Eastern realm there was never a hint of unnatural animals, only of natural individuals somehow under the control of a witch. One informant, assuming evil in a particular case, had heard that a specialist could extract a gonadal medicine from a female, and with it attract a rogue male to a neighbour’s byre. Despite the widespread fantasy and boys’ house talk connecting witchcraft with vulture-like necrophagia and a communion feast of witches, in actual cases and procedures Kinga seem not to deviate from their pattern of blaming an individual acting on private motives through objective if devious means. The witch who is killing you may be your friend. Kasitile freely suspected his wives as well as his fellow village elders. So far as we 140

are told his story, though he kept on trying he never really confirmed who or what sort of being was wishing him harm. He lived on for many years, possibly always in partially bad health, looking for his antagonists. Just as the wrong which Kasitile might have done to provoke such anger would have been inadvertent, and just as guilt and selfblaming remorse would have been beside the point, so their fault would have been in an isolated act not an unnatural state of the soul. Moral indignation often and justly arises, but usually is not magnified by infecting bystanders. As with any tort, if the affair entails no public hurt it stays private. This is not to say either that friendships are not spoiled by witchcraft or that interpersonal relations are radically ‘insincere’ and episodic in Malatan culture. But the conventions of Malatan discourse which govern such matters do not give support to the person who would interpret his or her experience in terms of a deeply moralistic worldview. Men don’t range from highly-placed and wealthy saints to poor and marginally-connected witches. The ‘natural scapegoat’ formula (blaming witchcraft on deviants, widows, and loners) won’t work here. Character attribution was not lacking among Kinga men and women in 1960. This girl was bright, that man stern, another not to be trusted. But character in that sense is up front, dressed as what everyone already knows. Many Malatan Christians in 1960 had acquired some taste for the modified destiny-imagery of the West—God, angels, devils struggling for the final possession of your soul. But Malatan thinking seems unmoved by our psychiatric approach to human destinies. Their presumption is that behind the socially accessible and civil personality there is a private ego prepared to act on its own behalf according to a bold calculus of gain and loss, owing little or nothing to civility. I know of no effort by Kinga or Nyakyusa thinkers to rationalize this view of our natures. We might class it as ‘homuncular’—positing an inner homunculus, a featureless form of the outer person, able to operate independently in pursuit of its own strategies. Malatan thought offers the image of the python, a necessary furnishing of any sacred grove, able to communicate wordlessly with man, and the idea of a witch with an inner alter-ego. In Christianity there is an analogue in the idea of temptation to evil which divides you against yourself. Atonement has to be achieved by an act of equal weight to the delict done, and in Christian countries the thought persists in the secular field that criminals can somehow be cured by making the punishment match the crime. Malatan thinking seems different in its initial premise: discovering a witch and forcing him to back off will stop him and put him back on his honour. The malicious act of a witch is neither sinful or sick, rather selfish and uncivil. Witchcraft isn’t 141

checked either by guilt or unmasking but by quelling. I give now in full detail Tunginiye’s experience, more briefly mentioned elsewhere. ‡

The case: As a young man Tunginiye was bewitched by a close friend, whom I shall call Atangile. I give the case here in full detail as I took it down in Swahili, and hope in doing this to avoid the sterilities of abstraction and allow some prying below surfaces. Use of thirdhand reports as empirical case material falls well short of the ideal, amounting to third-rate methodology. Even first-person accounts, when quite brief, are little better. Tunginye deserved his reputation as a sage. Once he had begun to accept my friendship, our common interest in the history and culture of his people brought us together intellectually. We talked at length about this traumatic incident of his youth. It is hard to say whether the hurt of betrayal was greater for Tunginiye than his sudden fear of death. Forty years later the two former friends had not exchanged a civil word. But love and its loss are phenomena of personal involvements not structure. Having survived the attack, Tunginiye was not called upon to pursue a course of vengeance (as he might have been in another culture) or exploit his subjective experience, as by apprenticing to the curer. Formally the episode was closed. Tunginiye felt he himself had fought and won a spiritual battle. Tunginiye: There was a great epidemic (uluswa death, mortality) in 1919 brought by our men returning from the war. By 1920 all was quiet. In 1921 there came a teacher from [Malawi or Zambia] whose name was George Nyasulo Ziwambali. He had been sent to Bulongwa by the British missionary MacKenzie. He gathered some students together, and I happened to be among them, though I was early married and had a son of six months. My wife at the time was Misoni Kyawula, who was not then a Christian. We studied reading, writing, and arithmetic, and the bright ones were always pushed ahead. I was first in the class, Yaseti Sanga was second. It was in January 1921 that the Teacher came to me and told me to leave off drinking beer and follow his way. One day we students were set to tilling his wheat field. In those days we tilled for wheat in January and planted in February. We split into groups by sex, dividing the field between us. Atangile, my friend, deserted the rest of us to follow the maidens. The men finished five banks while the maidens only finished two. I twitted Atangile for that and lo! he became angry. “You think you are so great because you are better in maths than any of us! The Master expects great things of you, you are to be both a teacher and a clerk. As for me I shan’t be counting on it! Ndilolelage isigono syoloso! [‘I’d be on the lookout all the time’—always a veiled

142

threat of witchcraft.]” Everyone was stunned to hear these sharp words. We went home. Atangile was the first to tell the Master that his field was finished, and he was given a shilling to purchase beer for us all. Atangile led the way to some beer he knew about, and after a time I followed. Atangile greeted me facetiously, “Vasite avavaha, Hail to thee, Great one!” He went on in but I stayed outside. He brought me out a small bowl of beer, which I refused. He put it on the ground. It was finger-millet beer and only the hulls, the dregs of the bowl. “I suppose you think we should have set aside the clear beer for you, the prince!” That was his comment. I took none of it but left immediately after that insult. Atangile called after me, “Inyage ndungava utye ulinyaluhala—tulolage! I don’t know if you should make yourself out to be so smart—we’ll see!” I didn’t answer but went home to my wife who had cooked some white potatoes, which she gave me. On the third potato I felt a stab in the back through my whole chest. I began to cough and I was spitting blood. I had to lie down. I was groaning and spitting blood. They called my father and mother, who stayed the night with me. When finally I got to sleep I dreamt of Atangile coming to me naked. He was entering my legs and climbing up my body to the neck. I woke up. After that I couldn’t sleep. My father brought his kinsman Fulano Ndalama Sanga. He shook his head and went to fetch a certain doctor, Ungita bin Kibidi. The doctor arrived about mid-afternoon of the next day, coming with Fulano, and took me out of the rondavel. Right away Ungita said, “They have pierced you with the Blood lance.” He explained. They have a medicine horn and with it a bright, exceedingly thin and sharp copper needle. They captured a footprint of yours and mixed it into the medicines in the horn. Then they pierced it with the needle. They spit their evil words upon the needle. “Tunginiye—ptui—this very day—ptui—he must die!” They keep on that way, telling the needle to pass right through the lungs, the heart, all the organs. At last they have run the needle all the way through the bottom of the horn. As they draw it back out they find blood on it and then they know the medicine has worked. Now when I heard all this I knew that the witch was Atangile, for he had been the favorite minstrel-dancer of Bulongwa before the Teacher came, and always at the dances he had worn two miniature horns at his neck, resting on his Adam’s apple. The doctor had brought along with him fresh healing leaves and roots which he had dug along the way. He had brought his own medicine horn and a little medicine bundle the size of your thumb, ihilisi it is called. It is tightly bound and the medicines it contains are kept secret. The doctor crushed the leaves in a little mortar, adding

143

water. He asked for a wooden spoon and fed me this milky juice. Then he called for the smallest kind of pot ikisaji. He spat into it words I couldn’t hear. Somehow he was twisting his bored earlobe with the one hand all the while. Then he added the roots with water, and still spitting in these words I was not to catch. Then he took up the pot, covered it with a shard, and ordered a fire to cook it. While this soup was cooking the doctor took his own medicine horn isiva, and fed me bits of the paste he had there, made with castor oil, while using a special little paddle. He called for a razor uluketo, and I had to show him where I hurt. He made incisions all around under the left arm and over the right shoulder, a line following the pain and circling to meet itself. Then he took more paste on the little paddle, smearing it on his finger and rubbing it into the incisions. I can still feel the sting! He took the razor again and made three incisions at each joint of every limb, but on my hands only the thumbs, and on my feet only the little toes. Then he made three incisions at the navel, the spleen, the meeting of the clavicles, the centre forehead, temples, and the base of the skull. Each time he made the cuts he would rub in the medicine. The doctor took up the ihilisi bundle. Fixed at the middle was a pretty little horn, a tiny thing no bigger than a carnivore’s tooth and quite empty. Ungita went into a sort of angry mood, spitting unintelligible words into the little horn on the bundle around my left upper arm and warned me that only he should untie it. Finally he fed me the soup he had boiled up from the roots he brought, and I finished it. That is when he began to explain. There are two kinds of Blood lance attack, one where they take your footprint for medicine and one made with the dust from the place where you have urinated. Only if you have antidotes stronger than their medicines, if you have powerful imyesigo, then the witch’s needle will break when he tries to pierce the horn, and right away he will give up. The doctor said I was very ill and near death, but he had some hope just the same, as his own best medicine was the little horn tied around my arm. He had got it at Karonga in Malawi. “If my medicine is stronger than the witch’s, when you dream tonight he will come to you naked. You will find a spear in your hand. You will cast it. It will pass right through his body and then return to your hand. Until dawn you will go on doing this, and tomorrow you will feel better. But if your dream is different, if the witch comes to you naked but it is he who has the spear, or if he has a knife, he will pierce you. You will feel the sudden pain and you will die.” My people cooked him porridge, for we had no beer, and he took his leave. First he scooped out some more of his medicine from the horn isiva. He put it in a cornshuck and gave it to my people so they could renew the medication through the incisions from time to time.

144

That night I had the right dream. Atangile came to me naked. He was covering his private parts with both his hands. The throwing of the spear went fast, very fast. I would wake up. When I went back to sleep again, there was Atangile. He would just appear, staring at me. I would throw the spear. That is the way it was the whole night long. It was the same night after night. My illness continued for a week, with the blood giving way to foam. The Teacher asked after me. My ‘father’s brother’ who was attending school told him that I was probably dying. The Master called Atangile, “Let us go see your good friend who is ill.” They came in the early afternoon. My wife and I were inside, and our mothers nearby. When my parents saw the Teacher come they joined the assembly. Atangile stayed outside. The Teacher greeted me and suggested that we pray. Then he had them carry me out into the light. Atangile sat behind me without greeting. The Master asked about my pain but I couldn’t answer. I was craning around to see Atangile, but he wouldn’t meet my gaze. The Teacher unpacked some medicine which he had brought folded in paper. He called my brother to instruct him. He was to make incisions where I felt the pain and pack the medicine there. When the Teacher went, Atangile followed. Still he had greeted no one. So he confirmed his own guilt. After some days the doctor Ungita bin Kibidi returned. When he learned that I was still fighting off the witch every night, he told everyone that his medicine had proved triumphant. I was to recover, he said, and it happened. ‡

The ideal of a lifelong friendship can be realized. Tunginiye himself was proud at seventy of certain lifelong intimacies with men of his generation whom he admired. But friendships can also be irreparably damaged. Time and bad fortune do not favour their survival. After the experience he recalls so vividly above, Tunginiye moved away from his native Bulongwa and made his life in the neighboring realm. He remained when I knew him quite as peripatetic, though, and as irrepressibly cheerful in manner as most Kinga men are. On his side, Atangile could only accept in a passive way responsibility for an attack which all their intimates came to know about. Still, though the episode was traumatic for both young men, neither was embittered. Perhaps to grasp the moral experience of taking the blame for witchcraft we might resort to the metaphor of twins—though I know I didn’t perpetrate so grave a wrong as my brother has done you, I am prepared to bear responsibility in your eyes, for I can’t repudiate my brother. In like fashion when Tunginiye so clearly recalls the feverish time of illness and delirium forty years earlier he will be aware of another self, clearly his own only in dream, who experienced the attack in full knowledge and won the secret right to survive. It is 145

a dream self, a sort of twin, whom Tunginiye could not repudiate. That is why it was Tunginiye who felt he must move out of the circle. His experience of the affair seems to have been infinitely clearer and more truly religious than Atangile’s—and unforgettable. The two old friends survived most of their peers but never spoke, though there was no civil enmity between them. Only a few survivors in 1960 knew why, or thought they did. It was an old tale but I think it told me where Tunginiye had confirmed his conspicuous faith in virtue. Witchcraft and friendship are the obverse faces of one mystical reality. Friendship as between two Kinga men will be reinforced by sexual relations in a spontaneous way, and it is understandable that the breaking of a friendship would be experienced on the sexual plane as well. Tunginiye dreamt that Atangile was invading him, entering first the lower body and mounting toward the head. It was a physical analogue to the strange invasion of self which had begun in the Teacher’s garden, when Atangile had used confidences freely given in intimacy—that is, Tunginiye’s very personal dreams of a special destiny, to spite his friend. It is through intimacy a man learns the conditions of another’s felicity and how to bring it down. The fact that men are privy to each other’s bodies in friendship has its dark analogue in the belief that witch-horns or needles are thrust, cast, or shot into the body of a victim. Add to that Tunginiye’s kinesthetic apparition of the other taking him over bodily. The nightmare—the mystical battle, the unassimilable fact of hostility as freely given as friendship—turned his illness into a onesided orgy of repudiation. Recurrent dreams, we are told (May 1958:53f.), must be faced down before the cure can begin. Tunginiye rebuilt his moral strength each night as he fought off his witch in dreams, restoring himself. It was in every sense a religious experience, a descent into hell and return. Disenchantment must be the universal experience of sensitive minds in societies which permit the survival of innocence into adulthood. Kinga may experience disenchantment as witchcraft because they are more deeply and emotionally involved with certain good friends than they are aware. The episode we share with Tunginiye happened in the years when, around thirty but ahead of his agemates, he had just been launched into marriage. But with a nursling son his wife would not have had him in the house, and he would once again be sleeping with his friends. The psychic resources to which Kinga could turn, to defend against envy, seem to have been fragile. The voluntaristic structure of peer-group life fails to offer the familiar securities of a well-cultivated kinship system. I suppose ambisexuality, whatever its advantages, is less unambiguous than its main alternatives. In the old court culture, when a man’s hetero146

sexual adjustment was programmed to set in late, the deeper psychological consequences must have been manifold. There was no ground to doubt that maidens and maidenhood were idealized by many young men in the bachelor stage in 1960, and I have supposed that pattern belonged as well to the old court culture. Marriage would then have been conceived as a privileged state, to be achieved in time. When that time came with the world still rather new, as with Tunginiye, I suppose one could predict insecurities. The man is separated from his age mates, taking on tasks and concerns they can’t be expected to share. He has a place to build. He trades his spear for a hoe and a bush hook. From doughty raider he turns to the peasant life. The more typical transformation was later in life and presumably better prepared, often being shared with friends who will be staking out a hamlet of their own. Then I suppose the simple structural conditions may be more important than the psychological. The burden of court culture was to extend the bachelorhood of men to the maximum feasible age, imposing this norm on a bush culture which had called for marriage (probably) already in the second decade of life, and allowed a considerable continuity of kin relations on the general pattern of unlinked localized lineages. Friendship in the bush culture had not been the load-bearing tie it became with Sanga rule and its culture of militarism.

The politics of dishonour Just as the court was kept on track by balancing prince with priest, the Sanga territorial expansion had to have been carried by two kinds of men. The principal would be the unyigoha hero of war, of the type who would be given a princess in marriage after some fine exploit and sent out “to rule” a new fringe community. He was a putative candidate for lordly standing in future. Most likely, he would be lucky ever to mount so high himself, or even clearly to open the way for descendants to do it long after his own death. Such a descendent would have to break free and consolidate an autonomous domain by strength of arms. He would have worn the Sanga mantle, and his descendents or other successors, as long as they played by the book, would continue to recognize the higher court of an ‘elder brother’ as a source of their ruling authority. Presumably, a fief became an autonomous domain at a stage where its ruler was getting strong enough to ‘send out’ some of his own men to settle his peripheries, so to further expand the Sanga hegemony. To judge from the lore I could pick up a half-century later, a prince would want to test an upstart’s strength at arms, but if he proved himself 147

valorous would validate the autonomy of a new local domain within the realm by accepting tribute. Treaties of this kind were not done in stone. The paths to glory were doubtless winding, and a prince who did nothing when an annual tribute was late would find himself on a slippery slope. Orders of precedence were written in sand. History might well be recast after the next clash at arms. As the former outpost could become an autonomous domain with its own barracks and games, the domain at a distance could take on a satellite of its own and begin to style itself the centre of a new realm. The details of such developments are forever lost, though the ongoing politics of the unsettled Eastern and Western realms in the decades before the pax can be sketchily reconstructed, and suggest such a zig zag, ad hoc pattern. The other ‘colonizing’ emissary of the Sanga court was the supernumerary priest or ritual apprentice, such as Kyalawe who worked from Vululile’s court in the West. He was the witch-finder found there by the Germans and spared by them on the ground he had been a mere executioner following the orders of his Prince. It is not difficult to see why such an enforcer would be wanted at any court aiming to colonize a peripheral area and bring its independent prior settlers into the Sanga system. The higher functions of the priesthood presupposed a full-bodied courtly establishment, calling for authority in respect of the agricultural calendar, active acceptance of the Lwembe cult, and a tributary cum redistributional network presupposing a charismatic, rich, powerful personal ruler. By contrast, divining for witchcraft was an undertaking for the ritual entrepreneur able to attach himself to such power structures as might be found on the ground. The witch belief was endemic, but the antipolitan solution (bush variety) was either burning down a hut to expel a witch or, if things weren’t stacked in your favour, pulling up stakes yourself. Witchwork would have been the starting niche for a new ‘doctor’ like Kyalawe aiming to help set up a priesthood under Vululile—displaced as that embattled prince was to the margins of Kinga civilization, he would have been needing Sanga-style ritual assistants. As political and social change progressed with the buildup of a new Sanga polity, so must lifestyle values change. The psychology of friendship and the phenomenon of witchcraft as a specific projection of it would assume new importance. Was the Sanga domain a congregation of witches and victims? No. That picture is skewed, exaggerating the folk concern with witches and their ways. If witches were the truly aberrant types we sometimes suppose, the frequency of incidents ought to be almost negligible. But without going much deeper than we have, there are at least two sound reasons for considering witchcraft a religious belief 148

in the solemn sense of the word. One is that witchcraft is conducted by spirits, not persons. That makes it irrelevant that people who think they are witches are aberrant and few. There is drama in the idea of the ‘whole witch’, the person imbued with evil. But when you look for them you are likely to find few or no empirical cases—granted you can’t count heads when dealing with spirits. The other reason for taking witchcraft seriously is that the witch belief is a system of thought which, once established, will always be easy to act on, and for that reason requires no expensive maintenance. The social anarchy on which it thrives is never quite lacking in any human society, though the self-proclaiming witch of today’s media world, unless he or she is out to sow fear, has little in common with the Malatan self-proclaimer we see in the person of a strong Nyakyusa chief. Today’s urbane ‘witch’ resembles more the Malatan priest who professes to use his powers only to tame the antisocial impulse in others. Probably there weren’t frequent executions at an established court. A prevalence of witches at any given time would most likely reflect other sources of disquiet. Where and when the Sanga system was working well, its own high ceremonialism (the calendrical planting and the Lwembe cult) was designed to sponge up anxieties. Kinga on the whole are not inclined to load themselves with psychic burdens they can’t manage. Phenomenologically, Lukasi (cited above) would have been right. There would have been a general sense that witches were many but when the secular system was working they would be cowed and keep in hiding. In the regional belief system witchery does represent a sort of entropy principle which needs to be held in check by the timely interventions of prince and priest. Most major religions teach no less about the evils men may be up to than witchcraft entails, though the imagery of the staid pulpiteer may be somewhat less lurid than tales told over the fire by a circle of Kinga goatherds. Evil is generally the negative foil to more positive doctrines. The main objection generally found to taking witch belief for one of the world’s important religions may simply be that the positive foil appears to be missing. Still, for ordinary Kinga beset by private woe the ancestor cult was in some ways less ‘positive’ (being less therapeutic) than its alternatives. Solidarity with one’s lineage past can have little intense meaning in a society generally so neglectful of lineage and so careless of the distinction between kith and kin as Kinga were. Where you can’t assemble your kin for the sacrificial meal, what return can you really expect from propitiating your proprietary dead, even if you do recall some of the names? Against that, a well aimed accusation of witchcraft can create a new sense of solidarity with all who will 149

shudder in sympathy. It hardly needs pointing out that the wondrous flying baton of the witch finder when it singles out one certain man in a crowded circle will have pleased all the others there by vindicating them. When witchcraft is compared to shamanism as a response to disease and other private afflictions, it is at once obvious that the latter is the more ‘psychological’ in bent. Witchcraft is rather ‘political’. In a general way the prevalence of witches in subsaharan Africa is reasonably well matched by the presence of developed systems of law beside it. The development of law has brought with it conciliatory habits and established procedures for securing rights. Without a prosecutorial bureaucracy, of course, the victim has the burden of pressing for justice. This has meant the facts have to be clear—both the case for the court’s intervention and the implication of the accused must be undeniable. I expect the reason I have never heard of a diviner being called to name a thief is that factual alibis would too often be available to prove the diviner wrong. That is never the case with witchcraft. It may be difficult but is useful to see the difference between the pragmatism of the law court and the seeming lapse of good common sense in the witch ordeal as a matter of shifting frames not losing logic. Kinga courts, being far less confrontational, are less given to augmenting injustice than courts in anglophone homelands. The institution of ‘friend of the court’ is so universally accepted and applied by Kinga that dispute settlement before an authority customarily has the air of a forum for participatory democracy. In 1960, and presumably a century before, formal dispute settlement was the central political institution in which laymen could make their feelings freely and directly known to authority. As for witchcraft hearings, I witnessed none directly—they were summarily dismissed by the native courts—and must rely on hearsay. But in my enquiries toward reconstructing such hearings I found nothing to disconfirm an expectation of a thoroughly ‘non-Western’ procedure—reliance on ‘unproved’ accusation, divination employing carefully randomized mechanisms which must be seen to be free of party bias, and use of the ordeal to find a ‘scapegoat’ in times of moral crisis. On the evidence, I find no ground for defending the justice of a Kinga witch trial. But as political theatre, a half-scripted morality play, it does present a dramatic semblance of justice. In courts of the anglophone tradition there is at least one area where the drama, the morality play which is the subtext of the proceedings, can remind you of an African ordeal and possibly help to disclose its nature. Setting aside the recently promoted genre of 150

mass-entertainment through the televising of actual murder trials, which is not yet fully approved as a model of legal justice, the settlement of accounts for a broken marriage in our courts mostly depends on confronting one fictional biography with another. The manifest aim of each adversarial voice is establishing unilateral blame. The credibility of a settlement flows from a balance reached in the courtroom drama. If the result can satisfy anyone, it must have used theatrical means to do so. We get a morality play, a piece of judicial theatre instead of the judicious exploration we want of a failed dyadic involvement. Can we not talk of the ‘politics’ of this adversarial system as well as its stunning ‘psychology’? The political value of the adversarial courtroom in a democratic society is the semblance of a ‘justice’ freely achieved without state interference. The same must be said of the Kinga ordeal. The divining procedure is as open and manifestly fair as the drawing of lots, only somewhat more dramatically staged. The decision’s authority is made impersonal. Blame has been laid and the decision legitimated: revealed, in fact. Ten point is easy to settle though sometimes neglected. That is, when threats are uttered in a ‘witchcraft culture’ they usually are not escalated on the spot into violent quarrels. In a society where everyone goes armed as a matter of course, witch beliefs might therefore be counted as part of the system for keeping the peace. How soon after receiving a threat would one have to suffer sudden affliction, to apprehend it as witchcraft? Tunginiye’s quick sense of certainty concerning the nature of his illness and its origin sprang from the seeming immediacy of cause and effect. Even then, the diagnosis was not fully confirmed until much later, by his reading of the suspect’s conduct at their meeting. The probability is high that a ‘veiled threat’ will pass over without repercussion. After a few days of peace, few such threats will not have lost their edge. I find there can be no pressing need for a depth-psychological accounting for the prevalence of witchcraft in antipolitan societies. It is an instrument among many, in ordinary times, of civil culture. Threat, when it will be apprehended as alluding to violence by witchcraft, can’t be answered in kind. Indeed, the gliding ball of fire which is the manifestation of witch battling witch is a sight no Kinga boy would like to admit he hasn’t seen, and a spectacle none can help but smile at. So witchcraft beliefs beget a form of indirect action which won’t veer in a quarrel to dangerously direct action. But in stressful situations, or when witch beliefs begin to serve partisan political ends, matters will complicate. After the transformation of Kinga religion from bush to court culture it would have been hard to argue that witch beliefs were functioning mainly in 151

the innocent capacity of baffles or safety valves, and but rarely as catalysts to passion. What is a royal court, what is a royal theatre, without heroic themes? It is when we once admit that religion need not be ‘positive’ to be effective, when we notice the power of negative thinking, that we may grasp the near functional equivalence of ‘neighbours’ and ‘shades’ as popular figures for the channeling and legitimating of blame. Ostensibly, when (rarely) a Kinga man ties a goat for slaughter to propitiate an offended ancestor, the particularistic principle of lineage exclusiveness has been invoked. Yet we have seen that in the typical event, that principle is unlikely to dominate, being overshadowed by the prior and more nearly universalistic principle of merging kinship, fictional kinship, and amity. The informal sharing of meat with everyone who has smelled it roasting is quite taken for granted. We may read it as evidence that Kinga aren’t willing to scorn a neighbour to honour their shades. A diffuse apprehension of neighborly envy is implied. Theoretically in a ‘descent group’ society we might expect to find ancestors associated with solidarity and longterm, translocal alliances under the umbrella of a maximal lineage or clan. Witchcraft in that case would associate with trouble in the clan—segmentation and agonistic fission. But the Malatan social structure emphasizes lineage only in relation to ruling élites, and even then refuses to create solidary local descent groupings. In the classic literature of African ethnography no people stands out before the Azande. But readers have not always completed their reading of the volumes Evans-Pritchard eventually produced. The Zande ruling clan, to prevent internal structuration and bloc rivalry for office, had resort to undermining the generative rules of family and lineage segmentation. This was done by condoning sexual inversion and incest, producing a radical non-alignment of offspring at the domestic level. It is doubtful how long that system would have worked, as the sheer size of the conqueror-clan was not being well controlled. The Malatan solution for the same problem is not concentration but dispersal of the royals’ offspring. The game is to hold their loyalty as they proliferate, but allow no internal blocs to form as factions in a struggle for succession. So we get prescriptive segmentation, imposed fission and meritocratic weeding out of rivals for the chiefship before it becomes available, all in the Coming Out of young Nyakyusa chiefs. And we get the ‘sending out’ of Sanga avapapwa—the disinherited royal males so named to suggest a ‘little brother’ relationship to the crown—to settle new country as they finally come to marriage age. The result in each case is by dispersing kinsmen to prevent their forming power blocs. In both these cases also the fighting men are mainly bachelors operating within a politi152

cally segmented, pyramidal protostate. This has the consequence that rank and file have spent their best adult years fighting as solidary teams against outsiders—and will subsequently be redispersed. In the case of the Kyungu of uNgonde, since the buffer regions around the central realm were settled by immigrant groups never assimilated and still ruled by their own non-royal nobles, the solution was directly to prevent the proliferation of royal offspring. All but two of the king’s sons were systematically put to death in infancy. This is not the usual practice of subsaharan peoples, whose ancestral cults are meant to sanction kin solidarity and expansion. In general, the Malatan emphasis on royal lineage was narrowly tailored to the sanctioning of hereditary succession to high office, thumping no drum for a generalized agnatic ethic. Some misunderstandings about the importance of (nominal) patrilinearity or residence rules in the region might be cleared up by a careful distinction between such rules when they are politically promoted and when minimized. I believe it helps show the logical relation of witch beliefs to ancestral devotion when the two systems of thought are viewed abstractly. Ancestor rites are propitiatory, witchcraft rites punitive. Since ancestors are not subject to salutary punishment, the troubled descendant turns to the only obvious alternative strategy, which our learning theorists call instrumental reward. Even for the very same act of aggression a witch, whose appetite is considered unpropitiable, gets short shrift. Ancestors, their bodies corrupted, remain tied for comforts to their living kin. In the case of the witch it is precisely those invisible ties which are corrupted, for he will prey on his near and dear while his physical presence remains untransformed. The expression in visible act of these two logically differentiated notions about the invisible world consists in a sanctioning programme. Ghosts are sought out and rewarded, witches sought out and pilloried. Always propitiation entails a feast. Always it is noticeable how little food has to be shared with the shades to satisfy them. For the rest, envy is laid to rest through glut. The punishment of a witch, requiring public legitimation, calls for an un-festive gathering, the divinatory ordeal. Here is then a sketch of the relevant semiotic field:

153

Ancestor Physical corruption Righteous indignation Propitiation Sacrificial feasting Solidarity

Witch Moral corruption Amoral envy Suppression Ordeal Repudiation

The sacrificial meal valorizes a propitiatory stance toward an ancestor. While ancestors are not universally and deeply popular— their interventions seem implacably unfriendly—there are enough human faces among them to evoke the kind of surrender associated with prayer. When that happens to a Kinga man, the religious tone of animism can be transformed in a flood of all the emotional complexities which can build up in lasting relations of intimacy. But I only got hints of that from a few men I knew well, and all of them had been exposed in some measure to Christianity or Islam. For the most part what controls a man’s relation to his ancestors is the simple sense that if you follow the prescribed plan you will get the results promised. When you are offering a feast to your friends, no strings attached, you can expect a supportive response. I would have needed a sex change, I suppose, to have any comparable information on the religious experience of women. One thing they don’t do is tie a goat and set off with it to make sacrifice to their ancestors. Propitiation, if you bracket away the feasting, amounts to a sort of spiritual delousing. The ordeal is a much bigger show. It legitimates the community’s repudiation of an equal-right member as possessed of an evil spirit—the kind of spirit one can only admire in the belly of a legitimate ruler. Elsewhere it is meant to be hidden and never shown off on public occasions. Presumably, it is just such egoistic indiscretions which used to get men into ordeals at the old Sanga courts. A sacrificial feast is wanted in the one case to dispel the sense of hostility between living and dead which will have been in the air from the time the shade is identified as the agent of dire misfortune. Witchcraft also may be handled quietly. To privately valorize the branding of a witch may require as little formality or public theatre as Tunginiye reports at the unmasking of Atangile. But where the court’s power is seen to have been challenged, a grand spectacle is in order: ordeal and execution. The minimal requirement—that responsibility be unambiguously allocated—can be met privately in camera where the purpose is only to show the victim through a crisis. But where the witch has been seen, or where the court wants him seen, as a threat to community the affair will take on a suitably theatrical style. 154

The storied ordeal Kinga elders would most often retail made use of the foreign poison unkali, on which the priests of the court held patent. In this (post-pax, retrospective) telling the poison was not administered directly. Accuser and accused must attend, each with his dog. Either dog, when dosed with the poison, might vomit and survive or might die. Accuser or accused, if your dog dies you are found a witch, while vomit means exoneration. In the event both dogs survive, neither man is more than he seems, and the case is disarmed. No informant could think both dogs would die. I supposed that would have implied something unlikely: a witch resorting to the court in his quarrel with a colleague. The found witch had a chance to repent but was otherwise banished by the ruler’s decree, or might be executed. To judge from reported uses of the given poison in mass ordeals applied directly to the whole adult community, the probability of vomiting was high, and a reasonably happy ending likely. But theatre has its own dynamic, and if the perils it plays on are only ersatz the theatre will cease to move minds. How many actual executions would it take to keep a public in thrall? Only Naïve says none. But Clever says in good times the frequency could be quite low. There would be an inverse relationship also to the dramatic resonance of the show the court could put on when it chose to. Kinga had certainly never known the true ‘African despot’ whose power seems to depend on lavish displays of unmotivated cruelty. The ‘cruelty’ of Kyelelo had in common with a despot’s the peculiar license to invert the rules of honour. He was not a man of his word, he was the worst of ‘younger brothers’, his obsessive ambition was there, naked, for all to see. But he drew good men to him and bent good men to his will. The morale of his realm was high. In an antipolitan world a leader is not expected to live by the rules of ordinary men. It is an obverse to the principle that ordinary men are not expected to keep pythons in their bellies. Nothing, I was assured, prevented a woman from having the instruments of mystical power uvuhavi. But women did not have dogs to submit to an ordeal, and no one recalled a case of woman charging woman. The question would always be turned to the way women might fight, scratching and clawing like cats with no weapons but their nails. A woman could be accused by a man or accuse him, but I judged that the old culture never encouraged women in the witch belief. We might want to reserve the possibility that it was in effect a religion of men, or at least that witch lore was their special preoccupation. So far as the witch belief revolves around the delights of domination, that must be so.

155

I reasoned there were two motives for adopting the dog ordeal instead of a direct procedure, which Kinga agreed they knew but which they under-reported. The first motive is the consideration by a ruler that he could show his strength as well by banishment (should there be a loser) as by execution, while incurring less disaffection among kin and friends of the accused. There are not many ‘loners’ in a Kinga community, and those there are will more likely be pitied than censured. The theatrical effect of the dog ordeal was strengthened by its miming the bolder form in which a fatal decision would entail its automatic execution. And the easy valorization of the labeling act (“Witch!”) would in any event rule out popular demonstrations in defense of the accused. A second advantage to be seen in electing the dog ordeal is the greater ‘objectivity’ of this procedure. That is, it becomes less a defiant confrontation of powers, more surely a cool demonstration of the witches’ impotence before the soothsaying medicines of the court. Cruelty is not always the most effective drama, and at least at the court of Ukwama the reclusive prince may have resorted to the ordeal less often than the priests as a group, to promote their prerogatives. The priests were known as managers among the court’s men, the avanyivaha. As such they were the tax collectors and knew they required an extra helping of credibility. It isn’t altogether irrelevant to our considerations that a prince, himself believing in witches or not, could scarcely discredit the belief in witchcraft without discrediting himself and his charismatic claims. Assume that a ‘well known witch’ has finally been brought to the direct ordeal, only to survive it: has this man not now put himself in position to claim his powers are greater than the court’s? Won’t such a claim put itself forward in men’s private reflections? The dog ordeal has less histrionics about it and proportionately less opportunity for heroics. Whatever its outcome, the forced submission to it put an accused witch down, striking at those phantasies of supranormal powers around which the image of unaccountable moral endowments would have been embroidered in the popular mind. Why should a man so powerful submit to any test at all? Yet once a positive identification of a ‘political’ witch was achieved, the terrible command of execution, Kantagi kuluganda! became an act of authority no one could question. The world returned to its secular frame. Majesty reasserted itself. Priest handed back to Prince. The ‘objectification’ of a religion must always be the work of a professional priesthood. The true evangelist may teach only the way to faith, but ritual and formulaic rite are easier to routinize and guard. In the Kinga case it seems a court could evolve from the heroic 156

to the ceremonial style in a few generations. Vululile, Kyelelo, and Uliluvilo on the Western and Eastern frontiers were princes-on-themake and have in Kinga lore the character of heroes. Mwemutsi and Mwalukisa, in the established Central and Northern realms, were puppets of their priesthoods not the stuff of hero tales. Proof is hard to find, but I collected a good deal more underworld drama in the sacerdotal (or ‘unheroic’) than in the ‘heroic’ realms and came to conclude that there could be a systematic reason for that. A handson, charismatic ruler like Kyelelo or a high-riding Nyakyusa chief can afford to leave no power vacuum at the grass roots within which men may imagine the free-booting witch will thrive. The principle at work here is that the strong-bellied, hands-on ruler, to underwrite his power, has to assure himself no sign of witches and their countervailing power can appear in his realm; but the reclusive prince relies on an almost opposite strategy, seeing his majesty magnified by just the stagy kind of theatre that witch hunts and witch ridding provide. Am I overimpressed here by the bathos I met at Ukwama at the close of colonial times? In a lesser degree it was to be felt then also at the Western capital, Kyelelo’s domain, another which the independent government of the (then) state of Tanganyika chose to put in new hands. Where it was least to be felt, and charisma most, was in Sangilino’s part of the Eastern realm. Here was the strong man who’d stood up to the British and earned their respect, traditional but not in the sacerdotal frame. In face of this unfazable political entrepreneur it is hard to imagine a witch who would not fly the other way. That seemed to be the general feeling about him amongst a people ever ready to notice the weakness behind a mere show of power. At a mundane level of analysis, one seeming advantage of a ‘witch’ diagnosis over ‘ancestor’ would be enabling the aggrieved or bereaved party to claim, not pay, compensation. It doesn’t work that way for an ordinary plaintiff. His whole objective will more likely be extricating himself from an existential crisis, usually domestic, than claiming compensation or revenge. Of course for a prince, finding a witch could legitimate seizure of property, so enriching court life. You might like to impute that as a factor in the balance of a particular case, but you would find the drama of the actual procedure and an eventual execution playing out in quite a different existential frame. It is the drama itself which becomes the issue and drives the matter forward once begun. Although my Magoma informants, discussing self-help, said a witch’s house would be surrounded by stealth at night, and he with his wife and infant would be incinerated with the hut, they couldn’t convincingly name or place a witch so treated. It is a dramatic enough scenario to have an effect on moral premises, if only in the semiotic reality of folklore. But this is an obviously prepo157

litical way of resolving a problem—what you would expect in boundary-sitting Magoma—and wouldn’t have been suggested at a Sanga court. Kinga of the Western realm, in spite of the bad faith between Mwangawa (Vululile line) and Kyelelo (successful rebel line), held that a witch found anywhere in Mwangawa’s domain would have a chance of exoneration first at home, but then at need would be taken under guard to Kyelelo. This would have been the work of the Western priesthood which, unless engaged in the actual making of ritual defenses against attack, could only look to the sacred grove at Ihanga as its proper centre. The first Kyelelo had taken that throne by treachery, but the Sanga order was an overlay on an earlier order, rooted in immovable chthonic shrines, which had served the prior bush culture, and which the Sanga resettlers had in each case found it necessary to reinstate. A witch on the way to Ihanga still had a chance to repent. He must promise to stay clear of his medicines and stay at peace. Then the priests could grant a conditional pardon. They would proceed to his house and burn it with all its contents, showing him where he might build anew. Then they would claim a small daughter of the offender as kisikimuno ward in pledge. She was to be raised as if a priest’s daughter to marriage age. Then the real father, if his scutcheon was clean, might be granted the bridewealth, the avanyivaha priests charging only the customary animal as imongo marriage tax. Doubtless the politics of domain and realm varied from place to place and realm to realm. The Sanga system had exerted a branching, non-confining effect on conceptions of social space. Certain rulers who could not claim higher status than untwa still claimed the hereditary right to execute witches. Such a claim, even if seldom made good, was an effective political tool; but so was an ‘appellate’ system an effective tool for cooling out a flammable situation, and it was much used in the Central realm. The intergroup ambience of the priesthood, and the readiness of that estate to command its own armed task forces in local policing, may have made the business of witchcraft a welcome preoccupation when more important matters were in abeyance. But we have no measure at all of the volume of such business normally undertaken. I’ll hold with the safer assumption that it was quiet in quiet times. What I seem to have learned in support of that can be formulated in three propositions: (1) A ‘witch’ so called, wrestled down like Atangile in private and deeply embarrassed, is put on his honour as never before, and that is likely to last for life. 158

(2) A witch, discredited in public trial, as in the dog ordeal, will be more than ready to accept banishment as the only practical means to restore his honour. (3) Even a charismatic chief can’t afford to risk his reputation for great belly magic in directly challenging a witch or priest, but must find ways of controlling witches and priests through the professional services of (other) priests.

Witch and state The three variants of the Malatan pattern of chiefship/ kingship evolved within a regional system of social thought which had prevailed in the bush cultures formed during the Bantu expansion into and through their part of East Africa during the first millennium of the present era. The witch belief survives from that prepolitical culture and offers a window on it. Nyakyusa are most specific: all men of power—all men equipped to override the judgement of their colleagues—have witch power. That is what sets chiefly rulers apart from ordinary persons. But it is also a privilege the rulers in a protostate context have to protect against encroachment by lay witches, claiming extra-constitutional powers in accordance with the norms of those prepolitical times. Ngonde beliefs are similar, though Godfrey Wilson was explicitly told there it was the Great Commoners close to the Kyungu, not the king himself, in whom pythons were to be found. Nyakyusa accord them to both the aristocratic prince and the lay village chiefs (Great Commoners) who take on the task of controlling him. Kinga in Kyelelo’s Western realm cleanly relegated one sphere of state power to the body of priests, the other to the body of the prince himself. But as with the Nyakyusa chief, the Kinga ruler without his priests and their medicines would have been vulnerable to attack from every quarter. So it is that we find in Mwemutsi’s Central realm an achieved office of leadership by Commoners has been set up outside the gates of the royal enclosure. The crown has become an uncontested, hereditary office confined, as it were, to the work of the harem and a kind of remote-control presence in the decisions of his court and the projects of his priesthood. It is the priests in each society who must be the final arbiters in the conceptual allocation of mystical powers among the several players in the political arena. It is easy to see that without the witch belief, and without their own firm abjuring of such embodied powers in favour of medicinal knowledge and skills, none of the three consti159

tutions would be stable. The priests stay hors de combat or they are nothing. This means scrupulously playing an impersonal game, attributing decisions to infallible procedure. The greatest embarrassment was the incident in the Kinga wars between Kyelelo and Mwemutsi’s court, when a priest was overrun in battle and taken prisoner in full possession of his chemical kit. Of course he was treated well enough by his professional colleagues in the other camp. But in their game, exposing personal powerlessness is a bad show. Assume that the major political burden of a religious system is to give the structures of government a grounding in Nature—in the Cosmos-before-culture, the world in which mannered Man is only a guest. Agricultural societies sacralize the work of gardening and its fruits. Kinga calendrical rites dramatize the natural integrity of a domain and the dependence of its people on the ruler as unkilunga owner of the land. That the ruler himself, being dangerous to the sacralized gardens, must stay back and delegate the work to priests makes a clear statement of the division of office within the ruling court, and of the grounding of that dualism in the mystical properties of the royal office. Kinga also sacralize the predictable events and phases of the human life cycle. Socially significant differences of sex and age are seen to be inherent in nature and so to enjoy immutable meanings. Witch beliefs are associated with the attribution of meaning to premature death. A quiet death in old age, such as in our own folk tradition we refer to ‘natural causes’, is thought of in much that way by traditional Kinga thinkers, though our idea of nature is not theirs. Kinga don’t perceive witchcraft as suspending the laws of nature. It is one of the forces in nature which can interfere with normal cycles of life. The figure of the witch as (rogue) leopard is familiar throughout East Africa, but the single case I collected (Eastern realm) of treating a leopard as unnatural reduced to a mass ordeal to discover the witches controlling the animal. Witchcraft substances are used by men destructively, just as nearly the same medicines may be used by priests constructively to fortify authority or produce a bountiful harvest. In such a society political power can only be expected to find its ground in the fundamental notions of Nature which nourish the prevailing folk religion. Consider how accusations of witchcraft in normal, apolitical settings serve to ground the structure of everyday life. A little boy of nine is feared dying of whooping cough. His friends have helped him to his parents’ hut to say farewell, because he has had a vision of death. After his friends have helped him home, an hour passes. Then the friends come back to report. He is dead and the elders must take charge. It is nearly evening, so the body is quickly and rather secretively buried by the father and father’s brother at a place on 160

the edge of the wood where others of their children already lie. Word goes out. A few women gather with the mother in her hut to keen through the night. By mid-morning there are a hundred men and women gathered. A number of fires have been made and beer brought. Food is prepared, and the keening grows lusty. One late comer is a remote agnatic kinsman. He has with him a squirrel-skin bag with tobacco and some whittling work to finish. He has hardly settled himself when an accusation rings out above the keening. Little by little it becomes clear this is the man being accused. He was seen passing by yesterday but failed to stop. A spokesman for the bereaved at length emerges, stationing himself across the courtyard from the accused. We hear denials and charges repeated. A few others join in. There are intervals of silence and quiet general conversation. The accused man attends to his whittling. A stilted interview continues for several hours. At last the man will have no more of it. He gathers his things into the squirrelskin bag, the first adult to leave. He knows they expect compensation and he has in effect bargained the amount down to ten shillings, all the while steadfastly denying guilt. People say he will send the money but only after a cooling time. Far from binding back together a (loosely defined) lineage group which has been drifting apart, the whole transaction puts a seal on schism—who would continue to attend funerals where he will only be accused of bad faith? The significant social consequence of the accusation has less to do with the choice of a witch than the dramatic assertion that a boy’s death is unnatural. A child has the natural right to expect a full life. So the premature death is malicious and avoidable, an affair of the elders who have access to powers over life and death, and which they sometimes use amorally to hurt one another. This is the way the world is. The compensation the boy’s parents demanded was no more than a token amount in a society where the value of a grown person, as measured by the prevailing bridewealth, might be 500 times as great. What has been dramatized is the parents’ assertion that it is they who have been objects of aggression. They have suffered the loss. Perhaps an observer of the surface of life would have said the nine-year-old would have been most grievously missed by his peers, with whom he lived, though they are assigned no part in the obsequies. The forms of the funeral teach that under the surface of life there is a deeper structure which recognizes the transitoriness of friendships, the permanence of kinship ties, the dominance of the elders, and the awful play of dissension among them in accounting for misfortune.

161

A further consequence of the funeral accusation is the socialization of the dead boy’s peers, as they move about unnoticed among the guests, into witch believing. They gain a sense for schism, an acquaintance with alienation within a kinship group, an awareness of the boundaries of trust. If a young wife dies in her first accouchement, an experienced young man will know it is her agnatic kinsmen who must bury her not his own. Still he may be shocked when the accusation rings out as he makes his early appearance at the funeral. With a penis he has speared her! He has no rights, the loss is not his but her kinsmen’s. They will claim in compensation the whole part of the bridewealth already paid, and in Kinga law it will be theirs. He can’t claim it back. In effect, the law proclaims him a witch on the evidence of circumstance. Only when a new wife has produced and nursed a child of his agnatic group has he the right to bury her and suffer the loss. It is through her attachment to a child not a husband that she acquires a new identity linked to his. The kinds of change which most closely govern our being happen in huts, conversations, intimate lives spread out in time and space. A religion of blame has a theatre of its own, in the courtyard of an ordinary hut, where truths may be told and debts added up which are otherwise not accounted. Witchcraft in the court culture was this phenomenon politicized. To claim monopoly powers in ridding the country of witches, to execute a citizen after finding him a witch, is to put forward in dramatic fashion a new set of premises. While they don’t cancel out the plebeian principles we have noticed at a funeral, the high theatre of the court leaves them in shadow. Witchcraft is no longer embedded in the hard-to-notice processes of change in kinship, friendship, and marriage ties and networks. At the Sanga court the witch has become a threat to the community at large. Emerging from the vast infrastructure of reciprocal debt and moral exchange—the invisible nexus of community experienced in everyday life—embodied power uvuhavi in a common person of independent or egoistic bent has become a political offense. What was a ten-shilling tort, worth in the traditional funeral moot no more than a spent hoe-blade, perhaps suitable for a small girl, has become—should the court find it right to intervene—high crime. In the southernmost country of the Eastern realm I talked with Masonzo (Madamwa Uliluvilo Sanga), nominal heir to the court of the vanquished Uliluvilo. The illustrious father had been lost doing battle not just with the Germans’ askaris and their Mahanzi recruits but with young Mwemutsi (Dembademba) and his men, whom the Easterner had soundly beaten in an initial engagement. Uliluvilo’s domain was big, he was a hero-prince cut from the cloth of a 162

Kyelelo, no recluse, and this history makes plain the balance of powers then prevailing there. The son Madamwa understandably failed to find high favour as a grown man either with the colonial or native Sanga powers, but had held local office at Igumbilo as the devastated region recovered. He had been a child when the Germans came and still too young to stay and fight when the country was laid waste in 1905/6. He had seen little of the old culture of the court after early childhood. He had certainly seen no mass ordeals with his own eyes, but could convincingly show me exactly where one had taken place. “Just there under the branches of that tree on the ridge the witch was left lying. The place is avoided still.” A witch must be judged, he said, by the prince unkuludeva (in this case Uliluvilo) but was brought in to trial and executed by the avanyivaha, acting as the executive and police arm of the court. Ordinarily reports would be brought to the unkuludeva of ‘a witch who is killing people’ in one of the outlying hamlets. The avanyivaha would mount a party to bring the accused to court. They would surround his hut at night and order him to pass out all his arms through the small windows. Thus they would take him disarmed. At court he was directly administered the unkali poison ordeal. When many people were dying in one area, the prince and avanyivaha would go there and round up the whole adult population. Everyone must take a dose of the poison, even after some had died. They were assembled for the ordeal in the bush, and where they fell they must be left lying. In other domains the witch, once discovered, was equally untouchable, having to be strangled by a good length of rope looped at the middle around his neck, the body then to be cast off the executioner’s cliff uluganda without any skin-to-skin contact. I have no reason to doubt this is the scenario I would have got in the same realm sixty years earlier. I have no way of knowing how often it may have been put into practice. I take it that in the process of a people’s politicization the nature of witchcraft will be redefined. For Kinga, what had been an act of desperate hostility anyone could fall into became, within the discourse of the court, an indelible quality of the person, justifying aversion and even execution. Thus is the institution of witch politicized. The institutions which define the new witch are the mortal ordeal and, among Nyakyusa, the autopsy. All that can be said of the ordeal’s history in uKinga is that it may have been coeval with the Sanga system of rule. It may also be younger—associated with particular critical periods in the routinizing of Sanga authority. I say this because in the cases I could inquire into in the 1960s the plot line dealt with use of ‘objective’ devices. The ingenerate witch, who can’t be disarmed and only by greater powers than his own 163

destroyed, appears in any event to have receded along with the Sanga system during the colonial decades, falling into the frame of idle folklore. The facts suggest that amity prevailed over distrust for most people most of the time in both Malatan cultures; that this would not have been the case where ordeals did no more than lynch the defenseless and aggrandize the court; and that it is not the sentiments of fantasy lore but of practical daylight discourse which offer the best clues to social history. Uliluvilo in the East, more than either Kyelelo or Vululile in the West, was engaged in the conquest-colonization of a broad realm, having claims at Contact extending well beyond Kinga borders (eastward into Mawemba country, southward into Upangwa) as these borders were later drawn to the liking of colonial administrations. The character of this self-styled prince comes through to us as Cromwellian. A believer in more than his destiny, he was eventually to lead his people into the only unmitigated disaster ever to befall a Sanga realm. Before the Germans, he had been an exemplary statebuilder. When he traveled out to a bush community, nominally at their invitation, it was he and his party who administered the ordeal, the locals who must undergo it. Submitting your village to such an ordeal, even in some measure inviting it, was accepting the very premise of inequality on which all authority finds firm ground. Ultimately in the fully established princely court, after this kind of proselytizing was done, warfare and dispute settlement, even on occasion the ordeal, would come to be less directly administered. War would be led by chosen commoner heroes, insured by priestly potions and rites. The court of law would be run by deputies, the ordeal administered to dogs or hens. But by then the elevation of princehood to an unearthly plane would have been accomplished, and the elaborate manners of court culture, with its special paraphernalia, would distinguish its style and emergent historicity from that of the bush. Witch belief has the effect of sacralizing the power to stand above forces which drag men down before their time. Kinga informants would resort to the paradigm of three fields side by side, planted and cultivated with the same care, two of them failing while the third produces a fine crop. If not witchcraft, what explanation could there be? The powers of witch and witch-finder, selfish and magnanimous magician, are of one sort. What distinguishes the witch is not his power but the use he is believed to have made of it. An ordinary person accepting the reality of magical powers will accept the existence of an entrepreneurial or hereditary category of persons not subject to ordinary misfortune and ordinary sanctions. Pakipande, the magical diviner most famous among Kinga in the 1960s, was one of those gifted individuals, quite outside the ruling 164

set. He was widely admired—no witch, because his talents were put to constructive use and never concealed. Compare the priest: he has no special talent, rather professional training and knowledge. Compare the reigning prince or local ruler: his special qualities have been passed over to him in adulthood by the priestly ritual acts of inauguration. No one, in fact, is thought to be a witch until someone is seized by misfortune and, to escape, must know who to blame. Should all three plantations in the Kinga paradigm fail to prosper, the evidence hardly implies witchcraft. General misfortune, having no beneficiaries, is to be explained in other terms than the particularistic, grass-roots theatre of blame. But the paradigm does nicely define the crucial role of envy in giving life to witchcraft. True that the wealth of a prince exhibits the same kind of selfserving power as a witch’s prospering field. But royal wealth is legitimated by magnanimous redistribution—the socially corrosive atmosphere of envy is absent. A magnanimous witch is a nonsense, as is a local ruler who is stingy. But so far as office among Kinga depended on network connections—the Sanga connection for the untwa or prince, the priestly fraternity for the budding ritual specialist—the ultimate definition of witch was a person armed with uncanny powers and without legitimate title. This is a person whom you cannot afford to trust. But Kinga in 1960 felt such a person could be effectively disarmed. We have to ask why they would not have so thought under the Sanga regime, and the answer seems to be politics. Mwangawa/Vululile, legitimate heir to the royal Sanga throne Kyelelo usurped, coopted local priests from among the Mahanzi where he was driven to settle. These specialists eventually established their place in the Kinga fraternity, though details of this story are lost. On the other hand, the priest-cum-witchfinder, such as Kyalawe, sent into Magoma or Mahanzi communities, represents the extension of the priestly network in advance of the warlike. He will be followed by a Sanga hegemony gradually imposed. Here the ‘evangelist’ witch-doctor would perhaps be recruiting a local ruler into the secular network as untsagila to the doctor’s sponsoring Sanga lord at home. Presumably, in the zig and zag of history lost, the Sanga expansion was sometimes led by the secular, sometimes by the hieratic network, the one always facilitating and legitimating extension of the other’s influence. I prefer to see the reigning motive on either side as ‘entrepreneurial’ than as ‘ambitious’. Priest and prince each in his own way have undertaken an absorbing job of creative management. Power is part of the job but inseparable from it. I find this implicit in the language used. The group of courtiers which we would tend to call a ‘power elite’ or ‘set of oligarchs’ is 165

similarly conceived in Kinga discourse, but without reference either to embodied ‘powers’ such as witches and sacralized rulers are thought to possess. We have instead avanyivaha, ‘big men’ or perhaps more literally ‘men of greatness’—a plural term indicating a social position not a particular social type. Within this crowd are found the priest (unteketsi = celebrant) with his assistants and apprentices, magistrate, chosen assistants to the ruler, and armed servants of the court, some with particular competencies—in short, the whole agency of the Sanga court. A special mystique was attached to the unkuludeva or prince of a Realm, as distinct from the untwa or local lord of a satellite Domain. Sacred (socially dangerous) things were at once his own, and especially dangerous to a reigning prince. He never entered the sacred grove belonging to his throne until he was brought in a corpse, there to rot in the air unburied and thus unthreatening to Lwembe. He must especially fear the double eyed hoe, forged on his behalf as a most-sacred offering, through which were confirmed the ties of the four realms one to the other and all together to the tyrant-like divinity Lwembe; yet he must be witness to the passage of this ‘gateway’ sentinel through his royal grove. The priests will thus allow him to steal a glimpse under the black sheepskin which hides the strange hoe and its secrets: the esoteric symbolism which (Suluali Memutsi said) only the priests understood. In none of the Malatan schemes was the secular ruler any sort of wizard fit to exercise overweening mystical power over others or demolish witches by personal zap. German accounts make it plain that witchcraft was no more than an auxiliary weapon for the Nyakyusa prince/chief. His persona was more deeply colored by his boldness in war, his authority as magistrate, and the magnetism and bounty of his court. He had to have the right stuff or his chiefdom would not thrive or rise to the implacable competition of other chiefs. Chiefs and princes throughout the Malatan region depended on ritual specialists at any turn in the pursuit of religious ends, and not least in combatting all those entropic manifestations of the human condition locally blamed on witchcraft. The secular ruler’s transcendent status was owing not to his mastery of the mystical art but to power of another sort. The figure of a living Lwembe provides a clue. He displays in his fits of anger and petulant refusal of gifts the prototypical character of the Ancestor: peevish and demanding, a creature whose power to harm derives not from his lore or earned merit but solely from his pitiful ontological condition. To be sure, an outraged ancestor demands above all respect, and it must be convincingly rendered as a condition for placating the 166

imagined anger which lies at the root of human suffering and ultimately accounts for it. But rendering respect, whether to a prince by groveling or to an ancestor by prestation, though it be a performance laden with meaning is hardly a transparent cosmological statement. When Kinga priests come to Lubaga to placate the Lwembe they do not bargain or reason with him. The play is more nearly comparable to that ‘reasoning’ which goes on between guards and madmen, trainers and lions. The prince’s power is a quality of his person, comparable in its nature to that power which, attributed to a witch by the outcome of a solemn ordeal, calls for his execution. The person of the prince must then be conformed to the job of showing off the power he is destined to need—the powers, in all particulars, of the office into which his priests have been prepared to install him. I think the bush culture’s witch was a mere dabbler in medicines. People would seek from him cures and sometimes remedies against attack. In the 1950s the witches found by the Ngonde finder Chiganga (Kyiganga) and his ilk were the same. I won’t say witchcraft had been simply commercialized and so trivialized, I will say witchcraft had been depoliticized. I argued early in this chapter that the politicized witch of the Sanga court culture was a phantasy transfiguration of the small-minded neighbour into a public enemy, a daimon. It seems to me that the Sanga prince was in like measure a transfigured man, dangerous in his person, and dangerous, not by reason of the lore and pharmacological skills which gave priests their standing, but by reason of his inhabiting a separate frame of being. A prince is in this respect like a witch with the powers of flight and outof-body experience. But the prince is not simply, in the Nyakyusa concept, a ‘defender’. The constructive powers he needs are greater than and different to those of a ‘good witch’—one who might claim to be on the side of community, health, and prosperity. The office of prince, whether managed by éminences grises at Ukwama or fed by ambition as among Nyakyusa and in still-unsettled Kinga realms, could establish authority on a trans-local basis and within its sphere that crucial sense among a people of sharing a common fate, which is the ground of political community. It is a transcendent human figure which is able to do that. All the right props and scenarios must be available.

167

FEAR

& BLAME, CHAPTER FIVE

Priests and their Princes Histrionics of belief The Berlin missionaries were quick to gain fluency in the local language but not as quick to grasp the situational subtleties of their encounters with new neighbours. As would always be the case, some first observations were the keenest. While still unsure of themselves they were looking for clues to the meaning of daily events; but early judgements can also be naive, and the scholar today has to practice doubt concerning the missionary societies’ domestic discourse as it developed over two decades. The problem with the Wilsons’ ethnography is quite different. We haven’t scraps but a detailed and coherent portrait of a culture, among the best descriptive ethnographies the twentieth century has left us. The difficulty posed is the inescapable one of struggling for the feel of a culture from a distance. In both cases I think the major correction needed is an examination of the observer’s premises with respect to Nyakyusa religion. After dealing with that I’ll discuss some others. But on the matter of religion what is missing is fundamental. It is recognition of the profound difference between animism and the aptly named ‘high’ (or ‘sky’) religions. The Germans actually supposed they could find the right Nyakyusa word for “God.” Monica Wilson allowed her description of a “Divine King” (1959b) to suggest a measure of majesty, of theatrical power, and of political importance which could only deceive an unprepared reader. Worse, her ‘divine king’ appears to be a centralizing and unifying divinity, the apex of a priestly establishment. Lwembe as myth may indeed be just that for the Kinga; but the living Lwembe maintained at Lubaga by the local Nyakyusa priests represents a different kind of cultural theatre. If Kinga priests use Lubaga to focus a sort of state religion, Nyakyusa do not have one. The protostate process tends to modify an animist worldview, but does it piecemeal. Animism is generally eclectic and non-didactic, easily absorbing new cults and enlarging religious discourse so as to naturalize the new. What always remains is the daimonic premise: that the supernatural world is no more centred in a singular power 168

than the natural one. Nyakyusa priests comprise no sort of college. This is what most clearly distinguishes them from Kinga priests: it is not that the Kinga priest is less of an auto-didact but that he recognizes collegiality; and it is not that the Kinga is not quarrelsome but that he expects, like a good academic, to reach an eventual meeting of minds. In religion as in politics the Kinga world had taken a further step toward central direction. Its political aspect is not the essence of any religion—grant that. But it may be the one easiest to get at from the outside. The great puzzle Malatan studies can help to solve is the prevalence not of witches but of priests. Bena society can serve as the standard for comparisons within the Sowetan area. Hehe were transformed before the pax into a militant ‘nation’—the pre-Arab, pre-Ngoni condition of the peoples brought together under the tyrants Munyigumba and Mkwawa is hard to know. Bena were involved in the enormously expanded Hehe political field but less deeply, and were not so radically transformed. They reverted by various paths to localist separatism. In brief, the Bena pattern was a multiplicity of independent chiefly communities, each in external relations on an equal and independent basis with similarly constituted Bena groups. So far as we can know, the pre-pax situation reflects their earlier political pattern, though there was much mixing of households. There was in the result no sort of hegemony as among chiefs, no secular central direction and no hint of ritual suzerainty. Taking the pyramidal model in reference, Bena lived in coequal pyramidal polities each of which had taken on the character of a pedestrian chiefly community of moderate to maximal size—altogether a congeries of small pyramids constituting a single, spreading linguistic and cultural ‘people’. Each chief could look outward from his kiya or longhouse toward peripheral client settlements constituting a buffer between Bena and other peoples, or between Bena chiefdoms. Considered as a phase in political evolution, it is one the Nyakyusa as well as the Kinga had superseded. The main Bena study is the Culwick monograph on the displaced riverine Bena resettled after the Ngoni-Hehe wars in a new sort of natural setting. These Bena comprised about one tenth of the full ‘Bena’ populations of early censuses, in chiefdoms spread over an extensive highland landmass and comparable in scale. We have to depend on the ethnographers to judge how far the dislocation of the riverine group affected politics. I give here in skeletal form the sense of the Culwick monograph. While there is much, much more on record, all to be sifted and discussed before one could reconstruct a probabilistic model of Bena society before the wars, this much is confirmed in a rough way by other sources. The chief (Mtemi) 169

was himself the ‘high priest’ but was paired with a ritual specialist (Mzagira wa Tambiko) without whom major public sacrificial ceremonies could not be reproduced. The system is thus ‘monistic’ with respect to the secular/sacred dimension. Chiefly ancestors were the recipients of public ceremonial address, and these ceremonies were strictly parallel (only writ large) to the private propitiation rites which accounted for the popular/traditional religion of ancestor propitiation. Each chief relied on lesser local (client) rulers in marginal ‘provinces’. The tendency of the system was to grow localist roots but without generating frozen hostilities between the autonomous polities. Comparing Bena to Hehe we see a congeries of marginally mixing and intermarrying polities on the one hand, and a fully developed system of grand-scale central direction on the other. Comparing Bena to Kinga, the Bena Mtemi matches in status the Kinga Untwa as the ruler of a robust pedestrian domain; the Kinga Unkuludeva, ruling as primus inter pares a realm branching over several domains beside his own, each with autonomous sovereignty, has no Bena counterpart. But setting Bena beside Nyakyusa the result is a confusion of the both/and sort. The main point for my purpose here is that the successful Nyakyusa ‘chief’ is, like the Mtemi in Ubena of the Rivers, at once more than secular and less than a priest. But the obverse of that is interesting: why in the Nyakyusa case does this seem to require a prevalence of priests, when for Bena one is almost enough for a population of 15,000? Before this little mystery is solved we should have a better grasp of the way translocal thinking can emerge from chiefly rivalry in a setting like the Bena. We are after clues to the genesis of our Malatan protostates within the Sowetan political archipelago. It is the Nyakyusa model not the Hehe which best displays an unforced process of political evolution. Is there an inherent sequential ordering among these Sowetan polities? I don’t think the question, put bare-faced, has a clear answer. Tentatively, I suggest looking at the possibility that a phase of ‘arena politics’ has to intervene between the ‘Bena’ model and the protostate. That means turning to the questions which led Monica and Godfrey Wilson to choose the ‘Konde’ linguistic and cultural community for study. Their special interest was in the analysis of social change. The Nyakyusa, they showed, were in a process of expansion and so moving toward a new constitutional system. Since their position has been clearly rejected by at least one subsequent study, I am undertaking here to rethink their case. That means getting more deeply into the literature focused on Selya.

170

The picture I’ll be painting of precolonial conditions in Selya will emphasize a politics of uncertainty and managed turbulence in which secular political authority is vested not in the office but the person of a ‘chief’ or ‘prince’. It is a picture of arena politics characterized by a good measure of theatricality and entrepreneurial ambition. Seen through Kinga eyes, the absence of a firmly established territorial map only invited political mischief. The ‘priest’ in this picture is a more stable but no less uncertain figure than the ‘chief’. But if my picture of Nyakyusa doesn’t duplicate or epitomize Monica Wilson’s, the evidence for the priestly role lies in her case records. Her portrayal of the role, so largely based on close relations with one such (Kasitile of Selya), is the product of direct observation, still feasible in some cases in the thirties. Her chief, on the other hand, is based on men more radically affected by colonialism. Government under the Germans made no pretense at non-interference, and British Indirect Rule made a point of legitimating only the secular claims of a ruler, while keeping a hands-off position in what they considered to be non-governmental cultural matters. In the sixties at Kinga courts my own reconstructive task, harder for both élites, was particularly difficult for priests. But there were so many commonalities with Nyakyusa royal courts that one culture could be illuminated by close study of the other. When you have seen one amphora whole you are better qualified to put shards together toward the reconstruction of another. The ‘whole’ in this case is the court as an institution formed by the fusion of secular and ritually founded power. Animists do have names for their divinities, but as you move through a country you will find they have far more local divinities than names for them, and no name at all for an almighty God. The Lwembe of Selya whom Monica Wilson describes as a ‘divine king’ is the incumbent of a sacralized office solemnized by dramatic, esoteric rites. The (living) Lwembe is no true king though, and the (spiritual) Lwembe who seems to possess him is a stranger to apotheosis. The more we know about this Lwembe, the harder it is to like labeling him, as the standing literature does, a ‘Hero’ figure. He is a shrine divinity to whom the Kinga respond strictly as to a territorial cult, but whose local role in Selya is harder to characterize. His counterpart among the Ngonde people, the Kyungu, did become a proper medieval king in the nineteenth century but appears to have lost a good deal of his ‘divine’ mystique in the process. The circumstance that the Kyungu’s office was hereditary, as the living Lwembe’s was not, is enough to set the two apart. That one enterprising Kyungu was able to convert his ritual importance into centralized secular authority gives us a better case for deeming the resulting institution a ‘divine 171

kingship’. At the same time, the quasi-kingly aspect of the Lwembe, as recipient of abundant tribute yearly from a full roster of Sanga and Mahanzi courts, gave him a plainly secular stamp. What is particularly telling in the Malatan region is the family resemblance of many such ‘semi-sacred’ offices within the broader Nyakyusa-speaking culture area. It is especially important to have in mind that any quite prominent ‘chief’ or ‘prince’ among the Nyakyusa could expect the same impressive mortuary treatment as the ‘divine’ Lwembe. This was aimed at preserving a dying secular leader’s charismatic identity, so as to be able to pass it on (within the frame of ritual theatre) to a successor. But it was the priest not the heir who owned this transaction, from the taking of ‘immortal’ body parts from the old prince to the eventual anointing of the new. What we have here is an ‘hereditary’ office in name, but one to which a plurality of claimants might aspire. Birth position alone is not enough. The chrism of office has to be ritually applied by the priestly establishment which guards it. A perspicuous social consequence was the chosen successor’s moral and ritual dependence on a nonchiefly interest group privileged to bestow or withhold the one gift essential to confirm chiefly success—a supra generational identity. It was a gift they could withhold for cause. Considering the moral weight this bestowed on local priests, and their impervious condition, I judge it as a gift the chosen one could only accept as a revocable contract. Compare the Kinga. The secular ruler of each domain achieves his semi-sacred air on the inauguration of his rule, which can only be staged by the priestly group of the domain—in many or most cases, I believe, with active support or intervention from their peers in other domains of the realm. The priestly privilege of winnowing out potential successors who may be unpopular or incompetent is the same, but the circumstance is not. In the Kinga case the Mkinga passes his exams early in his manhood as a monitored candidate, overseen by his priestly jury. If another son of the incumbent is to be preferred he will have been chosen well enough in advance for proper stage-management. In the Nyakyusa case it is evident the successor to a major title has to fight his way to it against his brother-candidates. The histrionics contrast. In political theatre, claims to ritual power are no better than, in troubled times, a sometimes fickle public will allow. This is a danger the priest as such is sheltered from but not the prince, however grand the role he may cast for himself or what mystical powers he may claim. When there is trouble in circles of power the buck doesn’t stop with a priest as long as his every move has been properly 172

staged to seem none of his own. The very essence of priesthood, which is shared with the peripatetic ‘doctor’ but never with a prince, is this disavowal of egotism. This Konde religion of blame relies as the Kinga does on implanting through public ritual inward sanctions to forestall transgressions of a consensual code of civility. For a number of reasons, any such religion will be most at home in an egalitarian community. For reasons special to societies which find serious meaning only in transgressions by males against males, the political uses of blame may include domination. But witch belief, like the cowboy’s sixgun, is a great leveller. The aftermath of confrontation is dislocation not structural change. Like the duel which European custom treated as an ordeal exempted from exogenous sanction, witch accusations return surviving entrants to a status quo ante. Ironically, witches, who are in the popular lore everywhere taken to be members of a dark conspiracy, are not seen undertaking class actions. They are not conspirators but, in the local idiom, owls—lone night-feeders. As an instrument of sanctioning close interpersonal behaviour, the witch belief fails where structural complications put whole classes of persons mutually out of range for mystical blaming. Blame is, as guilt is not, radically inter-personal, strongest in relationships which can generate deep resentment, envy, or fear of slight. If blame can’t embrace both sides of an impersonal barrier to shared intimacy, blame in complex societies will not gain the moral depth it has where inequalities aren’t socially buffered. How then does one set one’s self up as a mighty ruler in a cosmos where witches prevail? In his classic Yucatan study Robert Redfield (1941) found witchcraft in Merida thriving in the slums as nowhere else, virtually absent in the classes above. If your slave has to be feared as a witch, you aren’t boundlessly privileged as master, since your own daylight sanctioning power diminishes if you create deep resentment within an enduring relationship. If women are exempt from accusation, then the witch belief will hardly take hold in them. If they can be accused but find their own accusations scoffed at or belittled, their position is embattled, with the implication of resentment. But if women are at par with men, both genders have the same concealed weapons for averting hostile acts and tempering domination. The best evidence we have for witch-prevalence among the Malatan peoples is the compilation of cases which Monica Wilson made for Selya in the 1930s. There are 38 records. Where gender is 173

given, only one case involved the targeting of women, while six women were accused of witchcraft. One of these was accused by a husband of ‘bewitching the loins’ of her co-wives and divorced. No woman is mentioned as an accuser. In one case an accused wife was forced to undergo the poison ordeal. By successfully vomiting the poison she exonerated herself from the charge of bewitching her husband’s cattle; he then paid compensation to her male guardians. ‡ As for Kinga, in the decades before national independence it had become the custom to send the accused with his accuser to judgement abroad, so as to settle cases within the letter of British statutes banning witchcraft from local courts. The one special case I witnessed was brought to court by a sadly muddled wife who wanted to accuse her husband. The court wanted no more information but politely explained its incompetence in such cases. A few men seemed amused. There were, as usual, no women auditing, and I know of no sequel in that case. To my knowledge, only men were expected to take the difficult journey by foot and boat to the professional witchfinder new custom advised. I saw many men who had received his absolution but never saw or heard of a woman bearing the telltale facial incisions of his knife. As with the Nyakyusa, Kinga women are strangers to the court as venue for civil cases, unless they are brought in as accused. Anecdotal testimony says this was so before, when courts heard witchcraft accusations, but circumstances then were different: I think it unlikely that ordinary women would have had the chutzpa to face a major court with a personal problem. Where political evolution is the concern, witchcraft likes to move with development from the domestic level of relations to the external sphere of politics, law, and ritual. It becomes ‘political’. However that may have been, and whatever the subtler differences between Sanga courts and those of the Selya princes, the broad Malatan scene was characterized by a division of spheres on gender lines which will be familiar to readers of ethnographic monographs from many parts of the world. It is men who are expected to claim the loss when a child sickens or dies; men who suffer when their women die or misbehave; men who sacrifice to their ancestors on either side; and men, on the whole, who dare bring charges to the authority. All this means that a woman seldom or never enters the lists to confront a man in the name of justice among equals. And this in turn at least whispers that a man is more than likely to have met a woman’s resentment and want, should trouble arise, to displace or reflect any blame. In the 1950s Monica Wilson found Selya women’s resentment very near the surface. In For Men and Elders she paints Selya as a repressive society in which young men and married women are the spiritual victims of a world 174

made safe for children and their aging fathers. Separate was much more nearly equal for Kinga women then and, most probably, in traditional times. So it is prudent to conceive of the Malatan witch belief as a creature of male discourse, fed by the presocial logic of masculine egos in a world where demand for women and wealth is forever fated to overwhelm supply. While animism in itself may not disallow normatively sanctioned status differences, as a religion of blame it offers a matchless matrix for witch beliefs. The way this keeps a society (or its male component) roughly ‘egalitarian’ is logistical: misfortune is no respecter of persons. When it strikes one of his women or children, a systematically capricious system of divination is in place to tell a man who to blame. But whether divination reveals that the spiritual agent of distress is a Burmese nat (Spiro 1967), a Bantu ancestor, or a living person with powers or medicines of witchcraft, the universe of suspicion is small. Blame has to fall within a closed circle of personal connections, and even in the case of witchcraft there is blame enough to fall back on the person of the victim. Who is not diminished, even as the butt of justified envy, by the mere odor of vulnerability? As for scapegoats, if the term is taken in its original meaning, to find a real case I expect you would have to look under a more doctrinaire religious umbrella than animism. But then, in face of this leveling tendency in the mystical arena of witchcraft, how does the authority wanted in any translocal political system grow and thrive? The simple answer has to be cooptation; and that, in a nutshell, is the business of the priests. With their help, the witch belief can be made to succor secular authority. Setting aside the fantasies of fireside lore, a ‘real’ or ‘realistic’ Kinga witch acts on personal motives, just as ancestors do. This is the witch you will grieve against. Be they nats or ancestors, the spirits who are plaguing you have each a quite particular congregation, and it is the same for a witch. The circle comprises your world of intimacy, people who depend with you on primary mutual rights. When these rights aren’t confirmed in action, the circle can break. Anxieties increase. So a religion of blame makes it acceptable, when things go wrong, to fault someone who seems to have proved undependable. The message is, Take care! Atangile in our reading of Tunginiye’s sorry tale was testing his friend, who seemed to be getting conceited: Take care or you’ll lose your friends! In the event, he got the same message back more clearly than he could if Tunginiye had actually uttered it. Most often there is no such drama as, by medical misfortune, happened in that case. A man gets to 175

uttering veiled threats. The witch word comes back to him in gossip. He begins to take more care. This may be a feature of peer relations everywhere but it is sharp and double edged where witchcraft is socially real. Take care or you’ll lose your friends! That is the advice Kinga priests liked to tell me they had for an uppity prince or the lord of a domain. Sanga rulers had the inherent powers which go with majesty but weren’t armed with the ‘pythons in their bellies’ which Nyakyusa chiefs could boast. The reason is to be found in the difference between the two worldviews: since the Sanga ruler had to fear not night-witches but the medicines of witchcraft, his safety was in the hands of his priests and the other court trustees. Nyakyusa put their faith in the mystically armed ‘Defenders’ of their village. This calling was the distinctive property of those men Godfrey Wilson called ‘Great Commoners’—not just the ‘village headmen’ but the handful of peers comprising their trusted entourage. The counterpart in a Sanga court was the group of avanyivaha. Among them were the specialists I have called ‘priests’ but also the secular men of power associated with the court of law, peacekeeping, and tax collection. The bond of personal loyalty was the key to a ruler’s security in Selya. In Sanga courts the nexus was structural. The Lwembe complex, seen from the Sanga courts, is a replay of the private case of ancestor blaming on a near-cosmic scale. When the community is under threat, Blame Lwembe! For the Kinga case, this is made workable by setting one chief out of the ordinary. It is a difficult balance, in good part because the scope of authenticity for the ritual theatre by which this can be done is strained with each increment of political-territorial growth. Bena chiefs fought this by bringing young ‘semi-royals’ in for schooling at the capital court. Sanga courts circulated the best of the rustic young males of a domain into the capital for the barracks life and its more sophisticated pleasures. Nyakyusa relied on the age village. Given the witch belief, in the presence of purposive politics the range of a veiled threat or muttered sanction can be extended beyond the polygynous compound and network of peers to important relations in the public sphere. So Kinga ‘priests’ or ‘doctors’ traveled with the ruler’s other political lieutenants (avanyivaha) to collect taxes, and arms were seldom needed. So priests talked to the unkuludeva in his harem-sanctuary about mutterings among the men, and wheels were put in motion to stage a feast, or a rehearsal for war, a crowd-pleaser. Animism can lend itself to politics in this way because a religion of blame sets up a

176

grievance procedure based on the presumption that deep trouble always comes from beings on the other side of a mystical divide. Animism has flourished wherever there are those ‘little traditions’ Redfield liked us to notice persisting under the broader umbrella of a ‘Great Tradition’. Records of witch burnings during early modern times in Europe and the Americas have in this century variously been interpreted as indicators of a pathological turn within the Christian clergies and as clues to the persistence of ‘paganism’—a (distorted) animistic peasant religion thriving under the disciplined formalism of a Great Tradition given to abstract moral teaching. However those cases are read, unless real devils were at work it is plain we have to deal with an under-religion of blame, and a crisis of incompatibility between folk needs and clerical beliefs which was destructive of both. Animism will never be hard to re-invent. The Great Religions of the world tend to be grown into the secular cultures where they have taken form over centuries, but secular change can spoil the fit. Religious thought is likely to be ninety percent unformulated, living as well in confusion and doubt as in personal knowledge, and would fairly cease to exist if things were the other way around. A crisis in the hearthside culture may call for a grasp of traditional principles which aren’t easily agreed on. It shouldn’t surprise that animism, with its flexibility and transparency of principle, takes over turf in such times wherever a more rigid religious establishment has left it unguarded. Where trouble must have a moral cause, the shift from blaming one’s self, whether for sin or impiety, to blaming a selfish friend or god is not always difficult to make. ‡

Credos and credibility There are two points I find the three Malatan cases can help to clarify. The first point bears on the roots of credibility in ethnographic modeling. The Malatan is a region of state-building in a formative phase we seldom get to see so nearly in vivo. Archaeology knows these transitions have to happen in the formative stages of any civilization, but the sociology of such a transformation is hard to reconstruct from stones and bones. As we’ll see, wonderful as it is to have on record a few pertinent observations by early missionaries on the lifestyle they had found in a pagan ‘paradise’, there were few sociologists among them. More than that, however apt the missionary’s training might have been for the time, we now have the last-best century of professional ethnographic study behind us and still it is a struggle to understand how the Nyakyusa-Ngonde social systems must have worked. I’ll reconsider some excellent efforts in 177

that direction published by anthropologists and historians in the post-pax decades. My own revisionist view will have to be judged on its merits. But for each ethnographic region most of the social facts not forever lost are already on record somewhere. We can’t count on much fresh evidence. The business of Malatan studies then becomes a forensic exercise, pitting one probability against another, and theoretical models are essential. The point is to keep asking how the model has to change to account for the viability of a particular world we are trying to know. It wants touring a maze of evidence always doubting but never quite denying. What all sides agree is that with the colonial deep-freeze gone the old political order in Malatan life is broadly misremembered. The fragmentary memoirs which we have need careful screening. If some few happen to be vivid, and early missionary notes occasionally are, common pyrite can have the luster of gold when it is gold you are after. Monica Wilson’s last fieldwork in the region was done in the 1950s, and her interest then was largely in systematizing her information on chiefly genealogies, a task some of her informants took up with enthusiasm. Her final monograph on the Nyakyusa offers a profound reflection on the structure of experience—the moral career of the individual—in the society she studied in the 1930s. But that task doesn’t engage her in reconstructing the political sociology of the turbulent ‘paradise’ her kinglists so clearly fail to evoke. Should we blame the pretensions of positivism for this? The rules of method seemed to the midcentury social anthropologist to say, Trust only your data not your private sense of the probable. But when the data themselves contain only your informants’ own private sense of the probable, what then? How much detached understanding do most of us have of our own social lives? If the normal dose of such understanding were truly comprehensive, the ethnographer’s job would be simple—round up a few articulate informants and take dictation. The biggest jigsaw puzzle in the world could be easy to compose if all the pieces were numbered. R. G. Collingwood’s Idea of History is the methodological bible for reconstructive ethnography. The following passage will be found under the heading, ‘History as re-enactment of past experience’: The historian does not know the past by simply believing a witness who saw the events in question and has left his evidence on record...[Rather,] what he does to his so-called authorities is not to believe them but to criticize them. [1946: 282]

When documentary evidence is fragmentary its credibility may hardly rise above that of mute stones and bones. A professional 178

monograph, whatever the conditions of its production, reveals the degree to which there was a meeting of minds in the fieldwork. But if it isn’t often you would choose to criticize your oral informant face to face, it is seldom you should be content not to pursue the contextual questions any statement must raise. As it is with thrown bones, any batch of social facts from which you hope to reconstruct a past may be moved about on the board of the mind and reconstrued. What is special about reconstructive ethnography, when virtually all the facts you’ll have are in, is that no single account of their meaning can be empirically tested against another. Goodness of fit, like elegance, is a matter of judgement. With the vantage point of a student of the regional culture, my reconstruction of traditional Nyakyusa politics begins with the social situation of the chief or ruling prince in his court, which I see in contrast to the more established Sanga realms as well as to the more nearly sacerdotal Kyungu court in traditional uNgonde. The second point on which the Malatan evidence has a special bearing has to do with the nature and history of religion. It is a truism that one age in any civilization will be more religious than another, just as in a single generation one man may verge on fanatical compunction while a little brother steals from the poorbox. I don’t plan to estimate the intensity of religious feeling among the three Malatan peoples, but I think the distinctive qualities they display have much to tell us. My focus will keep coming back to matters of civility, because I think the enduring virtue of a religion of blame is sanctioning awareness of one’s own vulnerability. Nine tenths of civility is guardedness. Consider the formal emphasis on amity (as against the more common reliance on kinship) which the Malatan cultures developed. The formality of Nyakyusa manners stands out against a turbulent if not chaotic political scene and contrasts in this with the Kinga case. To say why this should be, you have to take in virtually all the differences between the Sanga courts and the scene the first German missionaries had found in Selya, which prompted them to write home about a tropical “Paradise”. The same language had been used by some earlier European visitors. The great puzzle will be to comprehend a seeming paradise which could be as restive as Selya evidently was at the time. In the public sphere, nothing can equal the feast as an instrument of atonement. But in a context of intimacy and reciprocal sanctioning, it is blaming which most predictably leads to disruption of the peace. Quarrels arise among intimates and run their course, and amity is usually restored when verbal pummeling has worked off the worst of the mixed frustrations which started the ruckus. Histrionic displays of moral indignation, having found 179

their proper audience, can subside. Kinga communities tend to be stable. Even when very small they are not self-isolating. The ethic of amity promotes here—as not in Selya—a wide-branching network structure without the all-pervading mosaic of in-group identities the kinship ethic always (but the amity ethic less certainly) favours. When quarrels erupt in a Kinga hamlet there is—again, not as in Selya—no joining in. Kinga bystanders to a quarrel tend rather to hold their ground and intervene only to cool parties down when dire deeds threaten. If fear plays a role, it is common-sense fear. The kind of fear the witch belief can introduce is of another sort. Instead of raising hackles it wants to subdue egoism, anger, and indignation. It is the kind of fear a priest or prince can use. It works on multitudes. The actual prevalence of witches is imponderable no matter who you ask, but the importance of belief in their magical powers can be seen to rise with protostate politics. This is quite certainly the business of priests not princes, however obviously both of them stand to profit. The Malatan region, where the European pax found three linked protostates differently placed in circumstance and stage of development, offers special evidence.

Witchcraft, statecraft The best place to begin a general assessment is in Selya. This is the Nyakyusa region Kinga priests regularly visited when fears about hunger and health had turned to organized anxiety at home in their four realms; and this region was the practical focus of the better part of the ethnographic work done on the Nyakyusa-speaking peoples. There are also some special procedural reasons for beginning with the Nyakyusa around Selya. The earliest close and reliable notes on Malatan regional culture were left by German missionaries who actually preceded the pax by a bit; and there is confusion over the reading of these reports, particularly concerning Nyakyusa religious belief and praxis, which ought to be cleared away. There are anecdotal clues to the puzzle how the signal social and political mechanisms reported in the later ethnography could have worked in practice. And there is some robust material on the politics of fear. But what is probably more important is assigning an order of credibility to the fragmentary evidence the earliest German missionaries left. The methodological issue is clear enough. Classical scholars from time to time turn up some bit of new or overlooked information which doesn’t fit the standing Wisdom. Occasionally this sort of thing may start an academic swing toward a new ‘paradigm’ or ‘disclosure model’ of the classical world. But eventually the dialectical movement turns again toward synthesis. Previous 180

views, also based on evidence, have to be reconciled. It is the same with reconstructive ethnography: until we have an integral grasp of a case we can’t even tentatively close it. In a linguist’s view, Nyakyusaland is remarkable for resisting influence from its neighbours, and the suggestion is that the language, having been in place for the best of two millennia, would hardly have succeeded so well without the help of a long continued and forceful incorporative politics. At a minimum we can conclude that the Nyakyusa political system has a long past, and a shining record of defending its communities against incorporation by others. ‡‡ The sociological reasons for this success are really not hard to find in the record, once a few ethnographic muddles have been got by. The problem is to reconstruct the institutions prevalent in Selya before the 1890s and ask what sort of social and cultural activity they would have generated in the region. By good fortune, some good armchair anthropology was done in the 1960s which made a fair start on the task of reconstructing the relevant past by sorting out documentary sources. The writers’ focus was on the changes which must have occurred in Nyakyusa institutions and worldview prior to the Wilsons’ thorough field study in the mid-thirties. The main work to consider is S. R. Charsley’s The Princes of Nyakyusa (1969). A critical reappraisal of some aspects of that work will be wanted when I come to evaluating the special character of Nyakyusa (and the related character of Kinga) religion. But on political organization the writer puts the right foot forward. Charsley makes two useful points about the way Nyakyusa political society worked, and offers two interesting speculations about Nyakyusa ideas and motives, on which some further reflection is surely warranted. About the society, he argues that a closer look at political succession shows Nyakyusa had no ‘chiefdoms’ in the usual anthropological sense of that term. To make the point, he avoids the term ‘chief’ in reference to their leaders, favouring ‘prince’ instead, as I had earlier done (1966) for somewhat the same reasons in dealing with Kinga rulers. While the two societies differ politically in two quite important ways, they show striking ‘family resemblances’ bespeaking shared origins. The special attributes Kinga have but Nyakyusa lack are these: a single ‘high prince’ claiming precedence and tribute in a pyramidal protostate, not a shifting political mosaic; and firmly fixed territorial spheres linking citizenship to residence. Among Nyakyusa, even the most successful ruler enjoys no sovereignty over a ‘country’. Entrenched boundaries like those of a Sanga domain don’t exist. Even villages, 181

considered as fixed locations, hardly can be said to exist, as the ‘village’ consists not of its buildings or fields but its people. These ‘villages’, no less so if they are princely courts, are prone to translocation. When a chief loses a group of his villagers he loses a tie to the land they occupy. Solidarity is through peer-bonding while leadership is through personal influence (and in that limited sense, propinquity) but ‘territorial’ identity is weak. ‡ The reason for this level of voluntarism is apparent when we consider Charsley’s second point, touching on the systematic turbulence built into the Nyakyusa political system. The ‘Coming Out’ ceremony, so well described by Monica Wilson, had been offered as a ‘rite of succession’. What Charsley makes clear is that there was no true succession to office or handing over of a princely title. Instead it was a fresh torch or banner passed from a still dominant older age set to a younger one. The system accords better with ‘age set’ examples than with ‘succession to power’. There was a regular period of about thirty years between one ceremony and the next in any given chiefdom. What we had from the Wilsons was a ‘how it is done’ scenario—better yet, ‘how it is designed to be done’. For some institutions this may pass, but for this one we were left with puzzles about the empirical results. Reconstruction has to envision the crucial practical implications of any formal arrangements recorded. The impression of abdication by ‘an aging chief’ began to look less credible with each new look. The impression of diaspora proved chimerical, as the main thrust was expansion. A whole set of untried leaders had been thrown together in a restarted political arena laid out in make-believe space. Where were the new women and cattle to come from if not by infringing on old rights? To establish himself in this arena on a par with a celebrated father or some other great chief, a young prince would have to attract and hold a lively following of commoners of his own generation. Cattle raiding was one path to wealth and following, and must be carried out by small groups of peers from a single village. But the calling of the chief was to muster for war. This meant gaining command of (and through) the young leaders of such peer raiding. But the conditions set by the Coming Out (ubusooka) meant that wealth, following, and readiness must be built from below. Our information on the war pattern is not systematic but suggests that actual battles were fought not for territory or inflicting damage but for glory. The game presupposed rival forces would be evenly matched. Raiding would have to be done on a village level, presumably at distance and by night. What sort of tournaments would have matched one young chief against another. or a younger force against an older as between chiefdoms, we can only guess. It does seem that 182

a young chief’s claim to greatness would have to be established by mustering a force fit to match and impress an already established neighbour. Comparing this situation to the Kinga and most others we know, the Nyakyusa chiefdom only comes about through radical change in the local political map. The turbulence set in motion by the princes’ Coming Out must at some point reach the relatively steady state of a re-unified chiefdom. The next step, it seems, would be the final one in the priestly charade which Monica Wilson describes in her Frazer Lecture (1959b) on ‘divine kings’ among the Nyakyusa. The new claimant to charismatic powers would be formally anointed with a chrism made of the mortuary relics of a prior great one—relics in the possession of local priests well disposed to the new man. This reconstruction is the most economical I can offer, taking into account what is known. I think one quite general premise is sound and to the point. All the evidence Charsley draws on, and a great deal more he might have drawn from the same missionary literature, supports a ludic frame for understanding chiefly rivalries from Selya to the lakeshore. Little armies come and go on translocal military fishing expeditions. The most perilous game, a collective challenge to the German guns in 1897, lures a great coalition into an instant massacre. Is the Nyakyusa spirit broken? Neither broken nor properly humbled. The German leaders were told jokingly on the morrow that the commander’s white horse had been the prize most wanted. ‡‡ Consider in respect to the motivational context of the Coming Out that this was an amity-based, male age-peer social organization, and one deeply colored by the same antipolitan ethos we found in the Sanga courts. At the Coming Out no scepter and throne were on offer. There is no mention in the relevant texts even of a formal transfer of wealth on that day, excepting the provision of a very young ‘chief wife’ for each candidate. What was on offer, in the form of a distant prize to the winner and his following, was something of a spiritual nature. When the scramble was over, if things went well, the appropriate group of priests would be prepared to announce the consensual choice of a battle-proven winner, in whom the mystical renaissance of an all-new chiefdom was to be realized. The credibility of any such reading depends on the way it illuminates detail. One of the problems we are left with is accounting for the rival princes’ recruitment opportunities. As background to that we have to consider the social situation on the morrow of the Coming Out. There are first the youngest boys, living with their mothers in one chiefly and perhaps a dozen commoner villages still 183

ruled by a polygynous older generation. Then there are the adolescents and young men liberated by the Coming Out decrees. Most of these potential recruits would be bachelors still. Their prime concerns now would be getting their own food. Their staple grains and bananas should come from their own gardens, and for this they would need to start accumulating cattle and wives. That is to say, each new village, chiefly or commoner, would have to build its fortunes on an accumulating mix of inherited and self-made wealth in cattle and women. A Nyakyusa man of a new generation only stands to inherit when the last of his patrilateral uncles dies. Any of the young princes would need time running into decades, better than average luck, and (it is hard to doubt) patronage from powerful neighbours to succeed in drawing enough men of the new generation together to muster a force of chiefly proportion. Along the way he would of course have to win the charismatic reputation which is the reward of growing a spiritual python in your belly. The basis for this interpretation of the Coming Out lies in the size ratio between the ‘chiefs’ who were strong enough in 1890 to claim recognition by the colonial government and the scores of invisible ‘chiefs’ still claiming that status in the mid-1930s. In the Coming Out, considered as a charade, the specific ‘loyalty’ of each man in the new generation is assigned, giving each aspirant to high office a fair number of his own, and a fair handful of ‘commoner village heads’ each with his own putative following. It is an arrangement for Cloud Cuckoo land. In the real world a full-blown chief could put some 6000 men in the field. The explanation has to be a scramble not so much for individual recruits but for whole commoner villages. The time frame is set, in the first place, by external relations, since desirable cattle and women are on the move and ill-protected. Men who are engaged in converting a property-less bachelor existence into a semblance of wealth and security will want to make quick decisions about future alignments, and it is just here the importance of the celebrated ethic of ‘good company’ can be appreciated. In effect, what this means is that a fumbling chiefly candidate may have his own small ‘village’ crowd with him, while by the same token each commoner headman will either lose his crowd or be carried with them into new alignments in a rapidly shifting field. Arena politics selects for opportunistic values. No proper ethnographic observations of the Nyakyusa social and political system had been put on record before the Wilsons began their visits (1934-1938), and no actual Coming Out was monitored by them. The documentary fragments from early British and German observers, with which all students of the region must be familiar, are helpful but hardly compare (for example) with the fine 184

and coherent work of Colonel Nigmann on the Hehe. To reconstruct from the 1930s ethnography I think the first requirement is freeing yourself from what is sometimes disparagingly called ‘the social anthropological method’, that is, picturing for your readers just the society your informants pictured for you. Discovering how your friend perceives his world is a necessary step toward understanding him, but not a sufficient one. Charsley does a great service here: having digested Wilson’s work and noted the puzzles framed within it, he turns to some documents which bring us back to precolonial conditions, wanting to picture the society as the early German missionaries might have pictured it for us. Now the page is turned and we have to comprehend the world as these Christian evangelists from a particular European country and sect were prone to see it. But still there are further steps to be taken. If we can’t claim the object of ethnography is the discovery of simple truths, it is nonetheless disingenuous to pretend there can be no ‘objective object’ at all. What could a very wise observer, an owl on the rooftops in 1890, have learned about the way things worked? It is a question reconstructive ethnography can neither properly answer nor ignore. One of the methodological problems with this is obvious: we have only indirect evidence concerning the perceptiveness and disinterest of individual missionaries. We have to judge their judgements—and so find ourselves ‘juggling facts’. But by improving our understanding of the regional culture we can still move a step or two closer to the scene which confronted the first German visitors. Like Charsley, I shall want to reinterpret the evidence the Wilsons introduce. But while this may mean rejecting their construals, it can’t mean rejecting their evidence. By working within a Malatan regional culture I am taking as given the commonality of Malatan history. Beyond the linguistic evidence (Nurse 1988), the cultural affinities of kiNyakyusa and kiKinga speakers are quite clear. It is with ‘cultures’ as with ‘languages’: you broaden or narrow your focus with an eye to the particular set of problems you want to study. And for historical reconstruction, you broaden to include several longrelated ‘local cultures’ because that is the most likely way to gain depth—something one tries to get through historical triangulation. Here is Charsley’s own summary of his argument on the matter of the Nyakyusa Coming Out ceremony and its indirect relation to chiefly succession: The ceremony itself seems ... not to have involved the two young princes succeeding to an existing position of authority ... but to have

185

concerned the creation of new positions and the attachment to them of followings. The hierarchy of princely titles produced, each with its own following, was manned by incumbents whose varying authority, power, and wealth were ... determined in the course of a continuing competitive process, on which ultimately depended the survival of the titles themselves. ‡

Charsley’s ‘competition’ sounds as Darwinian as political. But any reconstruction of the political scene in the valley just before the pax would have to deal with a princely or chiefly figure more like a political entrepreneur than a titled aristocrat or member of a landed gentry. The main fact here is that the Nyakyusa chiefdom was a personal fiefdom not the kind of established polity which could be handed over intact, in face of the internal rivalries which ‘succession’ in its fullest sense inevitably would awaken. But care is especially wanted when we come to consider Charsley’s arguments about the controlling political and religious ideas and motives embedded in Nyakyusa institutions. I want to examine in particular the social situations of prince and priest. Charsley’s prince is, appropriately, portrayed rather as a man with rank and high expectations than as the hereditary chief and defender of a settled territory. But his prince is, less appropriately, bestowed with a status lacking any firm charge or mandate for action. This is perhaps the kind of ‘prince’ Tolstoy would quickly understand. The problem is, such a floating rank system wouldn’t have worked at all in the Nyakyusa circumstance. You are no sort of lord in a feudal order when you cannot meet your rivals in the field with men at arms who are safely your own. It is the pervasion of the Nyakyusa world with manly rivalry, condensed and redirected outward by the bed-sharing loyalties of the age-village, which most wants recognition in the Wilsons’ ethnography. Notes such as the following, in Monica Wilson’s discussion of ‘values’, are not strong enough: Most Nyakyusa would hold that in the old days quarrelsomeness within the village was bad, that towards members of another village of the same chiefdom it was allowable, and towards members of another chiefdom it was good. ‡

Charsley does better than Monica Wilson by the test of approximation-to-credibility because he sets aside the ‘structural’ frame in favour of situated action. But his picture gives us few hints as to the qualities of a man who would succeed in turning the challenge of being chosen at a Coming Out into the full status of a translocal ruler. The social situation of Charsley’s ‘priest’ is no better sketched and seems to have only a bit part in the play. I’m 186

inclined to give priests more credit. There are two reasons for thinking, as Charsley himself suggests in the letter quoted earlier, we can hope to draw closer to the pre-pax Nyakyusa experience. First, the institutions in question are older, stabler, and more recognizable when they are taken to represent the broad regional culture of the Corridor region than when they are grasped only on ‘insider’ evidence and so must lie ‘outside’ probative history. Second, I take it as fair game to interpret them in light cast by my own fieldwork in another part of the same culture area. This is not so much because I can pretend to see the Nyakyusa version of the regional culture through Kinga eyes as it is because I can claim to have read the relevant professional ethnography with such borrowed eyes. What I read into Monica Wilson’s work is the importance of chiefly rivalries to the maintenance of ‘paradise’, and the importance of chiefly witch-charisma in the motivational set which keeps the rivalries going. This society thrives on what may occasionally pass as ‘friendly rivalry’. But to get into this frame I find I have to fly my own armchair back in time and ask who these chiefs and priests were, who their followers, and who their clients. In Collingwood fashion, I have to re-enact their past experience. I try to reconstruct, with the materials I get in this way, the prevailing ‘structures of experience’ in that past time, distinguishing such from any ‘structure of consensus’ directly implied by informants’ recall. This means looking as closely as evidence allows at the moral strategies which would have suited the several situations of chiefs and headmen, priests, prophets, and warriors; and fleshing out the broader moral context of these mainly masculine political roles. Moral strategies are not to be taken as data but models fitted to the career paths the evidence presents as data. Such a model ‘explains’ a specimen data-set on the basis of fit. Abraham Kaplan characterized the ‘pattern model’ thus: Rather than saying we understand something when we have an explanation for it, the pattern model says that we have an explanation for something when we understand it. The pattern model does indeed talk about psychological matters, but how else is knowledge to be conceived? Human knowledge cannot be anything other than an acquisition of the human mind, the product of human acts of knowing. The critical question is always whether the mere think-so (or feel-so) makes a statement so, and in the pattern model this is emphatically denied. The pattern is not constituted by our seeing it, but has its locus in a network of objective relations. ‡

Chiefly prince or princely chief? 187

Now we have evidence enough to make sense of the moral career of a Nyakyusa ‘chief’ or ‘successful prince’ under precolonial conditions. Start with the figure of the young prince at the Coming Out. His project is unusual for a traditional society: he is expected to make himself a leader of men. ‘Leadership’ is a buzzword in capitalist kindergartens and schools of business. It is a word whose etic meaning applies to peoples like the Tiv or the Nuer, where anthropologists speak of ‘predatory expansion’; but its emic meaning among the psychologically tainted teachers of kindergarten children is foreign to Tiv or Nuer discourse. It would be wrong also in respect of the ‘great commoner’ or village headman of Selya, though his career is in good part made on the battlefield. What counted in Konde warfare as it has been described by Western observers was individual daring and skill with spear and shield. But the moral career of the chief as I find it is a quantum jump away from that of his village headmen, because it involves and crucially depends on actively recruiting a following of village outsiders. My evidence for this is inferential. If you try to model a world in which ten ‘mother chiefdoms’ produce twenty or thirty ‘successor chiefdoms’ by going through periodic mitoses once each generation you can only make the model work by supposing that the mainstream political process entails a massive project of amalgamation and re-sorting of personnel, capable of reducing the twenty or thirty to a figure (say, twelve) appropriate to a rate of population increase (natural plus predatory) you can expect your expert colleagues to believe. Beyond that, if you want your model to suggest a less-than-chaotic political system, compatible with the African ‘paradise’ the early missionaries found, you have to assume that the process of re-sorting and amalgamation, seen as realignments of individual men of fighting age, would be massively de-randomized. That means villages remain solidary wholes throughout, and that is a condition all the ethnography firmly supports. To understand a single episode in the collective life of any human community you need all the context you can get, and to improve a little on the basic arithmetic above, I suggest looking at ‘the structure of experience’ affecting a chiefly career in Selya. I conceive the structure of experience as a fundamental feature of social reality though not a feature an observer can easily grasp or present as ‘data’. The easiest way into the structure of Nyakyusa (or Kinga or our own) experience is through the world of common, everyday experience, about which a people will build a predominantly pragmatic-empirical discourse. The idea of a ‘social maze’ is often introduced to help with this kind of analysis, but it is important in doing so that we avoid thinking of a ‘maze of rules’ or ‘roles’. Better to think in concrete terms: the personal and psychological compo188

nents of a decisive situation, risk management, ploys and tactics, misunderstandings and revelations—not a proper scheme of pathways but a tangle: hard knocks, prods and pitfalls. Consider how individuals feel out the moral strategies which keep them in phase with the institutions they will confront and manage to work-through in the formative stages of life. The walls of this maze are met as events not rules—immediate menace, fleeting gratification, not the memorized turns-right and turns-left of a rat’s cardboard maze. Think of friendships, marital ties, rivalries. Think of envy, pride, and responsibilities met or shirked. The experience which is the signal feature of this kind of social learning structure is the experience I have called the individual moral career. For the typical biography of a man or woman we may for convenience make do with a generalized account of a life-cycle as expressing and giving context to an ethos/ eidos we have found characteristic of some cultural group. But that is to suggest that role-differentiation is hardly important except as it relates to such categories as gender and age. The young prince from the day of his Coming Out is working within a social maze requiring the would-be ruler to build a viable polity of his own, starting with a small team of age-mates with whom he has shared the life of a bachelor boys’ village for as long as a couple of decades. If this is not ‘starting from scratch’ it is starting with about the same handicap as any of the rival aspirants for chiefhood matched with him by a committee of sponsoring elders. A special Rift Valley condition for making this kind of political game feasible is an environment which can be converted from grazing commons to farming land by the application of just the kind of industry such a team of young men excels in. An informing feature of the Old Nyakyusa world is that the importance of cattle was more social and political than subsistence-economic. Emphasis could be shifted at will. Their land-use pattern was a combination of banana culture, hoe culture, and cattle ranching which left open space between one major political centre and another, so putting no special premium on longterm settlement at a given site. The open space left was still in the 1930s enough to sound paradoxical in the Wilsons’ description. Explanation is to be found in politics as much as intensified agriculture. If the unity of a mature chiefdom, once gained, had remained entrenched as in the highland Sanga realm, this spatial isolation from neighbours would not have been needed. But it was a mine of freedom. What we have is a group of (say) a dozen senior stem villages each with its developing system of branch age-villages. Politically, the league is centred on the two most senior of the stem villages, both belonging to the chief. Each satellite age-village 189

comprises a mix of non-kin male peers. They are boys or young men, for the most part without women. Autonomy is out of the question, as the food they produce goes to the stem village, where it will be cooked and served out by their fathers’ wives. Dependence on the stem village of polygynous elders is total. In effect, the peripheral age-village is equivalent as an organizational feature to the central barracks of a Sanga court; while the married senior men (those huddled together in a stem village around the Nyakyusa chief or commoner headman) are counterparts to the senior Kinga barracks men who have been given a wife and ‘sent out to rule’ peripheral settlements. The two systems have a logic in common but use opposite signs. Peter Weber (1998) ably develops the argument that the acephalous political culture of the autocthones or first settlers (abilema) of the Rift Valley floor, as they lived before the coming of chiefly politics, persisted as an explicit state of mind into the decades of the pax. He sets out to disclose an underside to the latter-day Konde political culture, obscured mainly in the well-known chiefly areas but still showing at the edges. He sees in his reconstruction of the pre-pax condition that the ‘chiefs’ would have been resented by the ‘people’ (abapina). Thus he entertains the idea that German missionary annals should be read as supporting this ‘popular resentment’ of a system of authority still not fully legitimate in rank and file opinion. The main case put in evidence is that of a peripatetic medium, Mbasi, discussed in Chapter Six below. But that case mainly shows me that chiefs, priests, and ‘doctors’ of whatever kind shared in the characteristically entrepreneurial spirit of these people. It is true enough, as the reader will see, that Mbasi’s mischief was a thorn in the sides of princes and priests alike. When we talk of ‘extra-processual events’ which are inherently likely to recur in the conditions of a traditional community we are doing so to warn ourselves how little a neatly ‘processual’ account has told us about the object of our study. In that spirit exactly, Peter Weber argues that in uNgonde, the peripheral principalities are headed by ‘owners of the land’—autochthones actually presiding as chiefs. This he understands as evidence that the coming of chiefly rule from uKinga carried with it the banishment of autochthonous rulers from the Nyakyusa heartlands but not throughout the whole Konde cultural sphere. The conservative Saku, who stand somewhat apart from the Nyakyusa of the Wilsons’ monographs, actually reported to Fülleborn that their chiefly intruders from uKinga ‘banished the abilema’. That is, the status of ‘owner of the land’ was extinguished in uSaku, which is the small Konde community (a Nyakyusa subchiefdom under the British) 190

close into the Livingstone escarpment and the lakeshore—the ethnic area through which Kinga normally make contact with the coastal markets. ‡ Apart from peripheral communities like the Saku, who regard the ubusooka Coming Out as a reenactment of this banishment, Nyakyusa regard the rite as reenacting the settlement of a wilderness—that is, land inhabited only by acephalous, shifting cultivators. This comes to much the same thing, as the true ‘owner of the land’ in olden times would have been a pop-up leader like the Pangwa mkoyo or the Bungu ‘first settler’—not someone who distinguished himself by lifestyle or coercive authority from others, only a spokesman, primus inter pares. But Weber’s point is that the metaphors used (banishment vs. wilderness) put a revolutionary twist on the Saku version, where the suggestion elsewhere is only of colonizing available bush. The kernel of his argument is the signature meaning of the Coming Out ritual as we have it from the ethnography: “Everything shall be new.” Under the Saku, the chiefly element in the Coming Out seems to celebrate (by reenactment) the ancient banishing of the key institution of a pre-chiefly culture. That is the ‘revolution’ in point, and it lends new meaning to ‘making everything new’. The chief starts out as a ‘young Turk’ and, with luck, grows to greatness. It is a gradual and toilsome accession to power. Weber insists that the main social consequence of the Coming Out ceremony is that piecemeal and irregular accession is replaced by a regular generational periodicity, by which a new race for power only starts when the majority populace of a consolidated chiefdom is ready for it. We therefore have a central Nyakyusa chiefly arena, which is set up for creating and recreating from each generation a few warlike, centripetal chiefdoms of maximal/optimal scale. At the edges, where the owner-of-the-land tradition is the (‘pop-up’) basis of leadership, generational succession is not practised, and the local polities are always small, with linear succession to office. In uKukwe Monica Wilson caught the Hardyesque view of the marginal dirt farmer: The Kukwe chiefs found us here...It is this they commemorate in the ‘coming out’ ritual. We, the owners of the soil, were always here, the chiefs...are like you Europeans who fight wars; we owners of the soil think only of hoeing. ‡

Attention is then due to the kind of social tensions which must underlie the systematic forms of so robust a political culture. Perhaps the whole Rift Valley Konde community partakes of that system; but the ‘processual’ institutions don’t own the people. The 191

social structure as you might have met it ‘on the ground’ is more robust by far than the social geometry ‘in the heads’ of establishment informants. ‡ ‡ An appealing assessment of the ‘position of the chiefs’ for which I find evidence in Peter Weber’s work is that they were always in competition not just with each other but more densely with leadership aspirants among the commoners excluded from chiefship. There would have been two classes. First, there were the fickle headmen of the commoner villages which greatly outnumbered the chiefly. The Coming Out multiplied the commoner villages anew, and led to the older village heads taking on the ‘consultant role’ of commoner priest. This important set of contenders thus moved as they aged from secular leadership to sacred. A chief’s career was radically dependent on their continued support early and late. The second class is the ‘extra-processual’ set not taking part in the mainline structures set up at a Coming Out. Peter Weber offers even from his fieldwork in the 1990s convincing evidence for at least two types: shrine-keepers not part of the priesthoods, chiefly or commoner; and peripatetic mediums sometimes capable, like the Mbasi medium, of making a living by their enterprise. The ludic dimension of Malatan battles didn’t mean they couldn’t be decisive, only that decision must happen episodically, as in a gambling game. This view puts the chiefly aspirants to translocal dominion in the thick of a turbulent political arena, itself operating within the background turbulence of an affluent populace of high morale, responsive to every sort of external prodding from nature, neighbour, and stranger. It is a picture I find implicit if (predictably) seldom explicit in the descriptive ethnographies of the pax period. Solidarity is the prime value stressed within the age-village, and solidarity is the main product of the residential system which segregates young men from their female age-mates and even their own brothers within each village of the chiefly league. Boys daily join with their peers in working the fields of their elders, from which their staple food derives. Young men may nightly lust after the young wives of their elders, whom they will eventually inherit, and from whom they may already have learned to covet the heterosexual life. As warriors, youths take every occasion to display and catch the fancy of young women of other stem villages in the league. But these are women they will have to buy or steal, not women within their own fold. Intimacy with their seniors’ wives is freighted with social danger—fuzzy ‘incest’ inhibitions which impose high secrecy but can easily also lend to heterosexuality an attractive mystique.

192

The main political lesson is that there is solidarity inherent in the (compound) village but not in the chiefdom, which must be considered a loose league of a dozen or more such villages. What unity is to develop in the new league will have to be worked out by a new political set. It is the mandate of more than one young prince ‘sent out to rule’ to form an effective power group from within that set. Carried to its logical end, an aspiring chief’s life task will be to fill the whole space, political and demographic, of the old chiefdom as it was at its height. As demographic increase is slow, this would mean sidelining a few villages and assuming leadership of the rest. It is a moral career which seems to have entailed outshining and outdoing, but only rarely killing off, one’s princely competition. We should consider the way space was politically organized. The territorial principle laid on by the colonial powers masks traditional arrangements. Old chiefdoms can perhaps better be thought of as congregations than realms. A successful chief seeks to an ancestral shrine marking an important predecessor’s grave. It is the focus of politically important rites on behalf of his people, all of whom are in this way linked to a unitized past. The political space of the league of villages comprising at any time a chief’s people articulates about that past. At this point it is useful to ask how a ‘normal’ succession could take place in a society so organized. The short answer is, it couldn’t. Each stem village has reached a huis clos from which it can’t proceed intact, and therefore must disband, freeing its younger members (a mass of children and their still fertile mothers) to find attachments elsewhere. As if the oligarchs were determined to aggravate this problem, which arises from the elders’ near-monopoly of women, it is ordained that fertile widows are all to be collected by a surviving brother, or in his absence by a son assuming the elder’s mantle. It means the redistribution of women and wealth is reserved to the last. But each of the dozen or more stem villages has been fostering, through the system of food production and distribution, a plurality (as many as five or six?) of branch villages, and it is the all-male populations of these age-villages who must ultimately be targeted by the women for the attachments they will need. We have then in the model we have set up some 50 to 70 independent bands of adolescent and young adult males who must form into some eight or more viable new stem villages. The men will have to equip all these villages anew with buildings, gardens, and wives. It will be a long and turbulent transition. For the most part there will at first be scores of small groups, young men or boys, collectively scrambling for themselves. From the original mapping laid down by the old oligarchs at Coming Out, some sort of consensual order will eventuate on the 193

general model of the chiefdom as it was ‘ever since the heroes brought fire’. It will continue to orbit about the same sacred places, and recreate the same institutions. But it will probably not accommodate more people, and the system as designed doesn’t offer answers to the question, where the putative surplus must go. Typically the number (n) of men forming the core of a new stem village will be no larger than before, though the male offspring of any single senior village would possibly amount to something rather more like n² after thirty years of protected polygyny. Assume normal rates of natural attrition and some deaths in fighting. You would still expect a surplus of males in relation to available places within a manageable new chiefdom. These are not the conditions of self containment and political stability. But the Corridor region is not an island, and the great civilization of which it is a small part had taken over the better part of an enormous continent in the two millennia before. All the cultural groups comprising the Eastern and Western Bantu civilizations operate on structural plans which, realistically evaluated, entail expansion; and all require chronic effort to achieve it. The special turn of the Nyakyusa is to put boys and youths to the main work of the fields, doing the gardens cooperatively and leaving the women (once their husbands are aged) to a secure productive and reproductive life of their own in the closelybuilt polygynous compounds of the stem village. It is a perfect change on the Kinga pattern. In the royal Sanga village the main work of the fields is done by a close group of protected women—royal wives and daughters—whose bachelor male age-mates are mostly herds and barracks men. An apparent consequence should be a slower rate of expansion and a decidedly laid-back approach to the great masculine scramble for women. We may see the Coming Out against this broad historical background as a form of ‘world renewal’ ceremony which works by releasing the energies of youth in a general scramble to rebuild and renew. The whole new generation—as individuals—gets to re-sort itself on a largely voluntaristic basis. This is a brief and possibly glorious ‘springtime of freedom’ which comes to everyone once in his or her lifetime. The formal institutional templates of the Nyakyusa protostate predestine an organic cycle of about thirty years. Each cycle begins in free scramble, proceeds through turbulence, then consolidation, and finally reaches a point where the existing polity can’t proceed farther. At that point the oligarchs have to recognize the generational tension they have sown from the time of the first expulsion of boys from the parental village. A new cycle is decreed, the political arena is—nominally—swept clean. But I think it fair to assume that a random walk-through of the whole Nyakyusa region in 194

1890 would have discovered little to suggest what stage of this thirty-year cycle any particular village represented. Neatness, industry, and good humour were specialties of the Nyakyusa, and that would have been the case right through what I have called their scramble. It is this order of circumstance we need to grasp when we read the early missionaries writing of a veritable paradise in this unique Rift Valley bottom, and the same facts make it understandable that an unusually fluid social and political system could evolve there. The makings of a paradise are to be seen in the Kropotkin-style cooperation of young men who take demonstrative pride in the orderly plan and neat appearance of their little hamlets. These are men without women, bachelors never divided by the competing needs of ordinary villagers in cultures where men marry earlier, having to provide from their own resources for growing families, and generally must see most of their neighbours as rivals in a game which some fair portion always stand to lose. The maleness of the branch age-hamlets is matched in the stem villages of elders by an enormous preponderance there of women. At its beginning such a stem village is indeed a men’s community dedicated to collecting women; in its endstate it is a collective seraglio—dedicated to the gender segregation of the young right through early adulthood, but without a way of its own into the future. In the end, the whole project of this village of solidary peers takes it into a huis clos. In the impossible plan of the ‘processual’ model, every Nyakyusa male, as part of a tidy little boys’ village, sets out on a moral career which ends this way. Our puzzle is to get behind the oversocialized design Nyakyusa men held in their heads to discover what went on in the world of experience. A strict taboo of the Nyakyusa was that the father of a married man shall never see the face of his son’s young wife: that means the father, though he continue in a proprietary relationship to the son, can’t freely participate in affairs of the young household or growing village. While Monica Wilson treats the taboo as (subjectively) ‘incest’-like, its main social effect (following from the custom that the son will not marry until his own generation ‘comes out’) was political. In consequence of the rule, the elders who had always dominated a village no longer do so in the personal worlds of their one-time so loyal sons. In the new village old men are made to know they are intruders. The panicky dumb-show of avoidance means they belong to a dying seraglio, from which all still-marriageable women will soon enough be liberated by a welcome widowhood. It is a nice lesson in the pace and imbalance entailed in ‘social evolution’ that still in the 1950s Nyakyusa women were plagued by this strict taboo, having to fly for no good reason they knew, in a compulsory panic at the 195

appearance of a father-in-law. The rule had already become deeply dysfunctional, at least for young wives. Doubtless, it pleased old men and helped to slow down the pace of structural change. Though under the colonial pax (since 1926 under British Indirect Rule) the old political system had been kept alive, it was effectively castrated. Ironically, what this meant was loss of the countervailing power a young warrior generation had enjoyed vis-à-vis their elders, whose wealth and women had no other protection. Now under the pax the seniors might keep all the cattle and women with very little of the trouble. ‡‡ The autonomy of the commoner villages within the chiefdom had been a key order-making feature amid the many diversions of the Nyakyusa political arena, but had ceased to be vital by the 1930s. Commoner villages replicated the form and substance of the chief’s own, and their rulers would have to muster virtually the whole of the fighting force a chief could command. In this situation he always faced the challenge of leading these peers from personal strength. But the pax scotched this. Chiefs and headmen were now left with no true political, only administrative functions. But this was a new, and foreign, distinction. The balance in the old system had been morally sanctioned solidarity within the (two) chiefly villages and voluntary alliance of the (two sets of) autonomous commoner villages comprising a newly chartered chiefdom. Since at the Coming Out two or more such (compound) leagues were announced, each must have its (two) flagvillages. The legal fiction of aristocratic birth-right was a necessary card to be played by the old oligarchs to single out just a very few ‘royal’ heads flagged as the chosen ones of destiny, all others deemed eligible to lead being perforce commoners. Royal entitlement at the same time had to be firmly extinguished in any remaining (rejected?) ‘royal’ sons, though with a proviso: while they could neither qualify to lead a commoner village, sire one, or challenge any incumbent leader, they seem to have remained in a reserve position, should the priests need to produce a royal replacement. Such a contingency might arise as well for commoners when a stem-village elder died with many young or aging wives and no brother to replace him. To save the village from sudden loss of many households and the grown sons who tilled their gardens, the priests could legitimate succession to the elder’s social identity by a chosen son. We have to suppose there were many more slip-joints which allowed for gradual consolidation and for pragmatic political moves. Two critical provisions of the protostate constitution appear at this point. First there is the independence of the hereditary 196

section of the priesthood from the remainder of a ruling oligarchy. It is these priests who will (in the ‘divine king’ scenario) finally smother an old chief and bequeath his charismatic powers and identity to a single successor—presumably, the by-then foremost of the ‘royal’ princes. The powers of these priests would not be diminished but magnified during the interregnum in the political cycle, and it is hard to suppose they didn’t play their few cards close to their chests. What else are sanctuaries for than lending an air of destiny to hard decisions? The second critical provision of the charter laid down in the theatrics of the Coming Out concerns the nature of the league of villages comprising a chiefdom, and specifically the autonomy of the commoner village. This can be seen as a necessary feature of the old political culture. If commoner headmen were not ultimately free agents, able to speak for the legitimate loyalties of their villagers, the arbitrary realignment imposed by the old oligarchs would have been written in blood on stone. Two warring camps would have faced one another. The scramble then could only turn into a true war of succession, renewed every thirty years. It need not have achieved the famously fratricidal pattern of the Banyankole to the north, as their ‘succession wars’ concerned lucrative proprietorships in limited supply. Nyakyusa were not herders bound to a fixed network of client farmers. But fratricide, it is known, won’t do in Paradise. The actual political process by which a sole-successor is selected, as I here model it, is a bit more like a bidding war than a game of havoc, and matches the facts we have from the British missionary MacKenzie concerning the Nyakyusa war pattern. As with Kinga there was a theatrical pattern for wars, and a piracy pattern for raids. This combination means that when a fight has the nominal object of taking booty it will be limited to the small party launching a secret attack, and only those on the other side who can scramble to defense. But when a ruler’s prestige is at stake an appointment for encounter must be agreed, the two sides muster at strength, and individual bravado (a sort of fencing and feinting at a spear-throw’s distance) will be the style. The winner of glory in this contest is soon known, when a single warrior on the opposing side falls, and peers call for truce. This, transported in space and modified in detail, is the dance-like pattern of two facing ranks I got from aging Kinga leaders in 1962. Two of them had known it at first hand as war leaders. For Nyakyusa even more than for Kinga, the material winning that comes with chiefly glory is the defection of (mostly mobile bachelor) men from the losing side. Charisma, in this reading of the evidence, shines only dimly until the one true succession has been established. But it’s unlikely that, given the 197

normal strength in numbers of warrior-ready young men in the typical Nyakyusa village, there would not regularly have been expansion (by intrusive colonization of bush communities) when one young leader saw that as his best opening. That Nyakyusa were better organized politically than their neighbours is hardly to be doubted on the evidence of Saku, Kukwe, Nyiha, or Ndali communities. The Nyakyusa is a variant on the distributional pattern typical for East African protostates: an extensive, active political arena harbouring a network of developed and autonomous polities, and fully separated from other such centres of chiefly politics by buffer belts of smaller peoples enjoying the rather different rewards of localism. Apart from the ‘disgruntled individual’ scenario, there would be two major sources for defectors from the following of an ineffective young prince. For one thing, there are younger boys’ villages caught without clear loyalties at the time of a Coming Out. They are not ready for autonomy as a village or for marriage, but soon will be. They and the boys still too young to move away from their mothers are initially left in continued dependency on the now-declining village of their elders. Assuming they continue as an age-village on their own at least as long as their mothers are not widowed and taken elsewhere, they must decide, individually or collectively, which ‘side’ of the growing chiefdom of their ‘elder brothers’ they will freely join as peers. The second main source of recruits to a charismatic prince in this ‘bidding war’ would be whole commoner villages of young peers willing to realign. At the start of the contest, they may owe little to the young prince chosen by their elders to claim their loyalty. True loyalties in a commoner village are lateral not vertical on the seniority grid: they are in principle for peers who have safely lain together for years and will likely move as a group. It is in this context we must understand the primus inter pares standing of the commoner village headman, and see the contrast to the relationship a common man enjoys with his chief. How strictly must the rules be kept, to justify a description of high office as ‘hereditary’? Kinga can’t be sure who their next ruler will be until the priests announce his installment. Yet from that point onward, the sacred drum having spoken, there is no doubt—until, perhaps, a generation later the son of one who should have been installed stands for the office on the ground of righting a past deviation from the true line. Kyelelo, who seized office by murder, started the only genuine succession war I could discover in Kinga oral history, and lived on in legend as an heroic deviant. In short, royalty is as royalty is done by. Monica Wilson came to treat her catalogue of ruling lines as virtually real genealogies. So for her a chief was born to chiefship, just as he was for her Nyakyusa informants. A python 198

quite predictably grew in his belly, giving him dignity, authority, and personal power. This is the ‘princely chief’. A born leader to whom the manners of an aristocrat are natural. I have presented a different argument, for a man born a little prince, who achieves through enterprise in adulthood the status and attributes of a chief and commander of men—the ‘chiefly prince’. There were, as Charsley noticed, scores of princes living here or there who couldn’t claim much of a following. Monica Wilson talks of perhaps a hundred ‘chiefdoms’, the great majority of no consequence. Semantics aside, I think Charsley’s mistake was assuming that aristocratic heredity had a lot of significance for status among ordinary villagers, where the evidence suggests that ‘bloodlines’ without the proven ability to take and hold political office bestowed a rather modest social advantage. Both Charsley and the Wilsons underestimated the importance of achieved standing among men, the political career. Greatness, which is a corollary of influence, could have been but rarely thrust upon a Nyakyusa prince. For Kinga that is not the case. A great Konde chief might have thirty wives or more; each might bear two boys who grew to adulthood—we have sixty young princes. Suppose forty of them are too young to be chosen as leaders at the Coming Out of their generation. Still there are twenty from whom two, three, or four might be selected by the old oligarchs. They are chosen for chiefly promise and sent out to settle the succession in the manner we have seen. The narratives, of course, are missing. But it is unreasonable, almost to the point of folly, to suppose everything always went ‘as it should’. If, and only if, it did so does the model of the ‘chiefly prince’ match that of the ‘princely chief’ with his symmetrical kinglists and the timeless stability of a Nyakyusa world true only to social memory. Sometimes, I suppose, a chiefly line in Selya may have lasted a century. That would mean one prince taking the lion’s share each time, and early enough to be in commanding position when the old chief died. Then things had indeed gone ‘as they should’, and the new chief was made by the divine chrism of the priests into an embodiment of the old. But continued repetition of such a nice outcome is hardly to be expected. Chiefly success itself, begetting demographic increase, must lead to schism. I think the true history of any realm in Selya, as found in 1893, would have been infinitely intertwined and criss-crossed with others. Had we a carefully-made dialect map from 1900, we’d have far better chance of reconstructing the internal history of Nyakyusa speaking communities than we do from oral testimony. Symmetry is a simpler, more direct product of the human imagination than history can be.

Arena politics— ‘African despotism’ held in check? 199

The constitution of the Nyakyusa protostate depends on the fluidity of local political structures. Fission within a chiefdom is controlled fission thanks to the generational cycle. Because the chiefdom can’t simply be rolled over at succession—because it has to reconstitute itself as if from scratch—the size is automatically limited to what a skilful new leader can manage. Paradoxically, the Coming Out guarantees a chiefdom can’t grow beyond its proper limits of political coherence. It is most instructive to compare this arrangement on the one hand to the more formally stable Sanga constitution (most developed in the Central realm at Ukwama), and on the other hand to the Sangu and Hehe despotisms of the late nineteenth century. Two useful methodological paradigms combine here. Fred Eggan’s ‘close comparison within a region’ promises the identification of observable institutional procedures and those discoverable social forms the several communities have in common. Asking how and why each community has its own version of the roles of prince and priest, and their relation, can only illuminate each case. This will lead naturally to further questions about the way the different ecological settings of these four peoples offer clues to the historical paths protostate evolution can take. A special feature of the Nyakyusa system is its fluid orientation to on-the-ground uses of political space. Chiefs are not anchored in place until their final moment, when they are dead and a sacred tree is planted to mark the grave and commemorate a successful career. This contrasts with the established Kinga prince, who is immured in a palace in life, but rots on the wind in death. The contrast is real, because the two peoples share a single regional cosmology. The Kinga chief or high prince is immured because he is and must continue to seem endowed with socially dangerous powers. In battle he is represented by a commoner hero wearing his otterskin crown and bearing his arms. Kinga princes who were famous for feats at arms were those who, like the occasional Nyakyusa prince who was able to gain legendary status, found or made an opportunity to build the Sanga protostate outward, involving peripheral peoples in the process. A Nyakyusa chief by contrast is nothing without personal mobility. He himself leads in battle and in the jural arena, and he knows Lwembe—the one ruler among them who is immured in a fixed and sacred place—as a secular wimp. ‘Pure Kinga’ (non-Mahanzi) elders of the Western realm presented their traitor-hero, Kyelelo the Cruel, as if cut from this same chiefly Nyakyusa cloth. Even the sacred groves most often visited by Nyakyusa are scattered over the land without reference to such natural phenomena as secluded springs and dense growth. These groves 200

contain ancestral shrines meant for a domestic market. Translocal concerns such as we attribute to accidents of nature are more firmly and strategically planted at known portals to the netherworld. The chiefly shrines are like those of Bena chiefs. They allow making a public benefit of the personal ancestral hearing of a particular public’s current chief. Each marks the burial place of an important ancestor of recent vintage, and consists at core of only a few trees planted at his death. The custom says something about the political use of lineal kinship rhetoric in an amity-based community, but not about territorial boundaries. A group of interested priests may perform a sacrifice there, addressing the full circle of chiefly ancestors through the ghost who guards the grave. A bullock may be begged from a reigning descendent, and the object of the rite will be release from whatever particular trouble royal ancestors are blamed for visiting on their living descendants. One thing this says concerns kin groups among aristocracy: if they don’t constitute real lineages, the priests will have to construct mythical lineal links back from today’s hero-chief to predecessors. Something else is said as well: if it weren’t for the third-party role and the earned credibility of the priests in this transaction, how would the people know that their secular rulers are so fraught with mystical danger? It is not a thesis one can usefully propound concerning one’s self unless one is indeed a tyrant prepared to daily demonstrate the size and power of his ego in the rankly unjust but unapologetic ways of politically constituted fear gone mad. In respect of this point it is notable that a Nyakyusa chief for all his immanent power is, like the Bena chief, not able to make in the public behalf the simple sacrifice to address an ancestor, which any man privately might do. These rites on the public behalf must be produced by a priestly team. This is in microcosm the way public responsibility is structured. Going over the narratives Monica Wilson records, it is especially noticeable that many chiefs are youngsters from a priest’s point of view—chosen as like as not for candidacy by some of the priests who will be arranging their sacrifice. The same tutelary role is to be seen among Kinga and glimpsed among Bena, but is most deeply structured-in by Nyakyusa. It will be recalled that Mwemutsi himself is never allowed to leave off sexual contact with women long enough to be clean for a ritual act of any sort at Ukwama or indeed for entering his own lineage’s sacred grove. We know from early documents that a young prince may cultivate the patronage of a strong neighbour of an older generation, and it is fair to speculate that the client will be expected to repay favours at going market rates. We also know that on average over the centuries the number of princely polities in the whole Nyakyusa201

speaking area (estimated in the thirties at about a hundred) must have expanded at a snail’s pace in relation to the potentially geometric expansion of the number of young aspirants to high office. That means an aspirant stands no more than a fifty percent chance of success; and if the scale of success would approximate a bell curve when graphed, notable success would be rare. Since notable princes were quickly drawn into whatever turbulence erupted within their purview, and seldom on the same side as near rivals, a princeship however successful was no sinecure. Still, a young prince willing to accept patronage rather than fight for fiefs of his own might well have had an easy time. We just can’t know how many played cool and how many hot. What is clear follows from knowing that a Coming Out called for a shake-up of the chiefdom as a whole. The burden of political action was conclusively divided among many: the new commoner village leaders would have been nearly as important as the prince challenged to gain their loyalties. At starting, the prince could be no more than first among equals. A Coming Out was a big dose of deregulation. If my modeling is right, peer-group ties were left intact while interpolitical ties—the web of loyalties aligning each commoner village head with a prince—required intensive practical affirmation. Political distance was introduced not just between elders and the young. The many peer villages of the young were also sorted into separate segments, distanced from each other. Political powercentres were set up which addressed relatively small communities of unmarried men and charged them each with taking over the human and other material resources of the old chiefdom to build one of their own. The easiest assumption, based on what we know of Nyakyusa character values, is that a great deal of wheeling and dealing would have been folded in with a good deal of tougher gamesmanship, and occasionally with some foolish and some quite adroit uses of force. What were the intents and expectations of reward which would have driven such a system? Bear in mind that the Ngonde moved from very much the same sort of open political system toward a closed theocracy, probably during the course of the nineteenth century in response to significant external pressures. This systemchange has to be understood in political terms as a pragmatic shift toward central direction. There would have been the same process of pulling together which each Nyakyusa chiefdom set in motion with its Coming Out, only now applied on the larger scale. What Ngonde found, in their position of special vulnerability to the subcontinental turbulence of that century, was that the highly segmented, loosely-

202

knit ‘Nyakyusa’ protostate template couldn’t stand up when the surrounding buffer belt of localist bush communities was broken. It is not quite so clear that the Kinga, who proved themselves throughout the nineteenth century perfectly capable of the same open rivalry and charismatic warlordism as the Nyakyusa, had been moved by their exposure to external raiding toward a measure of central direction. But especially in their Central realm the Sanga ascendancy had achieved at mid-century a quasi-sacerdotal organization not wholly unlike the Ngonde. In the same period, entrepreneurial politics was thriving in their Eastern and Western realms, neither of which appears from my evidence to have been nearly so well dug in as the Central and Northern courts. Arena politics had hardly disappeared even in those two more staid realms, but all the political centres were reasonably well fixed, and at least in the Central realm the hegemonic order between peripheral and central domains was regularly re-endorsed by a flow of tribute to Ukwama. Of course, it was part of the business of the priests in their processional visits to other courts to air the grievances of their princes and cool down the atmosphere of mutual blame. One way to identify the special quality of the Nyakyusa scene is to call it an extreme case of arena politics in a segmentary system. Another is to note the singularity of a system which uses an elaborate kinship fiction as the foil to—not the template of—its segmentary political structure. To these political considerations we shall later want to add consideration of religious worldviews. But when the question is what ideas and motives were behind the thirtyyear cycle of consolidation/devolution which Nyakyusa age-village arrangements reflect, I think the love of freedom was paramount. I have dilated at some length in earlier volumes on the importance of an antipolitan ethic for the Kinga and their region. Kinga men move out when their bachelor days are done from centre to periphery, claiming back the relative autonomy of boyhood, which they had exchanged for the pleasures of sophistication. Nyakyusa boys are early put to work in the fields with their fathers, but enjoy as Kinga do a close domestic life with peers. Then they are living no barracks life but in small groups and miniature houses of their own building. Nyakyusa as they developed politically kept this pattern in their own way. It appears that the small age-villages of youth could be quickly removed and rebuilt elsewhere, as circumstance required. Owing in part to bugs in the bamboo, Kinga huts are unlivable after twenty years. In a hotter and wetter valley micro climate the expectation is probably less, and local contamination by waste a greater problem. That makes frequent moving for the young age203

villager a pragmatic option. But we have to assume this was not the case late in a man’s life, when the Nyakyusa elder would be settled down with a crowd of wives and children in a closed circle of many another such crowd, all surrounded by byres, trees, and gardens. Mobility in youth, roots in old age. It was the same in essence with Kinga men, as all through their middle years as bachelor barracks men they enjoyed great personal mobility, being at home with their friends both everywhere and nowhere. The great difference to Nyakyusa arrangements was a relatively cool version of arena politics. Treachery, assassination, and back-room dealing were there, but on-stage every succession to office was clean, final, and designed to minimize the need of structural change. Armed clashes were always possible especially where a given domain bordered another linked to a different princely capital. but multilateral hostilities seem to have been rare. In the Konde context the age village pattern with its thirtyyear scramble was special to the Rift Valley where nature afforded any village great freedom of movement. While the Sanga ruler staged spectacular war games in the spirit of sport, the declining Nyakyusa chief prepared a clean-swept arena for his best young princes, there to reinvent war in a small way and play at the great game of statesmanship. It is not convenient here to think of the political arena in the carefully circumscribed way it was developed by Marc Swartz. We are after ‘arena’ in a fuller sense: where power pyramids are defined and timely issues championed by contests open to all comers: unseasoned upstarts and old guard, political entrepreneurs and established rulers—case-related litigants and otherwise interested intruders. An arena is the pre-eminent scene of contest, and in particular of ‘staged’ contest. Of course, arena politics may come in all degrees of intensity. But there are other words which better describe settings for palaver or executive decision making. When it comes to trials of strength through violent action, though, war is an extension of nasty diplomacy. Kinga lords were assuredly good at that, just as their priests were good at cooling them down. Is there a certain climate of familiarity and gamesmanship which softens the clashes and gives its mood to a particular ‘arena’? A broader set of comparisons would be wanted to decide that. But the Rift Valley floor would not be a bad place to start. ‡‡ With the Nyakyusa ethnography in mind it is most convenient to place plottings ‘back stage’. In the arena itself we then find two sorts of group activity: direct confrontation between (structurally opposed) autonomous groups, and ceremonial gatherings of 204

mutually dependent subgroups celebrating ‘solidarity’ and, by doing that, advancing their commitment to the politics of incorporation which binds them. Nyakyusa funerals are such gatherings. Back stage we can look for the ritual acts by which individuals affirm their own social and political identities and watch for tactical moves by their ritual stage managers. Every interior, every informal meetingplace in a village is a ‘back stage’ place of its own, and the scenes played out there can range from contacts leading to adulterous elopements or shifts in personal allegiance to the weighting and working out of new individual moral strategies or commitment to special group projects. Ethnographers unfortunately get to observe and describe little of this off-stage activity. Kinship itself, considered as a set of committed loyalties beyond the immediate family, is on stage for Nyakyusa mainly at funeral dances. Young men famously make eventful occasions of them, turning amorous or aggressive attentions, respectively, to the non-kin females and males among locals of their age set. Though it be only kinsmen of the deceased who make up the visiting parties, all are very likely from different villages. This is because brothers and half-brothers, who once hoed together for their father, seldom belonged to the same age-village. As adults they will have gone different ways. The instant solidarity they celebrate at a funeral allows, as well as frolic, renewal of bonds and mutual influence on career choices. What is at stake here may be clarification of goals, decisions on residence, plans for immediate set activities, or commitment to longterm projects. A young prince or headman will be looking for recruits. Kinship ties in any society entail maximal claim bonds. More than friendship or village membership, kinship relations are permanent even when they are not regularly reaffirmed, and may prove emotionally binding. When the bond is refreshed, as it may be at the funeral of a father or mother or uncle, social leverage is created which a brother or sister may wish to use. In this way kinship can balance the demands of peer-village amity, offering at least an easy escape to restive youth. While an ‘on-stage’/‘back-stage’ distinction helps to set the scene, at least by reminding an ethnographer how much transpires in a village which can escape observation, a lot of the relevant public action overflows understanding for lack of the very keys which escape in this manner. Some special puzzles left by Monica Wilson’s corpus relate to character traits. How is it that men who were fighting you last night can be so cheerful about it all in the morning? How easy is it, in spite of all the press for ‘good company’ among peers, to pack off to new friends elsewhere? Independent ways are what one knows quite well among Kinga boys, and are plainly not alien 205

to Nyakyusa ways. In particular there is the complex figure of the Nyakyusa chief, armed with wizardry, habitual bellicosity, and bornto-rule self-importance—something of a self-made man, something of a titan. The Kinga example, Kyelelo the Cruel, comes to mind: a patricide and more, a tyrant for certain—and the most popular hero of Kinga folklore. For that matter, much of the Nyakyusa blood shed by hightestosterone warrior-youths is shed in petty fights at funerals and taunting raids on a far-neighbour’s cattle. If I am not wrong, such excitements would be stirred up less for the fun than for exploring or hastening personal realignments. Fights nominally for spoils or amorous advantage would generally nurse a tactical political aim of ‘testing’ and ‘softening’ a rival group—sometimes even instigating a full-scale confrontation at arms. If you are apt to be stealing a man’s wife or his cattle you must first be on bad terms with him. Interchiefdom war itself, so carefully was it sometimes staged, could be performed in ceremonial style, at once ‘confrontational’ and ‘incorporative’, with the outcome settled by coup-counting rather than the harder victory of rout. War patterns are by their nature interpolitical to the extent the parties retain their structural opposition, and Kinga evidence confirms that ceremonial warfare was a regional not a localized pattern. Structural opposition within a region is of course an absolute condition for the rise of balanced (non-despotic) political superstructures, and it is the loss of that balance which was threatened by the intrusive changes accelerating in the broader Sowetan area after about 1830. The most nearly understood case of a military despotism set up in the Sowetan region in the nineteenth century is the Hehe, first under Munyigumba who died about 1879, then Mkwawa. Plains Sangu despotism began a little earlier, and later the Bena were involved but never truly organized under a centralizing despotism of their own. The Hehe story, as we have it, is well known and bears only quick rehearsal here. Raiding on caravans, defensive war, and coping with or obtaining guns were in the air, and the many-sided ethnic relations in the Iringa highland area (later uHehe) provided a huge arena for intergroup contest. The degree of chiefly political development in the area in 1800 is not known but need not have reached the level of translocal authority. Yet by all accounts the ascendancy of Munyigumba and his lieutenants was accomplished in a short span of months or years, and a fighting force put together on the promise of quick, massive attacks on far neighbours, the victors returning with proportionate private booty. The difference to any earlier war pattern native to the region was that a full share of the booty (cattle, women, children, tools, and weapons) would go directly to 206

the man who took it. The crown’s share was to be modest. ‘Market forces’, in short, were abetting this revolution. How directly the scheme was based on the Ngoni example can’t be known, but this kind of ‘horde’ behaviour doesn’t match the way neighbour peoples had governed their relations. A full history of ethnic mobilization in inland Tanganyika in the nineteenth century will be long in coming. For the case of the Sowetan political archipelago it appears the ecological insulation of the Kinga and Nyakyusa lands were protective. These two peoples experienced political continuity with their own pasts right up to the pax. Ngonde fared much the same, retreating to a tighter political organization around a carefully kept sacerdotal throne and meeting the secular threats of the times by coordinated defense. But when we take the regional examples together, a rather simple conclusion almost spells itself out. Protostate development proceeds in a gradual, evolutionary manner under conditions of partial isolation from other protostates in similar stages of development. There was much learning from neighbour polities in the archipelago, but little need to match them. Each community followed its own historical path. But when isolation is not just ruptured but lost, we meet a form of contact which breaks continuity for the weaker side in war. War patterns by their nature spread with remarkable rapidity through a region of contact: if you can’t stand up to your neighbour, you must quickly learn his tricks if you will survive. Both Kinga and Nyakyusa were found at the pax in a stage of centripetal militarism. It was a vital part of the protostate process that they fight continuously within their own ranks, but in controlled ways. They thus remained safe against intrusions as long as they were strong enough to seal themselves off again after beating back a trial attack. In the Sowetan region the billiard-ball process went more or less this way: Ngoni barge in, beating about in all directions but never massively, and usually moving on after settling down for a season or two to recoup and regroup. Their war tactics and equipment were quickly picked up locally, and the Sangu under Merere, finding themselves safe from booty raids by Ngoni, themselves turned around 1860 to distance raiding within the region. In response, Munyigumba got together his massive ‘Hehe’ coalition and in the 1870s pushed Merere back. Eventually, Ngoni now regrouping south of uBena forced the Bena chiefs into coalition and alliance with the Hehe through a series of cattle wars. Raiders reached the Eastern realm of the Kinga but were repulsed, and no catastrophic discontinuities affected the Kinga until German times. Since Kinga smiths had perfected a two-handed stabbing/slashing spear, at least their 207

external war pattern seems to have been affected by their Ngoni contacts. But as a device calling for close combat to death, this sword must have been found useless in the internal war theatre where the prize was only collective glory. A single serious wound was enough in their ceremonial battles to take the prize away. For Nyakyusaland as well, the prior development of translocal politics allowed for the massing of formidable forces of defense—in short, for effective preparedness—against external threat. Looking then at the history of the whole political archipelago in the Sowetan region, it appears that Sangu, Hehe, and Bena were rushed into growth by external impact. This was due to their immature structural development at the crucial time of first contact with the Ngoni (and Arab) versions of a new and massively destructive form of warfare. From a stage of developed localism but still without a routinized system of internal war-preparedness, they were drawn directly into (or in the Bena case, toward) a despotic military state organization which enabled them to match the Ngoni raiders already so well organized. In this way of modeling the Sowetan scene, only the Malatan segment of the region had developed (by about the 1840s) a sophisticated system of translocal political authority. And while this authority was necessarily based in war, its destructive potential within its own cultural sphere was minimized for as long as the preferred system could be kept in balance. ‡ Have we discovered that arena politics is a necessary phase in protostate development of the kind the Malatan region used to boast? Have we shown that a prevalence of priests is wanted to tend to the stage-management which makes a political good of such quarrelsomeness as Malatan leaders displayed? What is exaggerated in the Nyakyusa case is not replicated in the Ngonde state as we know it, and Kinga history tends to flow in a rather less turbulent fashion at least in the better established Central and Northern realms. But a weak principle, at least, does emerge: multilateral rivalry does characterize all three cases, though I have not tried to argue here the relevance of this to Ngonde. If the detailed history of the forging together of a unified Hehe people under Munyigumba and again under Mkwawa were more perfectly known, is there doubt that the triangular play of one party against another for a third party’s gain would be there at each step of the way? Missionary reports from uKinga made much of sabre rattling and attitudinal pugnacity. Well into the pax there was rivalry and opportunistic politics in the scramble for colonial patronage, most particularly by the Kyelelo team in the Western realm. The Mahanzi and Northern realm Kinga who played a strong part in the Maji Maji massacre in the Eastern realm seem never to have offered 208

reparation or apology. These are simply not expectations in the context. The broad idea of ‘arena politics’ seems to apply, and the prevalence of priests is, I think, established, though Kinga can’t match in this their Valley neighbours. As with ‘red tape’ in a modern bureaucracy, ritually correct procedures, seen to by the avanyivaha in a call to arms, will effect a shedding off of moral responsibility for outcomes. Is the hand that throws the dice to blame for a bad fall? Even conspicuous success in this political arena is attributed to invisible pythons more often than to nobility of character.

209

FEAR & BLAME, CHAPTER SIX

A Malatan Cosmos

When the Berlin missionaries came to the country (1891) they had to choose one of the names for God which had currency among the natives, for their own proclamations. They chose the word Kyala. They could have chosen Mbassi. For according to Richter (1924:645) both words are equivalent, both are designations for sublime beings from ancient times. ‘They are presumably names of earlier tribal chiefs or outstanding heroes...’ It is unimportant whether Richter is right...for the orientation of Konde to Kyala is fundamentally distinct from their orientation to the shades of their ancestors, and the names which they generally still use for Kyala prove it. They call him (according to Mackenzie) Tenende = All-possessor, Nkulumuke = Immortal, Khata = Omnipresent, Mpeli = All-creator, etc. When they call on God at need they do it through the intercession of their ancestors. They pray to them directly, imploring them to pass on their requests to Kyala... God is the everlasting terror for Konde. That is why they pray so often, “Go far away from us, oh God, go to the Sangu, for your home is so broad and great!” This fact offers a shattering insight into the pagan condition: Angst. [Ludwig Weichert] ‡

The mischievous ‘medium’: a hard case I reconsider here a seemingly trivial but intriguing series of episodes from Selya in the 1890s. The scant facts became a matter of record when they got under the skin of an early German missionary. The point of stopping to look at this small tale (the ‘case of Mbasi’) is that it has entered our literature on Kinga/Nyakyusa religion as evidence for revising the main ethnographic accounts. I find the revisions worthy in turn of revision, but there is a healthy thrust in all the critical work toward reconstruction of precontact conditions. Our conception of subsequent conditions, as given in the Wilsons’ ethnography, can’t be trued without such a model of timezero before the pax made its changes. 210

It is no fault of the ethnographers in the 1930s that they focused on what they could discover by direct observation, or that interpretation was based on this well-grounded evidence. The noise of random events mustn’t distract from an account of common practice. Still, what is aberrant is often significant too. The reconstruction I’ll offer of the fragmentary ‘case’ as it can be extracted from missionary records is indebted to the original ethnography, the critics thereof, and in particular to Dr. Peter Weber, who has recently checked out Mbasi-survivals in the field. ‡ From a newly established missionary’s sometimes puzzled notes on encounters with a certain spokesman for ‘Mbasi’, a fragmentary story of divinity in action was brought to life and lively debate eighty years later—getting under the skin now of interested anthropologists and historians. On the archival record, I found the case uncomplicated. Discounting deeper ideas of the kind visiting academics were later to read into the evidence, what I learned from it was that self-taught gurus—entrepreneurial zealots—had a bit more prestige and leverage in Nyakyusa country than in Kinga. Here was a wandering ‘herdsman of Mbasi’ (traveling, it seems, with only a single acolyte to help with camp and campaign) able to alienate cattle and wives from rising young chiefs. The missionaries got involved when a ‘wronged’ chief got them to intercede for him. The great divinity Mbasi was, it appeared, angry at a young woman for deserting his herdsman and running back to her husband. The priests agreed to protect her. The herdsman himself was lawyerly enough (if quite a braggart) by day, but went about at night with a huge voice threatening mystical trouble if the woman were not returned, with perhaps an extra cow thrown in. There is no evidence the young seer had ever paid or meant to pay a suitable bridewealth. I saw no good reason for senior chiefs to support him, as some of them did, unless it was a matter of (a) patron-client politics or (b) fear of the leverage the young herdsman was getting over impressionable local publics. His trick was called ventriloquism in the records—loud tirades in a giant voice, always from the cover of darkness. Going over details of the case today I would vote for (a) + (b), but all of the principals remain veiled by the darkness of time. I promise not to read their minds. What I think I can do is use the case in much the way it has been used by colleagues as a ‘connect the dots’ game to produce a picture of Nyakyusa religion and politics at time-zero. My picture will not be theirs. I’ll get to premises later, but I want to begin with an appeal to common sense. The pretender to occult knowledge is a social type I had become familiar with in uKinga. In the several cases I knew the claimant was a youngish man of boastful manner and socially nomadic habits, rarely 211

taken as oracular by his elders. The young men I have in mind were would-be prophets in search of a calling. Since some individuals in the mid-century decades had been able to develop forms of divination good enough to earn a measure of fame and fortune, the calling was there. But Tunginiye was sure no practitioners of the occult unacceptable to the jury of court priests would ever have been tolerated. That the case was quite different in uNyakyusa became clear only from bits of documentary evidence I encountered in expanding my study from uKinga to the broader region. What was special about uNyakyusa was the disunited character of the ‘priesthood’—that choice of a name could itself be misleading, as the priests comprised no social entity. Still I was astonished to find Mbasi’s ventriloquist called in the revisionist literature a ‘priest’. I was also surprised, though not astonished, to find no recognition in the Wright narrative of the evident and persistent preference of the young chief’s young wife for staying home. If nothing else does, this should dissuade us from seeing this ‘Mbasi’ encampment as an established shrine or ‘cult centre’. Her side of the story, if we would dramatize it in the interesting manner missionaries and scholars have done for the masculine side of the case, I don’t doubt much would be disclosed. I don’t now think my initial take on ‘Mbasi’ was wrong. But it’s clear that historically for Nyakyusa, in absence of high courts and their centralizing dispensation, it would have been just such selfemployed diviners who might rise to the occasion as popular saviours in times of widespread trouble. The irony in this is telling: Mbasi the divinity is busy withering crops and magically killing healthy cattle; yet his herdsman, the only mortal he seems to love, is presented in the documents we have as some sort of hero. As it happened, troubles from abroad and from nature had been mounting up in the decades just before German contact, and epidemics, droughts, and pestilence were just hitting the region when the Germans were setting up their missions. It was natural for the missionaries to put themselves forward as ‘doctors’ in the root sense of the word, and so they were taken to be, by those who came to patronize them. Within weeks this led to some meddling in local affairs and ambivalence in their social situation. Enter Mbasi. Though he put the missionaries only briefly at risk, the ‘Mbasi affair’ made something plain which an observer ought to take seriously. The chiefs were in open competition with the ‘doctors’ in a great game of influence. The major tool was blame. If the missionaries would make themselves felt, they would join this game. When trouble was translocal and overflowed the ritually defined jurisdictions of chiefly ancestral shrines, the chiefs 212

themselves must turn to ‘free divinities’. A plague is no respecter of human boundaries. These spirit beings might be deemed ancestral, but only in a mythic and universalizing sense. As such they might logically be blamed for wholesale tribulation—disease, drought, pestilence. The German missionaries were not slow to put in their claims for a new and powerful free divinity. They had translators brought with them from Karonga (uNgonde). They picked ‘Kyala’ as the nearest word to Gott, and went right to the task of preaching the powers of Kyala. If this didn’t clarify their presence, it made them interesting, and important chiefs came to claim their patronage. ‡ For the Sanga courts, the relevant free divinities were, at home, the rainstones; and for ‘national’ concerns certain portal divinities, all to be found below the escarpment in spots which seemed to give mortal men access to an underworld. The main free divinity for the Kinga was Lwembe. Propitiation meant a near-yearly ceremonial act of histrionic humility in a sacred grove at Lubaga in the Rungwe district. In essence, the meaning of that rite was, “We— all of us—accept the blame for causing you to bring this trouble on ourselves.” But Lwembe (sometimes also known as Mbasi) was not hearing confessions. He is a chronically angry, accusative presence whose demands take the form of tantrums and must be met by symbolic acts of placation. What makes Lwembe, as the Kinga know him at Lubaga, especially accessible in times of trouble is the combination of a permanent location with the high seriousness of established ritual theatre. In my reading of the matter, the chiefly mind in uNyakyusa turns to Lwembe at Lubaga for the same reasons as the Kinga courts. To put it bluntly, the point is to pass the buck further up, without actually having to confess one’s own impotence. But that is the formula for the humblest private ancestral sacrifice. For Nyakyusa, Lwembe may have had an apical position among free divinities, but there were many contenders. Only the Kinga seem to have found him worth elaborate ceremony. ‡‡ But what Peter Weber makes clear is that chiefly authority, even where territorial jurisdiction was not the concern, always fell short in uNyakyusa of the universal relevance it might reasonably claim to have achieved in uKinga. The Mbasi we discover in the annals of the early missions represents the presence of a plurality of relatively free divinities whose persuasive powers could sometimes match in the political arena those of a great chief. The main difference, which always favoured the chiefs in a long run, was clout. ‡

213

To achieve conviction most Kinga consultant diviners depended on a display of objectivity. Only Hikadiseku (alive in the flesh in the 1930s) lived on in folklore as a proper spirit medium, and no one knew the learning source of her art. She said she carried ancestral voices—the whole ruling line of princes at Ukwama. Mediumship puts a turn on the subjective type of divination by acting out a dramatically effective dichotomy between the speaker’s ordinary presentation of self and the supposed divinity’s manner of speaking. It is not often independently invented and is not a regularly reported feature of the Malatan cultural region. In the Mbasi case, the spokesman for the divinity was a medium manqué. He didn’t present trance or other outward signs of inward (spiritual) transformation. This may seem to betray amateurism, but the routine was made effective by night-time rants in the numinous voice of Mbasi himself, and this may have been the routine of choice in those days for presenting this divinity. Mbasi was not new to the Corridor scene in 1892, nor was his ‘medium’ the first or last to speak for the divinity. We know nothing to guide conjecture about the herdsman’s recruitment. Self-recruitment would not surprise. The evidence is overwhelming that before missionaries singled out Mbasi as the main name of the main Shetani on the valley floor there were many more free divinities of varying stature, some with localized shrines and others more or less deracinated. Peter Weber maps thirteen ‘cult centers’ in uNyakyusa, each accredited as ancient through careful interviews in the 1990s, including one attributed to Mbasi. But the Mbasi the Germans had to deal with kept no local shrine. We have to keep in mind that each ‘cult centre’ would be steered by a mortal, fallible, and not always utterly convincing person. Mbasi’s spokesman in German times quite soon came to grief. The great survival advantage the free divinities enjoy derives from the longevity of a mythic name, and the fuzzy logic connecting that name to a face, a place, and a time. But as for shrines, like small businesses everywhere, they are not easy to pass on to a capable heir. It is not clear the Lwembe shrine, which was closed for business through the last decades of the British pax and only reopened after national independence, would have had preeminent standing in the Wilsons’ Selya, had the steady Kinga clientage not underwritten its credit. What is notable is the sensitivity of such a business to market forces. By the 1990s the trade in mystical revelations had enjoyed three decades of deregulation. ‡ Could anyone with the right talent aspire to producing the Mbasi oracle? The setup was not like that of a fixed cult centre to which a public seeks as if to a known healer. Here was a fly-by-night oracle on the move, opportunistically using or even stirring up 214

trouble for his own ends. Neither was the setup like prophecy, looking to what Fate has in store. This was an oracle offering to stay the hand of a vengeful Fate on a quid quo pro basis. Mbasi’s herdsman (a.k.a. Mwamafungubo) might have been a young man running a confidence game. When his warnings missed their mark he would not be around to take responsibility for failure. He had the advantage of the itinerant in escaping claims. I was not immediately impressed. For me, what eventually made the case significant was its mischievous career in scholarship. The revisionist story in its most inflated form seemed to demand we give up the whole sense of Nyakyusa culture as we had it from the particularly thorough fieldwork of Godfrey and Monica Wilson in the 1930s/1950s. Anyone familiar with that fieldwork could hardly be blamed for demurring. The evidence contra Wilson may possibly be taken as hard but certainly not hefty. Monica Wilson herself soon laid the matter to rest or to stalemate. But what hadn’t happened was a real expansion of our understanding of the way things must have worked for Nyakyusa in 1890. The ‘Mbasi case’ and the literature it provoked should have moved us toward a better sense for the developing protostate process among Nyakyusa, and the featuring in it of the (unfinished) cooptation of ‘bush religion’ to the advantage of the courts. Here was a clue to a continuing ‘process’ where we had had only a seemingly finished ‘structure’. The revisionist critique started out well but lost its way.‡ The position I am taking runs quite contrary to Monica Wilson’s most emphatic critic, Marcia Wright, who sees Lwembe and Mbasi as contemporary spokesmen for two ‘rival hero cults’. The implications are staggering. It is as if proselytizing ‘cult leaders’ were competing for congregations just as chiefs compete for warlike followings. Now I fear the devil has got into our house by the back door. Angry spirits, small-minded and abusive as is their style, are made into Titans. Wright’s fundamental premises about the lifestyle and beliefs of the Valley floor peoples are not mine. The historically linear picture which she develops on these premises is centred around tangential archival evidence which almost never can be thought penetrating, and can’t possibly be thought relevant and rich enough to confirm or disconfirm any far-reaching premise. Her retrospective time-scale (owing nothing to the ethnography) is cut short nicely to fit the disappearing paper trail. But as an exercise in collecting and connecting documentary evidence, her work is distinctly worth reading. It is comprehensive as a scholarly information source, the more so wherever it treats of missions and missionaries, not the peoples they ministered to. Where, as with Elise Kootz-Kretchmer in uSafwa, she has a penetrating intelligence to inform her of the world outside the mission station, her work is invaluable. So what sort of 215

counter-critique should I mount? I can begin by looking over the case she makes. Then comparing it with mine will lead into the matter of premises. I sample only a few pages to represent her version of events. I present her case under the following heads: (a) Shrines (b) Priests (c) Cults (d) Clientage (e) Narrative.

(a) Shrines. Wright would have this both ways. Mbasi has a shrine “only an hour away” from the new Mission station. But Mbasi is loose-footed: The Mbasi priest, who was not attached to any fixed ritual centre, then decamped to Lubaga. ‡

Wright consistently calls the self-styled (b) Priests. herdsman of Mbasi a priest. There are several objections. A priest is a functionary in some sort of stable social order. He belongs to the ‘processual’ not the ‘extra-processual’ sphere. A priest is a priest ‘in’ and ‘to’ some congregation or community, unlike a ‘prophet’ who has no fixed clientele. A priest is a trustee of property (such as Kasitile’s rainstones) related to his calling. He is a functionary to whom a case may be brought, not an entrepreneur who goes out to find paying customers. Even bonafide shrine keepers are best not categorized as priests. Mbasi’s spokesman fails to meet the criteria.

(c) Cults. In one sentence we have ‘the cult’ and in the next ‘the oracle’ referring, in each case, to the same object of popular skepticism or allegiance. If the latter term had been consistently adhered to we would escape being told of “the Berliners, who became engaged in a spiritual conflict with a cult of Mbasi in the neighbourhood.” Rejecting an oracle is a world away from fighting a cult. The time presence is other. An oracle is not a corporate group. What these passages and countless others in Wright’s two reports obtain to is a three-sided imbroglio over influence. The Mission has tried to quash an oracular herdsman whose veracity some chiefs have underwritten with loans of cattle and wives, while other chiefs prefer to magnify the claims of the missionaries. There is no evidence of a cult brought forward, if ‘cult’ means a group adhering to a cosmic persuasion identified with a dedicated divinity. In Georg Simmel’s terms, there is a rising potential for group conflict, but only amongst the chiefly polities, and in any case it is of the kind that ‘they need discord in order to preserve the relationship’. When the focal confrontation does occur between a mission and an oracle it is better described as a collision than a duel. Neither side could have 216

won on its own terms, both were lucky (or smart) to walk away from the scene. They didn’t after all speak each other’s language. So far as we know, the ‘cult’ disappeared with its spokesman later on, after getting into more trouble elsewhere in the valley. The persistence in lore of an Mbasi even a century later as a free divinity in the Konde cosmos was not compromised by the failure of an oracle using his name. But this is not saying an ‘Mbasi cult’ so persisted.‡‡

(d) Clientage. A system of clientage is implicit in all Marcia Wright’s accounts of ‘cults and politics’, but there is no evidence she understands its workings. None of us do or will know the details of 19th century patron-client relations, shifting as they clearly were, among established and aspiring princes (chiefs). But consider this bit of narrative: An armed party of Mwanjabara’s men came from the plains to force Mwaihojo to appease the god...

I set aside the careless use of the ‘god’ idea. What we know is that in the general confusion about the blame for a rinderpest epidemic ‘Mbasi’ was quick to claim responsibility, and a plains party was sent (armed as always) to investigate. We certainly don’t know that anyone was holding Mwaihojo or any other chief responsible for the epidemic. Mwanjabara’s lieutenant seems to have been passed over to the missionaries. There was a stand-off and withdrawal. What had happened? Clientage is a system by which direct conflicts between chiefs or headmen are turned to indirection—the play is to gain influence and clientele by discrediting a rival. Each player is thus bound to get into the game where the prize would magnify his political persona. Almost every narrative of chiefly alliance and conflict in Wright’s gloss assumes we can know how their system worked by reference to the way we like to play chess or war. But in this case we can’t even see the chess board. ‡

(e) Narrative. Historical narrative tries to serve as a roadmap in time, showing how people got from one documented condition to another. The usual ploy is to impute motive, as though history were a series of premeditated actions. Everyone understands the heuristic need for narrative, but everyone doesn’t condone ‘connecting the dots’ between tangentially known positions in a culture—a moral universe—which the writer is unable to describe. None of us is qualified with respect to pax-time Nyakyusa chiefs. Much as we owe to Charsley’s armchair initiative in taking up the puzzles in the Wilsons’ ethnography, it is regrettable that he had to ‘romance’ the task with a question-begging narrative voice. Marcia Wright only follows his cue in supposing she can reconstruct ‘Nyakyusa cults and 217

politics in the later nineteenth century’ from missionary notes and letters concerning their dealings with chiefs and oracles about whom half of what we know is their names. Here is a quick sample of Wright’s narrative voice: For their part the resistant princes now turned to the heroic tradition to legitimate and consecrate their resistance. For a prince like Mwanjabala, who had once supported the Mbasi priests, there was now no alternative but to support the Lwembe...It is unknown what role Lwembe himself played in the ceremony, or whether anyone represented the German enemy as a pestilence, seizing cattle and disturbing the people...The princes were under great pressure of public opinion, and the whole style of religion and politics was in crisis. ‡

There is in fact a great deal that is unknown. Scraps have been organized into coherent narrative. Unfortunately, coherence as such is not a sterling virtue in historiography. What isn’t unknown about the Living Lwembe is that he was not the star player she portrays in the political arena. As for ‘the pressure of public opinion’ I find that a particularly obtuse way to explain that, five years further on into this occupation of their country by hubristic aliens, responsible princes would finally turn to confront their enemy. I say this, of course, only to lend weight to my distrust of narrative and its rhetoric. But I am not fond of fighting rhetoric with rhetoric. Ethnographic reconstruction is a non-narrative approach to historiography, and I prefer it. It tries not to be seamless or didactic, its aim is like the fairy-tale carver in wood to unveil and free what was already there. It is my general premise that the religious style of traditional uNyakyusa was radically pluralistic and had never been organized around congregational ‘hero cults’. (Are there cults without congregation?) The only Nyakyusa congregations were those loosely organized about a chiefly burial shrine. All members of the relevant chiefdom, including immigrants, looked to the ancestral figure buried in the chiefdom’s sacred grove as the invigilator of their destiny just as if they were true descendants. Such was the power of peer pressure then. Regular communal rituals comprised the separate theatrical spectacles of each such congregation. But to ground my argument firmly in the ethnography a closer review of evidence is wanted. As Charsley used ‘Mbasi’ to investigate the chiefly dramatis personae in precolonial uNyakyusa, I would use Mbasi the ‘free divinity’ to investigate the social ground on which the pioneering of political authority had to find a footing.

218

The story of ‘the mischievous medium’ with very careful retelling and editorial spin might let us emerge with an acceptable sense for Nyakyusa history, building from tactical deconstruction of old and new models. I promise only a little more of that. My point of departure is the premise I take away from ‘Mbasi’ that Nyakyusa chiefs faced some loss of their legitimacy whenever they were faced with ‘territorial’ portents. Their only means for keeping face was the military one. The same would not have been the case in remote prechiefly times for the localist regimes headed by hereditary pop-up leaders whose ritual prowess, allowing access to parental shades, was home-grown and not translocal. Arena politics had not been invented. Territorial shrines, like Lwembe’s at Lubaga, grew up or moved into the Corridor with newcomers in a pluralistic fashion, matching the needs of people spread wide in an extended pedestrian area. To survive, the self-employed doctors must be many not few, because it takes time and chance as much as talent to see the end to a disease or a natural disaster. Mwamafungubo is an example of a doctor who overplays his hand. He is very smart but not cagey enough in the playing. It was the genius of Eastern Bantu cosmologies to recruit the victims of any standing ecological threat to the game of finding its cause and propitiating the free divinity who might confess to being at mischief. This is a common-sense reading of the pragmatics of a religion of blame. It is not the religious reality Kasitile (in Selya) or Tunginiye (in Ukwama) would teach us. But I think it reveals something that has always been veiled even from the foremost Malatan intellectuals, yet always was there. By juxtaposing Kinga (Sanga) with Konde (Nyakyusa) we have one system which has sacralized agriculture and centred its mysteries in the high courts, and another system whose chiefs are focused in war or keeping the peace through the resolution of private trouble cases, and remains radically dependent on apolitical agencies. When the two parties meet in the Lwembe grove at Lubaga, the Kinga priests are facing a far more powerful divinity than the Nyakyusa, its familiars, who treat it as one among many and possibly not infallible. When a Kinga priest at a Lubaga gathering told his Nyakyusa colleagues, ‘you are all witches’, I think he was aware of the multidimensional geometry of trust in uNkyakyusa and comparing it to a more simply triangular logic of trust in uKinga, where only a witch would question the court’s authority. Mbasi can represent, I surmise, the Konde voices Kinga would not have been attuned to hear on their visits in the 1890s, but knew of distantly. The title early European accounts used for all Nyakyusaspeakers—‘the Konde’—is especially appropriate in this review as I move to addressing cosmology, a subject only loosely tied to place; 219

and since, in turning to reconstructive interpretation, I need a broader reach: Selya has to be seen in the fuller context of a regional history. So I include the whole population of which Monica Wilson’s Selya chiefdoms comprise only one especially well studied part. It is important we have in mind the pluralistic identity of ‘Nyakyusa’ males as they are still portrayed by the Wilsons well after the pax. Bearing arms ‘for the Nyakyusa’ happened rarely. A man’s primary political identity was with his peer village and (most often through its commoner headman) with a chief able to keep his respect. Fighting was ordinarily against some (if perhaps beside other) Nyakyusa chiefs. The pax made a system look stable, which earlier had certainly been guaranteed turbulence. The deployment of force had become the organizing framework of Konde society long before the Ngoni and then the Sangu and the Germans came. But unlike the Kinga, the Konde protostate process outside uNgonde itself didn’t integrate religion with politics above the level of the chiefly segment. I have dealt with ‘the Malatan peoples’ as a regional culture including three distinguishable protostate processes. My main problematic has been the relationship between Kinga and (Selya) Nyakyusa. We are in position to consider these two as linked participants in the evolution of translocal social systems. The Selya evidence is the best I have for judging the maturity of the protostate process among the Sangas’ Rift Valley neighbours. Identifying ‘Nyakyusa’ narrowly with ‘Selya’ and shifting to ‘Konde’ offers a panoramic focus even while mooting automatic falsification of points I might make where ‘the ethnographic evidence’ of Selya is ambiguous or seems to conflict. Broadening the scope of detailed comparison to include Ungonde—Konde south of the Songwe River— if that were feasible, would bring in details of a splinter Konde community which had already drawn away politically from the plains Nyakyusa to form a compact pyramidal state. Using the available Ngonde evidence would allow concrete measures of difference within the Konde community. I’ve had to forgo that project as a step too far. But merely saying so much (and reminding a reader that all Nyakyusa-speakers were never claimed to be sufficiently described in the definitive ‘Nyakyusa’ monographs) makes it easier to focus on the dynamics of change in the Corridor region as it must have affected Kinga political awareness. Kinga, after all, were proudly the source of the great ‘Nguluve’ migrations which are supposed to have supplied the Konde with their royals. Konde communities have not all followed one historical track, nor had any of them unless it be Ungonde itself drawn hard boundaries and ‘settled into pattern’ before the colonial pax. The rest of the Konde lie athwart the Corridor, occupying its most attractive 220

ecological zone. All Konde speak one language with dialect variants, indicating a high rate of internal circulation. But it is essential to have in mind that, despite notable statelike developments, the Konde of the Corridor know very little of hard internal or external boundaries, both of which are salient, hard-won waypoints in political development. Konde boundaries are not lines between parcels of land but implicit lines between groups supposing themselves dependent on different chiefly ancestor shrines for their welfare. The Nyakyusa/ Ngonde difference is understandable as a function of political ecology: once Ungonde has become a fortress state, its politics changes accordingly. But formerly this was the part of the Malatan region least well isolated from external connection (Swahili/Arab ivory and slave traders) or telling collision (Bemba, Ngoni). As for the other (Corridor) Konde communities, since they were many chiefdoms they had the Hydra’s advantage with respect to defense: cut one down and two pop up. But what should be understood is that this kind of chiefly organization is inherently instable. It must be such that large forces can be massed at need, yet be a system in which massive force readily breaks down into politically autonomous, self-supporting, but potentially warring parts. The picture we have from the Wilsons can mislead a reader accustomed to the ‘one-culture = one-system = one polity’ format toward which ethnographic reporting used to lean. If as a practical matter anthropologists must make free to concentrate fieldwork on one segment of a grander historic whole, and if we will picture this segment as a self-containing, self-stabilizing ‘structure’, the logic of science asks us at least to note what we have cut away from the specimen we claim to describe. When a human community involves itself in the kind of systemic growth-and-transformation I call a ‘protostate process’, translocal relations begin to supplement and replace local praxis. For Konde, intermixture was accelerated owing to the delayed marriage age for men, which left males free to move and breach whatever casual borders might be forming. This was combined with a high rate of polygyny for older men, which collected females from far corners into a single extended household in any one of the six or eight senior villages which might be found in each chiefly realm. The cattle-bridewealth system had supplanted marriage choice on the basis of kin-partnership or prior acquaintance. Konde culture became a living mosaic, coherent by reason of common language and ideas but far from homogenized. It goes without saying, though, that local informants (especially after a generation of the pax) would unlikely picture the matter so. It has not even been possible on the records to be sure just which Konde communities actually would have had the signature Nyakyusa age-village system, with its climactic institution the ubusooka, in 1890. 221

Is oral history not the way to resolve questions of this kind? Interviewers encounter the same spectator paradox a reporter meets in reconstructing a street accident. It is not simply that some observers misremember, but that each of them has seen only what he or she was prepared to by nature, situation, and experience. Human wisdom is a function of the number and kind of private intellectual matrices a person has for assimilating novelty. When it comes to hearsay history, the differential preparation of individuals is hard to exaggerate. Rule One in the interpretation of turbulent cultures might be looking for laminations. That is, digging for more in any individual’s testimony on the chance of finding what is not just new but is revealingly incongruent with more superficial information. If you are prepared to think of ‘thick description’ as the essence of good history, you will know that even the best informant can’t supply what scholarship needs. Even a Kasitile (Konde) or a Tunginiye (Kinga) is no scholar from Mars. The ‘case of Mbasi’ well illustrates the difficulty of trusting information as if it were knowledge. The story of ‘Mbasi’ as it has been represented in the professional literature is wanting in the kind of interpretive key which turns chronicle into history, narrative into telling social theatre. But the play bears clearly on two issues: the expansive quotient which we should read from the periodical ‘coming out’ ceremonies—the nut Charsley set out to crack—and the importance of theistic patterns in Konde religion. These are both significant ethnographic questions where our aim is constructing a model of ‘unspoiled’ Konde life and thought. So far in the critical literature, differences in premise and emphasis have led to incongruous conclusions. Two matters of principle need attention: (1) Were the Nyakyusa in a phase of ‘predatory expansion’ in their Rift Valley habitat in the decades before the European takeover? The ethnographers seem to say yes, an historian says no. My premise is that Konde culture was clearly expanding, but that the mechanism is more like recruitment than predation—assimilation or ‘internal colonialism’ is to be distinguished from geopolitical domination. There was no ethnic stratification entailed, though the semblance of it may appear in the aristocratic claims of the princely estate from which chiefs are drawn. The telling point of insisting on an aristocratic privilege to take over the highest offices of secular authority was to limit competition for such offices and allay popular mistrust of men assigned to them. As much of this has been settled in previous chapters, it remains only to adumbrate some exegetic implications. I think the most important is that Konde culture was organized around political mobilization. I have suggested that political ecology offers explanation. To make that more explicit, I premise my 222

reconstruction on the view that we can’t comprehend the meaning of Konde village relations within a chiefly realm without relating it first to the ‘loneliness at the top’ which is created in the ubusooka or Coming Out ceremony, and perceiving the same effect only intensified in the division of any one chief’s realm into ‘his own two villages’ and a greater number of ‘commoner’ villages each with an independent head. What I find especially significant is the clean separation of the ‘Nguluve aristocracy’ from which a chief must be chosen, and the ‘commoner’ population from which a lesser village headman must come. This is achieved by rather special rules for cooling off aristocratic blood and its claims, for men not chosen at their ubusooka to be one of the two or more chiefs of the next generation. They are barred from consideration as ‘commoner’ leaders. This is precisely the opposite strategy to the ‘Sanga factories’ of the Kinga protostate, where ersatz ‘Sanga’ braves are sent out to rule the court’s periphery. It seems to me pretty clear that a commoner village in Selya or the Plains is always in position to detach itself from its prescribed chief when he proves unlucky or unworthy. We have that evidence from the ‘Mbasi case’: the young chief Mwakatungile who has made himself a client to the Germans is shunned by his own commoner headmen. How else explain the taboo on kinsmen of the chief aspiring to head a subsidiary village? Any contrary construction of the Nyakyusa constitution would have to account for the special fit this model has to the facts the Wilsons offer. The persistence of aristocratic blood lines in a system which actually picks its leaders on a merit basis does beg for explanation. So does the Nyakyusa proclivity for breaking up a chiefly realm when it seems to be at its apex. My explanation is that chiefly charisma in a wilfully turbulent social system can’t simply be passed on through priestly ritual but must be rewon with each generation by moving the actual person in high office—in the mobilizer role for set group activity—closer to the rank and file. A commoner man never comes to his chief for judgement unless brought there by his own village headman. Those who do come directly to the ultimate court are the chief’s own villagers. The chief has, with two villages, the extra fighting force to predominate in military exercise. Aligned separately with each of these two chiefly villages at the Coming Out is a short string of commoner villages each representing one (opposed) side. This is a setup for nicely controlled inter-village rivalries and for strategic recombination over time. Such changes proceed organically, not according to schematic plans: ‘pre-pax’ does mean ‘no pax’ and no steady state.

223

I have no doubt that in precontact conditions, when the elders of a chiefdom decided on a Coming Out of fledgling chiefs, their hands had often been forced. If there were peaceful years there would be turbulent times as well. Political mobilization was the key to success at arms, and in times of drought, epidemic, and pestilence dissension could prove an aging lion impotent. Political theatre empowers its star actors when it succeeds, and opens them to scorn or ridicule when it fails. The main legacy ‘Mbasi’ leaves us is insight into a unique African political model. The Coming Out on careful examination turns out to be the key institution in a strongly democratic system for succession to power. It is ‘democratic’ in that it provides for a heritor to power who is proven in an open political arena from a properly nominated set of candidates, and proven not by promissory rhetoric but a history of leadership. The system is doubly ‘democratic’ in that the arena of proof is organized around an antipolitan ethic: unpopular means unsuccessful. (2) What was the place of the free divinity Mbasi in Konde cosmology? I use ‘free’ here always in loose parallel to Godfrey Lienhardt’s classic monograph on Dinka cosmology: some Dinka divinities are assigned to a clientele by birth, as ancestors are in the Sowetan region; others are unassigned (free) and broadly available, as are all the so-called ‘heroes’ of Malatan literature. The canonical ethnography portrays ‘Mbasi’ rather ambiguously as a legendary Hero name but without a distinctive legend of its own. The one ‘hero’ easily conflates with others unless you notice that this ‘Mbasi’ floats also in political space, correlating with no fixed abode either in geography or lore. Monica Wilson found that ‘Mbasi’ was a spare divinity name often used interchangeably with some others better localized in the Nyakyusa pantheon. A revisionist historian added only a touch: Mbasi was a Hero on a par with others (notably Lwembe and Kyala) who were only mistakenly thought to be more securely lodged in legend and myth. This then took on the corollary that the scruffy (un-chiefly) spokesman/herdsman for Mbasi was a proper priest and a challenger to the priests of Mbasi’s fellow free divinities. The ‘Mbasi’ episodes in the collective Nachlass of the German missions have thus been held to afford us a unique, neglected insight into the political as well as the religious life of the region in traditional times. If the argument doesn’t ring quite true, it does point to the importance of ephemeral events and movements in affecting the Konde eidos. The lesson of this is of course that turbulence was not restricted to secular matters in Konde life.

224

I find that a proper understanding of the Mbasi episodes requires us to credit the Konde community with greater dissensus, intellectual subtlety, and imaginative genius than the relevant literature has done. Also required is a sharpened sense for the processual dimension in the Konde political system as we have been able to observe it. This is a lesson I would apply broadly to our understanding of the civilization of the Malatan peoples and of the Sowetan region. If I picture the Sanga protostates as better stabilized than the Konde, the distinction I’m urging is not one of black and white. I suggest that in our reconstruction of this civilization as a free-standing historic achievement we should reconsider the balance of eidos and ethos, beliefs and values, in the emergence of translocal political structures as they may be seen in regional context. Cosmological ideas tend to reflect and reflect upon the political. In particular it is obvious that Konde and Kinga shades have their afterlives in an underworld of no great splendor—up on earth bad luck is to be blamed on their jealous resentment as often as on that of a living witch. Giving their offspring trouble is the shades’ way of demanding help from their former dependents. Certainly, no shade individually has anything like the awesome mystical power of a well-entrenched chief. Mystical and political beliefs are symbolically joined in the high office, whether for the Konde ruler or the Sanga. They are also joined privately at the base, where an insistent antipolitan ethic keeps alive that disillusioning doubt which a high court’s political theatre is supposed to allay. What I have found arresting in the Mbasi case is the ‘lion tamer’ heroism of the itinerant oracle. In face of the wanton devastation of rinderpest, it seems the popular wisdom was prepared to accept that cataclysmic vengeance was justified by the niggardly behaviour of one young chief in refusing to lend his wife to a certain doctor nominated (he said) to speak for the divinity. Two features are strange. The more usual form of an oracle is theatrically effective mediumship. But here the ‘theatre’ is hidden in the dark, and the ‘oracle’ disclaims having been there. Then there is the public scale of the trouble matched with the private and personal scale of the grievance. Must we, after all is said, leave our inquiry into the Konde and their cosmos with the impression of a preposterous credulity? I prefer to see the scenario as ‘extra-processual’. What is clear enough is that the translocal rumour mill was running at speed. Translocal awareness would have been active since the arrival of the Ngoni, the Sangu, and the slave and ivory hunters from mid-century onward. The coming of Germans in an astonishing ship has given rise to new rumours. In this case general apprehension is stirred anew by reports of cattle dying and, ‘They say Mbasi is angry’. This automatically suggests trouble on a big scale, because it goes back to the 225

mythlore everyone picks up in childhood. The person everyone turns to when trouble comes on a grand scale is their recognized chief. The young Mwakatungila, for his part, is personally involved in a confrontation with the Mbasi oracle, and finding his position backed by an obviously powerful patron in the Mission leader, decides to stand pat. All the other chiefs are committed to chiefly action, but only the oracle has a plan, and this is the one they cautiously begin to follow. Marcia Wright thought just here ‘the whole system of religion and politics was in crisis’. An askari might have thought so, but the Nyakyusa thought Mbasi or some other divinity was angry again. My own opinion is the security of the system depended entirely on its ability to absorb Angst on a scale matching that of the troubles in hand. Nothing the chiefs could have done would stop the epidemic. This was not their view but it was their situation, and they managed to muddle through without, for the most part, losing what matters in politics, their personal followings. The chief least likely to lose a follower in this system is, paradoxically, the chief who has them all with him ready to fight whatever comes—provided, I suppose, there are windmills and lion-tamers enough to keep them busy. ‡‡

Sifting premises I find the observer premise which biased the missionaries continues to bias even current analysis of Konde religion toward a form of establishmentarianism which can’t countenance extraprocessual events. Students of religion often complain that observers have tried to judge religious institutions from the outside. Thus we often used to hear of ancestor ‘worship’ where ‘placation’ would have been closer. And in the case I turn to here, sage observers have taken all Konde divinities to have been turned out by the same stamp. All, including the faux-incarnation of Mbasi in the 1890s in Selya, are ‘hero’ divinities. Two things become obvious: ‘Mbasi’ as the proper name of a recognized divinity is a part, if infirmly, of the Konde cosmological scheme—thus, ‘structural’; but the ‘Mbasi episode’ wants to be read as extraprocessual. We are witness to events not hallowed occasions, involvements not structures. Part of the problem is encased in that word, ‘hero’: it strains logic. It makes poor sense unless you suppose that Konde society was obsessively competitive in just that more-than-political sense the english word ‘hero’ wants to carry. But this supposition flies in face of the very human evidence the German missionary records and the canonical ethnography offer us. More than one missionary took ‘Mbasi’ as the local equivalent of Swahili ‘Shetani’, while a ‘hero’ in Monica Wilson’s book amounts to no more than a divinity (spirit) 226

about whom there is a definitive myth as a ‘founding father’. A source of confusion lies in her treatment of ‘Mbasi’ as a hero name— for Mbasi in her texts lacks a myth (and so an identity?) unambiguously his own. Certainly she nowhere intends to assert any of her ‘hero’ figures are knowable from legendary days as actual leaders of men or founders of ruling dynasties. They belong not to history but cosmology. Would the fearfulness of a chiefly ancestor enhance a chief’s worldly power? Only in this sense: even a chief can pass the buck upward. There are at least three reasons for questioning the historicity of the ‘Mbasi’ narrative as we have it. One is simply that, though documented, it remains anecdotal and elliptical. This makes a superficial, connect-the-dots reading too easy. Another consideration is that the original, newly settled European tellers of the tale witnessed everything through the heavy screen of translation, and admit achieving little real insight into the events they recount. Fülleborn, who in his syncretic ethnography (1906) produced a grand-scaled account of the region, finds much to discount in his missionary sources. But the third cause for reservation is one which makes rethinking field evidence vital. With or without Mbasi, the Willsons’ canonical monographs are based on hallowed but arguably untenable premises. The ‘Mbasi case’ opens that door for us. Let me at this point retell the tale of the Mischievous Medium in my own narrative style. The tale concerns a youngish man named Mwamafungubo who moves about the country in an enterprising manner collecting a one-man (or was it a two-man?) string of cattle and wives in the name of Mbasi, a renowned but slippery divinity. The missionaries, who interviewed a mute ‘doctor’ through his (unnamed) companion and mouthpiece, took Mwamafungubo as a fraud and would have discredited him, but this “herdsman of Mbasi” proved too sure-footed. His ploy was to blame local troubles, real or fancied, on the anger of Mbasi, boast of his own omnific powers to mediate, then use the leverage gained to collect women and cattle to supply his independent living. He was an itinerant seller of cures who stumbled onto something big. From our bits of information it seems he was always on the margin of losing credibility, but managed to escape by moving house (complete with cattle, companion, and women) to some far locality, there to renew the game as opportunity allowed. He was not a spirit medium, as he didn’t present the utterances of Mbasi as issuing from his own belly. There was no show of trance and possession, no conversation with a spirit-being through a visible man’s mediumship. His art was a simplified form of ventriloquism. The project was to stage an histrionic evocation of divine anger in the name of Mbasi. As a general thing, Konde of the time probably knew 227

Mbasi as a Trickster figure featured in casual folklore and sometimes in legend. Associated with no fixed address, Mbasi had no specific portal shrine, no rooting establishment able to continue over the generations. The divinity was presented in the 1890s as an unearthly voice (was there a gourd-shell megaphone?) hovering about in the depth of night, shouting out in the high style of moral indignation, claiming vast powers of public mischief and offering to relent only if certain rather trivial material demands were met. The source of the bellowing was seemingly uncanny and (meant to be) invisible—i.e. the ghostly divinity himself, unmediated. How far does this account oblige us to challenge Monica Wilson’s premises? My position is that we should hold more stubbornly than she does to the fear-and-blame motifs of Konde cosmology. The Mbasi case specifically shows us that the fierce mystic powers of the python lodged in the belly of a chief were always at risk of exposure as hollow rhetoric. Mwamafungubo comes to us as a representative of a social type which had to be well known on the Konde scene. This is the clever, traveling entrepreneur who, once exposed, we recognize as a trickster or confidence man. The ‘Mbasi’ impostor almost pulls off the trick. If he had, and the new mission had been overrun, German history in Kondeland would have begun with a massacre. Considering that, we should have had to say the mighty band of Nyakyusa chiefs had been proven fools or worse by a ragged diviner. Do the chiefs in this picture know that their authority is under challenge ‘from below’? I don’t think so because I think existential dread is the source of all the power in the picture though none of the principals would have agreed with me. The herdsman of Mbasi is convinced by the very success of his confidence tricks that it is Mbasi who speaks through him and turns the trick in his favour. The chief believes in the python whose presence in his belly is manifested in events as impressive to the chief as to his cowering clients or chosen enemies. A chief in this picture is as fully entrepreneurial as the diviner. Both were attracted to the Mission in the way any professional is attracted to fresh opportunity. When untoward natural catastrophes intervene (beginning with rinderpest) the presence of the Germans in some sort of rivalry with a local diviner gives sudden importance to his frightening claims that the reason for the disease is Mbasi’s anger. The mystical danger in this case is affecting cattle particularly in the plains area. Plains chiefs accordingly pick up the gauntlet. They, with their priests, are as always ready to serve the needs of their people—it goes with the office. These questions arise from the critical literature on Mbasi: Does the itinerant diviner in some effective fashion represent the interest of an important un-chiefly social segment or class? Is this a 228

crisis driven by inflamed public opinion? Is the Mbasi oracle treading on the toes of the Living Lwembe? The answers I find justified at this point are these: The Mbasi oracle has a special appeal to the un-chiefly but is driven by self interest. Public opinion is intermittently aroused by chiefs, priests, and independent diviners in pursuit of their respective trades. Each of these trades depends on the manipulation of public opinion. It is a one-dimensional view of this sophisticated society which would deny such pragmatic qualities to its leaders and professional practitioners. The Living Lwembe accepts clients at their call and deals with them by traditional tricks of his trade—he will doubtless be glad to hear the last of the Mbasi oracle. But Mbasi and Lwembe are often thought to be the same divinity. An entrepreneurial model fits the facts better than an ideological one. There are no ‘saviours’ in the picture. The divinity’s brief appearance on the stage of history follows a familiar Malatan formula in all but one detail. There is the standard presentation of vengeful anger, the claim of mystical power to punish by plague and pestilence. But the deity’s claims on his hearers don’t match the majesty of his anger. Beholding his demands you get no whiff of a lordly presence. The anger is big enough but the mountain shakes only to claim a mousey share of placation. This Mbasi seems to want no more than a particular woman, perhaps an additional cow, for the comfort of his ‘herdsman’, and to get that is willing to set vast plagues in action. Fortuitously, the ‘Mbasi episodes’ took place during first-contact years, rich in such troubles as rinderpest and locusts. The dark-of-night performances were plainly staged to give leverage to the faux medium himself to negotiate (through the mouthpiece) by day, consolidate winnings, and take off to another place before being caught in a showdown. Eventually, as it appears, the showdown came and this particular play-out of an ‘Mbasi cult’ fell apart. The missionaries were betrayed by their assumption that Konde speakers had a generic word (if not a proper noun) for God. In their wordlist ‘God’ was ‘Kyala’. But ‘Kyala’ was the name of one locally notable divinity, ‘Mbasi’ another. Rivals they might possibly be, but not (as the mission would have had it) actors on the field of Good and Evil. The missionaries imagined the Konde cosmos was a confined one, rooted in ignorance, and distortive of quotidian human experience. But the cosmos we must consider is no simpler than the missionary’s own. Here is Monica Wilson on the theatrics of Kyala and Mbasi:

229

...The numinous quality of the heroes is conceived as contamination, not holiness. Like the shades, the heroes must be ‘driven off’ lest men go mad, and it is when they have driven them away that men are belu—white, innocent, free of anger. Men should be belu but this is never cited as an attribute of the pagan Kyala. It is the terribleness, not the goodness or purity of God of which the Nyakyusa are aware. This is underlined by the fact that a madman who, ‘when he comes home rushes off’, and ‘does not see his fellows are human beings,’ and who ‘has a loose heart’ (i.e., is passionate and quarrelsome), or ‘whose heart curls up like a leaf in the sun, or turns upside down’, is called ‘Mbasi’ or is said to have been ‘caught hold of by Kyala’. ‡

Here now is an earlier snapshot of ‘Mbasi’, this time as some kind of assumed alias for ‘Lwembe’. The author is missionary Hübner, whose tenure at the Berliner Mission Station Bulongwa (in uKinga) was certainly a distinct success. The excerpt is from his daybook as published in 1902, and looks backward from a fairly well established vantage point as a successful pastor. Hübner has got ‘recent information’ from his Kinga trustees bearing on matters (our 'Mbasi affair') already a decade past. He may seem confused to us, as indeed he did to Fülleborn who cites the passage, but we should have in mind an able observer trying to make sense of the religious discourse of his parishioners while fumbling in the dark for clues to the baffling cosmology it reflects. Hübner’s story begins by describing Lwembe as the actual grandfather of the present priest of that name residing at Lubaga (in Nyakyusaland), and as a wizard (Zauberer). The missionary recounts the tale of Lwembe’s magical talents and the amazing circumstances of his exile—all of it freshly retold in earnest detail—and continues as follows: Now that the wizard was gone, famine came to Kingaland, such as never before was known there; it was perceived as direct punishment for the expulsion of such an important man. Emissaries from the Kinga went after Lwembe with gifts and urgently bade him return to his homeland, which however he refused. As a return gift he gave the emissaries seedgrain from local bounty, wherewith uKinga was blessed with a rich harvest. This wonder produced a great sensation, and the famed (wizard) Lwembe accordingly became a most honoured priest. Henceforth at the start of each new growing season, emissaries were sent him with gifts of cattle and hoes, to ensure his support for the coming harvest. As with the father so with the sons, and in the old priest Lwembe’s place his successor was shown the same honour and credit by the people. After some time this man devised a new name for himself, saying: ‘God has newly appeared to me, now my name is Mbasi.’ This new name brought the wizard great riches. Yet he was no miser but slaughtered for the emissaries something of what they had brought, and thus these festivities continued without end. The gods as well were particularly

230

well disposed to him under this new name. There followed a richly blessed year producing such fruits that in Kondeland bananas decayed on the stock, and Kingaland as well was unable to consume all that was grown. Likewise, the wizard understood just how far ahead to set the time when an Enemy attack should be expected. Lo, all that Mbasi prophesied was fulfilled. Then came the report as well to uKinga, that Whites (our missionaries) had come to Kondeland and settled in to cultivate. The priest-prophet was asked advice, and he promised they would soon leave. But the prophecy was never fulfilled, any more than the assurance that all the cattle felled by pestilence would come back to life. These and other remaining unfulfilled prophecies gradually undermined the popular reputation of the once so respected priest, who lost his credibility. ‡

Hübner’s Kinga informants have put together fragments of hearsay into a narrative of their own. (Need we be reminded that all recorded folklore is first encountered as hearsay?) The missionary in turn has worked their testimony into a narrative his German supporters at home will find enlightening. The Mbasi affair has moved at least four steps beyond from the drama of a decade past: (1) Time is foreshortened and the mystique of legend lost; (2) Fülleborn’s complaint: the play is now about priest-wizards and their power not about spirit beings; (3) Lwembe has become Mbasi in this Kinga drama—awkward, but perhaps reflecting a special strategy Kinga often evince, taking a new name as a way of taking or pretending a fresh tack in your personal life; (4) the plot has shifted: where the drama had once seemed to pierce the curtain veiling Konde cosmology we are given a pat fable about false prophets. Is it the Kinga storytellers who are confused? They may be wildly ill-informed as to details of the ‘Mbasi case’, but the sole criterion of clarity in myth is that it shall pierce the curtains on what is locally acknowledged as a mystery. Kinga tellers of course must claim the Mbasi drama for Lwembe, their own divinity, just as some Konde said that Lwembe (in form of the living priest at Lubaga) had finally vanquished Mbasi (in form of his intrusive spokesman) in a challenge ordeal and so driven him off, taking back cattle and women. All the name-change in Hübner’s document means is that some Kinga are claiming this miraculous tale for the Sanga portfolio. The ‘new name’ business (since ‘Mbasi’ couldn’t have been deemed a fresh name by Konde in 1902) is distinctly a Kinga talemaker’s innovation and so not to be taken as a nominal substitution but as a naive equation of the two names. This is no more than we see in Monica Wilson’s methodological shrug wherever she has to cope with local Konde conflation of ‘Mbasi/Lwembe’ in her Nyakyusa monographs. ‘Mbasi’ was patently a new name to some Kinga folk in 1902. It was not in general use (as ‘Lwembe’ distinctly was) among Kinga in the 231

1960s, though it seems folk were prepared in the 1990s to conceive of Mbasi rather vaguely as a territorial (translocal and free) divinity whom they should contact when visiting in the Valley. Unfortunately he was unavailable. Hübner’s special difficulty may have been in expecting one cosmos, whereas the myth maker thinks rather in terms of possible worlds. Here is one of Monica Wilson’s summary judgements on the case, which bears the signs of information from interview rather than mission documents alone: During the years before 1914 a man, with two boys in attendance, posed as Mbasi and went round Selya by night, growling in a gruff voice that he was Mbasi, and seizing cattle and fowls and food. He was eventually taken by night, by some Christians, and died after seven months’ imprisonment in Tukuyu. ‡

Only one premise can be helpful here in sorting the facts, and that is easily summarized: popular theological discourse in a socially and politically turbulent society is unlikely to be self-consistent from time to time or place to place. Social memory sifts the ‘facts’ of history differently in different affected communities. There is no canonical form of a myth, unless long after its death it be found in (to scholarship) a convincing documentary form. The main support for the premise that ‘tradition’ blows on the wind is that this seems to be the case quite commonly throughout the world. The pertinent corollary is that in an animistic cosmos divinities of varied sorts come and go, flourish in popular fancy and fade, far more easily than in orthodox worlds. A broad application in the present instance is simply that a protostate process moves quite gradually toward the stabilization of a cosmological dispensation which will be a bulwark to political orthodoxy. I have been at pains to demonstrate that the Kinga had moved farther in this direction than the Nyakyusa had. The (presumptive) primary reason for this is setting: the Sanga régimes in the Livingstone Mountains had far less varied and troublesome social contacts to deal with than Konde chiefs had. Or perhaps I should simply say that the Kinga bush culture was relatively homogeneous. Peter Weber reports from the 1990s that a busy incumbent for the ‘divine king’ position is in place at Lubaga—a modern-day ‘Living Lwembe’. There is not only a certain traffic of Kinga visiting, priests and a varied lot of others, but counter-traffic as well. All is in place in what seems to be a spontaneous recreation and updating of tradition, though I have no news of a full parade from Ukwama being re-enacted with all its heavy symbolism:

232

Whereas the people who are living near the cult-center on the top of the hill assured me that until recently they were always waiting for the “big priest” from uKinga, on the other hand the chief-priest who is embodying the Hero nowadays told me that quite often he is ordered by the Kinga to visit them in order to make the rain in the Livingstone Mountains. ‡

The lay speculations Hübner recorded concerning Mbasi/ Lwembe may be taken as a warning against finding cosmic meaning in smalltalk about divinities. These (to us) timeless divinities are at the same distance—beyond a cosmic arm’s length—from the typical living teller of tales as secular heroes of the past like Kyelelo of the West or Manduba of the East are, and hold a similar fascination. In a society where oral discourse floats free, the sole anchor of narrative is myth: that means (transient) authenticity, which flows from the hallowed stature of the teller. Thus among Kinga, the Lwembe narrative has been stable at least since it was recorded in the Njombe District Book in the 1920s. If we ask why that should be, a simple answer comes to mind—the existence of systematic authority. Tellings like that Hübner recorded have been denied by those elders who have standing as keepers of myth. So far as I can say on evidence, the ‘Mbasi’ of the tale Hübner took down about then-recent events in Selya—an Mbasi who is plainly not presented as a divinity in his own right—never did achieve mythic status among Kinga. The contrary appears to be the case for Konde cosmology. Peter Weber found that one of the five ‘cult centres’ at Lubaga claimed Mbasi as its patron. He sees Mbasi as a plebeian hero: Mbasi was after all the divinity of the abapina—the ordinary people. Bwibuuka itself served, as already explained, as a place of origin, as the place whence mankind, not the chiefs, arose. A reprimand at Bwibuuka, the cult centre dedicated to Mbasi, was therefore enough to remind that the political establishment had not ruled the Nyakyusa since mythical times—that is, not quite always. ‡

Was Mbasi indeed such a people’s hero in olden days? As my reading of the ‘case of a mischievous medium’ and the abortive boycott of the missions called by ‘Mbasi’ offers no support for that view, I tend to think the proletarian Mbasi has emerged since then. It would be surprising to find a wandering ‘herdsman’ successfully claiming to hold the mystical key to thwarting the locusts today. It would be even more surprising to find that a territorial cult centre was selling its services in the very same way now as it was in earliest German times. The Nyakyusa mind in 1890 was not inclined to blame 233

the politicians for the vagaries of nature. Pestilence was the work of spirit beings like Mbasi or Lwembe, two figures often confused in popular lore. The boycott (in which Weber finds proof of the populist character of an ‘Mbasi’ movement) in the early months of pax was imposed by the chiefs and after an appropriate tarrying time aborted by them. The reason for Angst at the time was a growing ambiguity in the German presence—growing ambivalence. We may doubt the boycott would have been aborted if, as happened later on, the medium’s threat had been backed just then not by the annual return of vague doubts about the coming of rain but by the drama of pestilence: rinderpest or a plague of locusts. If it is insecurity that turns us to our divinities, the certainty we may find there is itself a turn of mind. Konde chiefs had many divinities to help them search out truth. Given the triviality of the social issues ‘Mbasi’ was raising in that early case (and given the great number and peripatetic habits of the Nyakyusa priests) his threats could hardly have got assured confirmation on all sides. In particular we learn that ‘Lwembe’ was opposed. The Living Lwembe was appointed by a thoroughly conservative synod of Nyakyusa priests from amongst their own numbers. His opposition means theirs. I see multiplex, not uniform, implications as the thrust of a Konde religion of blame. Where turbulence is the norm, let trial-and-error rule. It works mainly by adroitly avoiding confrontation while time and events move on. What reception would a Sanga court have given an aggressive ‘Mbasi’ prophet? In the actual Konde case, there was neutral ground for a traveler in magical notions to practise. The best chiefs seem to have preferred chasing him off at his price to chasing him down. In a Sanga court such a ‘movement’ would have constituted insurrection and, had it come to that, would have been dealt with accordingly. But we are dealing in both cases with societies which legitimated the antipolitan ethic, and that ethic normally invokes a safety-valve mechanism for escaping confrontation. For Konde, the peer-group ethic, which Monica Wilson found key to the moral strategies of men, was brittle. When it broke, the dissident was instantly welcome to join some other group. But it doesn’t surprise that, with kinship delocalized and in the virtual absence of the ancient land tenure rights which only vestigially survive, peer loyalty held. No such compelling ties bound commoner villages to their chief’s. The political system which the periodic Coming Out fostered was elastic where the Sanga was firmly buttressed. A collective shift of loyalties meant a major rupture—a Kyelelo, for example, and his endless war against the Sanga-led but still un-Sanga Mahanzi.

234

Konde cosmology could embrace innumerable spirit beings, any one of whom might through divination claim responsibility for punishing acts of some kind, visited on some unlucky individual or community. As a rule of thumb, I suggest a self-employed diviner would expect the range of free divinities available in a typical trouble case would be quite like the range of dead kinsmen close enough to a client to qualify for blame, or like the range of living kinsmen or affines who might credibly be accused as witches. There is always a handful of suspects, never less or more. We have to deal with a religion of blame not fear, though the distinction may be hard for a deeply Christian mind to accommodate. In religions of fear, a believer is responsible, in the manner of a child to an all-seeing parent, for ‘being good’. In classic (ideal-typical) animism divinities are not interested in your welfare but their own. If a divinity seems to have punished you, it is not for crossing your neighbour but miffing the divinity. You will not find a spirit ready to be your moral guardian. When spirits decide they want something you can give, they will bully you for it in their own, well recognized ways. It is particularly clear when invisible agents seem to be chastising the multitudes that the problem is to recognize which of the several hostile spirits is behind the trouble, and what sort of appeasement must be offered. This work is not going to be done by ‘multitudes’ but their agents. Religion in this way flows into politics. As for panicked publics, though, I find no evidence the Nyakyusa political system was shattered by the Mbasi affair. It is at least in some good measure pestilence, disease, and drought which drive a human community into some form of political organization which allows panic—or call it fear or dread—to go through channels. In the Nyakyusa case, this can entail some turbulence. That is the way a firmly agonistic system works. But what I see in the case of Lwembe and the Sanga chiefs is a neater compound of politics and religion, fear and blame. Where there is an original sin there is an Original Blame, shared by the multitude. Only the principals to the original acts against Lwembe can still his anger. The greater and more general the Angst, the firmer the court’s hold on its public. How far the Konde mind had shifted from classic animism toward a tenuous (poly)theism we can’t know precisely. Peter Weber finds support in the symbolism literature for taking Lwembe as the first Konde sky-god. For Kinga, in any event, Lwembe is certainly and radically chthonic. But from what the Wilsons found in the 1930s I believe any Konde shift from shades to gods, when Christian and some Muslim overlays are set aside, was more apparent than real. 235

The main free divinity then who seemed to be making a transition to a post-animist cosmos was Kyala, but that is only as one might have expected. He had been adopted by the Missions, and the rhetoric of deism is catching. As for late nineteenth-century conditions, I have set out my reasons for caution in using ‘cult’ to construe the Konde record. I find the term usually inappropriate for Kinga practice, where divinatory practice doesn’t lend itself to inspiring a congregation of believers but seems always to have been practised solo, without retinue or apprentice, for pay. But if ‘cult’ may be somewhat misleading as a tool for revising Monica Wilson’s work on Nyakyusa ritual practice, one facet of the word is enough to support its continued use: it fits the facts of religious pluralism. Given that one portal ‘cult’ doesn’t set out to vanquish another, still business rivalries must arise in a setting where trouble-clients go shopping about for the best in doctoring—just as we see modern Kinga doing in Peter Weber’s interesting account.

Konde/Kinga cosmology in perspective What we have been considering is a special strength and special weakness of religions of blame. As translocal contacts and ‘winds of change’ break up the consensual fabric of an acephalous pedestrian community, manipulative activity moves from mumble to rumble. One manifestation is the emergence of translocal authority roles. Another, always with a strong political coloring, is the ordeal— turning witchcraft fears into trouble cases which leaders of a new stamp can use as political spectacle. Witch ridding ‘cults’ represent a different, more transitory phenomenon and may burgeon in the lapse or absence of an effective chiefly executive. Wherever there is a coincidence of chiefly incompetence with tormenting times we should probably expect the appearance of rival claimants to magical wisdom or magisterial calling. I do not suppose in the Malatan case that witch-ridding cults were present before translocal politics began, and it seems they were new on the scene after colonial rule had effectively barred the ordeal. Witch-finding impresarios did thrive, unblessed by authority, in early decades of the pax, when salaried chiefs found their hands tied. The Mbasi ‘cult’ we glimpse in German missionary documents present yet another format for turning the special weakness of animisms to manipulative purpose. I’ve noted the word ‘cult’ is strong, heavily weighted now by its application to recent Western history. The vulgar connotation is ‘enthusiasm’. But where the ‘Lwembe cult’ is obviously not of that stripe, an ‘Mbasi cult’ easily slips over the semantic line—the word wants to 236

become too spongy for use without special gloss. I have tried to show that the ‘rival cult hypothesis’ implicit in the revisionists’ argument doesn’t fit the facts as known. Malatan geology makes Konde country good for ‘portal cults’, which grow around a netherworld access (one which, for instance, a python might frequent, whether for water itself, a secure retreat, or the small game attracted to a secluded spring). These are likely to be spots off the main track for tillage and human traffic. Look at the matter pragmatically. The portal curator having to make the most of his venue, an obvious way to that end is to offer ‘territorial’ services. That is, he must be seen to remain free of local political overlords and keep a practice open to pilgrims from a broad, translocal territory. On the evidence we have, an Mbasi cult is something else entirely. One might call it mystical street theatre. The oracle presents himself as everyman’s natural rival, a know-it-all. The lines could be written for Punch. The particular German missionaries who tilted with ‘Mbasi’ (in the scant few scenes we have to suggest their drama in Selya) were indulgent participant observers, like the anthropologists and historians who have followed their lead. The trouble is the record as it stands fails to choose between genuine and spurious aspects of Konde religious and public life. Native Selyans may have set aside doubt as they were drawn into the thick, but in the sequel (so far as we may know it) their plurality of frame becomes clear. Humanity everywhere survives by employing an ifthen/as-if interpretive frame. Without belaboring Sapir’s well-known distinction, I take it that the action of ripping off your friends can serve as a touchstone for what is spurious, while actions reaffirming such values as kinship and friendship may stand for the genuine in any culture. To assume Konde culture is not capable of spawning and supporting spurious (extraprocessual, anti-structural) performance is to deny a magnetic community due respect. In particular, an assimilation of all the divinities the Germans ran across to a stereotype of the ‘heroic cult’ offers little sense of depth either in time or character. I have wanted to make a clear break from the reigning wisdom of accepting the Mbasi narrative as evidence of a priestly ‘rivalry among equals’. I find unacceptable the picture of a ‘battle of the heroes’ staged by their respective priest-curators. Once you accept established thinking about territorial cults it seems likely the Mbasi episode represents the incursion of a fresh social type into that day’s politico-ritual arena, challenging local practice. It isn’t clear to me that Mwamafungubo (as the ‘herdsman of Mbasi’ called himself) had any more legitimate standing in the Selya neighborhood than the German missionaries had. Both claimed powerful spirit patronage, 237

but the advantage of the native Konde speaker was in experience and tactics. Mwamafungubo had apparently been living by his wits for some years as a roving presence (nuisance) in the Konde community: the evidence is that he had already collected for himself a small herd of cattle and a modest company of wives. When his original effort to present as a client to the Germans failed, he claimed them as his clients—he had brought them in. When next they cried ‘Shetani’ he cried it back. They were straight and he was slick. So I read the case. But just as we don’t style this Mbasi as a rival of the Mission, we don’t style him as a ‘rival’ of the established priesthoods. He is a thorn in their side. He doesn’t threaten to replace them. The ethnography shows the Lwembe, with his arduous taboos, as a priest among priests, not a free-spirited ‘doctor among doctors’.‡ To some degree the fault has been with the language of discourse. The notion of a chthonic ‘hero’ guarding each portal to the netherworld has given the land of the Konde a suspiciously Olympian look. Lwembe as Mbasi-the-Creative-Trickster was indeed some sort of hero in the (romantic) fictional sense, a protagonist in a tale of his own—ultimately defeated—adventure in political subversion. But seen in the context of ritual action, a ‘hero’ Lwembe is certainly not, for he is a magician of the most dangerous sort, difficult to appease. The Lwembe ‘cult’ is deeply misunderstood where the deity is pictured as an emissary sent to the Konde to establish new standards of civilization. This is to conflate Lwembe with the legendary Nguluve, said to have led an exogenous princely party to transform and rule the Konde. A Konde prince is no recluse. Lwembe was banished not ‘sent forth’. He was subversive, feared, and out of control. He could not be commanded, only placated at last with arduous collective effort and ostentatious expense, employing the special skills of knowledge-masters in the regional community. Olympian? One might think of Hephaestus; but empyrean rivalries on behalf of hero-herdsmen? No. What we have to consider is less romantic. We must judge the pace of a long-term movement away from a deeply established chthonic religion, always requiring self-help, toward ‘ceremonialism’—religious theatre. If regionally exogenous factors are not overwhelming, this kind of holistic change should be gradual enough to provoke no signs of marked disorganization. Yet in any historic instance the paths of change would be predictable only in principle, not detail. By the same token, the use of oral history in reconstructing past processes can only benefit from skepticism. Let us suppose the movement begins with migratory drift: in this case, the migrants are individuals and small groups with marginally 238

better farming technics, moving into more sparsely settled parts of the larger Sowetan region. We may picture loosely connected pedestrian communities embedded in a gradual flow of individual culture bearers looking, in the main, for a marginally better living but encountering resistance of the sort which intensifies a need for adaptive change. One result is the emergence of new political institutions. Another is the elaboration of new ways of dealing with the kind of private woe which stresses social ties. We have two novelties which are made for each other. I am supposing that the interworking of these two strains can lead to a protostate process giving prominence to secular rulers cooperating with priestly keepers of a powerful ritual theatre. It is particularly in this last that Kinga and Konde versions of statebuilding differ. Konde culture as represented in the Selya studies gives great scope to life-cycle/kinship ritual where Kinga does not. Communal rituals at the Sanga courts are those of an agrarian society whose crops are proclaimed to depend on the regular ritual reinforcement of fixed order among a known assortment of domains and realms, envisaged as a single, pyramidal structure. Communal rites of the Konde in Selya emphatically sanction political pluralism, where Kinga balance local autonomy with translocal dependency of the ‘symbolic’ or ‘genuflecting’ sort. Konde narrators often glibly conflate ‘Mbasi’ and ‘Lwembe’ in telling the tale of the Lubaga divinity. While some confusions may be due to colonial ‘noise’ the utterly undogmatic nature of animistic thought can easily explain all of them. Sometimes in Konde/Kinga discourse ‘Mbasi’ just means something like ‘Fugitive Trickster’. Monica Wilson heard the meaning ‘lwembe’ (‘razor’) in the high divinity’s name. This reflects particularly a facet of local knowledge of pythons, since the known way to kill one was with a sharp-edged reed or still better a palm-held knife of the sort used for shaving. These knives are made by the Kinga smiths and traded to the valley. So ‘Lwembe’ from uKinga becomes the python-killer, an unsaintly St. George. A python would have to be mastered by its meeting a hero with its own mystical kind of personal strength and armed with a razor as well. So the essential idea Wilson found conveyed by the name is that Lwembe should be understood as a divinity of higher order than the pythons of the Konde or Sanga courts’ sacred groves. There are two suitable meanings: Lord of the Pythons, or Master of a New Order replacing the decentralized pre-chiefly political organization. The old order comprised largely exogamous and ritually autonomous local settlements headed by a pop-up leader, heir to an original settler and Owner of the Land. I find no need to choose between the two meanings. 239

Divinities’ names can operate as concepts in legendary narrative. This hero of folk fiction is indeed a hero of legend to those who find themselves wrestling against tyranny—let us grant that is what, for some, his story could be about. But when Lwembe is approached in time of trouble it is not as a potentially friendly intercessor but as a fierce divinity whose anger has been roused. If ‘Mbasi’ seems to share all his magic but little of his majesty with ‘Lwembe’, there may be ground to grant Mbasi less weight. I fear we know too little about this, but we do know that the Konde Lwembe and the Kinga are not deeply one and the same. The territorial cults of (especially) Malawi and Zambia are contingent here, both geographically and analytically, though the weight we should give to south-north influence on Konde in the more purely ritual order is not clear. The influence from that direction which is clear, though, is that of the Ngoni. They imposed a series of sorties and long-lasting settlements in the Corridor region during the midcentury period and after. Mpezeni’s group was formed in the northern reaches of the Corridor after Zwangendaba’s death about 1845, but didn’t withdraw from the Corridor region for another quarter-century . Mbelwa’s kingdom, once set up in northern Malawi with borderlands on Ungonde, remained on a nonstop war footing, aggressive and defensive, sending out massive raiding parties into the final decades of the century. Since the objectives of the massive raiding parties were, apart from simply proving solidarity and dominance, booty in slaves, cattle, and ivory, the apparent success of the Konde peoples in holding their own has always seemed remarkable. What I find most interesting is that this was accomplished by following an opposite strategy to the Ngoni—not solidarity through massively disciplined militarism but something closer to guerrilla tactics. Each Konde village was an autonomous, moveable unit linked to others by deeply voluntary ties. Yet something these two turbulences had in common was an ideological de-emphasis of ancestralism. ‡ Margaret Read based her account on fieldwork in the 1930s in Malawi, and especially funeral ceremonies. She found: The attitude of the living toward the dead was one of respect and honour, but not predominantly one of avoidance and fear as among some of the neighboring tribes. ‡

Unfortunately, attendants at funerals anywhere else in the region must have professed such respect. Rau sheds more light, noting it was the Paramount Chief whose ancestors got public praise as ‘protectors of society’, and: 240

Mizimu (spirits) of ancestors were occasionally identified by vyanusi (izanusi: i.e. diviners, doctors, prophets) as causing personal misfortune or illness, but oral accounts indicate that such instances were infrequent. Rather than assigning personal problems to ungratified mizimu, the Ngoni felt that most cases of misfortune and unexplainable problems could be traced to witchcraft. ‡

It is not difficult to see that Ngoni, as an amalgam of men and women of quite various cultural origins, would find it impolitic to emphasize their diverse ancestry. The Nyakyusa again, having delocalized kinship, followed a distinct strategy toward the same goal. They gave funerals to the kin group while doing nothing to prevent fighting at the funeral between kinsmen as visitors and the deceased’s peer villagers as hosts. In this way neither kinship nor political allegiance was unduly reinforced, and the split basis of identity redounded rather to the benefit of individual freedom of choice. Ancestor cults are famously sensitive to spatial displacement. For one thing, it is hard to blame a spirit you may require weeks to reach—shades stay where put. A further consideration is simply that social turbulence breaks up ego’s close kinship networks, allowing non-kin ties and rivalries to take on greater importance. The often near presence and protracted threat of militant Ngoni neighbours was probably at least as strong an influence on Konde thought and praxis, as was the colonial contact which followed and eventually pacified the Ngoni kingdoms. We have seen that Konde social institutions had a direction of their own. But certain ideas and values easily overflow borders. Ngoni contact with Konde communities, though perhaps dramatically inconclusive, was arguably as important an agent of change as were other Ngoni contacts within the larger Sowetan region. Here are two special considerations:

Political theatre: Ngoni statecraft leaned heavily on the poison ordeal. We have to understand that institution as a quick-and-dirty way of converting private trouble cases which have begun to trouble the peace into a show of imposing political power. The institution of the ordeal is ancient and widespread in Eastern Bantu civilization, but ranges from almost bloodless to dramatically punitive. Ngoni influence by example would be to advertise and exacerbate the drama, but not its use in statecraft. In absence of a heavily institutionalized high-chiefly authority the poison ordeal would be administered by local priests at the level of a moot not a court. That is, spectacle would be subdued and blame off-loaded in a less prodigious style.

241

Prophecy: Ngoni kingdoms as we know them in pre- and postcontact times from Malawi and Zambia had incorporated so many diverse practitioners of diagnosis and prophecy as to stir up a true efflorescence in these occult arts. Considering prophecy as no office but a calling, recruitment to the role is not by routine but animus. Cults of affliction of every sort are reported not just for Ngoni but Cewa and many other unconquered peoples in the Central Africa region south of the Malawi-Tanzania Corridor. Most occult arts are everpresent in one form or another in East Africa as well, going through periods of dormancy and arousal in response to troublesome events. Where a protostate process is underway good statecraft means controlling (co-opting) as much of this action as possible. Famously, one Ngoni tyrant set a trial assignment—a trap, indeed—to all the diviners under his court, and executed most of them for toadying. Still the calling lived on. A prophet’s fame is proverbially far from home. Boys pass on magic tales from one end of the continent to another. When the times are out of joint, new prophets will pop up. Malatan culture, in the canonical version, tended to institutionalize prophecy through the portal cults, rulers’ ancestral shrines, and the carefully staged ordeal. Ngoni influence by example would have been, on the one hand, to light up ambitious young prophets everywhere; and on the other to point the benefits of cooptation for the local ruler. But Ngoni influence by intrusion would have been to heat up an already turbulent society, creating a bull-market in certainties. In the Konde case particularly and to a less degree for the Sanga courts, the Malatan world can be seen changing by multiplying alternatives, not by the tidier but tedious process of endogenous institutional transmutation. Animism with all its open pluralism is being stretched by interpolated innovations, and each community has its own evocative symptoms. But it is only in the very long run that pop-up innovations begin to warp the cosmos.

Kinga/Konde uses of ritual Recall that the missionary’s fragmentary narrative concerns a sort of prophet—a young man with some knack of ventriloquy, presenting himself as the ‘herdsman’ and spokesman for an easily roiled free divinity. There is something out of tune here: priests in Nyakyusa culture are not self-appointed and are in the elder category; yet we are asked to accept this self-promoting young man as a ‘priest’? The clarifying context needed is a life of the mind and spirit as richly colored and turbulent as the observable political scene I have been concerned to reconstruct in earlier chapters—an 242

inner life not even as plausibly called ‘observable’ as the scene already presented. Consider how important it has been, and how difficult, to be aware of the difference between informant-models of the Nyakyusa polity and actual (deemed most probable) practice. Now apply the same consideration to ideas held about religious matters. It is hard enough to be sure what those ideas were in the 1930s, and harder still to have anything certain to say about the generative ideas which would have prevailed in the 1890s. Then comes the task of reconstructing praxis: everyday life and response to crisis. Still, having once recognized that the record can afford us few certainties, we need to ask how we should deal with what remains. When all is said, that means consulting the empirical evidence, and it is not insubstantial. Witchcraft wants its mystique. I suppose it is the very empowerment a person may take from being suspected of mystical powers which makes successful denial unlikely. To be called strong is flattering. Our wandering ‘shepherd of Mbasi’ found a low-risk way to empower himself as a for-profit prophet. As a strategy he borrows from witchcraft but looks more like the crafty witch of fable than like the poor chap who loses the coin-flip of a staged ordeal. Have in mind that arcane powers are institutionalized in at least four ways in the Selya we are given by the Wilsons. Three of these ways make mystical power legitimate. A ruling prince, to effect respect, is supposed to harbour a python in his belly. Priests hold offices which give them legitimate access to the special paraphernalia and chemical instruments of mystical power; and their use of these instruments is subject to control only by the in-kind ploys of rival/ collegial priests. A third legitimate agent of occult power is the spirit medium or prophet. Here are some bits from Monica Wilson’s summation: Reality is thought to be revealed to men in dreams, by prophets who ‘go down thinking like the roots of a tree’ to the world of the shades, and by oracles. What men seek to know are the causes of misfortune—public and private—and the identity of witches...The contemplative is not honoured in Nyakyusa tradition; the man of spiritual power is a detective—a discoverer of evil-doers. ‡

Selya priests in the precolonial years would have known among others these three portal shrines: Usweve based to the north in the Porotos, Lwembe (alternatively Mbasi) at Lubaga, and Kyala (also and for the same reasons alternatively Mbasi) to the south on the lakeshore near Ikombe. Each is a special case, but all have access to supernatural power of the generic kind witches are known to use in harming others. Of Kyala, more later. Usweve the Selya priests would 243

at least have known through their Kinga colleagues. I have pointed out that the Mbasi figure refers to a traveler with magical powers, an animal maker in exile on account of those powers from his native community. Usweve, who was sought to for relief from locust plagues, was described by Kinga as an exiled rainmaker. He would have been an autochthonous practitioner, first banished by a new Sanga lord as a rival for authenticity, then by a successor on office urgently but vainly wanted back. Other rainmakers, invaluable for the very localness of their wisdom, in other domains did return, but Usweve was unyielding. What we have then is a paradigmatic expression of an animistic religion of blame, in which the narrative hero is first object, then subject of righteous indignation. As with Lwembe, Usweve proved himself a free divinity by plaguing his people. A witchlike figure, envied and exiled, he escapes with supernaturally magnified powers of terror-and-forgiveness over former friends and kin. It is a formula which must have been applied to many more divinities than we can reliably name. Joseph Campbell might welcome Usweve or Lwembe into the company of his omnificent Hero, if only they would desert their underworld abodes and claim the miracle of rebirth. I suppose what is really at the heart of the Malatan hero myth is the logic of social leverage. Our Mbasi impersonator was playing the very trickster character of the legendary narrative, using special powers to escape the trammels and injustices of the normal human condition. We are dealing still with jealous not generous divinities. The diameters of fear and blame, deprivation and feasting, anger and forgiveness are dramatized wherever misfortune offers leverage—a people informs its understanding of the cosmos, always hoping the better to sense the part which human will must play in it. ‡ ‡ To this little homily I would add that ritual intercourse with the spirit world in the Malatan region is esoteric. Since most people won’t have had close and direct contact with even one major cult centre but depend on priests for interlocution, common speech in any part of this orally grounded culture may have its own plan as to the importance, location, or proper name of address for any resident spirits. About certain things priests themselves are bound to disagree: chunks of narrative are borrowed and turned about, history is improvised and the teller is always ready to recognize his protagonist by some other name. If the name ‘Mbasi’ in the episodes we know was used to identify a spirit without a shrine and quite possibly without free-willing clients, there were certainly many applications of that same name and many not kept for posterity. If there is a single point to have in mind it is that Kinga/Konde culture is a manifold system, all of which will never be seen in the frame an outsider may regard as properly ‘evidential’.

244

Swahili’s two common terms for occult power, uchawi and ulozi, have Malatan counterparts. Kinga favour a cognate to the one (uvuhavi) and Nyakyusa a cognate to the other (ubulosi). Both words translate as ‘witchcraft’ in english, though our ‘witch-hunt’ connotations are usually misleading. Any account of the context of Malatan witch belief should start by inquiring into the cast of mind which equates legitimate moral authority with access to arcane power. Is this a simple form of rank concession, or is it a yet subtler way out of responsibility? Can the apprehension of your friend or your wife as a witch be seen as a begrudging concession of rank in the presocial order—the actual-cum-unofficial power scale—of your social world? That is the order in which egos swim or sink. Tunginiye perceived a threat from his friend which he could not withstand. By conceding deadly powers to the other, and turning to a doctor able to match them, Tunginiye experienced an estranging victory in his presocial world and a heartening cure in the fully socialized world he could return to. Had Tunginiye failed to survive, the idea of witchcraft would have had shadowy confirmation and would have attached itself to his suspected aggressor. Had his self-diagnosis led the ‘victim’ to a priest of the court instead of a doctor-friend, there would have been no victory for anyone in the (ensuing) ordeal, and the tale would have had no hero but the court itself. That is the kind of drama I think reveals an essential political value in owning narrative truth. I have not located any documented information improving on my scant field notes about Usweve. Kinga sometimes sent placatory gifts to him at a fixed shrine-place somewhere in the hills north of Lubaga. His pestilential specialty was locusts. Kinga myth makers worked him into their pantheon by linking him to Lwembe by marriage and character. Not quite consistently with the Kinga-Konde formula, they pictured him to me rather as a pre-Sanga type of ‘doctor’ and rainmaker than as a royal or a priest before he was ostracized by an early Sanga ruler. Konde narrative seems to have favoured batching such figures as Usweve in one syncretic tale of exile from the Sanga high court in illo tempore, along with Lwembe in a ‘still younger brother’ guise. Either way, like Lwembe, Usweve had been exiled from Kinga country and lived on as an immortal, keeping a jealous eye on his people. Kinga say he stopped with Lwembe, who gave him a daughter in marriage and sent him on northward with locust-control medicines to be forever, like his father-in-law, another of Blame’s idols for the Kinga people. In the telling by a panel of elders in the Western realm I got hints that the Usweve shrine had been patronized by pre-Sanga leaders and could be regarded as (in my terms) an add-on feature of the Lwembe cult, resorted to particularly when locusts (suddenly) threatened. Locusts don’t leave much time for priestly manoeuvring. 245

More I can’t say, but the measure of agreement by elders interviewed in several domains was confirming. It is clear that Usweve was considered in some circles to be just as tough as Lwembe, and I suppose he was retrospectively married into the Lwembe line to assign him a clear status as ally rather than rival. This may have been needed, as Usweve (in the 1960s) was not as well known as Lwembe in all parts of Kinga country. There are no documentary records for an Usweve pilgrimage such as there are for the Lwembe cult centre, and I know of no confirming notes from a supposed Usweve site in Rungwe district. ‡‡ Any suggestion that Lwembe ‘was God’ or was a figure of ‘Good’ or of ‘Evil’ is even less pertinent than the same suggestion about ‘Nguluve’. As an otiose creator or Adam figure, ‘Unguluwe’ (as Kinga say the regionally important name) could be adopted to play the role of Christian God without inspiring deep confusion in Kinga thought. It just meant worshiping one’s apical male ancestor. No serious displacement of the folk wisdom was entailed, as no Kinga institution is chartered by an ‘Nguluve myth’, though what clarity the missions might have won is not apparent. By contrast to this otiose figure, Lwembe remained dramatically real in the 1960s despite the languishing of his ritual; and he reappears in a new adaptation in the nineties. When traditionally Kinga made their expedition to Selya they wanted to ‘do business’ there and would be looking out for credible wizards in the Rift valley to the west of their highland country. The weather came from there, and with it the occasional hostile infestations of insects and disease sent by the same mystical agencies. We are dealing with an animistic worldview not doctrinaire theism—the invisible powers are many, as are views about which of them may wear what masks and bear what responsibilities. It may be because weather (with its rains or rainless skies and its clouds of locust) comes to Kingaland from the west that a responsible divinity had to be sought there, but that is the sort of thing you might hear offered by a scholar from uZungu (the West), not by a believer. Generous overlaps of name and function abound in a polyglot cosmos. In new places you will find new divinities to deal with. Kinga priests did not serve Lwembe but their individual local rulers. In a quite pragmatic view, the priests who make the pilgrimage to Lubaga need their Lwembe to serve their special interests. He is the main source of their translocal prestige, and so of the influence they can exert at home. Each priest came with his own portfolio of local anxieties to assuage. So there was a healthy amount of improvisation in the implementation of the Lwembe cult. As it wasn’t undertaken at scheduled intervals (though Mahanzi informants told Peter 246

Weber in the 1990s that it probably was annual) but in response to general ‘unrest’ of one kind or another, and as the main concerns of the pilgrim-priests were to make firm connections with professional colleagues all along the way, particularly within uKinga, the show seems never to have happened quite the same way twice. It is easy enough to accommodate all the conflicting stories recorded about details, if you bear in mind each has reference to particular undertakings and events of different date. Speculative thought accumulates in granular fashion in preliterate society just as it does in the less-literate segments of a more urbanized world. All the hard granules of theological speculation once established in the Malatan region may possibly persist in some version even today. Neglected divinities who live on in myth may come into fashion again. The case of Usweve illustrates. Consider that Kinga treated Usweve as a ‘territorial’ deity with influence over only one wide spreading pestilence. Locusts plague a given place at long intervals of years. How is such a ‘cult’ to be maintained between? If it is to be active, a shrine with such a far reach would have to be kept by locals, and that entails providing a far broader range of services for a regular local clientele. More likely is reinvention at need. As with any institution, shrine-keeping costs must be met the year around; and religious services are impossible to render where honour is malnourished. The keepers are mystical power brokers who need to keep the shop open and hand it on to trained successors. The known details of the Lwembe shrine at Lubaga prove the rule. Currently, the shrine is in revival. Peter Weber notes that both local and Kinga visitors are currently frequenting it, and that the sorts of grievance addressed to Lwembe are quite varied—in this version the shrine has a manysided business. The older version was languishing in the 1930s, with only a makeshift arrangement for its upkeep. It is at least clear that it was never kept up years-around only for the sake of Kinga and other such occasional pilgrims. Consider also that Lwembe the Prophet must enjoy more credit in far Kingaland than at home in Selya. The ‘divine king’ imagery lived in precolonial Konde discourse for both the Kyungu of Karonga in Ungonde and the Lwembe of Lubaga in Selya through the purity of belief which distance allows: for the vast numbers of believers the imagery drew its life from a hallowed text about primal times and places unseen. Supposing now that an embassy of Kinga priests had encountered Mbasi’s herdsman on a visit to Selya in the 1890s, what notice would they have accorded him? There is reason to suppose it would have been slight. This Mbasi made claims counter not parallel to Lwembe’s. Presenting himself in street-theatre episodes, not ceremonially, this Mbasi had no shrine-place, with supporting cast of 247

diverse priests, wherein to play the kind of solemn waiting-game required. The sociology of the Lwembe shrine makes clear that in response to plague or fear of drought the only politick solution for the court’s priests was to be seen still hard at work on the problem until the turmoil of fear had begun to fade of itself. Still, I do not put it past Mbasi’s oracle to have accosted the Kinga priests, and they would have been out of their own bailiwick. Failing intercession by a career priest familiar with the oracle, how would the career priests of uKinga respond? ‘Kyala’ in precontact times was another name to conjure with. Since local names (not deities’) were used by Kinga in reference to their visits to the lakeshore in the 1960s, the near-absence of ‘Kyala’ from my field notes needn’t mean that name was never current in lay discourse, as it certainly had been among priests. The regular business of Kinga with Kisi priests was doctoring the rainstones, and Kyala’s portal shrine would have been generally known, by whatever name, from such visits. But the rainstone business was one the Sanga courts had not fully co-opted. Rain falls just here and there and irregularly at the end of the dry season, and its care thus falls to local ritualists. Some if not all these rainmakers continued a pre-Sanga tradition of their own into the 1960s. Again the story-plot is one which accounts for placation of one’s public without disavowing a unique mystical connection to the hidden source of unnatural trouble. Each domain thus has in its sacred grove a deep pot, fashioned by women in uKisi, covered and kept moist by its location, containing a number of mischievous rainstones. The pot must be visited from time to time as the dry season looks to have run its course, to see that the rainstones are secure. Magically, as the rain-priest’s visits grow more urgent, the stones will be found missing, escaped from their dark and ritually secluded store. Like the water they symbolize, the stones ‘run down the mountainside’ heading back to Kisi country, a fifteen (plus) km. strip of the Malawi lakeside. Are they ‘feeling dry’ and running down for water? My rainpriest at Lupila wouldn’t say. He spoke of them as one might speak of a rare animal with a mind not to be read. What is special, on the Kinga scene, about the rainpriest is that he doesn’t belong to the Sanga court but to the older bush culture. His name will not be Sanga. He is of local descent and repute. He is not, and does not see his heritage as, at one with the avanyivaha. Before the Germans came to Lumbila to police lake traffic in ivory and slaves, Kyala’s portal is said to have been there in a cavern overlooking the Lumbila promontory. In Kinga folk memory the eviction of that shrine and its keeper was ugly, but I had no information as to the relocation of both to another cavern of similar 248

description (since called PaliKyala, Kyala’s place) between Ikombe and the market centre at Matema. Though in the longer run the eviction of the Kyala oracle, with its move much closer to population centres in the Plains area, may have been good for business, it’s hard to believe the status of Kyala in the regional pantheon was not affected during the early German period, in the 1890s. Then once ‘Kyala’ was taken up as the Christian name for God, the operation of the shrine would have reflected new ambiguities. But I think we should suppose the Konde, like the Kinga, rather enjoyed than deplored the license ambiguity could lend their cosmic speculum. Monica Wilson’s review of the Konde aristocracy’s mythological charter is to be found in the second chapter of Communal Rituals. While her purpose is clearly to link the main free divinities of the Konde to her chiefly genealogies, it is here she provides the fullest of all recorded statements by a Kinga priest, Kikungubeja, of the canonical narrative in its Sanga version. Several points bear clearly on the status of Konde/Kinga free divinities: (1) All the Kinga chiefs when praying in the sacred groves say the names of their fathers who have died, and also the names of Nkekete, Kyala, and Lwembe, in that order. (2) Nkekete and Kyala only made crops—millet and beans. We ask for fertility of the soil, and sunshine, and rain from them, but not animals. From Lwembe we ask for milk, and fertility of men and cattle, and snakes, and goats, and sheep, and all crops. Lwembe has everything. We ask for sunshine and rain from Lwembe as well. (3) When Lwembe went down into Selya he found the hill people and the people of the plain already there. They feared him because he had all things, and they still fear him. He was greater than Nkekete. They cut Nkekete’s throat, but they couldn’t kill Lwembe. ‡

In the rambling account (as we have it in translation from kiKinga through kiNyakyusa) there are also some interesting discrepancies. First, it is unlikely a Kinga chief would personally enter the sacred grove at his court—something has been lost in the retelling. Nkekete’s haunting-place is either ‘at Mbela’ or ‘in a stream with pools in it’ which is inferentially located at Lumbila. The two versions may be resolved by equating ‘Mbela’ (locus unknown) with the known ‘Lumbila’. Kyala’s preferred location, however, is always given as Lumbila; it was there the Germans shot at him (i.e., his shrine keeper) and drove him northward to an available grotto near the shore market at Matema. Are Nkekete and Kyala, like Mbasi and Lwembe, two casually distinguished faces of one chthonic hero? That remains a possible resolution, granting that such composite myth249

figures result from the conjunction of ethnically separate traditions with mutually-borrowed narrative plots. Or the matter might better be settled more simply: in a regional culture like the Malatan all traditions will be heard in many dialects and versions. One man in a fairly representative case might well in his lifetime subscribe to several of them without sensing a difficulty. Kikungubeja, as arbiter of differences, must feel compelled to give a syncretic account: so all these free divinities become ‘brothers’ within the Sanga ruling line. Other expert tellers splice in other names and events. The myth remains readable. What the myth says (one of its readings) is that free divinities are created when a superior kind of human being can’t be tolerated by the political establishment, and is inadvertently ‘kicked upstairs’ (or more correctly, down). Such divinities can best be understood as antipolitan heroes. It is thus that all the free divinities we have to consider for the Malatan region are tagged with the same narrative schema: magical feats—exile—apotheosis as a chthonic divinity. That the Sanga regime should have elevated such a figure to an apical position in its open pantheon is not hard to comprehend. Intramural warfare among Sanga rulers was an escalation of the wargames each domain exercised under the auspices of its court to wargames at the level of the princely realm. The nominal prize in this ‘great game’ was triumph not spoils; the real spoils which followed from a triumph were new and morally renewed adherents to the otterskin crown. The politan thrust of the Sanga system depended on the rivalry of the leaders for the loyalties of men. The symptoms of failure are to be seen in weakened morale and desertion; the quintessential politan show of strength is the vainglorious act of exiling a man of fabulous personal power. In the narrative, the UnkeketeKyala-Lwembe character is just such an one, an inevitable and irresistible strongman with whom no lesser personage could share power. Predictably, it was Kinga priests who were most inclined to construct syncretic versions of the magical strongman story. The same narrative pertains in Konde cosmological thinking, but with the difference that vainglory is displaced to the Sanga regime in uKinga. The necessary connection is the Konde Nguluwe fiction, by which Konde princes are made genealogically complicit in the gods’ exile and set apart from the Valley commoners for whose loyalty they must compete. Then for Konde as for Kinga, the readiness of the highest secular power to atone at the spirit-world level for a blameworthy past is the critical point of turning for the narrative of human misfortune and redemption. But to appreciate

250

the critical difference between Kinga and Konde settings it may be well to revisit the ‘case of Mbasi’ one more time. This time we ask why the local Nyakyusa chiefs were willing to lend credibility to Mbasi’s ‘herdsman’, almost to the point of dropping their own ‘great game’ of internecine contest and turning as one against the apparently benevolent missions. I take it as given that the herdsman’s claims, counterclaims, and reverses of field were not simply so adroitly done and so compelling as to explain the hostile boycott of the missionaries in which most of the courts joined. It appears the deciding factor was simply that the powers of the chiefs were not equal to meeting natural (territory-wide) challenges. These they would normally pass over to their priestly colleagues; but these colleagues were even less well prepared to act as a body than the chiefs. The Mbasi claimant was the man on the spot on whom to off-load popular concerns about rain, dread signs of change, and (eventually) rinderpest and locusts. Whereas the Sanga prince had a fully established priestly entourage to turn to when trouble arose which, of its nature, transgressed local boundaries, the Konde prince had to deal with only transitory boundaries and a plurality of distinctly individualistic priests, some of aristocratic and some of commoner hue, scattered among residential enclaves of chiefdoms of varying size and maturity. Crucially, the time-andspatial scheme of breakup and consolidation given by the ubusooka institution would often if not always have left the personal links between secular and sacred leadership tenuous. This stands in contrast to the self-integrated Sanga system, in which continuity of power was guarded by a co-opted priesthood whose every interest lay in keeping close to the secular powers it served. The sharpest image for making the difference clear between the two systems of ritual action might be a ‘ritual mazeway’. Assume that the point of ritual is to deal with trouble through procedural interventions which, by appearing to deal with its source in rational fashion, manage to sublimate the dangerous human passions which have been aroused. Ordinary pragmatism isn’t appropriate when trouble goes deep enough to portend a breakdown of order. In the Sanga case, ritual intervention leads up by solemn procession from domain to realm and, at need, onward by the royal road through the four realms, descending only finally to Lwembe. In the Konde case (Selya and Plains) the mazeways are uncountable, comprising a web in which the sense of crisis hangs suspended perhaps for years. The prize example is Kasitile himself, the most constant of the Wilsons’ informants but forever perplexed by his own ailments and forever pursuing new paths toward cure in the rich mazeways of ritual action open to him. Summary execution of a 251

witch, brought to the high court at Ukwama, had all the drama of high treason, as if to prove that if the ordinary ordeal had failed to find the witch, Destiny itself would take over. I find it hard to picture the equivalent in the Selya I can reconstruct from testimony taken well under a lifetime after traditional times. The standing of a ruling chief was not, as the Lwembe oracle was, set apart and high enough. Nyakyusa folk analysis in the Wilsons’ monographs had the chiefs holding critical power. Yet Monica Wilson heard mid-century youths complain that in illo tempore the chiefs had had to deal with democratic checks, which Colonial administrations (unwittingly?) suppressed. The skeletal facts of the Mbasi case, as recorded, ought to be seen in this light. Mbasi is playing one chief against another, even playing the missionaries against chiefs and local publics. That is pretty much the whole story on the Mbasi game seen (coolly) from the outside. The missionaries, when initially they took the herdsman’s antics for chicanery, were at least being true to what they perceived. I have elsewhere (1994: 204) discussed implications of the voluntary, unforced quality of the Nyakyusa ordeal. By Kinga standards, this is a mere ‘bush’ institution. Sanga succession was to a well entrenched office, while the Nyakyusa postulant received only the conditional mandate of the ubusooka—“Go forth and seek your fortune!” This left the logic of longterm political loyalties fuzzy, and made room for spirited outsiders. What the Mbasi narrative tells me about young Nyakyusa princes—would-be chiefs—is that they were unusually vulnerable to the ‘Mbasis’ of their world, because they wanted to claim, but had not yet got the personal assurance (or call it python-power) they would need to rule, a chiefdom. The evidence Charsley extracts from mission records is clear on the tutelary status of the fledgling ‘chief’ Mwakatungila, vis-à-vis his more established chiefly neighbours. It was Mwakatungila whose unlucky attempts to thwart Mbasi’s ‘herdsman’ (evidently more clever by far at wife-and-cow collecting) found the young prince swinging between chiefly and mission patronage for support, and even hiding out at the mission from his own ‘councillors’. These would be the commoner headmen and interested priests looking to find the right stuff in him for leadership, after a recent Coming Out. ‡ In the canonical doctrine, an established Nyakyusa chief possesses mystical (witchcraft) powers prepotent over any lesser man’s. Yet it appears that an uncanonical ‘medium’ can put the touch not only on a fledgling but on his chiefly patrons as well, for several of the most powerful men of Selya had ceded women and cattle to Mbasi’s herdsman. It turns out that a ‘case of a mischievous 252

medium’ has told us much about the state of both religion and politics in Kondeland. I have called it turbulent and stressed the contrast, if always in shades of grey, to Kingaland. Client-rulers in the Sanga system are given a piece of the political periphery of their domain to develop as they will. Their competition is in the business of recruiting strong followers, not besting internal rivals in power ploys. Client rulers (fledgling chiefs) in a Nyakyusa setting are launched into a harder game, either to win conspicuously or inconspicuously lose. I have argued elsewhere that turbulence is a natural feature of any social structure, since rules (or say, the local body of common law) are norms not reliable statements about empirical behaviour. The whole Nyakyusa political system, as we may now understand it to have been in free-traditional days, was built on the premise that it is by bona fide high deeds a newly chosen prince must transform his conditional mandate into a seat of entrenched power. The kind of protostate process the Nyakyusa knew was actually stabilized by a sort of ordeal of hard knocks, a periodic world renewal ceremony (the ubusooka) unleashing the competition of duly nominated fledgling chiefs for de facto power. If it seems paradoxical to hold that such a measure of chaos could be needed to stabilize a polity, the evidence is (as Charsley argued) that the ‘Coming Out’ of fledglings constituted not a succession but a hiving off which very likely left the parent chiefdom and its important clientele formally intact (and influential if no longer as theatrically ‘powerful’) right through an initial phase of the regenerative cycle. That the Wilsons could have collapsed all this drama into a ‘normal’ scenario of the old chief retiring in order to ‘soon die’ and leave two chiefdoms in place of his one does seem odd; but it follows from the circumstance of looking for structure of a sort that quells turbulence. ‡ The Konde system of oblique succession may usefully be contrasted with the well-known Banyankole system of succession wars in Uganda. There a fully centralized autocracy was instantly devolved with the death of the incumbent ruler to a condition of mortal contest. Eligible princes, each backed by a cohort of local warlords, were now fated to fight to the death for ascendancy to the vacant office. How is it that the Konde case does not conform to the pattern of an “African despotism” we find in the Banyankole and other states of the Great Lakes region? ‡ I think we should look at the argument from amity, the peculiar structural feature which sets the Malatan region apart. There is a brittleness in the amity tie which has no counterpart in kin ties. Only fail to affirm amity in a single episode and the tie itself is put in peril. 253

Political management strategies in ‘amity societies’ are quite different to those in kin-based polities. Because the Nyakyusa people as a whole were never left without firmly constituted central authorities (having many chiefdoms, never in sync), and since they were politically advanced enough to abhor a vacancy in high office, they seem to have achieved stability over a long run through switching loyalties, even though local pockets of low-level turbulence were a chronic feature of the landscape. The genius of the Konde system shows best in their age-village institution. By maximizing kith-based male solidarity, and doing so in an especially effective fashion, these rules gave light but durable weight to locked-in loyalties, greater weight to moral suasion. The peculiarity of this constitution can be called ‘bottom heavy politics’. That is, closed pedestrian communities ultimately made the choice who to follow into battle. Translocal power—power at a high level of protostate structure—was only to be achieved by competing with one’s princely colleagues for the loyalty of a number of autonomous commoner villages. For Kinga on the one hand, the antipolitan option of disloyalty was open only to unaided individuals. For Nyakyusa it was also open to whole villages of able-bodied men. Again, the ‘case of Mbasi’ provides the evidence that an aspiring chief could be deserted by his commoner villages as well as his mentors. I have styled the Coming Out as the critical ceremonial act in a regenerative political cycle. Consider the domestic cycle of a herding community, which has its critical point when the family head is prepared to hand over cattle to a chosen heir. Suppose that the brothers of this heir now have to look for their inheritance to him, not their father. Then add the proviso that the whole generation of that family head must likewise pass the care and ownership of their cattle to heirs chosen from among their grown sons. Then you have something like the Nyakyusa Coming Out. But when the political aspect is considered there is this special difference, that a grandscaled reorganization of life is taking place, which appears to put the security of the community in jeopardy. What begins with the Coming Out is a ‘political renewal phase’ which will continue for a decade or more, as a new generation of offspring are produced and a provisional settlement pattern is fully established. Through the little window on events we have called ‘the Mbasi case’ we have seen that the chiefdom hosting the first Berlin mission station was in that phase of the political cycle, and is vulnerable. A young chief trying to show strength effectively fails and loses his commoner headmen. His mentor, a local chief of the older generation, backs out of the matter. An important plains chief sends up a scouting party to see what is to be gained. And we know not what else. My argument has been that out of this turbulence will eventually come renewed strength. Behind 254

the scenes we see played out there are a hundred political huddles— steering committees, watching each turn and planning moves. Or call it a school. The stability Konde achieved was rooted in a kind of moral strength in solidarity which transcended shifts in political structure. Nyakyusa priests, some of princely hue and some commoner, were stalwart individualists, less precisely inserted into the political structure than priests in the Sanga domains. Probably the rainmaker role represented the core of early Kinga priests’ terms and conditions as adjuncts to an original Sanga court. A rainmaker divined the coming of seasonal rains by reading the appearance of his rainstones, and claimed no inner voice or power to control the heavens, only a special calling to monitor and control the errant behaviour of his magically active stones. We usually make a distinction in the study of divination between ‘objective’ and literally oracular (godspeaking) methods. The dramatic performance of a rainmaker, as described by Kinga practitioners, entailed the covert kind of manipulations which Malinowski called ‘magic’: do it right and it will achieve your aim. This represents a match to ‘objective divination’ in that the interpreter of signs may be the bearer of bad news but betrays no will to deceive. From another angle his situation compares to that of the bureaucrat who claims only to shuffle papers, not destinies. Nyakyusa priests as Monica Wilson describes them are also readers and shufflers of objective signs or symbols. Kasitile himself is a rainmaker with stones inherited down a long line. Though he styles himself one of the chiefly class, what that probably means is that he is heir himself to the (shamanistic) pop-up leader of a pre-chiefly community. As such he represents the stand which identifies with ‘owners of the land’ from ancient times. He styles himself rather vaguely as ‘an hereditary priest of the chief’s lineage’ at Selya and is seen to resent signs of disrespect from commoner priests there—they are veteran village leaders who turn to priestly (lawyerly) work when their soldiering days are done. The title of ‘lawyer’ fits as well as ‘priest’. ‡‡ But Mbasi’s ‘herdsman’ spoke with passion in his master’s name and in a style making him complicit in the power struggle which ensued. This is not the lawyerly style of a priest. The Nyakyusa ethnography offers no real close-ups of the for-profit diviner. The one Kinga analog I know to oracular performance by a spirit medium would have been the occasional ‘prophet’ or ‘prophetess’— politically unaligned—one might still meet in uKinga in the 1960s. These individuals worked as consultants not claimants, taking fees not prizes. They used spectacle dramatically not for personal 255

empowerment but for setting their persons at arm’s length from the problems their clients might bring them, and for dazzling doubters. Would a deity share his wife with a herdsman? It is easier with a Kinga prophet to accept that the wizardry, though mercenary and made explicit, comes from the spirit world. Tunginiye in the years 1937-8, while he was working as clerk of the main court set up by the British, had dealings with a witch-finder who was selling sweet cures to ‘witches’: This woman diviner was called Hikadiseku. She found many witches and always sold them the cure. She came into trouble when the cure killed one woman, and the foetus of another was lost. It was another person, a Pangwa man named Boimanda, son of Mjoukalala, who prepared the cures. He took the money for each cure (a half shilling) and split it with the lady augurer. When questioned [by Tunginiye] she said, “I create rain and rid the land of witchcraft.” She called on Lwembe and the whole Mwemutsi lineage by name. She said all were gods and lived in her body. It was they who gave the orders for all she did. She was receiving fees in cattle as well as money. She would slaughter by night and in secret carry a portion of the meat to Mwemutsi himself. She was a commoner but never harmed by Mwemutsi because he had confirmed these gods were building in her body, and she gave him proper tribute in the form of meat.

It is notable that her list of royal ancestors (god-spirits) was always given in full. All were working together: “Imyitsitsi gyavene Their shades have built in me.” She was a spirit medium (untsungwa) not an objective diviner (untung’anya). She would begin to shake uncontrollably, falling down in trance and kicking the fire about with her feet, scattering coals vigorously at and over her onlookers while uttering incomprehensible phrases. When she calmed down she would sit up, take generous time for greetings from the royal shades, and tell what the spirits had said in response to whatever problem a client had brought. Night or day she would divine in this way. She had many clients from uKinga, uSangu, uNyakyusa, uBena, and uHehe—there was a constant procession. Accused witches must be there. If they denied when found responsible in the case, they could only expect death to the whole family. She would name a price: “Your basin, your cow, your sheep—bring them!” She had become wealthy.

I would distinguish Hikadiseku in the 1930s from our Mbasi herdsman of the 1890s. Both are mystical entrepreneurs, successful at their trade. Both are first and foremost performers on a stage they set and manage themselves. Anthropologists have to be chary about oversocializing their subjects: the Malatan scene 256

is rife with social types of any stripe, people who play their roles for all they are worth, each at his or her own pleasure. The problem a star performer has is to establish the charismatic frame a good act needs if it is to work every time. In a protostate context that means pulling strings. In the Kinga case there was a clear path to legitimation. Hikadiseku needn’t play off one secular authority against another. She pays off the High Prince in the same tender he chooses to entertain and bind his own clients. But the ventriloquist cowherd Mwamafugubo (a.k.a. Mbasi) has no such simple option: political demesnes are fuzzily defined, the mazeways to legitimacy all deadended. He is not destined like Hikadiseku to die beloved of his countrymen. The difference is structural. General evidence suggests, as one might expect on ecological grounds and on the ground of known migration patterns, that Nyakyusa were more open to ideas from the south (Malawi, Zambia, Zimbabwe) than Kinga were. The one apotheosized ‘territorial cult’ intimately known to Kinga was that of Lwembe, located in the valley and shared with Nyakyusa-speakers there. So located, it was outside any Kinga ruler’s own domain. Kinga local priests worked primarily with ‘objective’ devices: rainstones, rubbing oracles, a poison oracle, and above all the reading of natural signs in light of shared lore. A spirit (immepo) diviner could never be mistaken for a priest, as Western observers have tended to do in the ‘Mbasi’ case. For Kinga, the Lwembe shrine worked not as a primary source of decision but as a final device of collective confirmation and redemption, following the prolonged and careful readings and ritualsacrificial preparations of local court priests drawn together from the entire array of Sanga-Mahanzi settlements. Only an ‘external shrine’ would have served Kinga purposes, which required that visits be rare and shared consensually among all the Sanga courts. If the shrine were not external it would belong to one court, not all. Nyakyusa, on the Mbasi documentation at least, can be seen to honour the more ‘mystical’ approach to the occult, which we find among their southern neighbours, and which there too seem to be associated with the need for quick-and-dirty decisions in a more turbulent political context. The Sanga pattern is not to off-load blame hastily but meticulously and solemnly, whether accepting the task of putting a witch to trial by a fiercely formal ordeal or discovering and mediating with a responsible spirit always in the same quasi-bureaucratic manner. Where the sacralized seclusion of a Sanga prince could set him hors de combat and above the turmoil of ordinary affairs—thus dedicated to the plainly religious ends of human and garden fertility—Nyakyusa chiefs were nothing if they were not engagés. 257

Only a forcefully assertive entrant could hope to prove himself in the open political arena born of the ubusooka process. In uKinga there are several famous cases of warrior-princes of the best Nyakyusa type, whereas in the canonical version of the four Sanga high courts, charisma and fierceness were for the youth of a prince-in-waiting, to be put behind him on his ceremonial accession to rule. The (extraprocessual) charismatic war-leaders best-remembered in 1960 were of the final precontact decades. Ngotwilwe (father of or at least predecessor in office to Sihudika in the East) had thrown the gauntlet to the high prince of that realm. This Eastern realm was the one soon to be shown on the German maps as Manduba’s, but the seat of the realm was in a small, hilly domain (Igolwa, only a half day’s walk from Ukwama), which was being dwarfed by the expanding domain to the east. Whether Ngotwilwe was set firmly on the path of secession, or (as I think more likely) envisaged claiming his own domain as the ‘true’ seat of the Eastern realm, we can’t know. This is the realm whose oral history vanished in the Maji Maji slaughter of 1905-6. In the Western realm the two Kyelelos, father and son, were embroiled in civil war for dominance over an expanded realm pressing into Mahanzi country. At contact, the elder (Kyelelo I the Cruel) had finally retired in favour of his son (Kylelelo II the Cruel), and was withdrawn to the older seat at Ihanga, just where he had in his youth assassinated and usurped the throne of his father’s brother to begin his career. Such seemingly errant cases are presumably echoes of the way the original Sanga system grew—here are the vital mechanisms of the protostate process, while the canonical constitution I had from my best informants is the normative ‘structure’ toward which each of the four realms may be thought to have been growing.

The viability of incongruous frames The ideal-type distinction I’ve made between two worldviews, the Konde and the Sanga, reflects a differential strain toward congruency of frame. In viewing their worlds, Kinga tend to reject incongruences which the Konde mind accepts as normal. Certainly the Kinga world thrives on rivalry, but it is of the ‘nesting’ sort we are familiar with from studies of another type of segmentary society based on kinship and clanship. Kinga segmentation is frankly political, though deeply premised on an antipolitan ethic of choice. Members of any segment have ceded a right of taxation to the same court for which the men were prepared to bear arms. Other important rights—access to land and equal voice in a court of one’s peers—are not subsumed under the mantle of executive authority but largely rendered concordant with it in use. It is probably important that the 258

Sanga courts didn’t internally segment their recruits or allow the formation of entrenched power cliques. Freedom of association was built into the bachelor lifestyles of both gender groups but always with this special ballast: dyadic withdrawal was not countenanced. Bonding within the peer tie was not to be exclusive. Konde rules of association dispersed young men into close-knit age villages but focused their work collectively in service to an older group of peers, nominally their genitors, in a long-standing peer village of aging males and younger women and girls—all as if kinship and descent were no basis of identity. But the law of property actually revolved about kinship and descent, quite as if the peer solidarity system didn’t exist. And substantial property in cattle was the key to wealth in wives, so much so that the moral strategies of a Nyakyusa man must always have had the accumulation of wealth in view, quite as though the deepest personal motivation of a man were not friendship or kinship loyalties but self-interest. Still it would be wrong to see difference between Kinga and Konde social arrangements simply as a matter of complexity. Youth of both genders in uKinga lived separately from elders, developing moral strategies specific to their sex. Pursuing a successful lifestrategy in either society, for either gender, would have been about as difficult. There is a clear difference in Konde and Kinga social structures, one favouring congruity and one spurning it. But from all the observational measures we have, it seems both of these structural mazes can be learnt, and even if the neurotic loads they leave may characteristically differ, the typical product of either life-cycle does it no discredit. Both worlds have room for the riskier form of divination—inspiration. Kinga make it conform to the monolithic political model of a segmentary state. Hikadiseku worked at Ukwama, the central domain of a Sanga protostate which imagined itself pyramidal, a segmented monolith. Whatever the divisions, however active the rivalries for power, the Sanga model saw itself emerging from the dust of battle in a pristine integrity. Within the Konde linguistic community the Ngonde system resembled in this regard the Kinga. The Nyakyusa, on the other hand, was aggressively pluralistic. Princelings inherited (by the indirect way of their ubusooka Coming Out) no fixed domain, only place to build a peer-village and a loosely loyal following of other such villages—and a challenge to build this to greatness. The difference in worldview conformed to this structural difference in protostate architecture.

259

We see the contrast in their main institutions. Kinga and Nyakyusa youths both must prove themselves in cattle raids not for their private herds but their elders’. But when it comes to war itself, Kinga men practice in war games, organized like a professional sport, not the flare-ups and skirmishes so frequently cited for Nyakyusa. Kinga and Nyakyusa both organize much of their lives primarily around amity. But Kinga never switch from the amity-base of youth to a domestic wealth-building phase of rampant polygyny and patria potestas. Nyakyusa do so with a vengeance, and create dual structures of generationally stratified kinship on the one hand and narrowly focused peer amity on the other. The two protostates fared quite differently under the pax. For Kinga there was a continual accumulation of decreed constitutions, the new one always jacketing the old, all of them subtly fused in a single ideal model to be used pragmatically ‘on the ground’ in sorting out one’s options according to practical opportunity. In this way time seemed not to erase the past but to continue and build on it. There was less talk than I expected of ‘good old days’ among Kinga, much more about great days to come. Past times in recollection conformed to the same template as these or future times. It was a challenge for the anthropologist to jog an informant back to the specific conditions which must have prevailed in an earlier decade. A little example would be the many informants who would assert that late marriage among men must always have been due to difficulty finding the bridewealth, all in the context of conversation stressing the opposite—the ease with which a young man in the old days might fetch himself the two hoes and goats he would need, once the need was felt. The anomaly reflected no simple lack-logic but the persistence of the feel of life today in the imagined life for an earlier generation. The prevailing premise was cultural continuity, and that is a self-fulfilling premise on the deeper levels of being. The opposite was the case young men made for Monica Wilson in the 1950s, when generational awareness had taken on the colour of class consciousness, and the young were restless for the autonomy a traditional Coming Out had once thrust upon them. The fabric of culture is rich or poor just as tapestries, texts, or musical offerings may be. Comparing Konde and Kinga cultural styles, it is especially the amity base they share which seems to enrich them. I have argued elsewhere that, since amity is structurally contingent where kinship is axiomatic, the two kinds of tie are fundamentally different. Cultural enrichment flows through intimate relationships of either kind. Contingent relationships enrich a person in the way we usually call ‘broadening’. Ties of amity to endure must be reaffirmed at every meeting; the ties of kinship, on which so many 260

Bantu societies are dependent, perdure the kind of fracture which ends friendship. The bright side of this shines in the wisdom of the West and may be taken for granted. Of the darker side it may be too simple to say most of the cramping done to human psyches occurs when an ego is ‘trapped’ in a relationship, but the remark agrees with general experience. Konde and Kinga share a way of life which keeps a male child from being ‘trapped’ in a too-close parental tie; Kinga provide this privilege also to female children. A community which allows scope to amity in its social architecture produces individuals who have established freely chosen intimacy with a series of other individuals on the basis of social compatibility. I know this gives rise to lively mind-games (songs, stories, skits, riddles, innuendo) among Kinga lads and among Kinga maidens in their private houses, and can’t doubt the same applies to Konde youth. ‡ Another kind of cultural enrichment is the one best known in the West, coming from culture contact and the selective adoption of foreign traits. Konde are deeper into culture contact than Kinga, and I think that shows in their cosmologies. Konde are the more eclectic, Kinga doctrinaire. That means it is comparatively easy to frame general propositions about Kinga beliefs, comparatively risky in the case of Konde versions of the Malatan cosmos. Animist worlds invite mindplay with respect to the multifarious agents of mysterious trouble. Whole-cloth acceptance of Christianity or Islam by Kinga men seemed hardly to cramp the cosmic space available. Part of the reason was that certain of the stumbling blocks one might anticipate were not especially relevant. Kinga men were not rampantly polygynist, as Nyakyusa used to be. The pressure was taken off bachelors by their chumminess in casual relations and their freedom from domestic duties. Monolithic implications in the Christian cosmos were not perceived as cramping to individual freedom, just as the antipolitan ethic had always left a way out of political impasse. One of my friends was a leading church person until he fell in love with a second woman. No one really made it hard for him. Kinga women found congregational life to their liking, particularly in the extended bachelor phase of their life cycle, which had always been much given to (informal) sodalities. Kinga maidens made the Lutheran church their very own club. But part of the reason Christianity could be accommodated without much grief and accepted as continuous with a pre-pax past was that it fused easily with their beliefs about their cosmos. Monotheisms are, in their own view, social contracts of the either/or sort. Animistic worlds make both/and an easier moral strategy. Kierkegaard, I expect, if he could have known the Kinga well, might dismiss them as humanists.

261

If what was new was the phenomenon of ‘worship’, what Lutheran worship built on was the Apollonian style of Kinga celebrations generally, their quick switch to gaiety, the lightness of their expressive style on jubilant occasions. The difference to Konde style is not abrupt but is deep. Social forms may look alike, but the dramatic side of a dance is less conflicted, more predictably amiable in outcome. Nyakyusa funerals are frenetic, Kinga come together quietly, keen mainly in low voice, and depart quietly as well. Ruth Benedict would surely find little of the Dionysian about them, though this is not to say drinking parties can’t lead to fighting of an unscheduled kind—fights “always forgotten by tomorrow.” ‡ Here is a remark of Monica Wilson explaining the intensity of a Nyakyusa funeral and their problem in handling anger: The Nyakyusa hold that merely to suppress grief or hatred is useless, and likely to increase it, and they constantly press on one another the need to ukusosya, to ‘bring out’, to confess, that which is within. The admission of anger against kinsmen and affines is the prelude to reconciliation; the open expression of antagonism between two lineages, which are then reconciled by the exchange of gifts and feasts in the Lugulu ritual of vituperation, is but the most dramatic instance of a common theme. ‡

Nyakyusa amity, as it has been described, is unclouded and unfeigned. This is the edenic side which so impressed early visitors. But it builds easily to rivalrous quarreling. Young bachelor men, as their elders say, are kept away from women to make them better fighters. Youths may be killed in the hot-house frenzy of a Kondestyle funeral, so repugnant to early missionaries and so intense (not to say ‘pagan’ and ‘erotic’ or ‘orgiastic’) in atmosphere. If the key to Nyakyusa self-display is ‘swagger’ the key to Kinga is closer to ‘frolic’. Kinga young men are not working for their fathers, are not in Angst about women, and tend to prolong their bachelor days, if not in law then in lifestyle, into the midlife years. A bit of evidence may suggest the way violence is evoked and socialized in the context of Ukwama high court. As it was sketched for me, though I never managed to walk out the actual scene, there is or was a dance ground at the court which has given its name ikivilila to a charming little chant beloved by children for dancing. The ground and the hill rising above it are called by the same name, and it is on the hill that a Dionysian revel was held after a successful cattle raid, probably on lowland herds. A bullock was started in full career from the top of the hill, pursued by young men trying to cut chunks from the desperate beast on the run. The clamor ended, of course, at the foot with a tangle of bloody men and badly slaughtered beef. It was 262

thought a great spectacle and high fun for the participants if not too much of the blood were their own. But as with a boys’ cudgelthrowing fight or mock warfare, no enmity arose even from an accidental death. Anger was simply not on the menu. In the feast and dancing which followed, the Dionysian element was gone. The language of discussion for the ‘deepstuff’ of cultural differences is necessarily improvisational and, if not imprecise, always notional. I’m ready to cope with that at need. But here my point is only that worlds are not always built around the strain toward congruity of structure. The alternative way of building a translocal world is not necessarily embracing chaos though it may seem to naturalize social turbulence. Are Kinga laid back while their neighbours in the Valley hills are truculent? The difference, because it goes deeper than that would suggest, is not such an obvious one but no less heavy. Doubtless, the relative remoteness of Kinga realms in the Livingstone Mountains exposed them to less constant threats from north and south, leaving them a marginally more secure existence. This and the willingness of Kinga women to do a lioness’s share of the gardening, plus all the preparation of food, would have meant greater leisure—if that is a suitable word—for the men and their state-building ventures. Animism is not special because it accepts plural frames of reference in the search for moral certainty. All systems of thought have to do that, just as all systems have their own way of warping a frame. What monotheisms particularly do that animisms do not is declare intellectual war on other peoples’ frames. One key to understanding the difference can be pursued under the familiar topic heading of taboo. What distinguishes a taboo from a moral law is the absence of rationale. It is not that the reasons for taboos are ineffable, not that fear of punishment for transgression is irrational, it is simply that maintaining the taboos of your society is the prime rule of membership in it. Toilet training is a good example to have in mind. Iconoclasm is another, and complementary. Rising in fury against the icons of your church is a decisive act because it is taboo, and may sometimes be the one direct way out.

263

FEAR & BLAME, CHAPTER SEVEN

Taboo and Law The Everyman narrative Let ‘the Everyman narrative’ be the story of a human life from cradle to grave, sketching what it means to pursue a normal moral career in a specimen culture, in either gender, anywhere in history. This is not the ‘moral’ concern of a morality play but a generic story about coping with forced social choices in a well peopled arena where the game is placing moral claims on others, and reacting to the claims they place on you. Kinga fall neatly into just two gender worlds if you allow for special rules governing the lives of a smith, the leaders of a court culture, and life within a royal enclosure. The uKinga of 1900 can be thought of as one single arena or as a set of smaller pedestrian domains pyramidally grouped in four realms. By the same token, taking a broader focus and considering only claims of the translocal sort, uKinga is only a segment of a larger arena of moral encounter. This arena has been projected here either as the Malatan or the more inclusive Sowetan region. The kind of moral claims a child can lodge upon a parent or peer are the child’s to discover through incident as often as instruction, though modern life has seemed to alter the balance with formal schooling. A lifetime narrative under classic human conditions moves through what Meyer Fortes called the internal domain of social relations and proceeds with growing sophistication into the external domain of public institutions. In each successive stage, the ordinary person of either gender confronts new problems in conceiving a successful strategy for conceiving and securing a good life. Included in the Everyman narrative are such choices as conceding or claiming advantage, honing or neglecting skills, keeping up with tasks or letting things go, hounding or begrudging or letting-be in close relations, and in-public policies of flagrancy or modesty, narcissism or comity. A thousand and one shades of moral value may be applied to anything you say or do in any community, from a sandbox to a city, and each choice colors your life. The idea that life in some far corner

264

of the world can be unmeditated, a sightseeing drift down river never lifting the paddle, is airy and not for Everyman. ‡‡ Let an anthropologist refer to the ‘culture’ of some human community, however large and diverse or small and homogeneous. This ‘culture’ will set up its own versions of the Everyman narrative, always in some ways unique as a set. Each empirical example of human culture is peculiar enough to tease out new soundings of that never-to-be-defined ‘human condition’ which haunts every revealing ethnography. Sigmund Freud made a famous attempt at framing the Everyman narrative a century ago, at a time when there was little help to be had from anthropology. Jean-Paul Sartre, half a century after, did a better job of dodging Western ethnocentrism, graciously sparing us Freud’s terrible infant, blindly driven by the twin daimons Ego and Eros. But Sartre’s hero of Huis clos (“Hell is other people!”) rather suggests adulthood is only a later stage of the same predicament, no less grotesque. I propose that a more simply narrative approach suits anthropology better. There can be no pat, culturally given life-cycle. Every life is a career. What we have to deal with is naturally governed existence within a series of culturally-given situational frames. The Everyman narrative watches children (often rather complacently) and adults (less often so) developing the moral strategies they hope will get them where they want to be in life— always, to be sure, with the special spin their corner of the world has to impart. At the existential level of experience, what and where a person ‘wants to be’ remains chimerical all the way, whether evolving or episodically cycling. This fate nags us everywhere: think of Eugene O’Neill’s pipe dreams, or only consider what novelists have done for our comprehension of ourselves. The Everyman narrative has been my tool for understanding the structure of human experience in the regional culture which seems to centre on the Kinga people. We are given a lens which lets us focus on the social structures a person encounters and negotiates in a full lifetime. We see the structure as one of experience not third-party observation. We are not concerned with the way a person turns out—if this narrative is a sort of Bildungsroman, we read it for the world our protagonist has suffered and enjoyed, not the personal fate that befell. Two caveats apply: (1) Think of the ‘structure’ as a pliant one (say, a wilderness forest not a hard-edged maze) but also a tricky one which doesn’t often turn out quite as billed. (2) You have to imagine for clarity that the person in each narrative experiences one social structure throughout, not a radically changing social environment. This last caveat goes contrary to fact, even in traditional social settings. But notice that the biography of a person who moves about among several classes or 265

cultural settings doesn’t disqualify as an Everyman narrative. We can still read it for the social structures encountered and the strategies evoked, not just for keys to individual character. But as disparate contexts multiply, useful comparison diminishes. I propose to compare the structure of experience in three cultures within a single region. The peoples are related through a long history of contact. Culturally the Kinga are most closely related to the Pangwa, who traditionally lived very much in the style of traditional bush-culture Kinga. Politically, Kinga are very much closer to the Nyakyusa (Konde). I make an initial declaration here that incongruities in the Pangwa-Konde-Kinga series are manageable. I find their differences instructive, not random, not baffling. There is no implication that I see the series as rectilinear. A better figure would be three paintings by the same hand, but on distinct themes—or the other way around. I deal with the same range of problems (‘core problems’?) in each case. All else aside, and acknowledging the ubiquity of doubt, there is one thing gained by this kind of reconstructive approximation of human experience in a small region over a telling stretch of its history. You trade off some of the hypostasis entailed in taking a personality type as phenomenally real. This saves some of the discomforts of matching ‘psychological’ with ‘sociological’ models. Your return on the trade-off may be achieving better empathy with persons encountering worlds different to your own. Is a traditional setting more likely to evoke successful moral strategies than a turbulent one? I see room to doubt even there. The kind of intuition entailed in moving from culture to mind is pretty private, and an observer’s judgement can be only as scientific as that proviso allows. The Konde scene was famously turbulent and famous for felicity as well. We are at the star-gazing stage of human astronomy. What I want to do is to pursue the implications of this analytical approach to the three peoples, Pangwa, Konde, and Kinga, considered as they were in the final decades of the nineteenth century. I treat them as an ordered morphological series representing three stages of developing protostate politics in the Sowetan cultural archipelago. I reject the idea that each of these political cultures had developed in isolation, even while I find them interestingly distinct. Each specimen culture will be a reconstruction of the precontact version, based on a critical reading of twentieth-century field observations and (virtual) survivor interviews. The point of the project is to watch, at each stage of progressive political change, the contact of social structure and individual experience through a lifetime. As a person moves through a 266

given social maze at the speed of time, person and world momentarily merge, emerging a moment later marginally altered. The point of taking the three cases together is to present the structure and experience of three historic versions of the regional culture, so aligned as to represent—as touchstone models—three successive stages in the phase movement from ‘acephalous’ to ‘pyramidal’ political systems. In this four-dimensional modeling the principal sanction of social order self-transforms through its motion along an ideal continuum from ‘taboo’ to ‘law’. But ‘taboo’, as the term is used here, does need special explanation. There is a lot of wisdom to dispel. Taboo is the customary basis of consensual social alignment within those social worlds which lack the institutions required for setting up a rule of law. Where courts of law do exist, though they sound the drum for consensus their main work is to make life possible without it. Where a new consensus does appear, it appears as it always has on the plane of taboo. The incest taboo is the most prominent consensual (‘spontaneous’) prohibition recognized around the world. The taboo is part of a consensual arrangement which prevents the formation of swollen, monadic kin groupings within a viable pedestrian community. It works in a stunningly simple way by outlawing (more correctly, by defining as utterly dangerous) sexual relations between close kin. The effect is to prevent the build-up of solidary in-marrying kin groups, owing so little to the members of other such groups that there remain to them no firm structural bases for translocal amity and (hence) for a pax publica. This can be stated as a rough syllogism which applies universally to acephalous societies, where domestic groups must be possessed of significant autonomy: Public safety in a human community flows from a premise of amity endorsed and sanctioned by ordinary, self-reliant domestic groups. A community which doesn’t prohibit incest allows ordinary domestic groups to opt for endogamy, with the substantial risk of their casting aside the premise of amity toward non-kinsmen. Therefore public safety is endangered when domestic groups relax their negative sanctioning of incest to become in-breeders.

Many kinship societies are organized elaborately in such fashion as to form a system of unilineal kinship ‘segments’, each of which breaks down into sub-segments until domestic groups are reached. When you come to the households in this way, it appears that the domestic group in such a segmentary society is simply the ‘least segment’. But another way to model these societies is to ask in what groups (what level of segment) you can expect to find a maximum emphasis on self-sufficiency. This is tricky. If you were 267

looking for symbolic expressions of solidarity, there are a number of well-studied societies where the clanship (highest segmentary) level would be an obvious answer. If you were looking for military adequacy, you might be led by the scale of fighting forces people say were most often mobilized in ‘the old days’. But if you looked for minimal social and economic reliance on (and responsibility for the welfare of) members of other ‘segments’ at the same structural level I think you would find yourself looking at domestic groups. In a school playground, who is going to watch your back for you, if you don’t? In a kinship society, however strong the ritual rhetoric of cooperation with kin, who is going to feed and manage your children if you don’t? These were not merely abstract concerns in Malanduku « [Twin Shadows], though the answers you would get there to those questions would scarcely point to any easily recognized, solidary ‘domestic group’. Still, the existence of domestic groups considered as solidary units can be asserted even for traditional Kinga settlements, though the family members are spatially dispersed. It is something we discover when we inquire into taboos on intimacy of the sorts commonly recognized as incestuous. Does dispersal in space, by reducing everyday intimacies, concomitantly reduce the call for strict incest taboos? It seems not. Better consider the possibility that the dispersal itself is an expression of such taboos. A comparison of Pangwa, Konde, and Kinga uses of taboo has to begin with the fact that the incest/marriage rules in all three societies are the same, though they present widely different social organizations. What they do have in common is a negative element. It is a sticky exercise to define ‘the family’ in any of them. In Fr. Stirnimann’s reconstruction of the Pangwa ilichumbi (homestead, Einselgehoft) you could say the family is the group eating from a single kitchen-house, though the two genders eat and associate at different fires, and it is by no means clear that the principle of gender segregation doesn’t go so far as to prevent easy intimacy between brother and sister or even husband and wife. Dispersed living arrangements begin at an early age for all Kinga children, and only a year or two later for Pangwa; in ‘the old days’ the same applied for both genders in uNyakyusa, though colonialism saw girls marrying straight from their mother’s homestead. Cooperative work between mother and daughter is important in all three societies, but otherwise intimate cooperation is among peers or at a more formal level supports inter-familial relations. Two significant generalizations seem justified. First, incest taboos do play a crucial part in anchoring the social (moral) identity of individuals in the ‘amity-based’ community. Second, a closer look at the way ‘taboo’ plays out in the lifetime moral strategies of 268

persons in either gender group will show that ‘fear of incest’ is only one aspect of the problem of intimate rights which haunts the Everyman narrative everywhere. Since I conceive the nature of taboo as both more interesting and more elementary than most do, I’ll try to make my notions clear in the pages which follow in this section, before turning to Fr. Stirnimann’s Pangwa. They were, in the image his informants elaborated of their past, a swimmingly taboo-ridden people.

Incest and the amity nexus. To understand how the replacement of a ‘kinship nexus’ by ‘amity’ bears on the matter of incest taboos I think we have to back up and start with more elementary considerations. Take the premise of universal envy, combine it with the premise of universal competition for goods in short supply, and look for the kind of rules which will minimize the size of the groups involved in such rivalry. We do this in recognition of a prime principle of social organization: the smaller the group granted autonomy in the pursuit of self-interest, the smaller the difficulties which are likely to arise from a collision of such interests. But an equally prime principle is that sheer anarchy doesn’t work. One set of rules which will do a fair job of keeping order in society, and for which anthropologists can supply endless examples, is a set which begins with tabooing incest within the domestic group, thus making ‘families’ dependent for spousal candidates on other self-reliant groups of the same order; then moves to exogamous rules which (a) enlarge the scope of the ingroup created by an incest-like sexual prohibition, and (b) invest in this extended in-group the responsibility for regulating ‘outmarriage’ in a consensual and systematic fashion which minimizes direct ‘insider’ rivalry in the competition for spouses. Political order is enhanced and stabilized in societies which move to this kind of kinship organization, but the move is away from the fully decentralized kind of order associated with taboo, and toward law. Here is a quick sketch of the possibilities. Because law wants universality and kinship is a doctrine of particularism, the growth of a state where kinship institutions are entrenched will have to elaborate a politics of devolution which allocates set powers to extant (particularistic) kinship structures, overseen by some sort of ‘federal’ or monarchial system of interethnic government. This is likely to have a more manipulative than legitimately authoritative stamp, if in fact it does not take the more direct form of tyranny. But where a protostate process begins with a strong institutional base in managed amity, a person’s kin-group or ‘ethnic’ identity doesn’t intervene, and government through local courts will more readily cultivate and spread a universalistic law.

269

But a viable political system which doesn’t have roots in ‘family ties’ remains to be described in the world’s ethnography. A good handful of societies are well known, which seem to keep family solidarity minimal by locating primary autonomy at the level of a local band comprising several small family groups in close intimacy. This may be thought of as ‘treating the band as a family’. The Kinga system is by contrast a micro-diaspora—tight family bonds which however are not allowed to compete with the members’ intimacy with unrelated peers. Now ‘tight family bonds’ arise from the web of taboo (ritual) not law, and constitute the primary frame for establishing the web of personal identity without which law can’t function.

Intimate rights. The obverse of an avoidance is a right to intimacy. So taboo is a window on the intimate use of personal space. A thief who violates your property has invaded your personal space. Having no sanction on the thief who won’t be caught, the law has no power to dissuade the secretly self-contained culprit from doing another person wrong. Isn’t it then miraculous that our property can be so nearly safe so much of the time—that the smashing of glass and snatching of bags doesn’t shatter our lives every day? There is no miracle, only taboo. Tourists soon learn where local taboos don’t apply to them, but find it hard to comprehend the casual attitude of local police. Police, for their part, are disinclined to bear blame, since their protection work is given credit for a tolerable crime rate within their own community. Operating on diffuse sanctioning, taboo is always local. Taboo is underrated as a universal safeguard of intimate rights precisely because it is so overrated (stereotyped) as a form of magic. ‘Primitive Man’ is supposed to have lived in such fear of breaking taboos that he dare never misbehave. Actually, as with the rest of us, people living under local tradition (wherever in the world) were dependent on the spontaneous learning of local ‘rules of intimacy’ by children, starting long before the idea of a rule entered their lives. The topic heading for a discussion of this learning process in the diction of social psychology is likely to be ‘socialization’ not ‘taboo’ or ‘civility’; some anthropologists prefer ‘enculturation’, which tends even more directly to suggest learning imposed by drill. Correction is, as in language learning, doubtless spurred by drill to some extent. Like musicality, ‘intimate correctness’ is more readily learned by some than others. But it is a kind of learning we share with other grey-brained creatures. We are a special case because the learning which starts with getting along as a nursling continues seamlessly into the acquisition of acceptable role behaviour in less instinctive forms of food consumption, proxemics, toilet tasks, and play—gentle and rough, fun and competitive, erotic and defensive 270

experience. If parents can’t depend on a child to do most of the work in the development of successful moral strategies in the early home years, it will take a mighty struggle to produce a half-way socialized adult. ‘Taboo’ connotes not only avoidance but safekeeping. Every episode of intimate interaction has a two-way sanctioning dimension. Is it not taboo to wound your friend’s pride? By its nature, the phenomenon calls not for mutuality but reciprocity— asymmetric but not one-sided. The kind of role behaviour which makes intimate coexistence possible entails kinesthetic, linguistic, and moral fitting of the self to the invisible forms of others. This has to be ‘learned by doing’ and like correct speech has to be intuitively understood, abhorring ‘explanations’. Identity also abhors explanations. A child learns who he is from learning who he isn’t, whose he is from whose he isn’t. Later on when the grown child has to ask the who-am-I, where-am-I-going questions they turn out to be about the half-remembered mind which did that learning. What else could they be about? Not what the child learned, as that’s easy. Not what the elders (the documents) would say if asked, as that is for un-persons. The question is how and how far and why the mind looking back is one with the mind which first looked out and around at a very small world. That world was defined, of course, by the conjunction of a set of ego-centred incest taboos each defining and limiting intimacy rights and obligations as between two persons, and each connecting those persons to a webwork of such ties stretching backward and forward in time to the ends of relevant social space. Rights of intimacy generate obligations of intimacy, obligations generate rights. This webwork is the structure of a moral world. Every member of that world, living or dead, occupies a tiny part of it and is its centre. Where there are not law and courts to apply it, only this webwork of privileged intimacy and obligation can be activated to ratify an inheritance, assign a motherless child to care, demand a delict compensated by a fugitive’s kin, or establish a basis for trusting a stranger. In the world of Malatan tradition, a growing child gets to know who is close kin, and of what kind, from kin-terms, privileged relations, visits and hospitality, and awareness of an extended incest taboo. The rights of intimacy are quite as mysterious as the avoidance rules, and are learned at an age when mystery carries a burden of high seriousness. So the moral map of a child’s world is deeply set in mind. It is because (to the extent that) each individual map overlaps and reinforces each other within a family group, and because each must learn to live up to his or her rights and obligations in accord with the common map, that individuals acquire and fix an indelible identity. The addition of incest avoidance to the map as a 271

child’s sexual awareness matures would regularly carry a new burden of meaning with it. A moment’s consideration makes it clear that as publicly sanctioned law develops it will tend to eclipse the autonomy of domestic groups and any close kin-based alliances among them. These kin-based leagues begin to lose their status as translocal sanctioning agencies. What is not immediately clear is how much and how soon the autonomous qualities of the domestic group will be annulled by the advance of a protostate process. In short, how far does public order continue to rest on the willingness of its effective kin-group units to maintain a taboo on incest? and how well can the ‘strong force’ of centralized power use naked rules of law to maintain the pax publica without support from the ‘weak force’ of radically decentralized—call it ‘internalized’ or ‘spontaneous’—taboo? Most of us will probably admit on consideration that rampant egoism and predatory gang formation within a large public is more likely to be associated with turmoil and the abeyance of legitimate means of social control than with a semblance of spontaneous order. The wellstudied association of Balkan familism with feuding is a case in point. So far as it appears that the maintenance of mutual respect among citizens depends on grass-roots support for a fundamental ethic of amity, and so far as it also appears that disappearance of that ethic endangers the legitimacy of government, it follows that law sans taboo won’t achieve minimum strength. ‡ Taboos are especially associated with life-crises and with the kind of individual experience which redirects a human career. Taboo is the essence of that ritual formality in court procedures which can impart authority to otherwise only-too-human pronouncements. And as effective taboos are features of ‘spontaneous’ consensus, of worldview, they won’t be as vulnerable to change as overt and practically motivated patterns of conduct are. All the same, the modern West has not been particularly strong in maintaining life-crisis rites, sanctioning altruism, or insisting on the just application of law. These are values which can’t be compassed by law in its secular mode. The most famous and nearly universal of our taboos, which is that on incest, has on almost every showing proved to be unenforceable by standard policing methods, and too often only eroded by the attempts of public welfare institutions to co-opt this responsibility from the domestic group. I believe the issues around ‘taboo’ deserve better clarification than the learned discourse of the West has so far put on offer. By looking at the matter in crosscultural perspective, we may also gain a sharper sense for what law can be and can’t.

272

Taboo has sometimes been assumed to have merged with and disappeared into law in the course of recent human history: taboo being taken as touchstone for a ‘primitive’ and law for a ‘civilized’ world. But quite apart from the burden of hubris implied in the assumption, it is not the case that taboo disappears into law. The microcultures described by anthropologists in the century past have ranged from lightly to strictly ritualistic. And where is our own to be placed? When we consider sociological studies of ‘modern society’ it is not only the military and ecclesiastical orders which look heavily ritualistic. Linguistic, political, and gender rules can be laden with penalties no mere, self-supporting jural system could maintain. Their sustenance is rather from the unquestioned acceptance of the subject population—Durkheim’s ‘diffuse sanctioning’. If there have been times in your own career when ‘crime in the streets’ was no public problem the reason was not better laws but greater ‘spontaneous respect for the rules’ (taboos) guiding civility—rules the most enlightened law cannot implant. If there have been times when unenlightened mob rule has taken over a community, the reason was the weakness of jural interventions meant to sponsor programmatic change in the ‘spontaneous’ (taboo-based) values of that community. The range of felt threats to social order anywhere, whether in the synoptic comparison of different peoples or in surveying longrun oscillations within a single culture flow, is only from light to grave, never non-existent. But if Hell and Shangri-la sound like storybook stuff, still the non-fictive differences within the range between can be monumental. Taboo and law are not alternative means to one end but separate formulae for social control which work together in organic fashion. One source of our problems has been the use of ‘ritual’ (as in distinguishing ‘the political order’ from ‘the ritual order’) as referring to the master class containing ‘religion’ plus ‘magic-and-superstition’. ‘Ritual’ was chosen (reluctantly, I believe, in the case of Meyer Fortes and his circle) in an effort to escape the knotty question where the line between two religion-like phenomena shall be drawn. The usage unluckily seemed to imply there was no ritual in political life, in marriage and adolescence, in the framing of moral careers—and no politics in the ordeal. For ‘ritual’ in just such context we may have to substitute ‘taboo’ to suggest how wrong the usage was which seemed to deny deep mystique to a Hitler rally or a very high tea. The popular romance of taboo among Western thinkers was nicely laid to rest by Franz Steiner at mid-century. We learned from his critical lectures (on the several ways scholars had treated the topic) that taboo is an aspect of the sacralization (ritualization) of 273

social danger. For ‘sacralization’ when applying the word to an institution one may sometimes read ‘casting in bronze’. Institutional wisdoms pre-empt individual liberty. As a sailor’s channel buoys are meant to narrow down his freedom of secure manoeuvre, an ‘incest taboo’ narrows the freedom of action between any self and certain others, and does so in a way we had as well call mundane as mystical. A taboo is a marker in social space which fends you off, just as perfume or a siren song may be a marker drawing you (perhaps with just the right degree of fear and trembling) into a zone of danger. Thinking of taboo as a bugaboo stirring instant trepidation is dramatic but more fanciful than scholarly standards require. Social markers may be the more effective the subtler their form. When the institution in question is as close to your roots as your natal domestic group, taboo merges enigmatically with identity. To see that social identities couldn’t begin to exist and serve their political purpose without the taboos that ordinarily protect and stabilize them, we only need consider the appalling breakdown of social order in a spate of ‘new states’ in Africa, starting in the middle and restarting toward the end of the twentieth century. I think particularly of arming preadolescent boys with the high-tech weapons of mass murder. Gaudy or even fearsome jural fences may serve to privatize but often at the price of inviting transgression. In my lifetime there have been sporadic efforts in the West to pull down taboo-barriers around public nudity. They have generally been dealt with softly and grown tame or petered out. So far, calling in the law has usually been reckoned an overreaction. That is in good part because law-making of this kind has a poor record of producing durable order. Body-modesty seems to be ‘spontaneous’ enough to reassert itself after a nudity fad wears thin. Asking why this should be so gives the meaning to ‘spontaneity’ which was intended in Michael Polanyi’s essays on ‘spontaneous order’. He argued that “such tasks as a system of free adjustment may achieve, cannot be effectively performed by any other technique of co-ordination”. A nudity taboo is by no means a universal in the annals of history and ethnography. It demonstrably isn’t a feature of ‘human nature’. But in societies where clothing has made inroads it seems to thrive by offering a valuable, controllable tension in the way we present ourselves to others. Only review what Marshall Sahlins in his Culture and Practical Reason had to say about our ‘clothing systems’, their syntax and the use they might have in making the several parts of a society less (or occasionally more) ambiguous. The abomination of public nudity has many sources and reinforcing conditions, from the aesthetic to the practical. But as with all taboos, the element warranting its persistence is its ‘spontaneity’. ‡ 274

My brief here is not that taboo is or isn’t a feature of human nature. I think the only taboo which scholars have quite often attributed to human nature is the ban on incestuous sex, and that scholarly position is controverted by clear evidence. The Zande ruling clan is one of those ‘swollen kin groups’ the incest taboo elsewhere militates against. We simply have there a politically deliberate exception to the rule, which tests and refines our understanding of it. Here is a clan which as a political expedient can afford to encourage brother-sister incest. The main gain in the king’s eye is perhaps to forestall the court’s young princes fraternizing with commoners, magnifying the risk of royals developing independent loyalties and ultimately obscuring in a new generation the divide between ruler and ruled. ‡‡ The recent history of Sanga exceptionalism illustrates. The penultimate ruler of the central realm made a first marriage to a half-sister. They constantly quarreled, until the marriage was set aside. The sibling-wife was sent to live apart with their two children, as no part of the court community. Tunginiye, through whose sharp eyes I could approach the narrative at arm’s length, did not see the turnabout as intervention by the powerful shades of a princely line. The simple fact of the marriage demonstrated the special nature of a royal sex-life: as there was no spontaneous alarm from within the group of kin, there could be no question of a taboo, just as there was no confusion of the huge royal enclosure with a domestic group. The trouble was ‘a marriage too close’, unsanctioned but unblest. It was commonly observed that the partners to a marriage ‘almost taboo’ could expect no happiness from it. But taboo unlike law is not negotiable: if it doesn’t arise spontaneously it doesn’t arise. It remains ‘man’s fate’ in its existentialist meaning never to be free of that wariness (whether of changing or holding course) which comes of the subjective awareness of alternatives—they range from a change of style to Exit itself. The little exits of life eventually lead to bigger, and days hardly pass in your moral career without your having to affirm or deny by some small act the private moral strategies your secret vocation builds on. Kinga women, particularly, reach an age of crisis gradually, where the joyful affairs of bachelor days wind down and give over to the more inward adventures of childbearing. The strict taboo which channels a maiden’s exit from the isaka of her youth is that institution’s prohibition of returning there, even for a single night, after knowing coitus. As with any taboo, the system tightness can be seen to depend on its (felt) specific gravity. The viability of the organization depends on the tightness being right. But the persistence of this taboo right through the broader Kinga community generations after Contact is an indicator 275

of the spontaneous rightness of the rule in isaka thought, for no other institution was normally involved. For maidens, the bachelor years were the peak period of private pleasure of a lifetime, and those gregarious maidenly years extended over the best two decades of a woman’s life. Does it seem paradoxical that the taboo of coitus, so negative in itself, should underpin an existence of high spirits, self confident friendships, and (often enough) confidently generous conduct? Readers who are agreeably acquainted with one of the Lesbian sisterhoods on scene today will not find it so. We have again the case of a taboo with ‘spontaneous’ staying power, even while the oblique appeal of Christian young women’s sodalities was making inroads in the 1960s. I think it behoves us to accept taboo as an effective mechanism of social control, ontologically prior to coercive law, and universally present as the rule-stuff of which all human institutions, beginning with the family, are made. The inherent privacy of individual minds is fundamental, but conversely too the very accessibility of culture codes of conduct. Taboos mark for an individual not a highroad to the good life but the pits and deadfalls in an otherwise open terrain—straitening the way by disallowing this choice or that but never making the critical decision or even truly protecting Ego from fatal errors of his own. All the time, or almost, Ego is scarcely aware of how much depends on others staying on track—just doing so for our own part is all the moral burden most of us can bear. What may seem to be ‘rules’ or ‘roles’ from the empyrean view of structural anthropology feels more like a jungle path from the inside. Do we each have our own, invented ‘taboos’? Grant that deeply deviant communities tend to develop within large-scaled societies, and grant that individuals in any kind of community can make free to act, go, or be ‘crazy’—trashing the rules they know only too well. But quite personal taboos contributing to the organization of ordinary lives? The word ‘taboo’ can seem extravagant in application to a person’s way of steering through and around the moral hindrances of day-to-day existence. These are the countless infinitessimal acts of avoidance and self-preservation which we learn from living—which accumulate in idiosyncratic patterns to form an ego and evoke a persona we dare present to a friend, a lover, a parent, a world away from home. Picture the process as the continuous psychic sedimentation inherent to the Everyman narrative—not a ‘formative’ process because never finished. We may talk of conscience. We may talk of the acquisition of tastes or standards. David Hume famously showed us at once their urgent privacy and public necessity. When and how does a standard of taste, an aesthetic, a form of art emerge from the privacy of conscience into generally sensible 276

existence? Isn’t it from private experience repeatedly coming to expression in a contagious form? and isn’t that most likely to happen when private intention moves a person toward the intuition of value? The ultimate result when there is continuity enough and generations of time is the emergence of a quite particular community’s quite particular culture. You hardly need read much news of the world to realize how little the private arts of living in amity with colleagues in a culture have improved in a century of stunning technological advances. The explanation we most hear is dissonance: dissonance in the cultures which have taken form in the century past, answering dissonance in private experience. It is a time when we are prone to seeing every conceivable taboo being broken. It is a time when cultural dissonance may seem to have filled the prisons of the richest countries in a trend beyond the scope of law to control. Does it make sense to ask if taboo, considered as the matrix of law, might have a regenerative power in such a world? (How many world leaders today are wearing the signal clerical garb?) So we are confronting a shifting balance of uncertainties. What we face today is peculiar to our age and place, but the predicament itself is timeless. This I take to be the universal setting of an Everyman narrative. Taboo, like law, rides the balance everywhere. Law, being inherently public, can’t serve as source for Taboo, though taboos, as in the case of divorce in the twentieth-century West, may be displaced by law. A Roman father enjoyed the right of life or death over his son, but this law of patria potestas did not make the exercise of that right morally neutral. Rather, it was the enormity of the act of filicide (still clearly taboo) which called for a rule of law to warn a wayward son of the real danger courted in disobedience. Here is Law paradoxically serving the function of lacing a dormant taboo of an offspring’s presuming upon a father’s love. Every changing society has to map the domestic group in ways which keep it viable. It was the Roman choice to set up a law making Father not State the ultimate judge of transgressions within the family fold—forerunner to yesteryear’s ‘my home is my castle’ and a far cry indeed from the mapping of domestic rights in successor nations today. With taboo is invented the notion of transgression but perhaps not (in an after-Freudian sense) its link to guilt. Break with a taboo and the danger is not usually yours alone, nor poetically sudden. The fallout strikes randomly on you-and-yours. Since incest by its nature falls inside that compass of you-and-yours, the putative danger of breaking with that taboo is doubled. Where guilt may be kept in hiding even from family or the determined attack of a skilled therapist, taboo is inherently a matter of group concern. In principle, the cast of characters entailed is known, though their 277

ideas about the kind of ill fortune at risk may be fuzzy enough. Stay out of danger and the problem does not appear. Break with your taboos and—only wait—it must. Each member of the cast has got to consider the moral options for action in redress. For each, the existential triangle remains a private space, but for the errant one the sense of privacy is shaken or may shatter. We see this obliquely in the way a court can protect the privacy of mind of an accused, holding that, whatever the crime, a suspect may not be forced to testify. Such rights of privacy, meant to spare the court the onus of forcing an obvious breach of taboo, at once guard the fiction that courts discover truth, and honour the perishable value of the human conscience. In it resides a person’s responsiveness to the rules of decency whose force is only born of Taboo—fear not of measured authority’s intervention but of wrong itself. Without the socially fostered attribute of persons which we call conscience, Law has lost that legitimacy of good faith which makes it viable. Who watches the watchman if it be not the people themselves? But they must do so, if they will, spontaneously. The possible plots of the Everyman narrative are infinite but always revolve about dramatic reversal. The application is always to the specific conditions a person will confront when pursuit of a wonted moral strategy has led to the apprehension of danger. At the appearance of choice in the narrative field, as we watch ‘social role’ or ‘character’ fade into ‘existential trial’, what we have to consider is now an everyday episode, now a solemn rite of passage, now a cataclysmic pouring out of passions heretofore scarcely given dramatic play. We have seen that in a ‘friendship culture’ a dramatic Peripeteia, often a deeply felt narrative reversal, may be built around betrayal by a peer. Such was Tunginiye’s recognition of enmity in the ‘veiled warnings’ of his closest friend. Such must have lain behind each case of ‘exile’ from a Nyakyusa age-village. Such is the tale which lurks behind the ‘Kinga patent’ technic for taking beer from the bowl your friend wants to pour for you: cup one hand to your mouth but always with your other thumb at the ready to flick aside the ball of poison. Do taboos exist because they are heeded? When they cease being heeded they are gone. Do laws exist because they are heeded, or only in proportion as they are enforced? You have to presuppose Kinga culture intact (even if all the while sporting the kind of infinitessimal novelties which lump up, gradually or sometimes with a start, to become the hard calculi of sensible change) in order to explain the existence of its rules. Kinga taboos relate to a culture which is not everywhere the same, but in a general way you still know Kinga culture from Konde, which is also not everywhere the same. 278

Nyakyusa taboos relate to a very different lifestyle to the Kinga, but the difference is not an ‘apples and oranges’ case—more like Courtlands and Gravensteins. There is a hand-in-glove, systemic relation between Kinga taboos and Kinga behaviour, the same for Nyakyusa. Of course we won’t insist this hand made this glove or vice versa. But glove is shaped to hand in either community, and vice versa, because in the case of culture and habit there can be no priority. Of course taboos are man made just as laws are, though the one may be lightly attributed to a god, the other to a Solomon. It is only misleading to link taboos any more than laws to a mindless ‘Tradition’. That would be to say people were able to make cultures in illo tempore but aren’t up to it today. The idea of a modulated culture flow at the speed of time is better than the cultural creationism implicit in earlier teaching. Must I defend the argument that taboo is maintained by usage not tradition? Taboo is a universal feature of folklore, one of the costumes tradition wears, one precipitate of custom. Moral strategies are privately worked out by individuals within the ample limits of choice not blocked by customary prohibition. Any strategy is in some measure a work-around, and one person may pull off a ploy another can’t. But the moral environment of these work-arounds is in principle the same as you move from one man or woman to another, always accounting for differences of status and social circumstance. Sex roles and age roles, rules of respect and avoidance or privileged intimacy, protocols of conduct in solemn or carnival contexts work in a system of social control rather like features of an inaccessible ‘machine language’, where law compares to a useraccessible software programme. Taboo and law set limits in different ways, but other mechanisms must decide the moral path an individual should take through the encounters of the day, or of a difficult life stage, or the full stretch of a personal career. Anthropologists have traditionally turned at this point to the subject of values. Seen as a maze-feature of the Everyman narrative, a scheme of values is a pattern of ‘preferences’ or ‘priorities’ which have been established in the gradual testing and cultivation of personal moral strategies. It has been common to propose that our cultures ‘introject values’ during (early) socialization. The process has sometimes been called ‘enculturation’, a mechanical clockwork of continual downloading and upgrading of a culture programme in or out of school. But in the model I have been putting forward, values appear as the labels we give to preference patterns (epitomized moral strategies) arising from individual choices, wherever we find patterns well enough established in a community to seem general. 279

For the most part, labeling popular tendencies in patterns of individual preference in this way (calling them ‘values’) does little mischief. It is a convenient shorthand which, however, reflects an observer bias toward dogmatism, as though values could not exist unpreached. The processual aspect of a culture is masked, leaving us with a puzzle: How are values formed? How do they change? How well can we expect them to stabilize community life? These questions are particularly relevant to anthropological studies, which usually deal with relatively non-dogmatic human communities. I have taken the position that one’s way in life is, and is bound by our human nature to be, found crucially by intuitive thought and interpersonal discovery not instruction. This is especially clear in Kinga and Nyakyusa cases, where children run their own homes. When proper conduct is ‘taught’ only in rule-of-thumb prohibitions, a way has to be found around these roadblocks, but there is no equivalent to our dogmatic moral principles—meeting bourgeois (godly) standards of acceptability and the like. Where values become explicit they are coded into daily discourse and doubtless have a conservative function. Where they remain implicit they pervade Everyman’s experience in his social world at the level of that stream of infinitessimal events which cumulatively give direction to a community’s persistence and change over generations. The Everyman narrative offers a frame of reference for the study of ‘personality and culture’ while setting aside the doctrinaire models of personality formation which have so often been found necessary in psychological accounts of human nature. Dealing with moral choice, this narrative approach easily blends with some existentialist thought, but without raising or begging the question of ontological priority as between ‘being’ and ‘existence’. I take it that in the empirical world neither one can either blindside the other or escape its iron grip. That part, at least, of the Everyman narrative which entails moral choice emerges only gradually in infancy, doing so only as the infant is exposed at a reflexive level to primary socialization. The evolution of individual moral strategies is the whole of the narrative alluded to here. Understanding the nature and importance of taboo is a key to following the narrative. Taboo constitutes a technique of social control which can and must be decentralized. The incest taboo illustrates. It further suggests the continuing role of taboo under jurally ordered institutions of social control. Equivalent terms to Swahili mwiko (taboo) are found throughout the Eastern Bantu languages and around the world. English had to adopt the Tongan expression when learned folk wanted to discuss the subject, because until the ‘age of exploration’ took the fancy of english-speakers no equally strong and felicitous 280

word had found its way into polite discourse. Even today the open discussion of taboo in self-reference remains awkward. If your moral strategies are based on acquiring and relying on the taboos appropriate to being who and where you are, exhuming them can mean losing hard-won self-confidence—the kind of being-in-the-world which goes under the rubric of ‘human dignity’. When psychotherapy ‘goes deep’ in search of the particular roadblocks a person has encountered in the pursuit of a successful life-strategy, a terrible dilemma arises: you are not asked to discard your ‘self’ because you cannot, but you are asked to shrug off some long-buried encounters with dread which have ‘always’ guided you—and with slim rational assurance of finding your own way forward again. Taboo marks danger, and danger marks us all. The experience need not be traumatic to go deep. There is no fixed time- lapse. An experience, to be morally effective, even to ‘haunt’ a person later, need not even be especially notable at the time of happening. That is what I intend by alluding to the ‘infinitessimals’ of human experience. How is it that you suddenly find yourself ‘free’ of some particular obsessive attachment? What accounted for it in the first place? And but for it what would you still have that you lost, what gained that you didn’t have? Persons come to differ in radical ways through the waxing and waning of wariness, of rueful and ruthless impulse, and self-reflective thought. I do not try to assimilate all difference to the dimension of ‘social danger’. What I have in mind is that the chief instrument by which ‘diffuse sanctions’ find their way into our moral sensitivities has to be understood as the inward marking or framing of social danger—that is, as response to taboo. And I argue that the process is, as with successful discipline anywhere, predominantly a subtle one, a matter of thrust-and-parry not one-sided battering, physical or rhetorical. The incest taboo illustrates how taboo works, as well as showing how taboo and law can make combination. Law, in effect, needs taboo. Crucially, this is because public safety flows from a premise of amity beyond the necessarily self-protecting kin group, formed around reproductive success, whose boundaries are performatively defined by rules which put endogamous procreation out of mind. Since such rules of amity are not enforceable without taboo, they have to be lodged within the structure of the domestic group, where the ethic of self-sufficiency is most effectively nurtured. The ‘teaching’ function of taboo is seen everywhere in the association of taboo with life crises. Special rules of propriety (decency and decorum, manners and mores) are part of the spectacle of life-cycle rites. Understanding of the universality of taboo follows from the work of Michael Polanyi, Marshall Sahlins, Franz Steiner, and other 281

recent writers, as opposed to the ‘magicality’ focus of much popular and anthropological thinking on the matter. The deeper aspects of taboo require us to be aware of the privacy of mind from which, through the centrality of moral strategy in the human career, a ‘system of public values’ is rendered possible. Taboo is maintained primarily by usage, secondarily by tradition. Crucial to this, taboo is reinforced by the privately harboured presentiment of ineluctable danger, without which external sanctions on behaviour are relatively impotent. ‡‡

Picturing the Pangwa world My main model for picturing Sowetan culture in the very distant past is uPangwa, the neighbour people southward of the Kinga. Apart from archival sources, which are lean, I rely on Fr. Hans Stirnimann’s monographs, which are not. Thanks to his exhaustive studies of Pangwa ritual practice, carried well beyond anything I was able to undertake for surviving elements of Kinga bush culture, and thanks to the obviously near kinship of the two cultures, I’m in position to reconstruct by approximation a pre-Sanga past the two neighbour cultures once apparently shared. As will appear, Pangwa practice as the good Father was able to reconstruct it from interview and example in the 1960s refers back reliably to times still alive then in the memory and imagination of older men. But owing to radical disruptions and dislocations of the Pangwa communities from the 1850s onward, the picturing is problematic. In a word, the portrait we have from the ethnographer’s informants is of an implausibly oversocialized people. ‡‡ Here is Stirnimann’s quick account of the history of troubles— here as throughout the translation is mine: For convenience, as there will be many citations, the location of the source for each is identified by volume (I or II) and page. Short titles are: Volume I — Existenzgrundlagen (1976); Vol. II — Die Pangwa (1979). From about 1855 on to the turn of the century the unwarlike, badly armed and hardly organized hoe-culturists of uPangwa were repeatedly attacked after harvest time by predatory Ngoni hordes from the south. These slew the men, taking the women and children away in slavery. In spite of this enmity, the Pangwa aligned themselves on the Ngoni side in the Maji-Maji uprising of 1905-6, against the German colonial regime—and must suffer therefore through nearly two years of unremitting guerilla warfare with the troops. According to contemporary mission documentation some three quarters of the [Pangwa] population fell victim to them. Only in British times (1919-61) did the country recover, mainly owing to

282

increased migration of culturally assimilated Kinga and Wemba into the depopulated areas of uPangwa. [II: 14]

I take it the main risks in using Pangwa self-description in the way I intend derive from (a) inevitable differences at the detail level in the pre-political cultures of Kinga and Pangwa communities, even setting aside problems of timeline discrepancies and the perils of imprecision (fuzzy logic) in all reconstructions; and (b) the specific distortion or mythlore built into any account of Pangwa culture prior to 1855. In mitigation with respect to (b) the oral evidence from the Kinga Eastern realm puts the Sanga influence there only in-process, hardly complete, in 1855, and probably stimulated about that time principally by fear of Ngoni raiding. Even in 1890 the Kinga of the Eastern realm retreated into a thornbush fort of sorts at the approach of a curious German party mistaken for Ngoni. This hardly matches the Sanga tactics developed in the Central and Northern realms, suggesting rather a ‘bush’ element in the military culture. As to (a) I learned enough of the differences persisting in the 1960s within the erstwhile Sanga realms to take seriously the variations within ‘Kinga culture’ at a detail level. There is, in effect, a Boolean overlap between eastern Kinga and neighboring Pangwa cultural communities, persisting into historic times. The proper interpretive rubric for us here is that ‘our Pangwa’ are not meant to be ‘pure Pangwa’ but are meant to be representative of older times in the neighborhood Pangwa occupied when they were studied in the 1960s. A review of Fr. Stirnimann’s material will show how thoroughly a system of taboos can serve the ends of social order in an acephalous society like the one a Sanga political class began to transform some generations before the Ngoni came to the region. A century ago such a transformation would have been termed a change from ‘sacred’ to ‘secular’ organization. It is worth asking what such a metamorphosis ought to be called, once it has been seen close-to. I am inclined to such pairs as closed/open, local/translocal, or with respect strictly to the use of ritual, performative/dramatic or proprietary/public. I suppose throughout what follows that Fr. Stirnimann’s account can be taken as a tidy and possibly even overly sympathetic recreation of the old lifeways, based on the best information a missionary could gather in an enviably ample period of fieldwork. What is of special importance is to see that Pangwa were a pacific people, not warriors; and that they were treated (in a strict sense, respectfully) as one of the bush communities into which the Sanga system, but for the intrusion of other forces in the nineteenth century, would normally have spread without the need of conquest strategies. The war games and ‘standing armies’ of the Sanga 283

courts were persuaders against resistance from the bush community as it moved from a first-contact situation to a fully fledged minor court. Sanga were, indeed, in the process of spreading by settlement across the Mgiwe/Kilondo river (variously named on various maps) which was the natural feature the British later used om setting up their ‘tribal areas’. That boundary settlement in 1926 ‘orphaned’ substantial ethnic-Kinga populations in the north of Pangwa tribal area and again in the trans-Mgiwe parts of Wemba territory, which were eventually allotted to uBena. Sanga rule, be it noted, lost nothing in the British boundary manoeuvres, as the new ‘tribal area’ gained more on other fronts than it lost in the east. What the new regime meant was that many men paid taxes and took their trouble cases to a new baraza. The new order did not officially or practically close any boundaries, but did put an end to the game of expanding Sanga rule by sending out client rulers to reorganize acephalous bush communities.

Performative and dramatic ritual The ideal-typical continuum an appropriate methodological device for imagining an historic series in the development of ritual and political sanctions in a region like the Sowetan. The independent variable in the series I shall project is ritual style. The series is pictured as a virtual time continuum. Over some three centuries the prevailing style is simply considered to move, at different rates in different ecozones, from an ‘early’ type toward a ‘late’. I look at three ethnographic cases appropriately placed along the timeline— uPangwa, uNyakyusa, uKinga—interpreting in each case the social and political setting of its ritual style as a set of explanatory correlates. Likely causal relations among the variables may be discussed but are not assumed. My three ideal types of ritual style are labeled (progressively) private, public, and ceremonial. Pangwa are most dependent on private ritual, Nyakyusa on public, and Kinga on ceremonial. But ritual of the three types is present in all these societies. To name the two poles of the model continuum, I distinguish performative from dramatic ritual action. As you proceed from uPangwa to uKinga the style becomes increasingly dramatic. Proceed the other way and the style becomes increasingly performative— each ritual has a definite project of effecting a specific change of status. Garden magic, cures, and rites of passage are frequent projects. The ritual if properly done is considered effective. But even where the project is to transform a private individual it may invite public participation, and miniature dramatic elements may be incor284

porated. A drum dance, or the like, may follow. The only sensible way to sort out three cultural communities on such a continuum is by scaling them—which depends more on dramatics, which more on performative symbolism. Modest as our information is in each case, I think it is enough to scale the ritual systems of our three Malatan communities. Kinga, for one instance, do make some use of puberty rites. The male rite, which is supposed to treat a group of boys from a given locality, seems always to have been honoured in folklore more than deed. There is no memory of a feastful dance afterward, and no circumcision even in token form, and no explicit ritual change of status. For girls there is token circumcision, a joyous ‘naked parade’, and spontaneous drum-dancing after. But if a change is effected in the girl’s status it is achieved through drama (teaching and mockmagic) not the rite itself. Pangwa are certainly more ‘performative’ across the board, but make dramatic use of song on nominally private ritual occasions. It will be seen that, if one grants the premises of this particular model, the three empirical cases I consider fall on the virtual timeline in this order: Pangwa—Konde—Kinga. There is no implication at all that the Kinga have been into the protostate process longer than the Konde, or that their political achievement is more impressive—only that Kinga have, in building a more ‘pyramidal’ (and so presumptively a more stable) statelike polity, depended more on dramatic style in their rituals. If Ngonde had been included in this comparative review, I can only guess it would place at the ‘most developed’ end of this putative timeline. I would expect it to, as I think ceremonialism is an especially appropriate tool for enhancing central direction in the political field. But this is not the kind of outcome to call without an inclusive review. The meaning of ‘private’ is clearer if we link it to the decentralization of authority structures. Privacy is inherent in domestic ritual, as the taboos required to sanction it—to put a ritual act into effect—have only a domestic compass. But the umbrella of domesticity can be extended along lines of kinship to the point of transforming toward (public) ritual theatre intended to aggrandize the social importance of the presenting group. In Konde funeral celebrations we see the grieving group projecting itself through dance and masque before a public of the deceased’s non-kin community. As ritual becomes ‘public’ it begins to depend on theatricality for its effect. The intention remains performative (accomplishing a specific ritual transformation, as with Malinowski’s ‘magic’), but requires and plays to an audience to achieve its effect. A Kinga maiden’s initiation grows only more public as it proceeds from an 285

esoteric seance for maiden and elder women only, to the painted ‘naked parade’ with its noise and whooping, calling on the whole world to attend. But the dramatic side is all there not only to effect the change of public status (to nubility and new personal responsibilities but not to a marriage commitment) in the maiden. The performative intention of the rite, when read as a whole, is to affirm the structural dependency of a young woman’s moral career on the sanctioning power of a matronly cabal not of her ‘mothers’ but her future peers. That is, the thrust of the teaching is that a woman’s move into marriage is not a move out of the autonomous circle of female peers and its intimate sanctions. So the intention is ‘structural’ not private. In ceremonialism the ritualistic performance is not meant to effect a change in the participants or their status in the community but to dramatize the importance of ritual itself, acting in the interest of the whole community which sponsors it. Here the drama must ‘transform reality’ (the world of common sense arising in local discourse) through the transcendent mechanisms of theatre. The drama (ceremony) fails if it falls short of the extra-ordinary. Ceremony presents a superincumbent version of the world known to ordinary experience. Here belongs the ‘flooding out’ experience of high dramatic art or of those moments of danger and exhilaration, when a person is ‘swept away’ by events, ‘caught up in the drama’ of an arresting situation. This kind of experience, being all but predictable for certain times of challenge or crisis, can be programmed. The staged event is memorable because it is framed apart from the ordinary run. Everyone loves a coronation, a dance, or a hanging if the ritual side is rightly done.

Intimacy and avoidance In classic ethnography ‘avoidance’ rules refer mainly to touch and/or address as between two individuals occupying distinct roles in one kinship network. As any two individuals who must avoid either touch or address could not a fortiori commit incest, it might seem that some ‘avoidances’ should be thought of as extensions of the incest prohibition. What this would miss is the essence of the incest taboo, which sets it quite apart. Intimacy is the normal condition of all the dyads to which an incest taboo applies. This is, in effect, socially prescribed intimacy, such as one finds in parent-child or sibling relations. So the incest taboo invokes an existential condition embedded in the conduct of primary (lifelong, maximal-claim) relations of kinship: break this rule and you can’t expect these ties— through which your society has constructed your individual identity and all lifelong claims and counterclaims implied—to persist. Your 286

social being, your everyday persona, implodes. Oedipus (not in Freud but in the ancient Greek legend) is an exceptional case which ‘proves’ the rule: in most versions it is not with the act itself but the actors’ recognition of its nature that damnation falls upon Jocasta and Oedipus for their incestuous marriage. It is a common thing in Bantu ethnography to find that a couple planning marriage may discover before the fact that they are related (through links they were unaware of) and before they dare marry must stage a ‘cutting’ ritual, mystically dissolving their blood ties. To my knowledge it has nowhere been shown that two such persons would be treated as pariahs for having known sexual intimacy before the link was known. As for ‘avoidances’ breached in ignorance, the problem could cause no more than a bit of undamning confusion. In a nutshell: an ‘avoidance rule’ cancels what in its absence would be a step toward socially privileged intimacy between the avoider and a kinsman or kinsman’s familiar. An ‘incest taboo’ sets the condition of sex avoidance on cross-gender relations within the kind of privileged intimacy group we usually call a family. It is a special ban within a bond. Where ‘avoidance’ is required between affines, a special logic applies. Where a new spouse is brought within the compass of a solidary kinship group, a new precinct of privileged sexual intimacy is created and set apart within it. In the Nyakyusa case, an addressavoidance taboo was applied to the bride with her marriage, forcing her to flee as if in terror whenever her husband’s father or father’s brother came by. That was feasible in traditional times, as her husband would not be living in or frequenting his father’s village. One vital consequence of this total-avoidance routine was to block the transitivity of the father’s authority—deeply ingrained over two or three decades—to other members of the son’s new family. Another consequence: though the father must feign to ignore it, he is an intruder now in his son’s private life. The son, after all, must be allowed to ‘grow up’ at last and till fields of his own. In parallel cases, but where a young husband and his father are in constant contact, such a rule would prove unworkable. The usual cure for securing the status of the ‘stranger bride’ is a group-spouse taboo—all other males in the group must (as it were) ‘extend the incest taboo’ to the new wife, so that normal social familiarity can rule. But this ‘group spouse taboo’ is not ingrained but removable. The institution of the levirate, often found in agnatic kinship societies, enjoins a man to marry a brother’s widow. In the Nyakyusa case the takeover is automatic, as a surviving brother inherits the other’s name and identity, and with that his wives and other wealth. The true incest

287

taboo is not removable, though the ‘extension’ of its rhetoric to intimacy with more distant relatives or affines may be. For Kinga, where a woman does not marry into a solidary agnatic kin group, no ‘group spouse taboo’ appears. A married man may take his unmarried younger brother into a sort of tutelary bed, a lit à trois, to redirect his sexuality and lean him toward a marriage of his own. By the same token, a woman’s right to a new pregnancy might see her turn to a husband’s brother when her spouse was absent or ailing. One point that will come up as we examine cases is that intimacy at both levels (including or excluding sex) can be made socially possible by the norms of avoidance entailed. The reason isn’t instantly obvious. You must consider first that stable relations of intimacy are not ‘natural’ to human kind. The nearest we approach that is in the nursing bond of mother and infant, which both Pangwa and Kinga custom extend almost maximally, but which is always eventually extinguished. Beyond this, a stabilizing social structure— a whole scheme of rights and duties ‘spontaneously’ sanctioned by participants—is a precondition of extended intimate relations. The ideas of ‘family’, ‘kinship’, and even ‘amity’ are meaningless outside of such a social structure. Only particular societies make use of rigidly formalized avoidance of address in their social architecture, but all use a set of ‘role relationships’, each with specific behavioral expectations attached. Some restrict, some enjoin the appearance of affection. All are sanctioned diffusely and show strong patterning which doesn’t encourage loose cannons. Where intimacy is episodic and fleeting, it is not a relationship. To be so, it must be stabilized by a systematic web of taboo. If a young Nyakyusa wife failed to avoid eye-contact with her father-in-law even in the 1950s, she would be introducing structural trouble—less grave than she might cause by incest, but parallel in challenging the way their game of family life is played. In this case, owing to the lack of age-traits as boundary markers, the ‘group spouse’ taboo is especially hard to implement as between this young woman and her husband’s polygynist father, who continues to enjoy a uniquely lifelong privileged relationship (parental authority, ingrained) with his son. By the same token, in a rampantly polygynous society a stable marital relationship will be hard to structure. As you think about the Nyakyusa case you will see that jealousy among a man’s co-wives would make marriage there a rather destabilizing experience if it had to be grounded in the possessive sentiments a monogamous marriage system so often relies on for stability. What the co-wives of an aging Nyakyusa man have in common is their hateful avoidance 288

of their husband’s ‘fathers’. By drawing them together, the system draws them emotionally away from dependence on the husband for intimacy’s undeniable rewards. The whole charade with each successive addition to the father’s plural marriage effectively narrows the door of intimacy in it, and for both genders weakens the fiduciary potential of the marital bond. As I read the ethnography, the ‘sisterly’ bonds among women are proportionally strengthened— though I confess to reading between the lines and looking for the structure of experience not of rules. A further point comes up. Structured social relationships are to be understood as ‘inherent’ in human kind. Since they depend never on actions impossible without mutual and congruent expectations as to the claims which must be honoured in a structured relationship, they presuppose deep socialization of the individuals who will be involved. Sandbox society is not self-stabilizing. Adult communities won’t make their way in history without more ‘grown up’ kinds of social relationships. If we have shown that in the absence of incest taboos domestic groups or ‘family systems’ can’t be stabilized as the primary units of self-sufficiency, a corollary is that unstratified societies depend on that taboo to furnish the ‘schools’ which will socialize children capable of useful lives and of developing viable moral strategies in an adult world. The situation remains about the same in principle almost anywhere one moves in social history in more simple or complex societies. I hope that the mysterious side of incest taboos need be deemed no more or less puzzling than the very existence of our kind. It seems obvious to most anthropologists that language, culture, and home-made social architecture set our kind apart from others, and in a quite special way apart from nature. It would be arrogant indeed for a scientist to predict the course of a (human vs. human) chess game and outrageous to predict the moral success of a single human career however staid its social setting. Yet if the human condition leaves us in one sense even less natural (predictable) than chess we are natural enough to know and fear ‘unnatural human acts’: they put the game itself at risk.

Pangwa particulars: intimacy and avoidance Incest and avoidance rules of the Pangwa can be made clear by citing and discussing Fr. Stirnimann’s accounts. Since his two major monographs are not everywhere accessible, I translate extended passages. I first want to make clear Fr. Stirnimann’s thinking on Pangwa sociology, as well as my own where it diverges, or where the passage 289

cited doesn’t speak for itself. A major divergence has to do with his assumption that the Pangwa once lived in localized exogamous patrilineal communities. Pangwa like Kinga and Konde do have patrilineal rules of inheritance but bilateral rules of descent. Thus their ideas about kinship are dual, combining ‘kindred’ and ‘lineage’ patterns of thought about personal identities. Pangwa call their political communities inana (sing. u-lutanana). As will appear, I do not gloss this major Pangwa political institution, the lutanana, in the same way Fr. Stirnimann does. I avoid using ‘lineage’, preferring the less specific ‘descent group’ as my rendering of the Pangwa term within a citation from his german text. Where I am myself translating lutanana into english I refer to it as a ‘deme’, or more specifically a ‘patri-deme’. In this I am following the reasoning of George Peter Murdock, whose gloss, dating back half a century, deserves citation here: ...The inhabitants [of a ‘deme’] are necessarily related to one another through intermarriage, although they cannot always trace their exact kinship connections. They are consequently bound to one another not only by common residence but also by consanguinity. ...Within such a group the only social structuring is commonly into families... The strongest sense of identification is usually with the community as a whole, which is viewed as a consanguineal unit in relation to other communities in a manner quite comparable to the attitude towards one’s own sib [clan] in a unilinear society. The widespread tendency to extend incest taboos throughout any kin group... naturally does not leave the deme unaffected. ...Exogamous demes, where unilinear descent has not become established, may be distinguished according to residence rules as patridemes and matri-demes. ...Exogamous demes constitute one of the two principal origins of [unilinear] clans. ‡

Here, for juxtaposition, is Fr. Stirnimann on the local descent groups of the Pangwa. George Peter Murdock would be looking for evidence as to how far the ‘kindred’ (defined by the incestuous flavour given the idea of marrying close kin on either the father’s or the mother’s side) had established its exclusive political rights in the lutanana settlement territory, and how far the rhetoric of ownership and identity had come to be expressed in terms of father-right: The common residence of the kin group in the patrilineal (väterlichen) lutanana demanded of the individual members not only mutual good will in the daily routine, but also readiness to assist and cooperate in house building, the care of the fields, support with food and seedgrain in time of need, participation in combatting harmful wild animals. For the clearing of bush areas and the care of the larger fields on the mountain slopes, the mkoyo [leader] of the lutanana

290

organized collective work groups, imikovi, in which all able-bodied members, as well as in-migrant settlers, must take part. ...Assiduous participation in the collective work of house building and cultivation was reckoned as visible confirmation of blood relationship and was of fundamental significance for the solidarity of the whole luxolo [kin group]. “We worked together, helped each other out all around, and shared everything with each other like children of the same father,” old men declared. [II 60-61]

Here are Fr. Stirnimann’s constructions of the ‘lineage’ or descent-group organization and the clearly rather ‘kindred’ sort of incest rules as they putatively were before traditional Pangwa culture was disturbed by the Ngoni, the German and British administrations, and the missionaries: In accord with their social structure, marriage is embedded in a frame of usage and taboo, which as first priority is aimed at maintaining the purity of blood of the patrilineal luxolo [kin group] in the lutanana [local descent group]. [ II 136] Intimate relations outside marriage, and marriage itself, are forbidden where kinship as close as four generations exists between partners. [II 136]

This last is intended to mean that two persons with any greatgreat grandparent (either gender) in common are too closely related to think of sexual contact. This is the kind of rule which best matches a quasilocal ‘kindred’ organization, not any corporate group which might be generated by unilineal rules of heredity. As will appear, very few Pangwa would have been in position after 1855 to collect the information needed for scrupulous clearing of all barriers to marriage under this rule. In my reading, this would always have been the case, although the ethnographer supposes the mistiness of genealogical recall must be only recent. Another point to be kept in mind in deciding what weight to give this ‘rule’ is that while no one in recent times has been held to it, Pangwa informants don’t find it anomalous. That is, the rule of patrilocal residence at marriage does give the man and his patrilateral male kin constitutional privilege in the lutanana. But this appears to be a function of local politics and ritually sanctioned orders of precedence, not unilineal consanguinity. We are dealing with a patri-deme not a localized patrilineage. Trying to see the lutanana as a unilineal local descent group makes the bilateral calculation of incest danger anomalous. But that rule begins to look more reasonable when considered as a justso story to cool the passions of the young and keep marriage in the hands of their elders. Protecting the elders’ grasp of ‘tradition’ is a 291

necessary condition of order in a society without formally vested political authority. The rule of exogamy for the lutanana, when it is coupled with the practice of estrangement as between intermarrying settlements, is a structural cinch which puts the destiny of a young generation in the hands of the old men, who are alone in position to find eligible brides for their sons. Fr. Stirnimann links this to the broad incest/endogamy taboo: Two exceptions from this broad-reaching incest taboo have nonetheless always been recognized: the sororate and geographic separation... Apart from these two exceptions, custom always allowed a man to fetch his wife from an unrelated lutanana [local descent-group], for as the saying goes, ‘New blood is only to be bought from another lutanana’. Accordingly, in the old days planning a marriage always lay in the hands of the father, who knew the several kinship networks. [II 138-9]

The first exception is rather slight. In this case the main implication of ‘the sororate’ is that a father of two eligible daughters might, when the first falls short of supplying her husband with the contractually prescribed three babes, be obliged to settle the matter either (a) by sending the second daughter after her, or (b) returning some bridewealth. This is clearly an ‘ideal’ model of an obligation some fathers might choose to resist. It entails a translocal contractual relationship calling for good will. The ‘sororate’ would only be an exception to an incest rule which held that accepting a bridewealth portion (as a bride’s sister arguably must do) creates a virtual kinship link, forbidding her (though it had not barred her sister’s) marriage into the groom’s kin group. As will appear, it is an author’s inference that this kind of link should set up an ‘incest’ barrier. An alternative reading would be that we are dealing with a society in which translocal political relations are mainly steadied by marriage contracts which are considered as irrevocable, lifelong partnerships between two families, but depend on each side having an investment in the progeny of the marriage. The second exception is more interesting. By ‘geographic separation’ is simply meant the two will have grown up as strangers. By ‘planning’ is also meant negotiating. It will be clear that one consequence of a bridewealth system is just this: wealth must accumulate, on the groom’s side, in the ascendant generation before a marriage can be arranged. In some societies, much of it can be brought in almost painlessly as bridewealth from another family by arranging a daughter’s marriage at the convenient time. This requires that marriages be thought of and acted on by the ‘fathers’ in businesslike terms. As bankers, they see the good of marrying off local young women to other communities at good prices, the better 292

to satisfy local young men wanting to get unrelated brides. Regular local endogamy would mean constantly robbing Peter to pay Paul, which means canceling credits and undoing the simplicity of a local exogamy balanced by the gender balance of birth rates. Unsimple systems of credit are full of snares. All the indications are that this applies in spades to the Pangwa. In accord with traditional jural norms the husband’s group acquires the right to the labour and fertility of the young wife, with the handing over of the bridewealth, as it is expressed in the teaching: Twikula umudala xulima na xubaba (We buy [literally, eat] a wife to work the gardens and bear children). [II 140]

Fr. Stirnimann is most acute in the way he chases down details of bridewealth distributions and their significance in his Pangwa model. He shows that, in effect, the ‘incest-avoiding community’ (by which I mean the group within which the taboo is a ‘spontaneous‘ categorical imperative) is recreated for individuals of each generation as they enter marriage, by the business deals made on either hand with a bridewealth: In contrast to the norms of many African societies, the bride’s father in uPangwa can’t keep for himself the bridewealth he has received from the groom’s family for loss of his daughter, but must hand over the one half to his own blood kin of the lutanana [local descent group], the other half to the bride’s mother’s brother representing the matrilateral kin. But this give-away of the compensation implies mutuality. All recipients of bridewealth portions are obliged by kin solidarity to put up equally valuable contributions to the father for gathering the bridewealth of a son. At first sight this exchange of gifts looks to be only a sign of solidarity within the lineage. In reality this mutual receiving and giving has a deeper dimension. It sets limits on the effective scope of blood kinship, and with it the incest barrier. [II 137]

Note first that a bridewealth which is seen to melt away as soon as received can’t cause personal envy even in the giver. This treatment of wealth is the quintessential one for (acephalous) communities of equals where neighborly amity among men is an ethical imperative. Next: The odd thing is that this kind of incest taboo ‘sets limits’ much too generous to match any possible system of agnatic or other unilinear reckoning. Where ‘one blood’ is the key to kinship imagery it has to be the key to incest imagery. But ‘blood’ just doesn’t run in ‘lines’ or ‘lineages’ in communities with bilateral rules of descent and inheritance. Bilateral reckoning joins two lines in the 293

issue of each marriage, so that each married couple represents a node in a centrifugally branching, all-consuming no-sided kinship network which generates no overall shape to the society: no segmentation, no possible sort of corporate kin group. To have corporate groups (as in unilineal societies) you must have unambiguous membership: each person recognized must belong to one and only one such group. This uses kinship rhetoric to create the kind of social structure we have come to call ‘segmentary’. But in bilateral societies there are as many radically overlapping networks as there are living marriages. For Pangwa the network of any set of full siblings is defined for them by the bridewealth arrangements made for their parents: receivers of bridewealth contributions (represented directly or in the persons of their heirs) become givers of the portions passed out of the lutanana community in the next generation—and so on. But it won’t be immediately obvious how such arrangements create political stability in the lutanana. There is one further comment due on ‘local descent groups’ and anthropological correctness. For a kinship group in the strict sense to own a territory it would have to be a corporate group, enduring as a political unit through the generations needed to produce enough people to make a biggish community. It would have to start with a new group settling a new place and making room over time for progeny and the progeny of progeny until the country was full. Then some segment of the big kin group could move out to start somewhere else. This is just as Pangwa informants know it must have been in illo tempore: The traditional worldview was simple and self-contained. The founding father of a lutanana suddenly appeared, took the land in use and thanks to his creative power brought forth not only the worlds of plants and animals, but also a numberless company of descendants. He allows the first-born of each generation to partake of his creative power.... The deceased fathers (mahoxa [shades]) live on in their graves and, together with the living descendants in the lutanana, form one single blood, a closed kinship group (luxolo). The visible group’s settlement area ‘above’ is identical with the country of the [ghostly] fathers ‘below’. [I 219]

But unless the kinship system is the sort that will build a regularly expanding corporate group with unambiguous central direction over land, labour, and the distribution of other necessities, it is misleading to say this is ‘a kinship group which owns its territory’. That means my use of ‘local descent group’ to translate Pangwa ulutanana is still not quite anthroplogically correct. As we shall see, the Pangwa kindred is a contractual co-op:

294

Then according to reports from elderly informants, such troubles affected kinship relations in the past, that the scope of the bridewealth distribution even within the lutanana [descent-group territory or settlement area] was always unstable. According to traditional norms, the bridewealth portion at a marriage was supposed to be distributed not only to father and mother, brothers and sisters of the bride, but to all patrilateral and matrilateral kin to and including great-great grandfather and his brothers and sisters—thus to four ascending generations. As to how this was possible, in consideration of the modest indemnity paid around the turn of the century (about six hoes and two or three baskets of maize, millet, and eleusine for brewing beer), the older informants pointed to the radical decimation of the population throughout the decades-long slave raiding of the Ngoni. Many now thickly settled descent-group lands were...completely depopulated, to the point that survivors barely managed a wretched existence in a few nuclear families... Among these remaining groups what was once a lineage group (luxolo) came down to the universal axiom: Where bridewealth is shared, there is common blood, and: The recipients of bridewealth shares are making alliance with the givers. For these, incest limits and prohibitions of marriage were for at least three generations. Intimate relations between cross-cousins, such as often take place today, used to be regarded as incestuous. Any kind of family disgrace was condemned as witchcraft (uvuhavi), and the guilty ones were expelled from the kin group and banished from descent-group territory. Should a married man be guilty of such an offense, they burnt his house down and, after taking away his wife and children, cast him out or killed him. They feared the wrath of the outraged shades (mahoxa) should he remain living in the community. [II 137-8]

I find it wise to discount this last bit of news on mob rule as only some more of the just-so rhetoric which needs to be taken with salt. House burnings belong to the high dudgeon mode of righteous indignation in acephalous societies, dependent as they are on ‘spontaneous’ demonstrations of people power. House-burning is a sanction executed by stealth and so uncharacteristic of the political rhetoric of chiefly society. What lurks behind this vehemence is the sacralization of amity within the ‘descent group’ community. Every member ought to share the ‘outrage of the shades’ and avoid stirring up the fires of envy. Taking a bride from inside the lutanana community isn’t bringing in a wife ‘to work the gardens and bear children’ but taking an eminently ‘salable’ young woman of the community off the market. In short, if some man whose son is in the market for a woman wants to object to the use of ‘our banked bridewealth credits’ to ‘buy a wife who is already here’ the taboo of incest can be evoked to quash the move. What has to be understood as the premise of amity in the lutanana is that it is structured as a ‘credit’ system with only the ‘books’ of memory. No man can boast 295

his holdings in the system outweigh another’s—the credit is nominally all in one pool, symbolized as ‘one blood’. Behind the scenes, young bachelor men are rivals, of each other and of some strong married men as well, for the use of bridewealth credits. It is out of this subsurface rivalry the real social danger of ‘incest’ arises. The elders must quietly negotiate around it. They must conspire. Looking for the ‘system on the ground’ through the lens of the ‘system of ideas’, our main conclusion should be simply that the virtually boundless nature of the ‘incest-avoidance’ group, in combination with the modest ‘price’ of a traditional bridewealth, has the consequence that while kinship (‘blood’) may rule at the level of conventional wisdom, a strategy of amity rules Pangwa praxis. The actual distribution of a bridewealth is determined ‘on the ground’ by a modest series of transactions, amounting to morally sanctioned contracts of reciprocal obligation, among men of the ascendant generation. If these men choose wisely they will choose to serve the ends of neighborly amity. Paradoxically, each such transaction injects a quantum of idealized kinship sentiment into an existing tie. Reciprocity is reaffirmed. Whether I am giving or taking a bridewealth portion I treat the contract as good and profitable, as a bit more than ‘money in the bank’. Here is an avenue by which ‘forgotten links of kinship’ can be made new. In short, it is a way ‘virtual kinship’ gets the ‘genuine’ stamp. The main trick entailed is masking the arbitrary element of choice in the business by letting key persons represent whole arrays of the living and dead ‘unto the fourth generation’. It gives the planners free hands. The most familiar instance is letting a ‘mother’s brother’ represent the ‘matrilateral kin’. A simple twist on that is this: whenever a local man is asked to give or receive bridewealth credits on behalf of the bride’s side of a marriage this man becomes a virtual ‘mother’s brother’ to her. Kinship which is theoretically established by dimly recalled lore about links is a fine and flexible tool for carving kinship in the sand of the public mind. If anyone they want to take into their crowd can’t be gotten in through the Pangwa door, it is quite likely to be a missionary or an anthropologist, if not both. ‡‡ But Fr. Stirnimann makes clear Pangwa marriage arrangements in an ideal case were bilaterally oriented, so creating a quasiaffinal in-group of persons within which civil conduct was afforded the positive sanction of kin-like amity. When I say ‘quasi-affinal’ I allude to the statistically high probability that every man and woman in a given lutanana, if they are not related by several paths of ‘agnatic’ or ‘cognatic’ kinship, are at least related by (passive) participation in a certain number of marriages. Implicit in the incest rules is the premise that ‘blood kinship’ always passes down through women 296

as well as men; and that, although the social ties in the ‘arranger’ generation are contractual, they are advertised as merging in a future descending generation to form ties of ‘blood’. It is a provision which illuminates the active significance of ‘virtual kinship’ within the Pangwa-Kinga region. But it can be made to speak also to the inner nature of the lutana. Stirnimann thinks of it as corporate. ...ulutanana (pl. inana) is the settlement area of a closed, agnatic kinship group (uluxolo). [I 57n]

If I consider this a good and true statement of the case my own conception of the scene certainly collapses. We have moved from a discussion of incest rules and their implications to a question about the main political institution of the Sowetan region’s bestdescribed acephalous society. If the incest taboos are somehow defined for Pangwa through the retailing of bridewealths and bridewealth obligations within the lutanana, it is clear enough that Pangwa reckon kinship bilaterally and so cannot construct a corporate kinship group. This means the lutanana membership list is not prescribed by rules but recruited and confirmed by the management of marriages and the obligatory associated ritual installation of a new household as belonging in and to the lutanana. Is a lutanana made up of people or homesteads? Appropriate phrases might be ‘local community’ or, to insist on the territorial aspect, ‘local state’. If, as it is often suggested, the essence of an acephalous polity is ‘organized anarchy’, and if the lutanana is something other than a structural segment, we may be prepared to call it simply a microcosm of the precolonial, pre-1855 Pangwa social order. Like any other acephalous polity, the Pangwa would have depended on ‘pop-up leadership’ at need, not vested authority. The larger ethnic community would be composed of an open series of inana organized by social coherence under a ritually sanctioned code of law, not political authority with its enforcement procedures. Nineteenthcentury anthropologist had a phrase for it: Custom is King. Now what we must do is deconstruct ‘custom’.‡ I have elsewhere shown for Kinga that a bridewealth system can generate power for men of the ascendant generation [Park 1994b]. That is the case for traditional Pangwa as well. Here it remains only to suggest that the illusion of patriliny and closure in this heavily bilateral and open (but necessarily patrilocal) society is fostered by the solidary style of male elders in the affairs of the Pangwa, and the well-earned but always personal authority with which they are able to conduct themselves. We have yet to consider in detail the household social structure which is the key to that authority. 297

Konde particulars: intimacy and avoidance My major sources for Konde ethnography are the Wilsons’ publications on the Nyakyusa of the Selya area. As with the Pangwa in this chapter, it will be convenient to label each citation as it is presented, with appended brackets. I list the monographs cited in chronological order of publication: A Godfrey Wilson, 1951 B C D E Monica Wilson, 1951, 1957, 1959, 1977 In the context of their study the ethnographers make a strong case for conceiving taboos of incest and ‘in-law adultery’ as versions of one social phenomenon. Here is Godfrey Wilson: One of the most stringent taboos in Nyakyusa life is that which separates father and daughter-in-law. A woman may never speak to, approach, or look at her husband’s father. If she does it is believed that she will die, through the ‘breath’ of indignant neighbours, a painful and lingering death. And the taboo is extended from her own father-in-law to all his classificatory brothers. The normal routine of life is constantly interrupted by women’s avoidance of these men....This taboo separating father and daughter-in-law is one of the reasons for the age-village system ... [A 269]

Monica Wilson adds this: The feeling that sons should not be cognizant of the sex activities of parents, and that father-in-law and daughter-in-law should avoid one another is not, of course, peculiar to the Nyakyusa; all we seek to show is that the extreme elaboration of these avoidances among the Nyakyusa is dependent upon the existence of agevillages. [B 160]

What needs emphasis is the ‘extension’ of this in-law taboo to a husband’s paternal uncles, the father-in-law’s brothers, full and classificatory. It is this facet which makes the meaning of this taboo seem to have ‘generational’ significance. On the game-board of incest avoidance, as one brother falls out, another takes his place and adds the dead man’s responsibilities, along with his wealth, to his own. Another facet of the same phenomenon: a sister’s son should treat all his mother’s brothers’ wives as ‘mothers’. Monica Wilson labels sexual intimacy in relations of this kind simply as ‘incest’. There are reasons, but this wants clarification. It has been an accepted convention in anthropology to define incest as sexual intercourse 298

between ‘born kin’. An extension of kinship sentiments to persons brought into ego’s circle by in-law contracts has therefore been considered as an aggravated form of adultery not incest . ‡ What is special about the Nyakyusa case is the vehemence of the in-law taboo; and since what is also special about this culture is the age-village, it is not surprising to find both natives and anthropologists linking the two phenomena: The incest theme elaborated in one fashion among people organized in clans and practising clan exogamy, and in another among people practising brother-sister avoidance, finds yet a different expression in age-villages, of which the overt purpose is the separation of sons and mothers, of fathers-in-law and daughtersin-law. [B 162]

Monica Wilson cites a case in which a man’s syphilis was blamed not on infection but the curse of “all men” and the shades for having lain with his father’s sister. There is a unique mix of amity and kinship rhetoric in this culture, where the ‘good old boys’ (called ‘the defenders’) of a village can speak and name sanctions in one voice with the ancestors of any member. It is as if they were one family: Very often the ancestors and the ‘defenders’ are believed to work together to punish wrong-doers—those who break avoidance and incest taboos, those who insult seniors, those who neglect traditional rituals. Amity between relatives is sanctioned by neighbours as well as by senior kinsmen...The values of village solidarity and of kinship solidarity are thus coherent. [B 164]

What we have to consider is a question about the nature and nurture of what seem to be the ‘core values’ of a community. Incest taboos belong in this category. They are experientially internal to the family they define, but externally sanctioned by the very ‘structurebearing’ quality of that performative definition: it is as if none of the claims family members may expect to put upon one another could survive incestuous sex—survive not the act in itself, but its discovery to an external community of neighbours and shades. It is this fatal ambiguity about what is wrong with incest (though inner promptings might seem to make it right) which gives the taboo its powerful, implosive/explosive quality. Is it actually possible that the same quality could be extended to relations with in-laws, late as they usually are in coming into ego’s life? The ethnography seems to answer, yes. It suggests even further, that the bannered peer solidarity of the age village behaves like a virtual brotherhood.

299

Consider once again the dramatic meaning of a taboo of dyadic facework which has young women bolting out of sight at the approach of any number of older agnates of the husband. The panicked bolting says at least, “I dread being caught in his sight because he is a ‘father’ of my husband.” If there is a good reason for considering this man as a danger to her marriage, it might be: “He could try to steal me (back) from my husband.” This is a slight twist on, “He will say his cattle were used to let me marry.” What this translates to in structural language is that father and son are, at heart, rivals for everything important in life: wealth, sex, scope, progeny. It is a rivalry suspended in thought. Like all the infinitessimal quanta of envy and resentment within a close family group, this thought is sustainable only while and because it remains unspoken. The little drama of bolting in fear at once acknowledges the impasse between father and son, and taboos the full recognition of the danger so dramatized. Like an ordinary incest taboo, this one defines a particular circle of intimate relationships within which carnal intimacy should and would be explosive, and must be suppressed. A young wife has entered this circle and can’t be expected to have acquired the intuitive cover, the ‘spontaneous’ reaction to social danger, required to keep trouble at bay. So she becomes the trouble and must herself be kept at bay. The difficult tie of amity protected is a structurally crucial one to Nyakyusa community life—the tie of father and son. In the split-up context of ‘family life’ in the Konde scene, it requires a dramatic public reminder to all and sundry: know the particular cast of characters in each young man’s standoff with the older generation, respect it lest we forget who is who in each of our young men’s invisible networks of kinship ties and expectations. Subtle avoidance in this case would be too weak to carry the structural load. A taboo which commands no spontaneous consensus can be no taboo at all. ‡‡ Here is Monica Wilson, dealing in thoughtful retrospect on her findings: According to Mwakyonde, the doctor: ‘In the old days a man did not marry the daughter of a neighbour in the age village. He said: “She is my child, her father and I eat bananas together,”’ but by 1934 this taboo was not enforced. What was considered impossible was for a man to confuse generations by marrying a half-sister of his son’s wife. Her full sisters he never saw—they avoided him; her halfsisters he might greet but not marry. [E 108]

What are the actual boundaries of this special, transactional form of sex-tagged avoidance? and how do these boundaries bespeak the sacred character of the elemental structures they defend? The danger group for incest among Nyakyusa is the same 300

bilateral network of known kin that we find among Pangwa and Kinga. On the intuitive level it at least taboos sex interest in anyone you know by a kinship term. In the supremely polygynous society we find in the Selya studies, their name is legion. Marriage was prohibited between the descendants of a common grandparent, or a great-grandfather in the paternal line, but it was just possible between descendants of a great-great grandparent in the maternal line. [E 107; see Case 37, 107-8]

The main thing this says is what the equivalent Pangwa rule said: you will have to consult your elders from whom you have your social identity and your bridewealth capital before a marriage can be made. I don’t think it impertinent to point out that if the elders can’t actually find a link, in the absence of record they can easily ‘remember’ the exact connection they may need. Memory may be the winning asset always left in an elders’ keep. They need it, as truculent as their sons would have become after a decade of soldiering bachelorhood, and as turbulent as the overall structure of Nyakyusa society was. What it takes to make sense of this case is feeling one’s way into the boys’ age village before the colonial pax. On the one hand there is the psychological permanence of the bond between age mates, and this is early-on reinforced by pretty promiscuous sex, first boys with boys and later wife-sharing where a few older lads are lucky enough to get wives with cattle at least nominally from their fathers’ herds. In fact, it is likely enough these would be cattle originally stolen by the same lads raiding a far chiefdom in a father’s name. The solidarity bred of fighting beside your fellows is proverbial. Then, when finally the Coming Out is called, each boys’ village becomes a little bastion to itself, looking out for its own interests, predatory, with time increasingly free of loyalty claims by the ‘fathers’ in the now post-ascendant political generation. Here is the situation at the start of the long and eventful career of an agevillage group: Since marriage between ‘brothers and sisters of the [inclusive] village’ is approved, courtship goes on between the youths of a boys’ village and the girls of their fathers’ village, with this limitation, that brothers and sisters, or cousins—the descendants of a common grandfather or great-grandfather—must not be present at a flirtation. Boys from the same age-village go about together, and they commonly flirt with any group of girls they meet, whether from their fathers’ village or not. It is then that the incest taboos are learnt. [B 86]

301

Bearing in mind that true brothers are seldom (if ever) allocated to the same age village, the incest taboos affecting one village member mean little to the others. They are not actual ‘family’ to one another, yet they owe each other lifelong amity. At an analytical level this comes close to the maximal-claim bond which is normal to the family system of other peoples. The difference is that Nyakyusa know no such family system in their lives. There are several points of importance. (a) Peer nurture: Though Nyakyusa boys are raised through their herdboy stage (six to eleven) living at their father’s place and herding his cattle, by ten or eleven they are living and keeping house with their peers. Girls stay ‘home’ (as we might put it) while boys, first with the herding group, then with their chums in a village of their own, spend their personally formative days with age mates. In this way, by nurture they are not very likely to be as close to a full brother or sister as to half-brothers and the offspring of ‘father’s friends’. As comradeship is confirmed by joking and needling, by petty rivalries, and by non-possessive sex—the kind where ‘nothing happens’—it is gradually deepened to a kind of male bonding (intuitive mutual understanding) able to support lifelong amity. Girls, first ‘at home’ and then briefly in ceremonial withdrawal with friends in the isaka—held to separate their sexual careers from their parents—live a supervised life by comparison to their brothers. But in the context of rampant polygyny, even a girl’s ‘family life’ allows little intimacy with her father. (b) Patrilineal nurture: While a young man after the group’s Coming Out is in an unforced condition of solidarity with his age comrades, on whom he depends for company and every sort of personal support young men can offer, the collective programme is to build a self-sufficient village. They will build the physical village together, seeing to the plantations and fields. Peer nurture (the ethic of ukwangala ‘keeping good company’) continues to the end. But without cattle and women the new village will never be self-sufficient, and that sets in motion a whole set of projects which must be pursued by each man separately, since each can only work his own line of credit with the ‘fathers’. The core of a man’s ‘fathers’ are the set of brothers of his patriline from whom he can eventually expect to inherit. A young man must have ‘credit’ for his years of working his father’s fields. But patriliny means a good deal more than that. It is a system of reciprocal leverage, of credit and debt. Patrilineal politics brings us into a quarter of the culture where it is not explicit rules so much as ‘case law’ which is decisive. Over time, the involvement of father and son draws in others. Here are some fragments suggesting the complexities of a system of credit and 302

debt which revolves about marriages, cattle, and patrilineal kinship, and provides a concrete measure of the kind of reciprocal trust though which the elders of a chiefdom build and maintain their effective command of public affairs: The heir killed the funeral cows and this act established a responsibility for dependants of the dead. ‘If my younger brother comes claiming a cow for marriage and I refuse he reminds me: “You are the senior, you killed the cattle,” and I agree. Or if my sister’s husband brings a cow when my father dies and I kill it, then later her son comes to me and says: “Give me the cow of mother’s brotherhood (ubipwa), you are my superior.” If I refuse he reminds me that I killed the cow, and I give him a cow.’ [E 38-9] If a man accepted the family estate as heir of his senior brother, all the property he held passed at his death to his junior brother, hence when he died his son did not receive what he (the father) had accumulated before inheriting from his senior brother. [Therefore] a rich cattle owner might refuse to accept the family inheritance from an elder half-brother on the ground that...his own son would lose all, even what he (the father) had possessed in his own right. [E 39] [A case] “I had committed adultery with a chief’s wife and was fined five cows. I handed over four but could not secure a fifth. My son, for whom I had provided marriage cattle, went to his father-inlaw and asked for a calf, which was given to him and he gave it to me to pay my debt.” [E 55]

The structure of credit and debt in face-to-face communities is not a system of semi-permanent rules and roles but a system of changing, lingering, sometimes deeply felt personal involvements of a contractual nature. Intimate role-relationships always foster such involvements, affecting the deepstuff in us. Urgent owings can range from the fleeting debt of recompense for an in-itself trivial favour or gift to transactions in cattle which may be remembered and continued through generations. In spite of the emphasis Konde culture puts on peer solidarity, a deeper kind of debt—bespeaking a more lasting order of personal trust—is the one tied to the maximal-claim bonds of kinship. Responsibility for a man’s debts extended in some measure to his father-in-law, though no more cattle than had been given for his wife could be taken there; it extended to his father or his father’s heir, and formerly to his brothers. ....Crimes were formerly avenged in a wide circle: the father, father’s brother, brother, half-brother, or father’s brother’s son of a murderer or adulterer might be speared ... and so might some village mate. The responsibility of a village-mate was less than that of a kinsman, for the relatives of the mate killed

303

claimed compensation from the kinsmen of the original...murderer or adulterer, not from the avenger. [E 56]

(c) Generational inheritance: When the local settlement has reached a mature condition and is almost ready for the ‘Coming Out’ by which it will devolve into its separate parts, it consists of a stem age-village crowded with women and small children, and a handful of branch age-villages mainly of boys and young men. Power and influence and wealth are stored along with the women in the stem village under the keeping of the village ‘fathers’. The rules of inheritance seem at first blush to run against the principle of village selfsufficiency. At a man’s death his cattle and wives, with some negotiated exceptions, go to the senior of his surviving brothers. If this does not mean the wealth is lost to the village, it means the village takes in this senior outsider. The ruling generations exercised political power and controlled much wealth; the retired generation exercised ritual authority and controlled a share of wealth. An heir took the social position of the man from whom he inherited. He commonly moved to the deceased’s homestead, took his name, and was accepted as a member of the deceased’s generation. The dead man’s sons became his sons, and this was reflected in the celebration of rituals and rules of avoidance, as well as in mutual economic obligations. The heir became tata (father) which was an office, not only a biological tie, and conflicts inherent in the father-son relationship were liable to grow greater in the relationship with his heir. [E 86]

But this rule delaying inheritance by the next generation, and this scenario for working around it, seem to refer mainly to that stage in the stem village’s developmental cycle when it remains the ascendant generation in the chiefdom and is actively promoting its own self sufficiency. This is also the stage of development at which a son (in the absence of same-generation heirs) might be offered the ‘office’ of his tata (father) and so leap forward (or is it backward?) a generation. It is clear enough that the strategy of the ‘fathers’ is one of preserving their privileged condition. That is to say, they are looking to the interest of their own part (the stem village) of the whole village complex. The importance of the vertical system of credit, debt, and trust—which ties individuals of the ascendant generation to individuals of the next—is more easily seen in this light. A son is required to treat all his father’s wives as ‘mothers’ while his father and father’s full brothers are alive, and intercourse with a father’s wife is treated as a heinous offense necessitating special

304

purification; but after the death of the father and his full brothers these same women become wives of their former ‘sons’. [B 161]

Taboo, in short, puts the women with the cattle in the elders’ keep until the root village begins to disintegrate from within, losing its self-sufficiency from aging. But if the ‘fathers’ have been making careful decisions in the handling of their wealth and influence, most of them will have been able to pass on, by distribution during their later years, substantial benefits in cattle and women to the sons they find deserving. There is no indication, in any event, that all a wealthy polygynist’s estate is destined to pass on to the next generation in one package (as it passes in lateral inheritance to a brother, or in quasi-lateral inheritance to a generationally elevated son). The big point of polygyny is progeny, and in a system of generational inheritance the keeping of your peers and your progeniture can be a difficult balancing act. Would the kind of filial inheritance so often found in other Bantu communities work out in this one? It would make age-villages not only pointless but impossible. (d) Turbulence and the war pattern: There is every indication in the ethnography that the movement of individuals into and out of age villages was surprisingly brisk. The movement of young women can be ascribed to the vicissitudes of marriage: first, the original marriage over which the girl seldom had much choice, then later owing to dissatisfactions of all sorts. But the movement of young men must have been owing to the very considerable sense of freedom the early bachelor years engendered. Though we have no data on this count for traditional times, we do at least know that the object of a newly fledged ‘minor chief’ or prince was not to remain minor for long, and that for notable success he would need to recruit a sizable new following. He must do this by dint of leadership at arms and in settling disputes, and by dint of much roast beef and beer. We know also that young bachelors liked to show off, get into fights, and attend funerals in full regalia. We may assume that then as later these same young men made friends easily and tended to find new special chums from time to time. Altogether, I think the picture of a preternaturally concentric chiefdom—with an unbroken ring of commoner age-villages—which can emerge from the ethnography applies better to empirical cases in their early and again their final years than in the rivalrous period of growing ascendancy through predatory raiding. We should perhaps have been reminded more often in the ethnography of the prominence of fighting in traditional times and of Lugard’s early comment that for Nyakyusa “cattle were the object of war.” ‡

305

The Konde pattern is not in its deepest implications quite unlike the Pangwa system for nurturing solidarity (amity) in the local group. Contingency is the basis for the original recruitment of men to a local group, only partially mediated by kinship. Amity is the condition for continued membership. The management of marriages and the political affiliation of progeny is the responsibility of male elders. Through a bridewealth system they are given the keys. A delicate web of dyadic personal alliances—ritually solemnized trusts—obliges these men to work in comity even while committed through kinship to partisan values. It is when we consider how without kinship peer solidarity would be a sole and tyrannical basis of a man’s personal status, that we find ourselves in position to appreciate the contribution great, if delayed, expectations can make to a young man’s freedom of movement. We have a clue, at least, to the often unclouded and seemingly unburdened character of his bachelor years.

Kinga particulars: intimacy and avoidance Kinga informants would not usually claim to know their sixteen ‘great-great’ grandparents but do hold that having any one of their eight ‘greats’ in common would keep an interested pair from marriage. The persistence of this bilateral definition of the ‘close-ofkin’ from earlier times reflects the comparative unimportance of systematic kinship responsibilities to political organization under the long-established Sanga regime. I approach the matter of sexavoidance (and other identity-marking observances which give form to individual social networks) by starting with the microsociology of everyday life as it proceeds at the grass roots. The clearest indication that the idea of ‘a family’ does apply, in the absence of cohabitation, to traditional Kinga is their adherence to special contact taboos. Kinga settlement patterns are far less massive than the Nyakyusa, much more intimate than the Pangwa. Since Kinga traditionally lived in hamlets of (say) a half-dozen homesteads within hailing distance, they were usually able to manage with one ikivaga for boys and one isaka for girls in the early years of a settlement. The story about settling a hamlet begins with a group of male friends setting out from the capital (’sent out’ by the court) with their new wives, and living happily ever after. If there is an element of truth in the tale, still valid for the life I observed in the 1960s, it would be that a young man hardly expects to settle as an adult in the hamlet of his early years. The older hamlets I visited were peopled by elderly couples or widows, and were soon to be abandoned. In a few cases, 306

the pattern of settlement was ‘sprawl’ around court, mission, and market. But until the new government legislated against living in consensual hamlets, few Kinga could have imagined being ‘villagized’. Where the semblance of a ‘local descent group’ organization is suggested in the Pangwa picturing of their distant past, Kinga think in terms of comity. Sons are more likely than not to keep a distance from their father and settle apart. To get from the Pangwa lutanana to the Kinga settlement pattern, you have to add-in a considerable talent for intimate relations among households. Where Pangwa depend heavily on the straitening power of ritual at every structural adjustment in network relations—and resolve problems of comity (quarreling) by adding physical distance between households—Kinga depend on greater fragmentation of the domestic group. Consider these three perspectives:

Male transience. If you take the adult sleeping house inyumba as the centre piece of a traditional Kinga household, and ask how often in a traditional community a man sleeps there, you would have to be guided by several precepts. The regular place for an evening’s concourse with other men, if only for taking beer, was the ikivaga serving as the neighborhood barracks. A man’s civic duties locate him there. His presence in the marital bed is wanted—demandingly—for conception about once every five years. He keeps private things in this house and always has entrée. He has an exclusive sitting place at the fire. Except when he is away, his evening food is prepared by his wife (or failing that, a daughter) at this fire. Heterosexual service is fabulously demanding, while homophilic relations are governed by quite relaxing norms. Female constancy. The only quite regular inhabitant of the conjugal house (inyumba) is the wife. It is where she sleeps with her infant until it is four and ready for weaning. It is where she will conceive and nurse a total of four children, providing all goes well. When she is in her menses she will need to have a hut or trench bed nearby, as she will fear spoiling the sleeping mats she regularly uses. Her man may claim the house as one he built. The woman has finished it and keeps it, just as a man may claim to have cleared a field and broken (half) the soil, but it is women who take over the growing, cultivation, and harvest. A man may absent himself for months, even years, without sending word. Men will not like a child about until it can fetch and carry. Don’t suppose this makes men neglectful. Women will not like a man about after dark until the child is ready. Peer nurture. While peer nurture for boys is not as strictly organized as it is by the Konde age-village, it is ever-present. For 307

girls and maidens in their extended bachelor years, peer nurture is the dominant force in their lives. After the break from her childhood peer-group (at marriage) a woman can’t be happy until she has established a like relationship with new (married) friends. When she has found herself anew, she will not be ready to leave if there is a way to remain. There is good evidence that re-establishment can be difficult for some women to accomplish. The good life is not to be plucked from trees. The Sanga regime by the 1850s had intervened directly in the male life-cycle to an extent the ‘prepolitical’ Pangwa would not have recognized. Yes, informants would say, at his father’s death the eldest or another married son might think of moving back if an aging mother would decline to move toward him. Would she need his help? She wouldn’t say so. Comity is most important, and conversation between generations seldom really free. (Kinga contrast in this to Magoma.) Have in mind that a son can’t enter his father’s house, and the fields his mother works are spread wide. There is no ‘farm’ or ‘childhood home’, no true local community to which a son would be returning. This is a society like the Konde in that ‘home base’ for a man is a group of peers not a place. Kinga men tend to keep with their friends through a lifetime but without the constancy of age-village organization. The circulatory system of the Sanga court meant that ordinary lives for both sexes began in one hamlet and ended in another, but always centred in peer relations. That is why older women were not often prepared to give up their independent ways. What was always important either for a man or for a woman was keeping in close contact with a few friends. For men this was deeply modified during bachelor years by freedom of movement and the broader comity of the barracks life, a pattern in good measure replicated, by the 1960s, in plantation work abroad. For a woman the universal pattern was a major move at marriage away from the place and people of her childhood and youth to a place and people where her husband was prepared to build and provide gardens. Living near your fields was a consideration in some cases, but it was only women who walked out daily through the year, and a woman’s gardens were always (and, for security of harvest, wisely) widespread. The life of a house is about twenty years, and that was likely to be near the full term for a settled hamlet. In older days, gentrifying your place with a raised and drained open plaza in front was not in style. Moving away from an aging place would always be an option, but it would not take twenty years of marriage before a woman’s bed became her own. An aging man needs a woman to cook 308

and bring his food to the ikivaga where he lives. Rights in land are in the long run pretty flexible. Services for which a woman needs a man are few once a fair house is built. His main tasks from a healthy young woman’s point of view will be organizing help from his friends in the fields at soil breaking time, supplying the necessary tools and pots, tending to the thatch, and the business of begetting, which need only come up periodically. It is well to have this picture of ‘home life’ in mind when considering what ‘rules’ and ‘taboos’ might be required to maintain an orderly system of personal identities and mutual claims, consistent with a grass-roots commitment to economic and moral self-reliance through domestic industry, and reasonably capable of reproducing itself. Since Kinga settlement patterns are not dictated by a need to live on the land you cultivate, the fact that settlement was traditionally in undefended hamlets spread about the slopes and ridges of the countryside may seem incongruous with the Sanga war pattern. If sleepers had been significantly threatened by human and animal predators, a ‘village’ pattern would surely have been chosen. I assume there were times and places where these troubles were severe. The one completely obvious function of the quartering of men at the courts was to make possible a rapid response to such alarms. At the political level, the settlement pattern is, in effect, ‘villagized’. A barracks-like ikivaga would have been found at the same three levels of government the Kinga enjoyed in colonial times: the realm and lesser domain capitals, and the several vicinities of a few hundred people within each domain. Only in the last was the leadership on a true primus inter pares basis. This corresponds roughly to the management level of the mkoyo among the Pangwa, though the office there was apical and hieratic. Domestic arrangements always reflect the strong emphasis of Kinga moral thought on good peer-neighbour relations, and the need for a certain ‘aesthetic distance’ between domiciles and generations. But they do reflect a pacific attachment of a small group of households to a place they can make self-sufficient and tangibly their own. Very often, the life expectancy of such a settlement is that of the small group who originally made it their own, so that the place names of a map made in German times were sometimes no guide to another map-maker at mid-century. From the start, boys will choose to build farther apart from their parents (up to a kilometer) than girls do. For both genders the need for distance is set up by the child’s desire for same-sex company, peers who must be drawn from different kitchens. A local boys’ group will visit all the kitchens in the evening, while the girls will 309

rather be helping, each at her own mother’s fire. Boys and men during their long bachelorhood will normally have less contact with a father than a mother, as the father is often sent food where he sleeps in barracks, while the son will be fetching his own with friends. Boys do the rounds of many kitchens, just as Konde boys do. Bachelor girls, for their part, spend the day with peers in their own gardens, returning to a mother’s kitchen with often ample provisions. Always, seasons and situations vary. But these are the prescriptive routines, and it will be seen that they differ from those of any true ‘kinship society’ by allowing great scope to amity and the free choice of associates. I have remarked before that, judging from its expressive culture, youth is the best of times for either gender. It seems to me uncannily rational of the Kinga to extend it through half the normal life span. Many cultures may seem, in Kinga perspective, to throw youth away. Amity groupings require to be reaffirmed on all occasions of face work. When they are not, the ties fall away. ‘True kinship’ is to be known from ‘virtual kinship’ by this: when a man settles far from his father, while the two may be content having little to do with each other, their ‘structural’ relationship doesn’t decay. Outward evidence of it may be as slight as their inability to meet on neutral ground emotionally, or shrug off the code of taboos which governed their manners from earliest years. But if a man moves away from the (fictional) tie of ‘proximal kinship’ he has contracted with a neighbour, the social elastic will probably break—what do they owe each other? ‡‡ In looking at the Pangwa and Konde cultures I have been obliged to treat taboos and avoidances as conscious rules, hypostatized bits of custom, following the lead of the respective ethnographers. For Kinga I have the benefit of my own ‘cross-examinations’ of informants. Intimacy and avoidance are nouns descriptive of sometimes subtle forms of human conduct which, when structural facts are reduced to ‘rules’, will too likely lose their subtlety. I asked Soda to reminisce about ‘beatings’ he had seen. He said mothers would beat children with frail bamboo rods, but a woman would not try beating a boy of more than nine years. Why not? The child would fight back. A smaller child would run away, whether from its mother or father, and would have to be physically caught. Then it is held by the arm and beaten on the lower body. A mother might refer an older boy to his father for punishment, but a child who managed to run away would just stay away until the next evening when tempers would be cooler. When a son is as big as his father,

310

especially if the son has already beaten up his father in a fight, the father will settle for verbal chastisement. Here are some cases: (a) A father was beating a boy about nine. He had him in the house and beat him severely. When the boy emerged he grabbed a heavy log and swung it at his father’s head as he emerged. The father parried the blow. The son threw rocks, but none of them scored. He ran away. (b) By the time he was nine, Soda and his friends were skilful at throwing rocks to fend off father or mother. Soda recalled fending off his father that way, hitting him on the leg. That was enough to turn the father back. I asked what would happen if a father was killed by such a rock. Soda recognized that could happen, but had nothing to offer on what the consequences would be. At the time, Soda himself was ready to marry and beget, but I couldn’t evoke any ‘Greek’ sense for parricide. (c) When Soda was of school age he had happened on a strange scene. An older youth was beating his own mother, holding her and beating her in the standard fashion of a parent thrashing a child. The mother shed all her clothes and screamed, “Very well, if you are my husband that you beat me so, let us go inside and have sex!” The lad only continued beating her naked, and she ran away. Soda took it that the mother was trying to shame her son into relenting, but failed. (d) Soda had heard of a father beating his daughter after the age of menarche, but had not seen it. Nam had shared boys’ talk with me about a girl’s coming-of-age rite. A less than dutiful girl might be beaten naked before the crowd, not by her father but a brother, at the behest of the local married women in charge. An aged veteran of those schools at Ukwama said an elder brother would be asked if the girl met the standards of modesty and decorum. If not, he would be given a stick to beat her. Soda as a Christian had no views on this kind of beating but had seen both fathers and mothers beginning to be wary of physical combat with a girl of nine or older, moving to moral suasion—scoldings. Own mothers are not of the party which stages a maiden’s schooling ceremonies. What can be meant by this turning of the celebrant ‘mothers’ to an own brother—who can only dread thinking of the sexuality of his sister—to bring the scolding home? It is bold, and may say more to the young peer groups, of either sex, who have been called to watch than to the unnerved principals. Rue doesn’t rule the day, in any event. The scene soon turns to one of great fun, a naked 311

parade, a comic procession louder by far than Lwembe’s and one that might make that dour patron-divinity of the Sanga rulers lighten. Hand-to-hand fighting is an occasional expression of stressful intimacy. Women fight more intimately than men. They might use sticks, but mainly they wrestle (Soda uses the same word as for hugging) trying to pull and push the other down. They can bite while hugging in this way, grab cheeks, and claw. They fight close, never at a distance as with rocks. Biting off fingers and tearing earlobes is special to a serious fight between women, perhaps co-wives where domestic emotional involvement overlies a stranger relationship. Men would never fight with spears, though they might with knobkerries, holding the light end out. The least intimate kind of fight is the boys’: they throw rocks, and in their roughest game they hurl heavy sticks, whirling them low to the ground. The game is rather open-sided, since targeting is inexact—the stick may hit anyone who is slow to dodge it. As for girls and bachelor maidens, Soda and Nam seemed to agree in their estimates: even when they might be feeling mean they won’t openly fight but hold to civility. For Nam, who was younger than Soda, this fine sort of meanness was the essence of the feminine mystique. Segregation is a way of limiting intimacy, as avoidance is a way of stopping or preventing it. Children learn some specific rules, and learn to be cautious in dealing with another person within the circles of intimacy generated by kinship and marriage. Soda was surprised when I told him that a young man of my acquaintance, a bachelor who had finished school, was currently sleeping with his little sister of thirteen, just becoming womanly. They were sleeping together for warmth. They chatted about it without any bashful demurring—they just got along well. I thought, “They are just not used to having to sleep alone, and out here they are each away from their kundi, their little peer group.” Soda thought, “Maybe his penis doesn’t work.” Soda himself was definitely against touching a sister’s body. He expressed aversion at the thought of sleeping beside a sister, especially if sharing a common cloth. What might happen in sleep? Greeting and touching were not disallowed. You could do that innocently. But casually sitting together with your sister after childhood days was not the kind of thing Soda would be seen doing. Why? It wasn’t bestial to chat with your sister, but Soda felt you should make your innocence of desire unmistakable. We talked about a young man who had been discovered by his father lying with a ‘mother’, a co-wife of the lad’s own mother. The father reacted with stagey astonishment. It was not a serious offense, but one properly discouraged by ridicule. It couldn’t be thought of as bestial (goatlike) since the son would like as not inherit his mother’s co-wives. 312

But as for adultery with a ‘mother’, the father’s cavalier treatment of such a transgression was roundly deplored by Mwanadyo, a majestic man still, and a long-time court-leader raised in another and more tightly structured realm. Only discover such a case in traditional times and it would be taken there and then for ‘witchcraft’—unthinkable. If it were not punished by death (the chiefly court so ruling) quite surely the son would be banished from the father’s life forever, and the woman sent packing, the brand of scorn upon her, to find another man. Partly, this difference of jural perspective goes back to precolonial variance among the several realms. In the East, the Sanga court was not so firmly established, the ‘bush culture’ not so seamlessly fitted to Sanga court custom, as in Mwanadyo’s Northern realm. Partly too the differing perspectives reflect the decimation of the East in the Maji Maji uprising, and its gradual repopulation since. But most I count rank. Mwanadyo was co-opted on account of aptitude to the highest position in a principal court. His project in his world had been one of closing the gap between categorical orthodoxy and merely mundane social practice. He would have seen the power of law used to enforce a taboo which spontaneous sanctions would not. Wisdom seems to stand astride. It seems parental incest is better blockaded than sibling incest. A boy learns directly after weaning (say, at four or five) not to touch even his mother’s clothing; not to touch the cloth she has used to sleep, even if his father has not yet shared it; never to overstep a parent’s sitting or sleeping place. A girl who might sleep with her mother on safari, sharing one mat or cloth, would be doomed to shiver all night, so little warmth could her mother offer. A boy of five who stole into his mother’s hut by night would be pulled out by the ear. Weaning is so final. Oedipal fixation of this kind is sanctioned by the ‘spontaneous’ fear of a mother’s seemingly fatal charm being taken as witchcraft—a deep thing, worse than the boldest crime. Little children are allowed in the Sanga-style woman’s house, which is her kitchen and sleeping house, shared from time to time with her husband. The little ones must always keep to their own ‘child’s side’ of the hearth, avoiding a parent’s sitting place or stool, and quite particularly shun any indirect contamination of the cross-sex parent. By the time a boy reaches puberty he has learnt he may not enter his own mother’s house at all, and must treat the house of a father’s sister or father’s brother’s wife with similar respect. While he has the freedom of his mother’s brother’s wife’s house, he must respect a mother’s sister’s house. These rules of respect are reciprocal, applying to the way the boy’s own marital house will be treated by his elders.

313

All this appears again when certain of a boy’s best friends marry. One of them is so related by kinship that the boy calls him ‘father’, but they are age-mates and only make fun of the paradox. Theirs is a personal tie of privileged intimacy, in Swahili ‘utani’. They go courting together and extend the joking utani familiarity to mockintimacy with one another’s girlfriends. But all this ends forever when one of the boys is married. The wife of a ‘son’ is a daughter-inlaw, the wife of a ‘father’ is a ‘mother’ and there are respectrelations. The boys who once shared a bed can’t even share the same room. The boy who once played with a girl’s cloths and sat on her bed finds the girl who used to heat his wash-water for him taboo. Any cross-generation tie within the great-grandparent range can be suspended among young peers but remains there in the utani mode, dormant. When a boy gets to know his older brother’s wife, he is soon into utani with her, badgering her for food and calling her ‘little wife’, but the wife of a ‘father’ (though she was formerly privileged) may offer him food or not. He can only stay and talk with respect. They are now of different (adjacent) generations and occupy compartments in the generational structure which must be denied intimate contact. Should we call this a taboo of ‘incest’? of ‘own-group’ wife adultery? It is cognate to both but a marginally different logical use of sex-avoidance to flag a structural feature limiting and defining the community of trust an individual needs in the delicate dealing demanded of kith and kin. Consider this as a case in ‘sanctioning theory’. There is a restructuring of Ego’s personal network by any marriages affecting it. A brother’s wife moves into Ego’s circle of familiars, and assumes there utani status, while a ‘father’ may oblige his new wife to put aside existing familiarity, assuming the manners of mutual respect. Are we in position to say what this is all about? It is certainly about defining Ego’s social space by taboos. There is demarcation of the morally sanctioned behavioral flow in Ego’s familiar network as between privilege and respect, and always with the requirement of reciprocation in the relationship entailed. The structural consequence of this is interesting. Ego lives continuously in a ‘spontaneously sanctioned’ world of familiars, even while at each new stage of life the players in the field may change. It will help to reconsider relations among age peers. In 1960, young men might have to be taught in court by their elders not to sue a brother for adultery. Peer tuition is limited by peer experience, and marital experience began late, especially for men, even in 1960. The exact boundaries of innocent intimacy have to be discovered on your own, by comparing your reactions to others’. The context may be story-telling, joking, or getting in deep. But some 314

of the rules (taboos) are part of early learning. The generational line which must not be crossed was learned while the child was young enough to be privy to his parents’ house: bedding, sitting places, and intimate garments of other-sex elders on both sides were untouchable. It was particularly dangerous to overstep a father in this way, the more so for boys than girls. A boy should also avoid such circumstantial intimacy with a father’s sister or a mother’s brother’s wife, and to a lesser degree with a father’s co-wife. When married he must cleanly avoid overstepping his wife’s mother. A man must greet his mother-in-law with great formality, avoiding direct address, keeping his voice in a child’s register. Both should kneel to greet at a distance. The intimate greetings of older peers entail prolonged arm contact and face work, repeating good words and eventually exchanging any news. These performances speak to the structural norms of the community in ways no child can misunderstand. At grandmother’s house (I was always told) there was only indulgence and great fun, no carried-over tensions. Even Christian men in 1960 would dread entering a mother’s sleeping house, or a married daughter’s. The effect of these formalities echoes the effect of weaning itself, which thrusts a child emotionally upon own-sex peers. Where Ego is male, the dependence on peers can be almost total. Then how is it that a free-roaming young men’s kundi never turns into a freeroaming ‘gang’? Do we have to look ‘beyond’ the system of sanctions or only look more carefully at it? Who, in fact, performs the weaning of a boy? Here is a favoured scenario: The mother has the suckling brought by his ‘brothers’ to the ikivaga they have made for themselves. In the evening she brings his food there, leaving the little kundi of ‘big brothers’ to console him with the food and convince him that he must follow them. If he is apt he soon learns to despise the little boy who once followed his mother like a girl. In most cases he will already have been teased into aspiring just this move into a bigger world, and all he must learn now is that this adventure puts an end to nursing. The structural significance of suckling, which has put an extended stop to parental coition, is not something he will have considered, and it will be another generation’s turnaround before he does. For the time, his only easy familiarity with elders will be at a grandparent’s or his own mother’s brother’s community. There is a seeming paradox in this absence of ‘fear’ for a mother’s brother, as if he were no sort of ‘parent’. Parallel cousins are very close to Ego, at least on the agnatic side. As children, these cousins are ordinarily quite like siblings. Ego’s father’s brother’s children are likely to live near and run in the same little makundi (boys with boys, girls with girls) Ego does. The boys’ ikivaga would have two 315

sides for sleeping. Big boys, being sexually mature, must on no account sleep with younger ones. When Ego is seen to have matured (there has been a night emission) it may be his own elder brother who takes him over to sleep with the manly set. For practical purposes, this is his ‘rite de passage’, and it must always be memorable. For girls in their isaka the transition is not so explicit; they are declared nubile by the women’s ‘school’. But they return from the ‘school’ to the same isaka and remain through their mature bachelor years closer to their ‘sibling-cousins’ and other friends than to any other human beings, unless it be their own mother. The mother-daughter tie is disciplined. It is reaffirmed daily during the girl’s childhood, when she follows her mother to the fields and grinds corn for porridge. But later she will have her own fields and till them with her kundi. While her mother lives, a girl never owns a cooking fire in her own community. Nam would always remind me, too, that grown maidens are thicker, more social, than grown lads. When we ask what these observations tell us about the social structure of a small Kinga community, it is pretty clear that the two little makundi of a typical hamlet do have kinship significance, even when many of the dyadic ties are ‘virtual’. Siblings and parallel cousins are not terminologically differentiated. A core group who call each other by ‘sibling’ terms and ‘fear’ the same ‘fathers and mothers’ constitute the controlling group within each of these genderspecific primary groups. A taboo of cross-sex intimacy is maintained ‘spontaneously’ even while non-possessive, same-sex intimacy has to be seen as the raison d’être of each group. A paradox? Perhaps it only seems odd, that the mother’s brother’s house is a free place, because own-parent houses there are not free to the minors who belong to the place and daily depend on it for cooked food. Cross-cousins are not called by ‘sibling’ terms but by terms which have a distinguishing vowel. The cross-cousins a visiting young person would meet at a mother’s brother’s place would not have free entrée at a parental house there—they might see a bit of a paradox in watching a distant cousin of their generation taken into a house they ‘fear’. But they would soon enough understand, if only from their own experience abroad. The point of the visit is not, in the first place, the intention of the young person but of the mother, who is visiting her home place and reaffirming connections. Her child must be welcomed by his or her peers as one with kinship rights to their hospitality. Their cross-cousins will over the years be taken into their circle of trust. A web of amity, under the name of kinship, is spread abroad for the new generation. Interestingly, the broker in this business of making a halfway house between full-kin peers and mere friends is an adult who can act like a peer. 316

The ‘free house’ of a mother’s brother can be, like a grandparent’s house, free of ‘fear’ because, structurally, it is remote from the day-to-day world of intimacy wherein the rights and wrongs of conduct are defined in a young person’s deep experience. As with Ego’s home kundi the ‘spontaneous’ balance of sexual license and taboo which prevails among his cross-cousins and their friends will be assimilated to his own. The status of sister’s son to a mother’s brother will remain, for life, one of Ego’s kinship resources where mutual claims must be honoured. It takes only a moment’s reflection to see why, if your mother’s personal close is not open to her own child, that of another ‘mother’ who feeds the same little kundi cannot be a free house. The very presence, for boys, of many kitchens and ‘mothers’ to beg from, the ‘grand tour’ boys will make each evening in twos and threes with their begging cups, extends their world. It comes to overlap, in a disciplined way, with the worlds of other neighboring makundi. The ‘system’ in domestic life wherever it is found is not spontaneous: it is, interestingly, rather the source of that special spontaneity which leads its members to keep the rules of the family game. Despite the relatively casual feel of agnatic loyalties among Kinga as compared to the Pangwa in Fr. Stirnimann’s portrait, these two neighbour communities have much in common at the subpolitical level. Both use the rhetoric of kinship (-kolo/-xolo) to forge close and lasting ties within a local community. Both also use a dormitory life for boys to propagate lateral ties of peer loyalty, even at the expense of a vertical inheritance of social identity. Both use the indirection of bridewealth negotiations to give a generation of elders the control they need in settling a younger generation into the structure of adult life. I am not close enough to ‘real life’ in a Pangwa community to carry the likeness further toward a reading of the deepstuff of that culture. But as I turn now to looking at the patterning of moral strategies in Pangwa, Nyakyusa, and Kinga contexts, some final implications may turn out to have been elementary, after all. My argument has led me toward the position that taboo (moral spontaneity) is as keen a force as law (obedience to rule) in giving form to these human communities. What I think has been little understood in our ethnographic literature is that Law, universally and correctly seen as the strong force in any system of social control, has to be paired with a weak force of a quite different nature. Our name for this is Taboo. It is the active principle in every system of social cohesion. Only consider the Old Testament meaning of ‘Thou shalt not kill’—it says naught about the treatment of ethnic enemies. Contrary to the popular stereotype of taboo, the only kind 317

of ‘fear’ it is likely to provoke is the kind which is a necessary ingredient of respect. When a Kinga boy says, “Ndidwada udada” (“I’m afraid of my father”) he is speaking to the moment. When he tries to describe the quality of this, his most difficult relationship he is more likely to say, “Ndikundwada udada”. The noun udada refers to one’s own father and others who stand in the same sector of the boy’s kin network. The verb -dwada means to fear. But when it is used indirectly (as in the latter of the two phrases cited) it carries the meaning of a Biblical fear, ‘My father commands my respect.’ We have been considering the warp and the woof of a system of rooted amity. The vertical strands hold nothing without the lateral, nor can the lateral take orderly form without the other.

318

FEAR & BLAME, CHAPTER EIGHT

Given Terms of Existence About the deepstuff in history When Herbert Marcuse turned to appraising the relevance of knowing human nature to knowing history, his somewhat flashy title was Eros and Civilization. It would not have helped much if he had substituted Thanatos for Eros—any way you may stir up Freud with Hegel and Marx is not going to give you a clear soup. I mention Marcuse partly to disclaim pretensions like his to grand theory even when it seems to be likeably ‘of the left’. I use deepstuff to suggest the deep implied in a generic notion of depth psychology but without partiality to any packaged brand thereof. The use of stuff is not casual. It makes the principle clear that content is to weigh more than modality. I’m wary of the conceptual tendencies tied to the term personality because it has lent itself to a mischievous analytical strategy—defining a fixed set of types, then sorting specimens (persons, cultures) to fit. Some readers may recall how, at midcentury last, Henry A. Murray’s elaborate efforts to provide a language of content for the study of personality miscarried. Followers cannibalized his exploratory vocabulary for tags which could be used as one-dimensional personality types with which grand explanatory ‘hypotheses’ could be exploited. Wide-angled inductive exploration ought to come first. Have in mind the option of ‘chiaroscuro’ vs. the ‘hard edge’ in graphic art. It strikes me as right to prefer the former in the case of a subject with no edges showing. ‡ This chapter undertakes to explore the interplay of structure and sentiment in the rise of the Sanga protostate. My strategy is first to consider the as-if historical divergence of Pangwa and Kinga during the formative centuries of the Sowetan political archipelago, roughly the 1700s and 1800s. The next chapter considers KingaKonde differences with an eye to defining their scope. Throughout, I make the assumption that there can’t be radical structural change in a community without radical change at the level of the culture’s deepstuff. The observations which flow from this assumption try to focus on what Murray called ‘regnancies’—motivational proclivities 319

which distinguish for each culture the dominant moral strategies of men and women. In considering what Murray meant us to understand by ‘regnant processes’ bear in mind that mid-century academic psychologists usually thought that by studying the highest of the ‘higher organisms’ they were covering everything in human nature relevant to their interests. Partly for this reason and partly, I suppose, in diffidence to critical readers more conversant with a ‘harder’ version of science, Murray doesn’t want to be taken for a student of thought but of ‘brain processes’. But in saltier language, he is interested in what is actually and deeply ‘on the mind’ of a subject when he or she is doing his or her thing. A regnant process is a special concern, evident in a person’s conduct, bespeaking a characteristic mode of involvement with the world. Regnancies are functionally at the summit of a hierarchy of subregnancies in the body. Thus, to a certain extent the regnant need dominates the organism. ‡

It may help (to excuse such intellectual body language) to be reminded that Murray’s book was dedicated to Sigmund Freud, Carl Jung, and Alfred North Whitehead among others. Murray knew about the deepstuff but didn’t want his insights to contaminate his method. If he had talked about moods and obsessions, hopes and fears, twists and tricks, the wrong people would have read his stuff. ‘Regnancy’ was one of the inventions he used to disguise a reliance on insight, and it is a useful, well-defined notion helpful in exploring regularities in the way cultures work to produce norms of conduct. Inherent stability is a feature of folk communities so characteristic as often to be taken for granted. It is reasonable to suppose stability sans inflexibility is no illusion and can be explained in good measure by assuming that the social system of a typical folk community ‘works’, and does so smoothly enough that internal contradictions are kept to a trivial level. Changes are not noticeably disruptive in a short run. Is there anything further which needs be said? Only that the ‘equilibrium theory’ based on field observations of social stability is faulty: there is no evidence that its real scope extends beyond the evidence of observation on which the theory was generated. Equilibrium is just another greedy theory trying to clean up more than its market share. The reason small-scaled social systems ‘work’ has a motivational aspect which skeletal theories of social structure ignore. I have argued this point at great length elsewhere. It is recapitulated in capsule form in the notion of ‘moral strategies’ which this study of the Kinga in their region has proposed. Just as hunger and dread might haunt one man and thoughts of escape and vengeance another in the Gulag, there is 320

normally a link between a person’s objective and subjective situation, but not a determinate one. In considering regnancies in relation to culture, though we can’t speak confidently about individual cases we can about well-defined social roles. That is the premise of my notions respecting moral strategies and the moral career. They let me model the systematic connection between persistence and change that I need if I am to deal with the rise of a new cluster of institutions associated with protostate politics. I look first at Kinga, later and more briefly at Konde institutions. The ‘prepolitical’ institutions of the acephalous Pangwa represent the foil against which protostate institutions can be modeled. As a foil to the Nyakyusa, a valley people such as Ndali or Nyiha might have been more appropriate, had the pertinent studies been done. But with the ethnographic close-ups Fr. Stirnimann offers, we achieve a certainty about the ritual structuring of acephalous societies unequaled for the region. In my assessment of the evidence, Pangwa are representative in many particulars of the acephalous peoples of the mountain slopes ecozone in the Sowetan region. Linguistic evidence does make clear that Pangwa would be less representative of proto-Konde communities in the Rift valley. But I can assume in both ecozones there were Late or Post Iron Age farming and herding settlements, predominantly Bantuized, spreading with a newly intensified and adaptable agriculture across the land, and that local settlements were small, politically autonomous, and unwarlike. In considering the rate of spread it is well to have in mind that iron tools were chronically in short supply among Sowetan peoples still on the eve of the colonial take-over. The likely time period for the peaking of a ‘prepolitical’ culture is the few centuries before a.d. 1600 which seem to represent in this region transitional times between Iron-Age style and early Medieval settlement patterns. The adjective ‘medieval’ in Bantu studies generally implies developed chiefly authority on a local scale, and increasing reliance on the spear in translocal relations. But marginal areas throughout the second millennium permitted migratory drift as a function of political localism. The Sowetan archipelago appears decidedly late when set against the time-scale of the betterstudied political development of the Interlacustrine region. Did ‘stimulus diffusion’ from the north prompt protostate development so far south of Lake Victoria? News-lore would have been coming in from all directions and would have helped prepare some sort of shift away from localism. But a study of Malatan development can best deal as directly as possible with conditions and processes internal to the regional culture. It is pretty clear, for example, that the political news which might have stirred the Kinga didn’t so stir their cultural cousins across the river in uPangwa. ‡ 321

Deep divergency: Kinga/Pangwa An early version of the Sanga regime brought uKinga into the Sowetan political archipelago. Sanga, though continually expanding, never tried to dominate another protostate, as in the latter half of the nineteenth century the Sangu and Hehe protostates and several Bena chiefs did. One of the reasons was geography. Kinga lived in a cooler, less lush part of the Southern Highlands surrounded by natural borderlands. They had few cattle to target, and goatrustling is, on all accounts, not worth trying. They were protected from the slave-trading routes, buffered to the north and east by the thinly populated lands of the acephalous Wanji and Wemba, buffered to the south by the similarly situated Pangwa, and protected to the west by the Rift escarpment and, below it, the militarily solid but self-involved Konde communities. Why in this remote circumstance did the Sanga protostate arise and achieve its considerable sophistication? I propose a plausible storyline. (i) An early important step in the process was perfection of an exceedingly efficient, very-long-handled hoe, a tool which made deep turning of the soil more effective (if perhaps not less strenuous) than a ploughshare. Dare I say this excellent hoe liberated the Kinga woman? It gave her practical economic independence in her daily life by giving her scope for the bounteous cultivation of fields, many and widespread. Another excellent tool, which we usually call a ‘bushhook’, would go with her to a stand of trees and return with something like a hundred kilos of firewood elegantly balanced on her head. The perfection of the Kinga hoe and the industrial capacity to produce it in quantity would have made for a gradual intensification of agriculture and export trade, both centred in the more fertile parts of the country. In short, centres of cultural efflorescence would tend to appear over time, and in a spreading pattern. As Pangwa had similar tools (some Kinga-smithed) but never got chiefly institutions, the technological means to achieving economic development helps set the stage for a programme such as the Sangas’ but doesn’t account for the drama. (ii) The great change, and presumably the priming one for Kinga, must be put to the account of Sanga politics. This was taking men substantially out of the fields and into a barracks life as cooperative guardians of peace, subduers of the forests (for clearing, for timber and wood, or perhaps a smithy’s charcoal), protectors against two and four legged predators, and labourers on private and public projects. In early days, before the land was so nearly cleared or put into secondary growth, many men were regular hunters, using dogs 322

and the bow and arrow (the unanticipated forest-hunter’s weapon which seems to have cleared the Ngoni out of Kinga territory at midcentury). To accomplish this change, a man’s nagging demand for wifely sexual services within the long period of her nursing (a demand which Fr. Stirnimann finds problematic among traditional Pangwa) must have been convincingly abrogated in the Kinga case. (iii) Completing this transformation wanted only consummating the emotional independence of women from men and men from women. There are several points. The new division of labour and its affective complement could only be pushed on as courts were developed at the new centres which needed and could afford a hefty share of young adult men’s time. The trademark Sanga pattern emerged: circulating young men inward to a power centre and older men back out to marry, to settle and police an expanding borderland. The Sanga lord or prince was differentiated from commoners as the only strongly polygynous member of his domain. As the embodiment of man’s heterosexual nature, the proper monarch was withdrawn completely from the barracks life and from fighting or policing, being devoted entirely to furthering the fertility of his wives and, through sponsored calendrical ritual, the far-flung plantations of his domain. Many wives afforded him many hands in the fields, since pregnancies were far apart, and a multitude of royal maidens were kept at the royal fields through their best years, raising crops and preparing food for the barracks men. Keeping the royal wives and princesses apart was guaranteed by their seclusion within a regal enclosure, and by deference taboos. (A man must step aside and verbally salute even the youngest princesses when they were met on a path outside the stockade.) Making all this possible was the acceptance by youths and bachelor men of easy homophilic relations within one of the two age groups of the ikivaga. As we have seen, the age separation there was based on de facto generation. Genealogical generation only became a basis for deference after marriage. And lest it be thought that the whole domain was somehow drained of its male youths to swell the ranks in one central barracks, I note there were isivaga for adult men at every level of organization. For schematic purposes one can picture three levels within each domain: at least one ikivaga at the central court (capital village), a formally linked ikivaga then for each recognized segment, and sometimes temporary local isivaga wherever splinter groupings were to be found. The system was meant for the easy association of men with men, for their circulatory movement within the domain, and for maintaining a residential presence of government close by any local settlement. To these must be added the multitude of isivaga in the pre-Sanga style which served the domestic needs of bush-culture hamlets. 323

In briefly outlining this as-if historical agenda for carrying out the Sanga protostate process I may seem to have thought many generations of Sanga innovators—all the while spreading out over more territory and occupying more separate, autonomous domains—had nursed in their heads a neatly programmatic scheme of manifest destiny. But the real narrative is less unlikely and I think it will bear examination on evidence. We need to model in some detail the moves by which a people as determinedly traditional as Fr. Stirnimann’s Pangwa could have gone through the structural changes the scheme suggests. The technological starting-point of the two societies was about the same, as the two populations took form with the same migratory wave, with iron-age Bantu replacing (absorbing) an earlier, less sedentary, pre-Bantu population. Perhaps Kinga enjoyed some marginal advantages of environment and disposition. But it is the political and psychological changes for which we have sufficient evidence to draw conclusions. I take it on myself to present enough of Fr. Stirnimann’s findings in his own terms (always in my translations) to provide the evidence we need of the way ‘tradition’ was maintained by the Pangwa in the absence of distinctive political institutions. Tradition is not automatic. It required in the Pangwa case an enormous and whole-hearted effort on everyone’s part. Even the tiny suckling was not exempt.

Pangwa particulars: life cycle and life style The picture of Pangwa culture we are concerned with is neither current nor recent. The ethnographic present for this exercise is 1850, before a series of radical disruptions and displacements began. The ethnographer can feel safe in his observations of traditional ritual, inasmuch as continuity has been strongest in this aspect of the social life. On the strength of observation and interview in the 1960s a fair portrait of the old life-cycle can be contrived and compared with changing conditions. What is more difficult is contriving the kind of insight into day-to-day matters, the quality of marital, kinship, and neighborly relations, which participant observation offers. Specifically, I find in the Stirnimann account contradictions between statements about the mutual isolation in space of individual households, belying kinship ties, and statements suggesting hamlet-groupings, even utopian solidarity within the major politico-ritual unit, the lutanana. Of particular concern is the quality and importance of casual association— occasions of phatic communion—among men. Assume that men met in amity for other than the solemn purpose of placating dangerous ancestors by sharing meat with them. Where and how often did what groups of men meet for phatic communion? A closer acquaintance with the ishyengo would clarify. It matches the ikivaga where Kinga 324

men ate and slept. Its openwork walls betray its origin as a protective house for goats, where herdboys had their fire and slept on guard. Detective work is wanted on its enlargement to serve higher purposes. The same questions about sharing space and affirming ties of amity can be approached semiotically by asking how valid was the claim of kinship among men. Are good neighbours closer kin than bad? Or is it simply that closer kin are better neighbours than more distant kin? These are questions about the microsociology of an historical society a fair century past at the time of fieldwork. The oral chain of communication is three or four lives long before the ethnographer records a tradition about how things were. Ethnography at such distance is perforce an exercise in ‘meta judgement’. Obviously though: finding context in the filial culture of the 1960s can be used to advantage and should be. Fr. Stirnimann sees Pangwa life-cycle rites as collectively performative: that is, each rite at once advances an individual to the next rank on an age-status ladder, and in its collective aspect insists on the social reality of such a ladder: The community is articulated through rank and status distinctions. In the course of the life-cycle the Pangwa [male] person moves from status-less child in mother’s arms to ranking household head who exercises authority over his issue. The life-stage rituals bring the transfer from a lower status to the next higher into public knowledge and acceptance of the associated rights and duties. The teaching during the menstruation ceremony, at the giving over of the bride, at the institution of a household, at pregnancy and parturition offer striking examples of the expectations of society for the member advanced to the new status. [II 292]

Social integration in this picture appears to come from the weaving together of a male status ladder, associated with a doctrine of agnatic descent, and a cycle of rituals either concerned directly with procreation, or otherwise employing phallic symbolism to create an (obscurely) integrative world-picture: Working together in harmony and peace in the lutanana [patrideme settlement] has been the basic condition for genuinely communal social life at all times. Just as the soil ensures the local lineage’s material existence, so does the survival of the kinship group ensure the fertility of the younger generation. Fertility serves along with patrilineal blood as the community’s highest good, which is conveyed directly by the shades to adolescent youth. That is why semen and ancestors are always associated in ritual. Loss of

325

fertility is thought of as severe punishment from the shades. No wonder then that nearly all the important events in house and field are symbolically ‘identified’ with the cycle of procreation. The most important forms of specialized knowledge—iron smelting and rain making—were associated with marital coitus, in the same way as daily activities were: the fire-drilling, grinding, pestling, winnowing, eating. [II 291]

Harmony and peace? These are ends everywhere espoused and almost everywhere partially realized, but Fr. Stirnimann reconstructs a lifestyle in the lutanana which can credibly be called a recipe for social stability, so long as the comparatively defenseless Pangwa had been left undisturbed. The pattern differs from the Kinga in important respects. Four divergent tendencies in the two cultures deserve extended discussion: (a) sharing, (b) sexuality, (c) genderpeer solidarity, (d) sanctioning rights.

(a) Sharing: Although both societies know the same pattern of sleeping houses, Pangwa are more restrictive. Each family has its own set of houses: kitchen, parents, boys, girls, stores. Even where (as only rarely occurs) a man has two wives both with children, the two wives have parallel facilities located well apart. But while it is not entirely clear I suppose it unlikely a man would build two of the kind of house used by boys, since that place has a number of more general uses. We are given this picture: The Pangwa in past times knew no close, village-like settlements in their lutanana, which might cover several square kilometers of acreage, but only scattered homesteads, surrounded by gardens and bamboo stands. The homestead, ilichumbi, consists of the house for man, wife, and small child; further of sleeping huts of sons and daughters (post weaning), and the storehouse, ishyuvu. The custom of scattering settlement over the lutanana was explained by elder men not just as necessary for protecting the fields from wild animals (apes, wild bores, hares, antelope and such), but also by the fear that living too close together might easily lead to breaking the peace and accusations of using black magic....On the same grounds the houses of co-wives and of married sons were scattered over the lutanana. [II 64-5]

The author goes on to discuss the high probability of feuding and counteraccusation within the deme which could cause the hiving off of homesteads and founding of independent small settlement areas (new demes) at a distance. Peace and harmony, in short, was 326

achieved mainly by disengagement. The same structural principle was applied to the sharing of personal space within the homestead and in the work of the fields in which several households would participate. The homestead was a school in microcosm for later life in a systematically self-privatizing world. Since the ethnographer’s picture is necessarily a pastiche based on informant reconstructions, care is wanted in our readings. In the painterly rendering of family life which follows, only the bits on herdboys in the bush are suggestive of the Kinga scene—close family life in the Pangwa sense was beginning to appear in Christian Kinga communities in the 1960s. It had played no significant part in the court culture in Sanga times, though it remained a recognizable feature of bordering bush cultures throughout the mountain slopes ecozone. With the sons in the boys’ house, the father was the uncontested authority, calling for respect and obedience from all, and who saw to it that all mothers sent the necessary food for meals, as well as that peace and harmony ruled the group....After breakfast the boys led the goats into the bush in the grasslands and spent their time there in games, with trapping bush rats and snaring songbirds. But the father controlled the herding regularly and should the boys let the herd wander into tilled fields, they would be given serious warning and in the case of another offense would be punished by withholding supper and caning. When they were rough-housing and fighting over small things, the father would not intervene... The evening food is the main mealtime of the day. The little sons fetch the laden eating baskets from their proper mothers and place them, without touching, before the father....No son, big or little, was allowed to touch food before the father or a guest present had been satisfied. They sit on quietly in the background and wait patiently until their turn comes. ...The father sees to it that the little ones get their portions just as the big ones do, and are not discriminated against. After supper the father sits with the sons in cheerful conversation by the warming fire, inquiring after the day’s experiences and success in the small-animal hunt. Later the mothers, nurslings on their backs, and the girls arrive at the shyengo [male sleeping house], but sit at a separate fire for conversation. When it is time to retire for the night father, mother, and girls take their leave...Father and mother go to the elders’ sleeping house, the girls to their hut. [II 109-10]

The reference to a plurality of ‘mothers’ and offspring may be misleading, since polygyny was rare. A homestead teeming with children, given the long spacing of births, also must have been rare:

327

In earlier times, when the child was nursed until loss of the milk teeth, six or seven years after the birth, the nursling remained until weaning under the direct care of the mother. Only afterward did the boys remove to their own sleeping house (ishyengo), exchanging the mother’s nearness for increased contact with the father and other boys of the settlement. Little girls stayed during the day with the mother, but slept in the shared sleeping house (inanda). [II 109]

Another matter wanting probabilistic judgement is the simple one of understanding the actual uses of the ‘Knabenhaus’ (ishyengo; cf. Kinga ‘ikivaga’). It is quite clear that it was the scene of daily contact between father and son(s). It was also the scene of a daily reunion of the whole household (around separate fires, for the two genders) after the evening meal. It functioned as the boys’ goathouse: Through the day goats remain in the care of boys and girls of about 6-12 years. Now and then in thinly settled parts today you may even find one single talking-and-sleeping house for boys in the lutanana. The goats are in any case kept apart in one corner of the ishyengo hard by the sleeping platform of the boys. Should predators show up, they are chased off with glowing coals from the hearth fire. [I 251]

I discount the possibility that a boy of six ever was left to sleep alone in such a place, with such duties. The probability vanishes on consideration. The house would have been in use for boys of some limited number of families. From the evidence of recent settlement patterns this would be a group of ten or twelve households. It is obvious that a child of six or seven of either gender would not be expected to sleep and keep house alone. Women (on the testimony of learned elders) would have produced no more than one child in about seven years. Supposing half are of each gender, ten households would likely have about ten boys and ten girls to contribute to the sleeping houses for each gender. If there were then more than one house for each gender in the settlement cluster, it is unlikely there would be more than two for girls and two for boys. Considering that the spacing pattern of the homestead was meant to keep the three parts of a family separated, it could not have been meant thereby to leave very young children without older siblings or cousins to ‘herd’ and tend to them. If we must admit some doubt still as to the size and composition of the community, it is because one account seems to imply the boys’ house was an unshared structure, one house to the boys of each homestead. They are charged with herding the goats, which they do collectively, and guard them at night. Packets of food are 328

sent in via the smaller boys, to serve all. But the father(s) of each uterine sibling group will normally be taking the same meal, while boys obediently wait for their share in the background. ‡ Assume for the moment that each household head wants to build an ishyengo for his sons alone. Then a critical point wanting attention is the probability we should assign to a father’s presence overnight, sharing the ishyengo with offspring during the long nursing of a younger sibling, when the father ought not to be about begetting. There is reason to suppose that, after the evening concourse including all genders and ages, mothers with their sucklings would retire to their ‘parental sleeping house’ while the father and boys remained to sleep in different ends of what we should be regarding as the family’s common talking house, used overnight as protection for the goats and their herdboys. Given the ‘hands on’ approach of the father toward discipline there, I take it that was most likely the Pangwa custom. But the arithmetic says several households would typically share an ishyengo, with half a dozen boys at one end and their fathers (two or three men) at the other. Avoiding sexual intimacy at the one end would require the kind of discipline the ethnographer describes, and at the other end its continuation in self-discipline. There is an implication of strictly sanctioned sex-avoidance. The ethnography describes a ritualistic social environment which lends that credibility. Fr. Stirnimann offers no clarifying explanation for a system of spacing children six years apart, should the husband continually share a bed with the mother and her child. At the same time he does describe the six- or seven-year hiatus in reproduction as a notable form of birth control. As I follow his discussion of the Pangwa social system, I think the Pangwa way of assuring such an hiatus would be to use avoidance taboos. Some features of a reconstructed past have to be asserted as interpolations: there is ample evidence that Pangwa folk memory packaged the past in a series of ‘stories’ or ‘frames’ not as a seamless panorama. It would be curious if they had managed to do otherwise. While we have no statistics on family size, I think Fr. Stirnimann’s guess that a woman could expect to bear six or seven children is high. A woman had thirty-odd years from puberty to her withdrawal under strict taboo from marital relations, which allows time for four or five young if they are nursed until the milk teeth go. All this is without taking in the vagaries of nature. The reference to a son’s congress with ‘other boys of the settlement’ can only mean that, on leaving his mother’s house, the boy was ready for herding and its free contact with other herdboys. Would a lone, small household have its own private ishyengo? We have 329

said no. Would such a household expect to flourish so well in the normal course of things, that it eventually required its own ishyengo? But the purpose of such a house is night-time defense for the herds. This is a house with the strong open-work walls proper to animal pens, and a smudge fire to control flying pests. As we are told there were regularly two fires, it can only be the biggest roofed house in the compound, and the most expensive (in firewood) to maintain. As will appear, both of a polygynist’s wives and children attend an aftersupper concourse in the one ishyengo. But if (as I am now supposing) that house is not the private property of one family, all the girls and women of the sharing families would be there. Nothing in Fr. Stirnimann’s hard evidence suggests that brothers and parallel cousins would not have grouped their households, at least initially, so as to permit using a common ‘boys’ house’ and nanda ‘girls house’. The question then is how the ishyengo was conceived by Pangwa in 1850. Specifically, how many adult men (‘fathers’) slept away from the boys and the goats at the other end of the ishyengo? I hate to leave that question iffy. Consider the odds. The term ishyengo is rendered now and again as “boys’ sleeping house” in the ethnographer’s text, where that rendering serves a reader’s convenience. The house is in fact much more than that, though much less than the communal mixing place the Kinga counterpart, the ikivaga which I translate as a barracks house, had become. There were only a few of those still in traditional use in 1960, but there were isivaga everywhere of the sort used for boys and goats. I don’t suppose the Pangwa ishyengo cannot offer us a fair picture of the original ikivaga, and explain the dual use of that term by Kinga. A boys’ house (called ikivaga in some communities at least) is closed to grown men—‘fathers’ and the like. But the men’s barrack house in many communities, such as not-modern Maliwa, is shared by men at their own end and boys at the other. There will be two fires, as nights are cool to cold in these mountains, but as with the Pangwa neither fire is for cooking. That is done by women at their kitchen houses. It is well to have in mind that a Pangwa family compound was according to informant tellings not jointly used even by co-wives. Two such families might jointly plant and harvest, but always the fruit of any field would go to its owner’s store. What was ‘joint’ and what ‘separate’ is a study in itself. Every tilled field in the lutanana has its individual owner, i.e. usufructuary, who planted it with [his/her] own seed and harvested it as well. Should a man have more than one wife with children, they

330

don’t live together in a distinct settlement but in their own houses near their own fields. [I 82]

All this, with but one exception, could be said accurately of the Kinga but has quite different implications in their case. There are no bounded ‘farms’ in a slash-and-burn landscape. Any suggestion in the citation that there may have been such in uPangwa I think should be discounted. The Pangwa/Kinga divergence is not black and white. Land holding was not land ownership but usufruct right. Fallow years were many more than tillage years for any field except ‘bottom land’. Patterns in ‘personal space’ are not defined literally ‘on the ground’. But Pangwa acted them out there quite explicitly: Both in burning the dried vegetation mounds before the first rain and also in planting the gardens and taking in the first fruits of the year, the age-ranking of the field’s owner must be observed... Before puberty the father apportions each child a field. When after the first rain it is time to plant, the children must begin this work according to their own age rank, not their mothers’ standing. Married daughters, young women, follow the age-rank of their husbands. In the patriarchal lutanana the brothers, sons, grandsons of the mkoyo observed their age rankings precisely, since their conviction was that the seedlings of one’s older siblings would fail to germinate and sprout, if a younger one, or even the mother, planted first, thus overstepping age-rank. When a man had several wives, in October the eldest child of the head wife, whether a son or a daughter, kindled fire in the first heap of dried vegetation in his/her field [and the others, with the mother, followed in order, taking coals to burn their heaps]. [I 93-4]

I want especially to note the ‘outsider’ status of a mother, over whom her own child takes ritual precedence in the custom of the old lutanana. Fr. Stirnimann is certainly not wrong in thinking the spirit of the lutanana is ‘patrilineal’ even if the sociology is not. In the practice of a collective soil-turning mkovi (Kinga ungovi) the Pangwa avoidance taboos are actually acted out on the ground. The silent rhetoric of avoidance coded into Pangwa kin relationships, as Fr. Stirnimann with ample if indirect evidence insists, was woven about social dangers of the sexual kind. Custom clad the women in a fashion bound to flag the danger of close encounters, which must be avoided especially in a cooperative soil-turning which brought a number of households together. In the ordering of the mkovi in the field the appropriate avoidance taboos must be observed, so that the collective work of men and women of different ages won’t give opportunity for seduction or

331

adultery. In early times women and maidens wore only narrow barkcloth strips loose on a belt, which would flutter in a light wind, uncovering the genitals. ...Therefore, lining up the work force for an mkovi in the field required observation of certain avoidance taboos. [I 91-2]

A complex mazeway is constructed around this fanciful danger. A son must not be working near his mother, the co-wife of his father, the paternal aunt, the mother’s brother’s wife, or near his sister or a female cousin. A daughter must not stand near her father, father’s brother, mother’s brother, brothers or male cousins, in the hoeing row. A daughter-in-law must be separated from her father-in-law and his brothers, her mother-in-law and brothers thereof, and her mother’s brother with his male cousins. A son-in-law must avoid his mother-in-law and her brothers, her female cousins, as well as avoiding his father-in-law and his brothers. A father-in-law must beware of the son-in-law and his siblings or cousins; also the daughter-in-law and her sisters. A mother-in-law must be separated from her son-in-law and his siblings and cousins. As we have seen, strict kin-based linguistic and proxemic taboos among Kinga only begin to apply after marriage. The same lexicon of avoidances survives there without being interposed and elaborated as a discipline of youth. Kinga boys graduate into a disciplined relationship with peers rather than from it. But for girls, unless they were burdened with the task, and its freedoms, of herding the family goats, respect for their elders was ingrained to the point of seeming ‘natural’. Working so close to anyone else as to let hoes clash together was for Pangwa to be avoided at all costs. In short, interpersonal relations were minutely defined with respect to the social dangers of intimacy on every public occasion. Inasmuch as after such an occasion there would be a festive drinking party, the mention of sexual ‘opportunity’ is perhaps more pertinent than might otherwise seem. Rigidity and brittleness are never total strangers. When informants were talking to Fr. Stirnimann about a community of household heads—men—as we saw in the previous chapter the emphasis was all on integration: men shared everything “like children of the same father.” To me this bespeaks an ethos of male peer solidarity like that of the Kinga or the Konde. I sense it in the very semantics of kinship. Both in the past and also even today in districts with less mixed populations the uluxolo [family, archetypal kin-group] was reckoned the paradigmatic social unit. It comprises patrilineal relatives of 3-4 generations in one or several extended families and stands under

332

the authority of the ‘great father’ (dadi mxomi or mkoyo) who is the direct descendant by first-born standing of a recognized founding father. [II 57-8]

This usage belongs in the Sowetan regional culture to an institution of land-right allocation powers to a certain ‘owner of the land’ who first settled there. Land thus ‘descends’ through time to the several households of a community in the way property might descend in a culture where land titles exist, only not by inheritance but contractual cession, person-to-person. Here is the way a new lutanana must begin: As soon as the first hut is erected and fields tilled, the new settler is recognized as land- and house owner, mwenye inyi [owner of the land] or mkoyo. [II 58]

Everyone in an established lutanana knows who is currently regarded as mkoyo. As the eldest functioning male in the luxolo whose ancestral burial grounds hallow the land, he is celebrant in any ritual addressed to those shades. What isn’t clear is how far, politically, his powers exceeded those of celebrant-spokesman. Could he— would he try to—prevent or otherwise influence the kind of division by ‘hiving off’ which any society of this kind is prone to whenever the main game of procreation is successful? Converting a neighbour into a follower is the essential magic of chiefship. If and when some mkoyo did so the event must have been felt as a bid for political power— unorthodox. But as we are perhaps inclined to think of ‘gerontocracy’ in relation to the luxolo institution as described, it is well to consider that radical departures from orthodoxy are really not often attributable to old guys. In his political aspect the mkoyo commands the kind of routinized fear we call respect. In his liturgical aspect he is the broker in a religion of blame, passing on responsibility for trouble to the shades, and representing their inalienable rights in the matter to their descendants. What can now be made clear is that anyone coming in from outside to take up land in the claimed area of a lutanana accepts his ‘civil rights’ there through recognizing that his right to till new fields or retill old descends from the original ‘founding father’. In this way he is jurally assimilated to actual blood descendants of a recognized if not directly remembered founder, since these descendants have their rights to land in exactly the same way, through negotiated cession from a spokesman (mkoyo) for the ‘trust’ or ‘corporation sole’ through which the original personal ownership lives on. That an immigrant must turn in crisis to his neighbour’s ancestors, where in ‘ancestor worship’ this would make no sense, is sufficiently explained 333

by the firm association of these spectral ancestors with the soil they once tilled. This gives us, then, the sociological basis for the semiotic merging of ‘neighbour’ with ‘cousin’, kith with kin. We shall see that this is fundamental to Pangwa social organization, not a merely opportunistic fiction or ploy for allowing non-kin settlers in, against tradition. In effect, the greatest thing a man can share with his neighbour is kinship. Just as a Konde man moving to a new chiefdom understands his destiny connects him to a new chiefly ancestral shrine, a Pangwa in a new lutanana understands he is part of a community inhabited by the invisible landowners interred in the soil he has chosen to cultivate. Sharing ‘kinship’ whether genetic or contractual remains the most profound way of sharing identity the Sowetan cultures have established, and we find it here in a form special to peoples living from land their ancestors still own. The most likely scenario for the in-migration of kin-strangers to a lutanana in ancient times is just at the point where a few householders feel forced to hive off. They are setting up a new community in uncleared country much in the way Kinga men did in their bushward recirculation under the Sanga untsagila and Nyakyusa men did on a far larger scale at the ubusooka Coming Out. The big difference is that Pangwa could assume no politically backed support. But where some particular group led the way, others—particularly land-hungry young men of a neighboring settlement, being connected by marriage—could follow. The political and economic preconditions for setting up corporate agnatic local descent groups are simply not met in the uPangwa we can reconstruct on Fr. Stirnimann’s excellent account. As we have seen, that is what the quasi-patrilineal deme organization, defined by Murdock, is about. For Konde quite explicitly, and for Kinga in a less world-shaking manner, the essence of this shared structural experience is the liberation which comes with a fresh start, a small-group version of world renewal. Where Kinga can conceptualize a new settlement in light of a ‘westward movement’ of the Sanga regime, Pangwa make a clean break from the old community and Konde find their village cast into a new political arena whose pattern of alignment and opposition has to be felt out afresh. About the sharing of food it may be noted that the distribution of food is policed by the Pangwa father in the interest of fairness. We might want to call this ‘obligatory sharing’. Its basis is distrust of the children’s and even a mother’s willingness to share fairly. There are suggestions that a Pangwa father is not above policing his women by inspecting the food baskets they send to the ishyengo for the morning and evening meals. ‡ 334

Kinga food routines assign it a radically different meaning. To begin with, it is the woman who is in charge of all the food. She takes responsibility for feeding men and boys and rarely will face complaint. Boys come to her kitchen in threes and fours, gratefully take from her, and go off to share it before going on to another ‘mother’. Every boy seems to have a dozen ‘mothers’, and though many are unmarried they can be begged from during the day, as when they are met on the trail with edible roots or other finger food. As there is only the one prepared meal a day (in the evening), women are off to the fields in groups early in the morning, usually before men are stirring. A woman and her daughters keep common stores, and passers-by may be offered food from a woman’s fire. Food is part of the language of openness to sharing—the rhetoric of amity. As for personal space, the centre of early life for Kinga children of five and onward is a sleeping house where friends of friends are welcome, and there are two main sleeping mats, one for the young and one for the older. Peer socialization achieves discipline but grounds in amity not fear. Games, conceits, stories, lampoons—if they seem to be endlessly amusing it is because they are endlessly reflective of local incident and character—these are not hallowed tales told at grandfather’s fireside. Words before sleep can form the cutting edge which deepens friendship and the tutelage an older boy can afford to offer a younger, or a bright girl offer a slow. Of course, my remarks are based on only a few good friendships and visits in the field, and are biased by the same awkwardness of the anthropologist—asking for confession and getting testimony instead—which led Fr. Stirnimann to his homilies on peace and harmony among the Pangwa. But I could never report that peace and harmony were imposed on the Kinga by any sort of patriarchal fiat. When you see fun going on all around you, though you may have no clue really to its deeper springs in personal experience, the one thing you know well is that it is spontaneous. It is particularly striking to see men and women of every generation enjoying the prodigiously energetic work of turning the soil, the women clearly bound to outdo the men, in the Kinga version of the Pangwa umkovi [Kinga ungovi]. In the part of uKinga where I joined some of these ‘hoe-ploughing bees’ it was the custom to hold the beer until evening. The party was as lively as the work; but the work had been as merry as the party and had managed it all without beer. Sharing is a major dimension of human sociality. In looking at divergence between old Pangwa and old Kinga lifestyles I have entered into the sort of interpretive documentary which is more convincingly done by a filmmaker than an ethnographer armed only with a treacherously unreliable Kodak. A delicate kind of detective 335

work is wanted. The particular truth has to represent the general, because the writer, be it Fr. Stirnimann or Dr. Park, has asked himself the meaning of behaviour he can best know only in his own experience of others, not in the tracks left behind by that experience on paper. Granting that any step of inductive reasoning risks error and the need of correction, the divergence between our two models of the two Sowetan cultures is real enough, however differently two unlike observers might paint the particulars. Even as I left the field I would have said of older Kinga men that they were strikingly fond of beer, and of older Kinga women that they were impressively self-reliant. On reflection I would have to downgrade beer to give it less importance than the wide net of long-lived personal friendships men feel compelled to maintain. And for the persistent freedom of widows from dependence on men I would substitute the lifelong reliance of women upon the company of the peers with whom they share their days. If sharing personal space begets a sense of sharing a common fate, it is easy to see that the microsociological change toward greater reliance on peer amity in both gender worlds of Kinga, which occurred under Sanga dispensation, would have been concomitant with a sweeping macro sociological transformation. What is most obvious is the diminution of shared time and responsibility within the nuclear family.

(b) Sexuality: Fr. Stirnimann offers two firm clues to Pangwa sexuality. The men strongly deny and abominate homophilia in any form. The women are quite content to devote themselves to nursing a child until its milk teeth fall. How individual experience falls out in a social maze so construed, we are left to ponder. This is (is it not?) the ‘real maze’— these are the rules of the game, defining where social danger lies and what ground is safe. As to paths followed ‘on the ground’ we do know this: there are no sins in the theology of animism, the danger of taboo transgressed lies without not within. That is, as long as danger is unfelt in the event, it is unfelt in the inner sequel. For pondering the male side of the Pangwa ledger a proper scene to have in mind is the ishyengo—see-through walls, offering privacy only after the evening fire has died away. Probably boys learned to sleep in the distant presence of their father from about the age of five or six. Then from the beginning there has been the repetitive intervention of a ritually sanctioned institution called

336

xukida umwana ‘shortening the child’. What is done was described thus by a (Christian woman) specialist practitioner: “School children of both sexes bring me a little basket of sweet corn as payment with the request to be ‘shortened’. The child lies down on its back, whereupon I remove some skin, of course cutting according to gender. When bigger girls come to me I also remove the hymen, with which I may draw for the first time a little flow of blood. If a maiden were not opened, the bridegroom couldn’t achieve coitus. Nevertheless many Christian maidens today are refusing to be opened.” [II 166]

As there is no male circumcision in the Sowetan region, it is not really odd that Fr. Stirnimann interprets this mild form of mutilation, for boys and girls alike, as a ‘preparation for marital coitus’. That both sexes are treated alike, that the operator in both cases is a woman, and that the operation though mild is certainly painful, says to me that the message for the children is along these lines: ‘Genital organs are more important, more dangerous, and more special in quality than children are equipped to know’. But the message to adults is subtler and possibly more revealing. There is an elaborate ceremonial prelude to the ‘shortening’ of a boy, at least equivalent in ritual-theatrical terms to the later puberty rite for a girl, but this is a rite he is not destined to recall. For Pangwa the sexual worlds of adjacent generations are mutually closed, but with one peculiar point of contact. This contact begins with the celebration in a male infant’s first year of the manly potential of his little penis. It entails months of ‘nursing’ the penis itself (as well as, in interludes, its new owner). It is the kind of rite that ‘says’ one thing only in order to ‘tell’ you another. Fr. Stirnimann offers an exhaustive text based on detailed observation and interview, with tape recorder and transcript. I can offer no more than a synopsis—there are twenty-odd pages of particulars on the one rite. But I want this event to be seen in the full catalogue of Pangwa life-cycle ritual, as I know no better way to suggest the weight of ritual intervention this community bore. To that end I reproduce the relevant section of the Table of Contents of Fr. Stirnimann’s second volume. The section is devoted to ‘Die Riten des Lebens’ [II 153-280] and runs as follows (as usual, in my translation): 1. Introduction: Worldview, Rites, Symbols.......2. Preparation for puberty (a) Rudimentary circumcision (b) Unyago [sexuality rite] for the masculine suckling (c) Puberty rite for girls........3. Marriage (a) Recruiting the bride (b) Proposing alliance (c) Managing the bridewealth (d) Giving over the bride (e) Installing the new couple in the homestead (f) Marriage customs in change......4. Teaching on

337

pregnancy (a) Pregnancy taboos (b) The pregnancy girdle.......5. Birth and early childhood (a) Birth (b) Mother’s return to house (c) First purification and integration ritual (d) Naming (e) Child’s presentation to its grandparents (f) Final purification of the mother (g) Weaning the child. Appendix: Abnormal birth......6. Burial and mourning rites. Burial of an mkoyo [ritual leader] in 1922

We are concerned with section 2 (b). I abstract (i) the action (ii) the native teaching which accompanies it (iii) some of Fr. Stirnimann’s clarifying interpretations. (i) The ritual as action can be presented in five parts: preparation, sacrifice, drill to mature the child’s testicles, repetition, policing.

Preparation. A number of ‘stage properties’ will be wanted. The young father of an infant boy must be armed with two specially prepared pieces of wood, each made from a particular species of tree, must hunt a game animal (or tie a goat), and must cage a cock. His wife must cage a hen, and must be furnished with three special water-holding baskets, one of them large, furnished with clear springwater; she must fetch smoked fish from the seaside. Plenty of firewood will be needed. Material must be readied for raising a solid partition within the house, making an inner and outer room. The young man’s mother must organize an mkovi—plenty of beer and garden food must be made ready. The main celebrant and organizer is likely to be the young husband’s father’s sister, past menopause; she is assisted by the young man’s father’s brothers and mother’s brothers; and by the young woman’s father’s sister and mother’s brother. This party, the inner chorus, must supply the special bits of root with semen-like white sap which must be put to soak in one of the small calabashes. In addition there is a ‘second chorus’ of men and women associated with the marriage contract or with preparation of the feast. All these direct participants must be in sexually cool condition. Others in the neighborhood, kith and kin, will join only in the all-night merry-making. Sacrifice. The place of the sacrifice is the grave of the eldest agnatic male forbear, or another suitable grave which is nearer. The celebrant is the oldest agnatically connected male available. The usual prayer or spiritual contract accompanies a small offering of beer, meal, and meat: in effect, ‘We haven’t forgotten you, give us your blessing’. Cock and hen are pointedly kept apart throughout. Tokens (cock feathers, fur) are taken to adorn the infant’s father later on. He has dedicated his sexual powers to service of the lineage. 338

Drill. The celebrant-teacher is a learned woman past her menopause. Such women are tabooed sexual relations. (Presumably her husband sleeps in the ishyengo just as Kinga men do.) There is a learned chorus inside the hut, listening with care to events behind the heavy screen which the teacher has erected to shield the sanctuary where the crucial rite unfolds. There is a protocol to be taught concerning the three containers of wash-water. The parents must be taught to pull back the foreskin and address the opening of the glans. There are several steps in the procedure of ‘spraying’ the infant: the foreskin, the calabashes, cleansing, hand-milking right and left breasts to spray into the penile orifice without touching the breast to it. The husband’s full attention and assistance is wanted at each stage. Milking should continue in each session until no more milk presents. There are insistent teachings and repetitive songs from the ‘first chorus’ following each step by ear and urging the ‘labour’ on. The emphasis is on application to an onerous but necessary task. An outside chorus supplies more song. Repetition. All through the night, as the mother gets more milk, the ‘work’ must be repeated. (Presumably the babe is fed also orally on demand.) The ‘labours’ and teachings are reiterated, the chorus attentive, until the second cock crow and daybreak. But outside there is an all night party with an assembly of the entire community, excepting the own parents of the couple, who must absent themselves. Throughout, both mother’s side and father’s side are well represented and made to feel responsible for the success of the performative aspect of the ritual, which is its overt purpose. Policing. All through the night of the esoteric ‘labours’ the husband has been abjured to monitor and police the continuation of the ‘labours’ nightly with no unnecessary omissions. This is to continue until the child’s erection shows the celebrant-teacher that the job is finished. This will take some further months of the same co-operative ‘work’ repeated just as it was taught the first night. The elders eavesdrop and check for spilled water by the sleeping platform, etc., to be sure the ‘work’ is kept up at pace. (ii) Some details of the native teaching. When the husband while hunting with weapons or working with tools injures himself, and blood flows from the wound, he should forthwith get home and warn his wife. She must empty the three calabashes in the bush and leave off the ‘labours’ until the wound has healed. First then may she refill the calabashes with fresh springwater and take up the ‘labours’ again. The case is similar when the wife injures herself or menstruates. But should the husband have a night emission the interruption is only for a day. [II 177]

339

[Verbatim excerpt of a taped interview with a celebrant-teacher:] When a husband is having frequent intercourse with his wife, the part of the semen which he gives into the womb is turned to building and nourishing the child. The other part gathers in the breasts of the wife. When a maiden reaches puberty and is not soon married, her breasts grow limp again. But if she is married, and with daily coitus there gathers a rich supply of semen in her womb, her breasts begin to swell again, so they will be full at the child’s birth. When the mother of a suckling is nursing, paternal semen flows with the milk into the child, whether it be a boy or a girl. Our forefathers have taught us, however, that we must inject semen into a male suckling along with the mother’s milk, letting it gather in the testicles. When a maiden is still at her father’s place and has no contact with a man, there is no semen in her womb. When she marries, though, and frequently accepts semen from her husband, she becomes pregnant, and later a child is born. The semen given by the man to the woman in intercourse was planted in him by his mother through the injection of milk into the testicles. We inject no milk in the maiden. Later on semen will be coming from her husband and pregnancy will occur. [II 186] [Part of the parting lecture given by the celebrant-teacher to the young married couple as the rite is closed at dawn:] Your grandmother performed this ‘labour’ for your father, and your mother, young man, has done so for you, exactly as we have done tonight for your son. If your mother had neglected the ‘labour’ you would never have been able to father a son. You both must observe this custom daily for many months, mutually helping and encouraging one another, getting up at night, waking the child, lighting a torch, and spraying milk for the child. This is the only way the little son will gain manly strength, in order to beget children of his own. Should the wife be lazy, then the husband must take her to the elder women. But should the man have no torch ready to kindle and be unhelpful to his wife, then she must accuse him. This custom will cost you a lot of pain, but if you neglect it the great fathers below will punish you. [II 182-3]

(iii) Finally, two interpretive comments on the meaning of the rite from Fr. Stirnimann : The right side symbolizes...the man’s side of the family, the left side the woman’s. From the right breast accordingly flows semen for the later forming of sons, from the left breast for making daughters. [II 180] This unyago rite takes place by dark of night. Night is holy time, as after sundown the mahoxa [shades] rise from their graves to walk the land. The sleeping room is a hallowed place, as it is here the shades work together with the parents in intercourse, calling

340

descendants to life. No more could the shades enter a brightly lit sleeping room than they can appear in daylight. [II 179]

What the whole, enormously elaborate piece of ritual theatre says, we have heard. It is a teaching about the importance of the younger generation continuing by ritual means the work of procreation and societal maintenance. It says the (alien) woman is central to the continuance of the husband’s agnatic line. “Everyone is watching you two, that you stay on the straight and narrow, for their sake.” It also tells the young married pair that their honeymoon period of dyadic withdrawal is over, the real ‘labour’ has begun of providing for the future, and this requires a six-year period of abstinence. It tells them that the man’s onerous work in building the baby is done. It has been taken over by the woman, through her milk, which ritual medicine and his manly labour has given her. Her milk is now her version of the agnatic semen the child still needs. It tells them that the central presence in the ‘parent house’ of the homestead (ilichumbi) is now a new couple, mother and child, entitled to their own extended dyadic withdrawal. Unless the Pangwa have some secret system of birth control the good Father did not discover, the mention of nocturnal emissions during these months of sleeping together confirms the reading I have offered. The Pangwa practised a successful taboo of intercourse right through the nursing period. Many or most other societies expect such a taboo to hold for a matter of months. My take on this is that where a wife can be tabooed to her husband for more than a month or two, it only takes screwing the clamps on firmly to keep it going indefinitely—the man is already broken in. What is impossible in a culture of laissez-faire need not be so where the clamps are firmly applied, and it seems to me Pangwa have all the clamps that could be needed. Since no such elaborate ritual theatre is called forth by the birth of a girl, it may be reasonably supposed the nursing period, if she were the first-born, might have been shortened in practice, but I have found no confirmation in the record. The suggestion in the ‘teaching’ favours something closer to a premise of equality between the two sides of a marriage in their private dealings, and with that between the two genders. Yet in their public conduct, orders of precedence seem to belie this. For Kinga women parturition entails an abbreviated application of breast milk to a boy’s penis, a modest period of seclusion with assistance from older women, and a ‘test’ a boy must pass (by following the movement of an arrow across his line of vision) before 341

the father will accept him home. These are charades of transparent meaning establishing the limits of claims and exemptions from claim in the marital relationship considered appropriate to the occasion. The sexual expectations of a Kinga marriage are legitimately met by the pregnancy and the building of the foetus to the point where the child is felt as a presence in the womb. Men cede control of the cycle, from conception to weaning, to the woman. Withdrawal from the coital relation, whether for nursing or to reward the timely sterility of menopause, appears to have been taken in stride from traditional times by the husband, and to have been deemed not unsatisfactory for the wife. Begetting is a duty around which privileged sexual access is laid on by Kinga rules. For the most part, Kinga men and women outside the Christian dispensation have seemed to approve. But what is most significant to my view is that only a ‘spontaneous’ avoidance taboo would have protected the long period of nursing wanted in either society—law is not enough. Pangwa young women undergo, after the long period of ‘shortening’ has led to quiet defloration, an elaborate rite de passage usually combined with an equally ceremonious marriage, rife with instruction by women no longer sullied by sexual contacts—women past childbearing and who have had no ‘mother’ role in the child’s life. A similar ritual school is held with parturition and entails weeks of seclusion before the woman can be restored to her house and a start made (if the child is a boy) of preparations for her laborious ritual transmission of potency to his testicles. One can only notice that these schools are peculiarly effective because, being imposed on life crises which put the subject at risk, they offer operators the special leverage of mixed excitement and apprehension. Kinga young women probably knew ‘shortening’ in some form as late as the 1960s, as even Christian girls in the Western realm went regularly to Christian elders of their gender for open-crotch inspections, complete with righteous advice. It was also commonly known that Kinga boys and men had no word and no clear image for the hymen—it had no part to play in their lives and little in the lives of Kinga girls and women. The ‘shortening’ operation is definitely part of the Kinga bachelor girl’s puberty school. Men are said (by the operators) to have trouble penetrating if nature has not been improved in this way. Men accept this sympathetically. They have problems of penetration of their own, which they resolve by learning to wait for a small flow of lubricant before they begin. As married men are easy as well about the occasional hard work of begetting, always smiling in retrospect as they reflect on woman’s demanding nature, it is clear there is more than duty entailed for them.The encounter compares, after all, to a Westerner’s second, third, or 342

fourth honeymoon. But the period of renewed intimacy need not be long. In a few months a man’s wife will feel sure of the pregnancy and cease requiring masculine attention. It is perhaps worthwhile to remember here another semantic curiosity of kiKinga, which results in the use of one word to designate either a woman’s birth canal or a man’s rectum, depending on the sex of the speaker. There are three words for the three main human crotch organs, but only one is used in the same meaning by speakers of either sex. This is the word for penis, isava. Both sexes agree males have one of these which they carry in the forward position. Men say men and women will both have something their regional neighbour peoples such as Bena would render as a ‘cock’, that is untsogolo, and carry it behind. Women say both sexes have a rectum, unkundu, behind. But women say they themselves carry a cock, untsogolo, before. Men deny it, saying what women have in the forward position is unkundu, something women say they have behind. There are obvious psychosocial implications but the major implication for us here is that men’s and women’s small talks have diverged. The genders live separate lives of intimacy from an early age, and manifestly have done so for many generations. If we ask what institutions the Kinga community will have lost over those generations, accepting the Pangwa of 1850 as cultural stand-ins for the Kinga of a century and more earlier, the answer in a phrase is ‘grass-roots ritualism’. Here are some notes of Fr. Stirnimann on the metaphysical basis of Pangwa ritualism: The shades live throughout the daytime hours in the deep below (panyi) and appear to their progeny in dreams, their clothing covered in dirt. Nonetheless, they are taken as manifestations of the life principle. In this a certain ambivalence shows: Life and Death, fertility and pollution come out as opposites. Still they almost seem reciprocals. Between them lies as a protective layer taboo with its sanctions: bodily and spiritual sickness. [II 284] In all rituals the close connection between ancestors, masculine potency, and fertility are quite generally emphasized. They form the claim of kinship, free for a youth the first flow of semen, in the maiden first menstruation. They help the young husband with the sex act and are actively present at conception. From the semen of the husband they build the infant’s body parts and join them artfully together, determining thereby the sex of the child as well. The field, symbol of woman ready for conception, is fertilized through the rain, symbol of the shades and masculine semen. Semen and shades are in this context ‘identical’. ...

343

Various rituals find their expression and climax in marital coitus, thus the installation of the bridal pair in the new house, the great mourning for near relatives, the purification of the husband after the death of an infant and the reconciliation of quarreling marital partners. [II 283]

Kinga group themselves as gender peer groups in the ungovi festive work party, each group keeping its own pace. Their positional relations are not kin-generated in any important sense. The ‘shortening’ of boys doesn’t have a counterpart unless it is that men still retain the ritual authority to confirm a boy baby’s fitness to come in out of the quarantine of the birthing hut. Though Kinga maidens are monitored by women of wisdom as Pangwa are, the ‘policing’ is mild. Though wee Kinga girls are mother’s assistants their affective involvements are all in a world of peers. The father-son relationship runs, as it were, by the same norms Fr. Stirnimann attributes to Pangwa. But ‘disciplined’ is not a word which describes Kinga sons. Hunting, the male individualist’s major sport in uPangwa, had become a minor one under the Sanga, traded for soldiering. For Kinga women not Christians the ritualization of the life-cycle—puberty, marriage and domestic installation, pregnancy, and childbirth—is lean, and the elders less domineering. Father-daughter discipline, which one senses in accounts of uPangwa, has resolved into avoidance, as is the case with father and son. Kinga talked to me of marriage being, not long ago, imposed by a father’s fiat; but that was always in story-teller times. The convincing cases-in-point were not forthcoming. By all accounts of courtship customs in times known to the teller, ordinarily strongminded sons and daughters found their own spouses. Considering only the mature age at which marriage normally took place in Sanga court culture, regular arrangement by parents is obviously out of the question. But here we are dealing with Sanga court culture. Its spread was still proceeding outward on the margins through the first half-century after Contact. I found older women, especially in the Eastern realm, who had married soon after menarche, under a bush culture ethic. The role of the elders in court-culture did strengthen in the British period, though. ‘Fathers’ came in, not just to build translocal family alliances but principally to work the magic of credit in financing a new style of bridewealth contracts. ‡‡

(c) Gender-peer relations. We are in position to compare the moral strategies fit to succeed in Pangwa/Kinga societies for women and men. Grant that 344

our knowledge of the typical biography in either culture is sadly wanting for case studies, approaching the matter through the frame I’ve chosen offers some safeguards. Moral strategies are formed by men and women pursuing a living and making their name within a social maze we are able to study and describe with some precision. Fieldworkers gain a sense for the drama in lives they have shared, if only in short term. Fr. Stirnimann and Dr. Park had their own gender in common with only half the people they might have pestered about all those small matters one finds missing in one’s notes, and special gender avoidance patterns are built into each of the four mazes entailed, the two ‘European’ schoolings of the observers and the two ‘Bantu’ solitudes. Still there is much that we know from induction and much to be gained by deduction if it can be guided by common sense. What is the character of the mkoyo? He lives at the top of an age-respect hierarchy. What scope of authority does that generate for him? The answer is different if the question is taken in the frame of ‘norms’ (say) or of ‘cases on the ground’. Normatively, Fr. Stirnimann rates him as the prime initiator of collective action in the lutanana. On the ground we should imagine that he ranged from the honorific spokesman who could delegate work to a good support team, to the man of wisdom whose calming hand and canny influence could be sorely missed when he died. The structural fact, the mazeway he represents, allows that much range, while demanding of him only that he be there with the competence to sort out problems threatening peace, and preside over public occasions as they arise. As with monarchs and high bureaucrats everywhere, once in place at the tiller this man has only not to rock the boat and the system will keep him in place. Seen narrowly in this frame the mkoyo is not unlike the ‘great commoner’ of the Nyakyusa, except that he is not part of any translocal organization of political responsibility and power. He is responsible only to the people he serves, and the legitimacy of this service is not considered contingent or revocable. This man is not the creature of a ‘career path’ special to only a few, such as are the positions of smith, priest, or petty chief in the regional culture. The authority of ‘elderhood’ passes to a man or woman steeped in tradition through praxis. The mkoyo has no rivals, as political leaders always ultimately do. (Translocal politics only arises in a rivalrous stand-off among several local polities.) In his being unrivaled the mkoyo offers us a model of the male career: the cognitive sharpening which rivalry fosters in other societies is traded off for social recognition with each advance along a well-marked status path embodying an agerespect ethic. A similar gradient of increasing respect appears to 345

apply to the case of the female celebrant-teacher, who has full license to bully a young married couple. In her person she represents the category of elders of her gender who have managed to pass through all the ritually celebrated dangers of a woman’s career from puberty to menopause and earned the sanctuary of a taboo which bars her husband or any other man from her house and bed. But the mkoyo also stands at the head of another kind of respect hierarchy, where the colour of respect is fear. Here is a thumbnail sketch of the mkoyo who died at Ludewa in 1922, taken from an account of the obsequies which Fr. Stirnimann found on record: As proprietor of the shrine of Ludewa and executive of the rainshrine, Xayembele enjoyed such respect that not only the Kihaule group from Ukoko, but many other householders sought help from him in times of hardship. He also had a number of medicine men in his service, who prepared the fertility medicine isembe and the war medicine lihomelo for clients near and far. News of his death was not allowed to spread over the country, as public recognition of his death would mean famine for the country. The grave was dug overnight, the burial itself taking place at dawn. The medicine man M. spread a mixture of river slime and strong medicines over the open grave. Matching the rank of the deceased, he was carried by four esteemed householders. [II 276-7]

Pangwa are best known by Kinga from tales of their magical chemistry. The other side of ‘respect’ and ‘rank’ in the sketch is fear and a window into the deepstuff of Pangwa life. Anyone who is famous for fertility medicines will know of a rich market for a contrary kind of medicine, which will kill off an enemy’s wives and offspring, leaving him a hapless survivor. Was any profit ever made by any real Pangwa medicine-master in such traffic? There is no way at all to answer that with confidence, except to say the game is not played for real stakes. Kinga believe and act on the belief that an ambitious brother of the high prince about 1950 bought medicine from a Pangwa medicine man hoping to wipe out the ruler’s many wives and offspring in order to inherit the throne. The belief, based on easy inference from the sequel, is that the medicine lashed back on the buyer owing to superior defenses supplied the Prince by his own (priestly) medicine men. There is a culture of Angst as well as a culture of esteem, and it is witchcraft which offers us a window on the deep, inner world of dread and blame which remained much the same for Pangwa and Kinga in the twentieth century. The one mighty difference in their situations was that the political theatre of the Sanga courts, as it was before the colonial trespass, owned justice 346

in the medicine game. The mkoyo, without that claim of majesty, owned only the tokens the game was played with. Stories about necrophagia and night-flight in the world of witches are everywhere in the region. They dramatize the moral basis of a witch’s powers. But when it comes to imagining actual uses of witchcraft by one of the kinsmen or neighbours you know ‘on the ground’, there is always a medicine horn or equivalent ‘chemical’ vector implied. One consequence is an absence of charismatic villains, as it is the mere possession of the right weapon which distinguishes the witch from his friends. The units of the Pangwa social system are patrilocal families ensconced each in its own acreage. The frequency of alienating clashes with others may have been dampened by this distance, but the frequency of moves to re-establish amity after a break would have been dampened too, perhaps in about the same degree. Kinga peer groupings as they can be read in the settlement pattern were far smaller and tighter. On the evidence as Fr. Stirnimann reads it, the evening meal and conversation put a few boys in a close and disciplined relation to their father. Supposing that ‘on the ground’ there was often in old times as in the ethnographic present one ishyengo for the families of several brothers or even several men more distantly or vaguely related, the discipline would have been family discipline, and the free, uncensored association of boys with peers met herding (which is at the heart of Kinga sociality in boyhood and continues in advancing forms through life) is decidedly crimped. Though we hear nothing about spying, vexing, provoking, and interfering between boys of two Pangwa neighborhoods, it is not hard to imagine that the dissension and antipathy which Fr. Stirnimann reports among adults has its beginnings in boyhood. His observations pertaining to times half a century after peaceful resettlement on ancestral ground are consistent with those pertaining to old times. Fear of witchcraft, which runs through these accounts, is said to be greater than fear of the dead, as reckoned from a count of accusations. The signal difference between the two fears, for the sociologist, has to be the affective response an accusation begets. If Ego’s bane is an ancestor, Ego can only bow to power and atone. The dead man has already played his hand, and only he can redress. But when Ego accuses a living person, though he nominally concedes rank by doing so, it is the kind of play that murders sleep. I think the absence of a court of judgement or communal moot in Pangwa tradition meant that small antipathies and sometimes-groundless suspicions could fester to the point of triggering direct retributive action and openly devious ploys. The resolution most probable would 347

be the antipolitan choice, moving away. The power play in the end is going to be deciding who is the chicken. In Tunginiye’s case it was he who moved away. In the compact ambience of the Nyakyusa agevillage, it was the one chosen by ‘the breath of men’. In the Pangwa lutanana as for Kinga, the odds-on favorite to leave would be the less openly assertive of the two men. Nothing like the formal ostracism of the Konde was known to Kinga. When friendships broke they were replaced by quiet avoidance, as such a break diminishes both men. When it comes to scope, the lutanana is a primary face to face community, neither as firm or as large a political unit as the Sanga domain. The looser, acephalous system of the Pangwa can respond almost seamlessly to the central challenge of political ecology, micromanaging the spatial adjustment of a population in relation to resource distributions. I have dealt with this in Kinga and Konde cases as a matter of contractual loyalties within a local system of political authority. I have ascribed an implicit antipolitan ethic as the main adjustment mechanism in those cases. The Pangwa situation is simpler and in that sense reasonably frictionless. Where men build apart they grow apart. As long as they remain ‘ingroup’ to the old lutanana they will be marrying ‘out’ as before and honouring old agerank relations, as if physical distance were irrelevant. It is a system such as the Mawemba and other acephalous back-country peoples were continuing in the 1960s. It can be understood as an adjustment to a marginal environment through micro-migratory drift. I find it likely enough that the normative ideals expressed in some memories of the historical lutanana are just-so stories which would have been told in the same words in 1850. Normative frames are indispensable if only for retouching—putting a spin on—the world of experience. Although human fertility is so highly prized, the Pangwa have always practised a sort of birth control through observance of an extended nursing period. A woman was under normal conditions hardly able to bear more than 6-7 children. When you take into consideration the high infant death rate, it is understandable that the population only grew slowly and that free land always was available for new immigrant groups. [II 291-2]

For Pangwa the ideal type of the self-disciplined individualist fits a man’s moral career better than it does the Kinga. Why do Pangwa men move ‘proximately away’ from their fathers and brothers while hesitating as long as means allow to move far? It is likely because at ‘proximal distance’ they can avoid rivalry for resource-access while enjoying mutual aid. The pattern of land settlement for Pangwa was heavily weighted toward allowing space enough for a number of fields surrounding the homestead. Kinga 348

made it possible to promote peer amity as a basis for spatial settlement by developing stepped fields (often mistaken at a little distance for terraces) where the necessary fallow period is shorter, turning the soil there massively with annual planting. By remaking the soil annually in this way over an entire hillside, they built up a series of ‘condominial plantations’ in promising areas, slipping the need to bring them close to every user. After the massive soilturning work, women did the planting and cultivation, planning with peer-partners what plantation was to be worked each day, then walking out together with hoe over shoulder and everything else on the back with the baby to garden (each woman all day on her own plot) in concert. This part of the Kinga adjustment is obviously to be attributed to women adopting an appropriate moral strategy which could gradually remake the maze the better to suit felt needs. Sanga men could perhaps claim some connivance in giving women their own heads in this, as the greater productivity allowed men more time for soldiering and ‘bringing things together’ in the many ways they were wont to do. I find the suggestion repeatedly in Fr. Stirnimann’s account that Pangwa men like to ‘rule their roost’. This is an inversion of the extroverted consciousness of peer rivalry. Kinga men don’t crave, or at least don’t put it high on their moral agenda, to dominate a wife, a daughter, or a son. I expect they would find a Pangwa man tied down—sleeping at home every night, closehandling the boys, monitoring every task the others perform. That is distinctly not a picture of the good life among Kinga men.‡ But evidence does suggest Pangwa men got their heads together more often than their women, who would have been working their own fields apart from others except on umkovi occasions. Men cooperated in hunting just as boys did in herding. The change as you pass from uPangwa into uKinga is then most prominently toward peer-solidarity among women, and it is clearly a prolongation of the isaka beyond menarche that opens the door both to lesbian loving, visiting boy friends, and independent decisions on marriage at a mature age. Pangwa men express horror at the very thought of homosexual intercourse, and we must accept that this thought dominates if it does not absolutely rule the lives of herdboys sleeping together. They are, after all, privy to the affairs of their goats. Fathers, it seems, see to it they are married off by the time they are fully grown. Magoma are the only men in the region I know who relish talk of their youthful lovemaking with girls. It began (said the men and boys together) as soon as potency came and could be hard to stop even with marriage. Boys would simply vacate their sleeping house at dark, leaving the young ones back. They were always welcome at a 349

girls’ sleeping house and always able to get home before the second cock crow, keeping up appearances. It is quite a different picture to that of the Kinga. As a starting point for the evolution of the Sanga system, where the Pangwa dispensation is credible the Magoma is not. It should be remarked that, for all the lively concern of Pangwa for the anatomical facts and moral significance of seminal fluid, they think lightly about midnight emissions—the moral significance is that the man remains ‘hot’ for a day and must not participate in ritual. I confess I’d expect any male living within the ritual maze Fr. Stirnimann describes to be on familiar terms with night emissions. Perhaps a simple reading of the facts just here is good enough. In Pangwa as in Kinga social thought, sexual body routines are not ‘consummate’ when they are casual and unburdensome. Both cultures make the act of begetting into an unforgettable Herculean labour, in no way a simple delight, and mazeways lead to radical discontinuity of sexual contact in a marriage. The deepstuff of the two cultures diverges in massively important ways, but here is a line of psychologically radical perseverance. A simple assumption would be that boys explore sex alone and with each other but learn to have their sex in deniable, indeed forgettable forms—“nothing happens” if nothing is planned, nothing said, nothing owed, no one committed to a relationship. It was Kinga elders who wanted me to understand their sex relations with other men. For boys it was a game that went beyond words, enhancing amity as only fun can do. As for Pangwa men, seminal fluid is sacred to the shades’ work of recreating human kind. Except in that work it has no cosmic significance. The moral path of a Pangwa adolescent boy leads straight in to marriage and the ‘labour’ of begetting. In both societies women are sole owners of the marital bed after their fertile years are done; for Kinga it can be earlier. In both societies birth-spacing radically reduces the number of pregnancies and supports a woman’s independence. Kinga men treat coitus as something near-sacred to women. A woman has such a strong need and right to bear a child that a man who cannot be her Hercules must allow her to find a lover who will. By the standards of a civilization which produced and long adored psychoanalytic theory, this sounds to me unusually grown-up.

(d) Sanctioning rights: Mechanical sanctioning, depending on diffuse and ‘spontaneous’ aversions universally shared, was the dominant moral system for Pangwa, but a fall-back provision for ‘intervention’ was ready where taboo failed. Fr. Stirnimann’s account of an mkoyo makes him 350

the pop-up kind of leader who only takes authority when the occasion calls for it. Though he was accorded respect as the eldest direct descendant of a founder, there would have been some measure of self-selection in his succession to title, from among the small elite of experienced ritual practitioners who might qualify. Obsequies for an mkoyo do not suggest more than an important man and do not make provision for installment of a ‘successor to office’. The political localism of the Pangwa was closely tied to the kind of ‘Bantu egalitarianism’ which favours the indirection of ritualism and the imperfect social cohesion it fosters, over the kind of vested authority against which a self-reliant man dare not raise his voice. ‡ The politicization of Kinga by the Sanga regime has been reviewed before. The voice of authority there is a feature of a courtbush dualism built on the detachment of youths of ‘military age’ from dependent status at home, attaching them to the court as its servants. The Sanga dispensation is dependent on the separation of male and female worlds, joining them firmly (as with the Pangwa in my estimate) only in the work of procreation. Peer sanctioning in the maintenance of taboos is more formally organized by Pangwa. Next to them, Kinga look relaxed and even undisciplined, though this would not have been the impression given by men of the court barracks in the heyday of Sanga expansionism. Kinga have de-emphasized agnatic loyalties. The fabric of life in the lutanana was tightly woven with performative rituals at each life-cycle crisis, an intricate pattern of avoidances and orders of precedence on public occasions, always embroidered with homage to the ancestors—with dire warnings against offending or neglecting their vital interest in the conduct of their offspring. Pangwa elders kept their juniors aware of the shades and their nightly presence around the scattered homesteads. Pangwa speak of white, faintly seen ghosts walking the land which is closed to them by day. Kinga work their fields in an obviously comradely fashion wherever close co-operation is called for. But there is no theology behind this. When Fr. Stirnimann writes of the wonders of cooperative work, on the other hand, the rhetoric is all but biblical: The deceased fathers embedded gardening activities in a sacramental praxis, aimed at the integration of the community and consolidation of the social system. Under their dispensation gardening work was to be led by the ‘great’ fathers not individually but whenever possible in collective fashion, and carried out with high endeavour in a cheerful manner. This demanded considerateness of individuals, obliging the members of the luxolo [descent group] to accept rank-ordering in the interest of the community. The same applied to partakers in the traditional [first fruits] ceremonies

351

which set the framework for the ritual intercourse of the living and the dead. The annual execution of the rite became a religious experience, through which the kinship group not only became aware of itself as a united society, but also the consciousness of shared values came to life, and each partaker came awake to the responsibility which had fallen to his share in this society. The rite furthers peaceful, generous co-operation among men of different ranks, and cements the kinship group. [I 221]

I’ll confess I supposed on first reading that this kind of prose was full of kindly empathetic wish fulfilment. But the two volumes, read through and carefully reconsidered, show the skeptical ethnographer that intentional communities are not an invention of the modern world. We are indebted for that revelation to a chance happening. A skilled ethnographer, obviously looking for such a place, found it. The Pangwa version is no more workable, it seems, than any of our own. But as long as it was all they had, they kept the faith. Fr. Stirnimann’s extended fieldwork came at a time when the whole great calendar of their sacraments was still embedded in Pangwa praxis, but (as we are given to understand by many notes on change, particularly in the 1960s) was unfurling in the presence of alternatives introduced under the British and Independent governments. It may be that we have much to learn from the Pangwa about the phenomenal success of the animistic world while mankind was still a rare species. I was taught at Cambridge in the 1950s that little communities like the Pangwa lutanana just kept on spinning like tops as long as they were left alone. It is a superficial view. What we needed to know was how much social discipline could be incorporated in the ritualism which was sometimes even then dismissed as no more than a prescientific way of dealing with life’s imponderables. Fr. Stirnimann has much to say about symbolism as a vehicle of values or deep-seeming ornament to the ritualist’s presentation of a normative worldview. Phallic or other natural symbols such as bleeding are certainly illuminated by considering the echo of ‘concreteness’ their use might lend to an otherwise vague or abstract idea. It is thus with the abstract notion of ‘purity’: As with most African peoples, so with Pangwa the sacred means first and foremost purity. They express this idea through the colour white. It means being free of guilt, especially that due to witchcraft. Every kind of stain and impurity awakens disgust and fear. A dead body is contaminated and everyone who comes into contact with it becomes ritually impure and must wash after the interment. The Pangwa seeks always the restoration of purity, which is associated with health and fertility. [II 292]

352

Is colour symbolism ‘natural’? What is special about it is perhaps that colors are (in good 18th-century language) ‘secondary characters’ of a thing. Does every white thing (a ghost, teeth, a dry bone) look pure to Pangwa eyes? Better say that ‘whiteness’ is an idea abstracted from visual experience of the world and put to semiotic use in the language of (Pangwa) ritualism. Colour symbolism supplies and enhances the rhetoric of discipline. Ordinary Pangwa do not need to ‘understand the symbolism’—we are not dealing with the cogency of an encrypted worldview but with the cogency of a teaching. Fr. Stirnimann almost agrees: The primary kinship group, i.e. that of full brothers, coincides usually with the residence group, whose segmentation into new inana [demes] is prevented by the loyalty and fellow-feeling of the paternal lutanana [deme]. Mutual aid and cooperation were always put very much in the foreground. The rituals on the one hand articulate the dependence of the living and the dead, but on the other strengthen the members of the community in their mutual dependency. [II 292]

What this does not quite do is make clear how ritualism works. What is missing? I find that ritual here, as in so many of the midlast-century studies Fr. Stirnimann has used as his models, is described in the language of dogmatic theology, while the practices in question belong to a belief system—animism—which requires its own appropriate analytical scheme. Animism is not a religion of faith, hope, and charity but of prudence and of rules. Pangwa moral strategies are not abetted by Pilgrim’s zeal or the good Father’s urge toward peace and harmony. There is dread not of doing but of going wrong, of being blamed. What the insistent, all-pervasive teachings of Pangwa ritualists are doing is building and fortifying a social maze. The material of the walls is psychologically opaque and strong. Our best word for it is ‘taboo’. What rituals are about is mending wall. What they say is, Do this thing right or the spirits will be angry with all of us. What they tell you is, Do this thing right or your neighbours out of fear will be angry with you. What person has the right to sanction another’s conduct? An acephalous society would be a Hobbesian terror, a hive without a queen, if no one had that right. The human solution is that everyone have it. I’m doubtful that mechanical sanctioning, where the right is dispersed (‘diffuse’ in Emile Durkheim’s usage) among all the married men, could be successful without a radical ritualization of the moral careers of both men and women. When I call the lutanana an intentional community and find it not doomed to failure, I have in mind that the spirits are, one and all, shaped in the image of Orwell’s Big Brother. Their eyes and ears are everywhere, on the marriage mat, in 353

the fields, in the sexual ‘shortening’ exercises of childhood, particularly at night as the fires die down. The collective freedom of all depends on each one individually staying on track. The whole is only as sound as each of its parts. Kinga would have moved through a period of escalating ‘transactional’ sanctioning before developing a system of chiefly authority which largely pacified life within a domain in favour of a politics of fear, acted out in belligerent rivalry among domains. In transactional sanctioning there is a minimal ‘rule of law’ which is not so subtle that people don’t spontaneously know when it has been violated. The right to sanction at that point goes to the party victimized. In some situations reported the longrun stability of peaceful living is not forfeit because someone from over the hill carried off a bunch of our cows or a woman. But to prevent the peace being forfeit when turbulence grows, the peoples involved in rivalrous possession of terrain may move toward chiefly authority. Chiefs can negotiate peace, take hostages and bargain with them, give credibility to a political compact, and generally manage affairs in ways their people can be expected to approve and not find it worth running from. More than that, chiefs with their secular approach to trouble can lighten the fearful load of taboo. I have in mind particularly the deepstuff associated with the taboo of violence among intimates, which in every society is subtler and deeper than the taboos of sexual intimacy. The idea that ‘one of us must be the witch’ can be described as a heavier version of ‘one of us must be the murderer’ as in a modern thriller. Suspicion of witchcraft is deep suspicion. Procedural intervention by the secular power system acquires loyalty in the same way as with civil ‘trouble cases’—by settling the matter. I have pictured the Malatan protostate as a place protected from gross tyranny by the antipolitan ethic, the right of the yeoman to vote with his feet. In the history of the Bantu or of the modern nation state this ethic has proved a perishable asset. I turn now briefly to look once more at some differences between Kinga and Nyakyusa protostate architecture which may allow us to glimpse the genius of antipolitan thought and understand its staying power as well as its perishability.

354

FEAR & BLAME, CHAPTER NINE

Kinga/Konde: Logics of Political Ascendancy What marriage implies Doing a marriage in acephalous societies will entail procedures meant to prevent some households (primary groups of kin) from rising through self-aggrandizing kinship politics to gain dominance over others. The household of a Sanga High Prince, with his dozens of wives and scores of offspring, was so huge as in itself to guarantee the householder political advantage over other men. Such disparities of quotidian power and privilege as are required in an acephalous society must be set up within each household. This is because the crucial requirement for stability in such societies is ‘equal voice for all households’. In the Pangwa case the voice is clearly patriarchal; but a married son counted, with the birth of a child, as among the avanyaluhala, people of intelligence, people with a political voice. Whatever rules of deference must be applied to meet the needs for a spokesman on special occasions (initiating assemblies, presiding at moots, staging public rites, dealing with trouble cases and defense) will be such that the quotidian parity of the community’s households may be set aside but not unhinged. The reasons for holding to these conditions in a subsistencelevel community are not hard to find. The society won’t prosper if its households are not set up to be equally self-reliant. The optimal constitution for ‘organized anarchy’ grants equal voice to its autonomous components, and obliges every individual, of whatever age or gender, to operate as a member of one such. Contractual relations between households—the most important being marriages of offspring—need to be made between equals to avoid complaints of strong-arming or exploitation. Reciprocity has to be enabled so as to permit free-willing cooperation to further the common good. Not only attempts at domineering but sullen withdrawal of individual households from communal activities are at once predictable and sensibly threatening to the common good. This is the consequence of an all but unspoken taboo structure which characteristically 355

manifests itself in the fear and blame drama of ‘witchcraft’ incidents. The logic of witchcraft in Sowetan discourse is to locate public danger in the private behaviour of certain others. These others are ordinarily to be found among a self-defined victim’s own kith and kin. In ‘amity societies’ where the claims of friendship are pressed wide among own-sex peers, it is the more likely that one of them will be found to blame. Egalitarian economics dictates that everyone have the same kit of productive means. This eliminates felt differences in wealth, so that wealth almost ceases to be a factor in community life. A strong polygynous ideal is anathema to egalitarian norms, as it creates a market for wives. Women become wealth—accumulated capital, sometimes interchangeable with cattle as in Konde culture. In acephalous societies, polygyny can’t be a systematic principle, only a fallback provision for keeping widows in a ‘voiced’ (male-headed) household, or for quietly supporting a politically central household such as that of the Pangwa mkoyo. These considerations help us comprehend why Pangwa do not allow co-wives to share their fields but set them up with separate ‘farms’, avoiding the impression of concentrated wealth. Sanga rulers, of course, did otherwise. I have argued that Pangwa men quite likely slept most of the time away from their nursing wives, in an ishyengo shared by several households, along with their goats and boys. If plural wives are infrequently found, it is because bridewealth rules put other values foremost. Community elders, not potential husbands, have to deal for marriageable women in stranger settlements, with which a marriage creates lifelong mutual responsibilities on a translocal contract basis. The token-wealth proceeds of a marriage are so well distributed as not to threaten the equal voice ethic even while establishing dense and overlapping networks of ‘special interest’. As for marriage age, the most obviously workable rule for an acephalous (egalitarian) society is early marriage for both genders. The Pangwa patri-deme is normally exogamous, in keeping with the spirit of a broad bilateral incest-exogamy rule. Marriage is virilocal— wives are brought in from other patri-demes, and marrying back in the next generation is disallowed by the bilateral taboo. The logic of this provision is obvious: longterm kinship ties are the basis of patrideme ethnocentrism and must be kept to that purpose. Early marriage for boys is (short of introducing polyandry) impossible where it is not established for girls. Inclusive rules of early marriage are expedient where youths must quickly be established in their own households to avoid loose cannons. For Pangwa, the authoritarian 356

relation of father to son (or daughter?) is potentially explosive once puberty intercedes. There is in fact no ‘lineage’ backup for keeping the peace privately. Further, when the typical age for starting a family is shortly after sexual maturation, a marriage can be managed by elders without much fear of refusal on either hand. This is the case particularly in a rite-ridden society like the Pangwa, where children of both genders are used to being good puppets. We have seen how the marriage itself begins with a series of massive maze-building projects to keep the young couple from dyadic withdrawal—from ignoring the well-tempered voice of wisdom. It is a wonder of Fr. Stirnimann’s ethnography that he makes so clear the earnest voice of the elder teaching the young. It is not as though the ritual were being done for its own sake or for the ancestors’ approbation. It is performative drama, like the blessing of a ship lest a board should spring. Early, arranged marriage gives its interested producers their best chance of seeing the couple make a go of it. What is needed is not devotion but ‘buffered amity’. Heavy affective dependence of the pair on each other—amity without buffering—floods out the voice of the teacher. Arranged marriages everywhere imply continued monitoring of a son or daughter (nephew or niece) through the decades of child-rearing, until they in turn are monitoring children of their own. If the result seems to be ‘equilibrium’ it is because control has been made seamless. For Pangwa and Kinga the control aspect of bridewealth provisions is perhaps less obvious than for Nyakyusa, whose traditional cattle bridewealths were as ostentatious as store-bought Kinga bridewealths gradually became, under the pax. Heavy dealing in transferable wealth can hardly go on for long without creating ingrained inequality. Obversely, the premise of inequality, once established, will soon be generating a massive traffic in valuables. But the Pangwa show clearly enough that bridewealth marriage systems do not require the invention of wealth. An ethical premise of equality is a key to the way of life Pangwa moral strategies presuppose. The quick dissemination of bridewealths is corollary. The premise of inequality in agrarian societies is often associated with late marriage for men only. Usually this is the direct expression of male dominance in combination with age-rank polygyny. This kind of asymmetrical late marriage can work by allowing property rules to eclipse the paramount character of the individual household as a unit of political and economic self-reliance, able to owe nothing and so free to move away from trouble. The meaning of wealth is quite different in transhumant pastoral societies. There it may be that the greater your wealth in animals the less you will be 357

prepared to let someone else call your moves. The difference wealth makes does not tie you down in space but in other ways. The antipolitan ethic in agrarian and nomadic societies is not deeply different, though. It encourages self-reliance. Late marriage for women-only won’t work as the prime rule in any society I know of. It does work for men-only, largely because they can be segregated from women of their own generation without damping the population’s fertility. The prime need engendered by a rule of delayed marriage of this kind is an effective segregation of young men away from women. Women having been made wealth by the polygyny game, young men have to be kept out of the marriage chase and kept nominally chaste as well. Age-grade and age-class systems for men serve the purpose, as does the age-village stratagem of the Nyakyusa. The generational displacement these provisions impose carries forward, of course, as the once-young— trained to think in terms of peer identity—displace the now-old in the running of affairs. In the Konde context, the explicit logic of extending their bachelor phase was that your fighting men need to be single-minded. Late marriage for both genders is rare enough that I know of no parallel anywhere to the Sanga dispensation. Biology seems to work against a taboo which excludes conception through the years of a woman’s youth—years naturally blessed with health. Common wisdom in the West is that women would have trouble conceiving. To that, Kinga would answer that they do, and consequently demand a particularly high level of masculine performance from a normally absent husband. And to that, Fr. Stirnimann would point out that the Pangwa start early in life and yet demand even higher sexual performance levels than Kinga. So let us set biology aside. Late marriage for both genders presupposes several special rules. Women will require ‘quality pregnancies’ entailing holidays from coitus right through the long years of nursing a healthy baby and child. It is hard to imagine a ‘nuclear family’ system which would find this a trouble-free option. The Sanga solution is to suspend family involvement for the young during their first decades of biological maturity. This does require some kind of institutional replacement— a special culture of youth for both genders extending it through young adulthood. In the urban West, extended bachelorhood has lately become reasonably popular for both sexes. Advantage is taken of reliable birth-control devices and the gradual opening of equal opportunity to women in the work force or in professions. This kind of bachelorhood is compatible with nominal marriage or as-if marriage relations. Couples are not usually designated as ‘families’ but dealt 358

with as John-and-Jane units. Doubtless, a fair proportion of our bachelors don’t plan to marry and ‘have a family’. But for Kinga women no alternative moral strategy will serve, and Kinga men are ever willing. Yet the Kinga formula seems to stand alone. It has in common with today’s Western scene that Kinga women have ‘fulltime work’ but differs radically in that a Kinga woman experiences no structural conflict between work and children. This is possible in her circumstance because she rears her four children serially, one at a time. Each child leaves the mother’s close care only at weaning (having reached what most Western countries take to be ‘school age’) and starts a second stage of socialization and a quite new kind of schooling by joining older gender peers in their separate house. With this the ‘primary care-giver’ withdraws to herself. Her project will be to conceive again and dedicate another half-decade to raising a second, third, or fourth new baby safely through infancy and its early childhood years. One thing that is clear about this social maze is that it doesn’t meet the necessary conditions for acephalous societies. Though buffered amity is one condition that is met, the Pangwa style of self-reliant households—politically buffered by an established culture of rules placing every individual in a fixed, pyramidal, and patriarchal status order—won’t be found within the ambience of the Sanga courts. At the same time we have seen that the women’s regime of the Kinga echoes the Pangwa in striking ways. Women free themselves of their marital sex relation at menopause at latest. Nursing is continued for four years (Kinga) or rather longer (Pangwa). The same phenomenal all-night effort at begetting is the norm in both cultures. Children after weaning live in small gender-segregated dormitories. Marriages are virilocal but not made locally—brides are brought away from their home communities. Yet despite the obvious historical connection, the Kinga are wholly affirmative about homophilic relations, regarding the two sexes as having only partially compatible natures. Pangwa are negative on homophilia and show little candid interest in the social values of sex—apart from begetting, which must be done according to the explicit teachings of the shades and even, it appears, with their active complicity. Too ghoulish for Kinga? For the younger men I knew, at least. What sets the Kinga sex life apart is emotional segregation of the two genders. They live together in symbiosis. Deeply supportive interpersonal involvements are most comfortably confined within a gender-peer circle, commonly wider for men, closer for women. The price of this ‘Sanga dispensation’ was not just the kind of family life which centres on a single hearth. The Pangwa family required at least three hearths but seems to have operated as a solidary unit. 359

Logics of the age-village system What we have come to understand about the age-village system is that, as older men insisted was its logic, it was an effective mechanism for keeping young, energetic, and sexually willing bachelor men out of the sleeping houses of their aging fathers’ young wives. This hardly tells us how the system evolved, since the elders’ stem village, comprised of age-peers, had to be there before any group of young boys could have been segregated away from it. Step One is to see the continuity between a patrideme system and the stem villages of the Nyakyusa as we know them. Both of them use fictive kinship of the bilateral sort to mask the contingent character of alliance based only on amity. Here is Monica Wilson on the peer village as she knew it: The forms of behaviour primarily appropriate to kinsfolk are extended in some measure to neighbours who, owing to the agevillage organization, are very often not related at all. All one’s father’s mates...are ‘fathers in the village’, their wives are all ‘mothers in the village’, and their children...are ‘sisters (or brothers) in the village’. ‡

We can get a little farther along by looking at some of the subtler logics. I begin with cattle. They lead us to the shortcomings of a transactional system of sanctions where the main forms of wealth have legs. Finally we have to consider the structure of political space which made the age-village institution unbeatable.

The cattle factor. A major cultural difference between the mountain slopes ecozone (uKinga and uPangwa) and the Rift valley stems from the viability of cattle in valley conditions. They are not absent from uKinga-uPangwa but don’t thrive there. Trophy cattle are bred if possible, but few men have succeeded in keeping herds, even under policing conditions which spare the owner the fear of theft. The common view is that goats (now even sheep) are a better alternative. In the valley, though, cattle are regularly bred as the prime form of ‘storable’ wealth, and the Konde bridewealth system, like so many in Bantu ‘cattle zones’, entails bartering cattle for wives. The logic is that, the identity of a cow and its history being readily ascertainable, a ‘hostage factor’ can be effectively coded into every bridewealth transaction. If a wife doesn’t stay and reproduce, the cattle may be forfeit. The wealth in a Pangwa marriage, once transferred, instantly dissipates. Though a wife there must have borne three children 360

before a marriage debt is counted paid-in-full, only well-buffered amity can make such a debt collectible. Konde cattle, situated in the turbulent Corridor itself, are themselves complicit in the major ‘buffering’ system which dampers conflict between local groups. A cow itself can be retrieved wherever chiefly jural authority can be made to prevail—a cow makes a good counterpart to a wife. Though residence is virilocal, age-village mates are not often brothers, kinship is not a basis for local solidarity, and local exogamy doesn’t follow from the residence rule. Konde bilateral incest/exogamy rules mean little more than that a young man must be aware of his kinship identity as he seeks a potential wife. What he thinks important is hustling the cattle for a bridewealth, but in this he is a rival of all his father’s or father’s brothers’ heirs. The familial relation is thus one of clientage, set up by kinship rules which imply no intimacy. A young man’s career strategy revolves about a prudential interest in pleasing his patrons in the elder generation and a shorter term interest in something closer to sport. To please fathers, young men work their fields in the same teams that come around to their mothers for cooked food. To please themselves, they want to shine especially in such activities as dance, banter, flirtation, cattleraiding, and war. Case data on cattle theft, to be relevant, would have to be true to traditional (pre-pax) conditions. But if the data are lacking one useful principle is inferable. To fund a marriage from stolen cattle, a man would have to be able to present them to a canny father-of-thebride as free of lien. That means, from outside the horizon of jural responsibility of the negotiants. This circumstance brings with it a logic of its own.

Logics of transactional sanctions. Consider a proto-Konde people sufficiently pressed by the difficulty of protecting its cattle from human predation to look expectantly toward the idea of chiefship. Chiefship becomes feasible when translocal politics, a level of organization above that of the Pangwa-style patri-deme, begins to broaden its scope from the buffering of inter-deme marriages to dealing with torts and sensitive trouble cases. Cattle, in the significant numbers which make them units of wealth, offer such troubles. The proto-chiefly way of meeting an outbreak of trouble between autonomous demes would have been a system of sanctions which kept acts of violence to a ‘private war’. This is the essence of transactional sanctioning. Direct action can be taken to settle a translocal grievance, in cases where the idea of transactional sanctions is culturally established, without escalating to open feud. The mime of retaliation by the aggrieved party must be framed to say ‘this and no more’. But buffering violence in this way becomes progressively 361

less likely when the level of ‘private’ violence becomes a focus of social danger. The proto-Konde people we have to consider would be reaching a degree of structural turbulence that would prompt wronged individuals to abandon the buffering code, leading to broader, deeper, untimely patterns of conflict. It is worth taking a step backward to be clear about the difference between domestic (local) and translocal delicts. Within a face-to-face settlement of pedestrian scale, violence is buffered by treating it as aberrant. Redress is aimed at restoration of a wellunderstood status quo ante. Fraternal taboos of violence pervade the situation of treatment. The finding of a moot is spontaneously taken to be unanimous. When either the plaintiff or the accused can’t accept it, there is stand-off. The ultimate sanction inside a patri-deme is obligatory exile. Always it bears the implication that redress was not possible. This outcome is facilitated by an antipolitan ethic which allows a man to get out of manly quarrels by decamping without prejudice. A jab of witch blaming is often entailed but is universally known to be two-edged. Pangwa use the poison ordeal in the same way that Konde do in matters of malign intent: it is a system which imputes equal possibility of witchcraft to both parties in a serious accusation, and requires no strong-arming to consummate its finding. Understandably, it is an institution treated with respect on both sides. Konde chiefs apparently preferred to stay clear. But it presumes a context suitable to the lawyerly style of indirect action. Kinga, under Sanga court culture, had magnified the ordeal by an adroit use of political theatre, sometimes abandoning the two-edge mystique in favour of authoritarian intervention. In the absence of cases I can only say I was told that summary execution was used when the high oracle pointed to guilt. But nowhere is the ordeal equal to stopping a war. It remains a domestic instrument. ‡‡ Transactional sanctioning is ‘political’ when it is trans-local. Weapons are borne. The action must be ‘surgical’—quick, quiet, done by dark of night. To meet that standard, the target must be undefended. The system works well in an out-and-out kinship society, where any persons or property of the culprit’s lineage, easy to target, will serve the purpose of retaliation. Patri-demes are not always suitable, since they are not ‘pure’ kin groups tied one and all by maximal-claim ties. Compare the difference, in microbiology, between a multi-cellular aggregate and an integral organism: they may look alike, but the former comes apart into autonomously living, unicellular components. When a patri-deme is new it must be far more like the still-forming aggregate than like an integral organism. Then you can’t retaliate for a delict unless upon the individual culprit 362

himself. Retaliation has lost its sanctioning power, which depended on the spontaneous, endogenous sanctioning qualities of the target community. Taboos are above all a product of much time spent in significantly reciprocal relations in a company of others. But the Konde political culture we have to deal with in late traditional times knew the rule of law as well as taboo. Chiefly courts were enabled to deal with trouble cases which crossed political borders. It is worth citing at length Monica Wilson’s mature summation of the political constitution. We have already seen that certain points need assessment. Traditionally villages were grouped in a number of small independent chiefdoms led by hereditary chiefs. Each chiefdom contained many kinsmen, but there was no fiction of common descent of all the men of a chiefdom, or of chiefs and people. There were thus three types of social group: villages occupied by [male] age-mates, who shared common land rights, and herded and fought together; lineages bound by common interest in inherited cattle and common rituals; and chiefdoms in which the bond was occupation of a common territory and allegiance to a common chief. Relations between the three groups are epitomized in the traditional law relating to theft, adultery, murder, and witchcraft. A man was a member both of a village and of a lineage and both groups were responsible for his behaviour. An adulterer or cattle-thief was speared by the injured party and his kinsmen, if they could catch him. If they could not, they might attack any member of his age-village. This commonly led to war between the villages, but short of this, the kinsmen of the dead man claimed compensation from the thief or adulterer and his kin, not from the avengers. Both lineage and agevillage were therefore corporate groups, the members of which were mutually responsible for each other, and it is significant that both senior kinsman and village headman of the defendant were required to accompany him when the case came to court. Murder was treated as a private delict, the kinsmen of the dead attacking or claiming compensation from the murderer and his kin....Therefore the lineage was not the unrivaled principle of social grouping among the Nyakyusa. [C 2-3]

The problem is, how can we have both chiefly authority and transactional sanctions? What keeps a chief in business? A simple answer would be that we are looking at a transitional stage in a protostate process. Chiefs have not yet reached the power position they require for making effective decisions on divisive issues. Translocal delicts and ordeals are both transactionally managed. But while ‘transitional’ is a label that fits, just putting a suitable categorical frame on an institution doesn’t bring it a lot closer. The best way 363

forward would be studying the quality of chiefly/princely power in Kondeland. How secular, how sacerdotal, how militaristic—how charismatic? How translocal, and how come? Chiefship is a famously context-sensitive institution. It says too little to say its context in Kondeland was transition.

Political space. Suppose that our proto-Konde were generally settled in patridemes. In the region’s acephalous societies they are common. They would contain some scarce-related households together with a significant core of agnates. But all would think of themselves as ‘one family’. Amity as a back-up basis of mutual trust would have been established but would remain implicit under the cover of fictional or retroactive kinship. Successful communities would have cattle which they exploited in the usual ways, and fathers who looked to increase the cattle in return for marriageable girls. They could feel reasonably safe from cattle raids from near patri-demes because the inherently kin-like patriotism of the demes meant that transactional sanctions could be effective. Lost cattle could be traced soon enough to the deme of the rustlers, and an equal number of theirs taken in counter-raiding. The transactional system depends on local patriotism; in kinship societies it is a forgone conclusion that even a spatially scattered lineage would regard itself as a solidary unit for purposes of self-help. But under other conditions ‘self-help’ hardly applies and the transactional system for sanctioning violence or theft becomes suddenly clumsy. The hostage principle melts away if retaliation doesn’t satisfy the requisites of revenge. Amity is an ordering key which can displace kinship in the spatial patterns of settlement. Patri-demes lend themselves to open use of the amity principle, particularly where migratory drift and frequent resettlement are features of the region. I suppose those two conditions were from the start characteristic of the Corridor generally and of the open Rift valley floor in particular. It is worth imagining its history in some detail. A likely narrative of the Konde protostate process is this progression: 0 Exogamous patri-demes are scattered about, already with some cattle and new-world cultigens. Quid-pro-quo politics are typical of acephalous societies, focused and epitomized in bridewealth contracts. Social cohesion is sanctioned through congregational rites. Supererogations is oblique and tied to special knowledge of medicines, not the magical power of ‘witchcraft’. A propitious environment favours demographic expansion; demes reaching critical size go through ‘hiving off’ without the trauma of rupture. Lineages are already spatially dispersed.

364

1 Loosening of the patri-deme structure results from a growing pattern of colonizing expansion, entailing assimilation of marginal peoples, less effectively organized. Political pragmatism broadens the ethnic horizon, permitting translocal alliances. Transactional sanctioning is applied to trouble cases among neighbour demes, but increasingly leads to settlement by amercement. A modest level of turbulence favours a political agenda of strengthening the deme through attracting young men who see the opportunity to settle new land, win cattle, and take wives. 2 Latent agnatic alignments among neighboring patri-demes give some special advantage over others. With increasing population densities and turbulence, the main political agenda becomes formal recognition of leagues among neighboring demes. Within the league, trouble should be settled without bloodshed. Agnation gives place to amity as an organizing principle when the demes in league represent several unrelated patrilines. Since trouble cases internal to such a league can’t be settled by crude transactional sanctioning (direct action), mechanisms for procedural intervention (indirect action) are developed. Chiefship appears concomitantly, first in the form of charismatic leadership within a league of demes. Patrilineal kinship ceases to be a major feature in the internal structure of individual demes. 3 Chiefly organization proves itself in rivalries for local dominance. The deme of the chief takes new form as a polygynous village, and the distinction between princely and commoner villages appears within the league. Each deme-village remains ‘historically’ autonomous, cooperating under chiefly leadership in defense of cattle and in longer-distance raiding on less well organized settlements. The principal political agenda now becomes rivalry among the leagues as represented in the persons and reputations of their chiefs. Rivalries within a league are largely sublimated to the level of external (interpolitical) rivalry. 4 The equation of cattle, women, and wealth leads to expanding polygyny, first as ‘commoner headmen’ adopt chief-like behavior in ‘commoner villages’ replacing the scattered hamlets of the former deme organization. Segregation of unmarried males is required as each village takes form as an ‘oasis’ in open country with its fields set in patches well apart. Boys are set to herding calves, youths to tilling, young men also to guarding cattle, while young women are kept back in the elders’ compounds. They process food and care for younger children. At a fair age they are married away, bringing cattle in. But boys and youths continue to sleep apart in houses of their own, as is done in both Kinga and Pangwa communities. 5 The age-village and Coming Out system develops from this arrangement by involution. The agenda is to maintain the elders’ discipline while schooling youths in the valorous life of men without

365

women. But unlimited polygyny means that the reigning married/ marrying age-group of men meets a point in aging where they must place the political future of the league (chiefdom) in the hands of a new generation. Then there is the Coming Out, which sends adult young men out to build village centres at a distance from the stem village. As this must be replicated by all the commoner villages (erstwhile demes of the league) as well as the chiefly one, their world becomes a bit of a circus for a while, but always an enterprising one not to be trifled with. As with the celebrated Nuer of Sudan: power is generated at the local level, where men ask full loyalty from one another. This makes quick defensive alliances easy to form in crisis, after which a fighting force falls back into its several solidary parts. These parts in early stages of the developmental cycle are the villages formed by young men with their new wives and progeny. In a later stage a ‘village’ becomes a ‘village cluster’ as boys in the next generation are set to building for themselves. 6 Kinship has become the key to the inheritance of wealth, and amity the key to local solidarity. Chiefly ancestors become the jealous patrons of all members of a chiefdom (a league of autonomous villages), thus fusing the two organizing principles of the community at its symbolic apex.

The result is an unbeatable arrangement. Each village, chiefly or commoner, is surrounded by camps of spear-carrying young men. They are engaged always in solidary groups, if not in the gardens of their fathers then in herding and protecting cattle, or trying to rustle some more. But owing to the non-congruence of patrilineal inheritance lines and residence rules, property rights in the village herd extend outward in a network reaching well beyond the chiefdom. Notably, the system of lateral inheritance, which means that brothers will be ‘rivals’ for property, is buffered by the planful scattering of brothers among different villages of the chiefdom. Village mates are not rivals, amity is not systematically undermined. In addition, as bridewealths are in cattle and retrievable with cause, affinal connections create a second and parallel network of property interests. This means the herd is comparatively safe from predation. Major disputes are most likely with unallied strangers, and selfhelp—transactional sanctioning—is only appropriate. Two special mysteries remain unsettled by this exercise. One concerns the peculiar profile of Konde chiefs in the final decades before the pax. They seem not to be understood as rulers without first being seen as stars. The other mystery is the signal Konde phenomenon, the plethora of priests. They are related.

366

Logics of autonomy One piece of the puzzle Nyakyusa culture has posed to ethnographers has generally been overlooked because it disappeared with the coming of the pax. It is the seeming flexibility of chiefly venues. Missionaries report a given chief here one year, elsewhere the next. Revisionist readers of the record (seemingly without second thought) suppose they have been ‘driven out’, but with no evidence of great popular dislocations. Can a chief be ‘banished’? Move one of his villages away from the others, still remaining chief as before? Acquire new welcome (take refuge) in a distant place? Inherit and hold a new office without relinquishing the old? I think we have to ask how it is that established chiefs and priests move about, seemingly without much regard to political boundaries. Priests, perhaps, are only following business interests, or possibly shifting their political alignments owing to some intuitions of mystical animus. But chiefs: can their business be no more bounded than a rainmaker’s? Very likely we’d comprehend the system better if we knew more about the durability of on-the-ground political allegiances between village headmen and chiefs under the turbulent conditions preceding the pax. Over a chief’s career, is it likely his villages will stay spatially compact or will they disperse? Social space among Konde seems to be reckoned more by way of the tabooed trees marking an old chiefly burial than by any artificial boundaries. On my reading, the record says I can’t assume strict territorial sovereignty was established above the village level, and shouldn’t assume even there that it went beyond the groomed area of an habitation. Since several distinct villages can be packed together within such a habitation area, I assume garden plots interdigitated and grazing was managed within common lands. We must have in mind that Konde villages are as moveable as Kinga hamlets are. Kinga priests cross borders constantly on state business while a reigning prince traditionally was bound to his own domain. In a Sanga realm the princely court constitutes a fairly fixed centre and is so conceived historically as the place from which client-rulers were sent out, some of whom founded their own ruling lines with the autonomy of satellite domains. These domains might be thought of as protecting the marches to the central domain, but occasions on which a Sanga domain or realm was subject to invasive war were rare even after 1850. Aidan Southall’s figure of a broad sphere of suzerainty spreading wider than the fixed borders of a central core domain almost fits. But in the Kinga case we have at least four princely realms, with clear certainty only that in a staged battle between two princes no ‘brother’ prince would intervene. One core domain plus three rivals for the honour? The mountain-slope ecology 367

favours a schismatic politics—maximizing the autonomy of viable units as they strain the geographic limits of pedestrian oversight. ‡ Ecological/demographic space is otherwise given and contrived in the Plains region of the Valley and its hillier northern marches. Let us suppose that a reigning chief would rarely find it expedient to abandon an improved village with its lavish plantings, taking his people, their houses, and stores to build and plant over again in a new place. While such a thing is always possible when there is cause, it becomes increasingly improbable once a new generation’s compounded village settlements are mature. While at an early stage in the maturation of a new stem village it might be done several times for tactical or strategic reasons, an established chief should be reckoned to have an imposing place and firm plans to stay with it. Yet we know that as the men of any aging stem village enter the final stage of their community of peers they are said to ‘draw back’— presumably in geographic as well as social-political space. Kinga houses, when they need replacement, are generally built on a fresh site. In the much hotter Valley country this gradual kind of domestic migration would have been sensibly continuous, particularly as the rapid growth of new plantings altered the landscape. The Western mind in its nostalgic mood is too much affected by the association of villages with ‘settled’ life. Here is the model I favour: Once a personal chiefdom has been firmly established and buttressed with a second generation of men in warrior status, it will not be long before the time is ripe for a Coming Out. When it is done, the ‘old chief’ (or his lineal successor) will normally still be in near-full vigor and unprepared to expire in the rather gruesome way (well-meaning torture and stifling by his priests) Monica Wilson supposed. As Charsley showed, the Coming Out is not a chiefly succession. To the contrary, I find the best name for it is an explosive transformation. What happens to the ‘old chief’ then is a shift from direct to indirect dealing with the daily affairs of the most trouble-prone half of his population. His workaday responsibilities halved, over time, the ‘old chief’ will see the youngest agevillages of his ‘old’ league hiving off, whether whole or piecemeal, to align themselves with one of the new chiefdoms. It is pretty obvious they will tend to be pragmatic in their choices of alignment, and that their decisions will help one or another of the new chiefs to consolidate a major chiefdom. Suppose the time this takes would be a decade. During this time the ‘old chief’ is not withdrawn but at the height of his specifically political powers, his influence courted widely. His administrative business diminished, owing to the explosion of new villages, his presence may yet be felt geographically well beyond his earlier sphere of influence. From being a chiefly ruler bound by 368

antipolitan norms to playing ‘first among equals’, he has become a genuine ‘politician’ able to trade on his reputation to continue promoting his wealth and mystical powers as patron to the clientrulers he has ‘sent out’ at their ubusooka. If someone is ‘driven out’ in the course of this turbulent ‘political’ decade, it is more likely his favoured client than the old chief himself. He is, we are told, no longer a leader of men in the field. But we have no evidence that privilege is denied him if he is inclined that way. The next move is rethinking what we know. We have certainly evidence enough to imagine ourselves into political space as Konde would have experienced it. The problem can be cut down to size first by noting that while a ‘chiefdom’ in the Wilson model might contain a dozen villages, only one on each preordained ‘side’ would ( jointly) have the ‘chief’ as its headman. Even there, the major quotidian duties of a local ruler would have been more or less devolved on others: younger brothers, skilled commoners. Commoner headmen rule as primus inter pares. Should we expect them to ‘take orders from their chief’? Of the village head Monica Wilson said, “It is through him, and in consultation with him, that the chief gives orders” More explicitly, Godfrey Wilson had written, “The chief can never make any important decision effective without securing the consent of his [village headmen] for it.” It is that consent which gives a ‘command’ the legitimacy of authority. That means even at the highest level governance is a matter of ‘first among equals’, which in turn means the appearance of a consensus is the key to that all-important task of making decisions stick. But if that is so, how do we account for the importance and special magical powers accorded a chief? We have to wonder if, in a politics of fear, anyone will properly ‘fear’ a chief who counts as no more than broker in an amity-based, egalitarian system of power. ‡ Even when we’ve given due consideration to the theatricality of power so characteristic of protostate politics the world around, we need to amend our sense for the structural position of the chief. Here is Monica Wilson’s model, as most succinctly laid out in a chapter on ‘the maintenance of order’: Village headmen neither have, nor had, power to enforce their decisions; such power lay with the chief alone in the old days—he would send some of his junior kinsmen and men from his capital to execute judgement—and he alone could order the parties to a dispute to drink the poison ordeal, from which verdict there was no appeal. Nor is a village headman or other arbitrator paid a fee—a pot of beer or a hoe from the plaintiff to ‘open a case’, and beer (or

369

occasionally a bull) as a thank-offering from the party for whom judgement was given, were the perquisites of a chief alone. [B 137]

The work of justice is thus maximally devolved, and procedural intervention effectively scaled to winnowing out lightweight stuff and defusing passions. Respected neighbours will be sought-to first. If settlement isn’t reached the matter is taken to the headman. All this allows for a reasonable cooling-off period. Serious cases can be moved off eventually to the chiefly venue, but the court in such case is handled by a panel of all the village heads (and I expect certain men of special influence close to the chief, or close to the circumstances of the case). A system of this kind for deflecting direct action is cumbersome enough to provide an excellent psychological heat-sink for the principals. But we have to set this in the context of a prevailing self-help system: where violence colors the delict, violence will colour the settlement. The path of colloquy applies to matters of contested rights, usually in relation to the contractual management of property, and witchcraft. In either case, the chief’s ante and/or the challenge of an ordeal are heavy enough to dissuade all but the determined plaintiff. In a well-established chiefly capital, the ruler will have distanced himself from the hands-on work of justice by the interposition of reliable kinsmen and trustees among his age mates. In all these respects, the structural positions of Konde and Kinga ‘chiefs’ are alike. Yet the reclusion of a Sanga high prince—his sanctity—is the greater by a quantum step. Set this fact in the context of taboo: the greater the weight of taboo a man lives by, the greater the charisma—the mystical fear his office and person inspire. And in the Kinga case there is the prodigious fate of a man captive to his harem, endlessly ‘hot’ from begetting. Visitors to Selya in the 1890s found well-governed villages, each headed by a man authorized to settle disputes, lead ritual, and take his men to war. There were also ‘chiefs’ ranging in age and standing, who could speak for the men of several villages. We have seen from the literature on the ‘Mbasi affair’ that young chiefs were clients of older, more established (and perhaps ‘charismatic’) chiefs, and it follows from what we know of the renewal ceremony called the Coming Out that there would in any neighborhood predictably have been fewer patrons than clients. The situation was the classic one for producing intense political rivalry and turbulent ‘arena politics’. Some clients would have to fall by the wayside. Where patrons scramble to back a winner, not every moral career can lead to stardom. But by the same token, those who do win will have proved their mettle, and will have done it very much in public.

370

Dramatic models are better than structural for disclosing the moral strategies by which a ‘client ruler’ becomes a chiefly one. The model I want to adumbrate is essentially theatrical. The client who turns out to be princely in the fullest sense is the one who best makes use of his initial position of authority to magnify a political persona. He must be opportunistic in using the arts of clientship without incurring the losses which accrue from abject dependency. Always his underlying strategy will be fortifying his chiefly persona, a thing best done by accruing wealth and putting it to tactical use in building a following. He must come to hold public sway well beyond its initial scope at his Coming Out, when he was a mere client ruler within a father’s sphere of influence. In this respect his career is parallel to that of a successful ‘chiefly’ priest in the same arena. Each has been given a mandate to pursue a career of influence, but each in fact must prove himself a man chosen not by his patrons but the shades. Consider once again the case of a junior chief (Mwakatungila) in Selya who lost a wife (Kinyolobi) to an adventurous medium (Mwamafungubo) and sought to get her back by leaning on an available patron. The scene is not the nicely framed stage the literature on it may have suggested, but an open arena. Chiefs from the Plains come marching through on military fishing expeditions. Oracular fiats ring out in the night, setting a neighborhood abuzz. ‘Big’ chiefs take stands on issues which may work for or against a given client. Leaders are seen to fall out with their followings and retreat from home turf. Short-lived alliances dominate one scene, only to be gone in the next. An ominous enemy has appeared in the Ngoni, doing business with the Arab slave trade. Behind this, the spectre of overwhelming German military power is advancing darkly. The possession of guns for the first time becomes a factor to be reckoned with in the rivalry for princely scope. Missionaries become embroiled both as patrons and intruders. Reputations are risked by established chiefs, who first join in placatory moves toward these new patrons, then retrench and repudiate the mission when a vacillating, self-appointed medium of the great Mbasi divinity finds better cards to play. The key to understanding this drama is, of course, to see that the political status quo will always be in jeopardy when pestilence and insurgency-by-proxy drive the action. Public opinion, aroused, is volatile. But one needs to know also that in the Konde protostate process all external boundaries are drawn in sand. The grounding of power isn’t so much constitutional as personal and, by virtue of that, can be translocal. On the surface, ordinary life in the village continues orderly, the people remain cheerful and politically disengaged while high deeds are done in a leadership arena by priests (cast as chorus) and princes 371

(committed to heroics) whose moral fortunes are avowedly at stake. The politics of high deeds is everywhere inclined to play itself out on a plane of its own, an arena imaginatively laid out by the local makers of myth. Leading roles are ordinarily assigned to persons who stand above the rest in wealth, influence, and destiny—in the Konde case, cattle and wives, patrons and patronage, and high birth. Our most common mistake, as Tolstoy might remark, is supposing that the theatre of high deeds is about history. It is closer to smoke and mirrors. Konde chiefs are not quite like Melanesian ‘big men’. While both depend on having wealth enough to be generous with, and both have their most telling years of fame and influence before they are slowed by age, Konde chiefs don’t lose their élite standing by losing wealth and power. Even if they may lose, piecemeal, whole measures of trust in their contact with followings, they enjoy a hallowed position as heritors of greatness in the form of medicines of special grace linking them to past rulers. A chief’s ancestral claims can be read from the grave-shrines he honours and (feigns to) dare not closely approach. Priests, since they apply these medicines, will not belittle their power. Priests also must decide what chiefly oracle to consult, with what slant, in a given contingency. It must be a decision which will carry the main publics concerned. Of course, only a positive response from a great ancestor is likely to help the career of his postulant grandson. But a good dodger can deal with bad news as well as good. Konde chiefs have a flare for backing away from trouble. Losing face in this society is not forever. The man himself is absorbed into his persona. This is a character in Konde political theatre who can sulk, rant, curse, and come on strong. Can we say that in the frame of political theatre he may have a very chiefly authority, but one which inevitably softens in conjunctions with his age mates? I am minded that Kinga priests reserved the right to warn and chastise an errant Prince. As in both countries the antipolitan ethic rules, the price of unpopular acts is losing one’s following. In Konde conditions as I understand them, this can happen without any man having to move house, as whole commoner villages can withhold support on some particular issue while remaining safely in situ. It is a state of things which makes chiefly authority iffy. Here is Godfrey Wilson’s account: The two most effective checks on the power of a chief are migration and witchcraft. Military strength used to depend upon the number of a chief’s adherents; ... men are quite free to move, and there is a flow of adherents from unpopular chiefs. But unpopularity is most feared in its spiritual form of witchcraft. ... An adverse current of opinion is believed to act invisibly, as the ‘breath of men’, against an unpopular member of a village and to make him sick. So

372

also the adverse opinion of a single village or of a whole chiefdom is believed to have power against a chief, to bring sickness or even death upon him. As in the single village, so in the whole chiefdom, the great commoners are the prime sources of this spiritual power; they are believed severally to protect the members of their own villages and jointly to protect their chief from the wanton attacks of witches; but if ever there is a just ground of offense they are believed to join their own power of witchcraft in the general attack. I know one chief who dare not give a judgement in a dispute between two of his villages for fear of the indignant witchcraft of the losing side. [A 287]

Where the nexus of power is witchcraft, authority gives way to spectacle. The stage-managers of the continuing drama are priests, an open mix of the ‘chiefly’ and ‘commoner’ kinds. Like their priests, chiefs can enjoy a peripatetic existence without fear of losing their identity or entrée into the arena games. But a masterful chief must have achieved the persona of a god-in-waiting. What must the little priestly secrets be, about how and when the ‘divine king’ body-relics of an ancestor will be made to bestow unworldly being on a great chief? They are secrets now kept, I suppose, forever. But in the priests’ view the hero is their creature, just as with Sanga princes. This in Konde entails the ‘python’ metaphor, which is not quite Kinga: Kinga don’t equate their princes with Lwembe, though Konde allow it. Monica Wilson could insist that the Lwembe and the Kyungu were not the only ‘divine kings’ in Kondeland, but she couldn’t make clear how some of her ‘chiefs’ would have earned the priests’ chrism through which a sacred identity passes on to an anointed heir, while others would not. Some secrets are doubtless half hidden even to their owners. In a world of formal statecraft by consensual rules, how likely is it the best of priestly informants could give a good account of Konde policy for high political contingency? Chiefs are unambiguously important when they are buried, and tabooed sacred groves have been allowed to grow upon the sites. Godfrey Wilson continues: While in the families of commoners an ancestral spirit is only believed to affect the lives of his own descendants, the ancestors of a chief...are supposed to influence, for good or ill, the fortunes of whole chiefdoms. Rain and good crops, the fertility of women and cattle are dependent upon their good will, which in its turn is believed to depend upon a recurrent sacrifice of cows. And for this sacrifice the chiefs, their living descendents, are responsible. But, owing to the continuous division of chiefdoms every generation, several chiefs are concerned in each sacrifice; every dead chief is a source of misfortune or of blessing in the countries of all his direct descendants. ...

373

Traditionally, sacrifices are said to have been usually prompted by prophecies of coming misfortune which was then averted by them. In each chiefdom the ruling great commoners [village heads], the hereditary priests, and some private persons as well are believed to have the power of dreaming of misfortune before it comes. And the chiefs are said always to have listened to these dreams. [A 287-8]

Lest this leave us with a chief who is no more than the local representative of his corporate ancestry, have in mind that an eminent ancestor’s shrine, if only for a limited public, operates in parallel with the translocal cult centres such as Lwembe’s at Lubaga, and without the protective privilege of reclusion for its living representative. His two sidemen, the ‘senior village heads’ of his chiefdom, may have led the men into battle, but their importance was translocal only in that indirect sense. Where Kinga ‘chiefs’ can opt for reclusion to escape just those threats of popular disgrace which haunt a Konde counterpart, and where Kinga rulers had ‘enforcers’ (their avanyivaha) to take tax and tribute to feast the court with meat and beer, Konde rulers had to depend on private funds and individual political skills of an ‘in the face’ style. In this matter of style, as in demographic scope, most Konde ‘chiefs’ compare not with the unkuludeva (prince or high prince) of a realm in the Sanga system but the untwa (lord) who ruled only over his own pedestrian domain and lacked the option of reclusion. But Monica Wilson must be trusted that some Konde chiefs do achieve high mystical standing. We see it plainly in the special importance of some chiefly burial groves which serve as the sacrificial centres for particular chiefdoms over many generations. There is a major institutional difference between the two political cultures. Sanga political space was organized about a militarily strong central court and dispersed peripheral settlements, all in a surround of unclaimed forest and scrub plainly isolating one domain from another. The lifetime circulation of men was inward from the periphery to court and back again to a perimeter. External relations during stable times were also concentrically patterned at a higher level around the princely domain which claimed the realm as its own. As there was a flow of traffic on the footpaths from hamlet to court within the domain, there was a flow of higher business (tribute, ordained prestations at harvest time) along the higher ridgeways leading to the court of the realm. Konde chiefdoms revolved about personal space in a world where ordinary pedestrian contacts created, particularly for men, spheres of personal acquaintance reaching into all the neighboring chiefdoms to one’s own. Distances and ties among the several villages would allow an instant turnout of a chief’s whole fighting force. The only solemnly 374

sanctioned boundary in a chiefdom was an internal line between two ritually planted trees, dating from a chief’s Coming Out and forever dividing the two ‘sides’ of his country, to neither of which he solely belongs. Only the chief, by rules of state, lived with both sides. His problem very often must have been showing impartiality. It is for that reason, in particular, that he must look to devolving power within his own two villages. His moral purpose must be an heroic not a villainous career. ‡ Still we should turn to the priests and their construction of the political world, as it is they, with their magical biochemistry, who claim to make chiefs possible. Insofar as the Konde community of Selya and the Plains shows the ‘theocratic’ bent which became explicit in uNgonde, it is theocracy in committee form. Both aristocratic and commoner priests are tenured in their positions and accustomed to working together, if jealously, on projects with more than private significance. By contrast, chiefs are ‘sent out’ as young men to seek their political fortunes. Their sidemen are armed with magic horns “to attract people to the country.” Behold the adventurous career of a political entrepreneur. We have looked before for reasons behind the Coming Out. The most compelling in this context is the community’s need to find a way out of the huis clos which looms with the aging of a charismatic leader and his close cohorts. If chiefly succession is proverbially precarious, it can make sense to make the scramble an orderly exercise carefully supervised by elders. What is a priesthood for, if not to find the lawyerly solution to internecine quarrels? ‡ Staging a pro forma succession to charismatic status is to routinize a mystery—in this setting at least, a nonstarter. I don’t mean to suggest a Konde chief has no loyal followers or that their number is unimportant, only that he patently needs to work for something more. To rate as a charismatic, you need a public. For the Sanga ruler, perhaps, his publics are captive. Not so for Konde. Konde young men whet their taste for public attention at gatherings like the carpe diem dances after an important funeral. They attend, coming ordinarily from some distance, primed for manly display and the challenge fights it may generate. But a newly named and anointed chief gets his first taste of populist political display in the grand climax to his Coming Out, when he is supposed to charge out of seclusion and lead his merry men off to raid for cattle and women and a brave new hero’s destiny. It is this kind of charisma, a universal game-winning talent, which is selected for in the Coming Out procedure and its years-long sequel. In the preordained rivalry with his fellow chief (or chiefs) in the Coming Out, the prizes are for a show of strength, quick command, meat, beer, and winning ways. 375

Much will depend on a man’s own success in external raids, but much also depends on his handling of transitory tides in public opinion. Konde settlements are at once dense and open to pedestrian traffic from any direction. It is an ideal situation for the rapid mobilization of public concern in response to rumor, prophecy, and ominous ‘signs’. Interpreting such alarms for a chief is largely an internal matter in which local groupings of priests with their medicines and oracles are likely to play an indirect but decisive part. But Godfrey Wilson made it a particular point that the political scope of a prominent Konde chief always overran—must overrun—the concerns of his own chiefdom. ‡ In both the political systems we have considered here autonomy lives at the local level. Because the centring power in the Sanga system was more systematically authoritarian, autonomy lived in the hamlet, in the local ikivaga, where formal political transactions hardly appeared because they were hardly needed. Because the age village of the Nyakyusa was a place of pressurized friendship its consensual autonomy had to be maintained by formal moots and an appellate venue for torts; and by a mystical safety-valve—naming and driving out the ‘witch’ in a manner we may describe as ‘surgical’. The lurking threat of schismatic realignment within a community of ‘lifetime friends’ is avoided. Catharsis is achieved through a dramatic show of unanimity. The deed done is final, the circle of friendship closed again. It is only natural, though, that after a sufficient time a repentant culprit might be welcomed back. Witchcraft in both societies is a matter of bad blood between friends or kin, not inherent evil.

A Profusion of Priests Commoner priests are ‘former headmen’ (or self-selected leaders) in the ethnographic model of the Wilsons. Hereditary priesthoods are positions supposedly initiated in an earlier chiefly generation. The holder of office will accordingly be dealing with more than one chiefly shrine and cooperating with more than one cluster of commoner priests. While there is no sign of a ‘scramble’ for the heritable positions, it is clear some laying-on of hands would have been involved, as ability and temperament right for the work would not show before mature adulthood. Godfrey Wilson called appointed village heads ‘great commoners’. As a descriptive term it can as well serve for most respected elders regarded as trustees in a village, and conceded rank in matters of judgement there. Monica Wilson appears to shy away from calling a youthful village head by that term, however, and with reason. 376

The mischief here is failing to recognize the gradualness of career achievements in societies seemingly governed by pat rules of thumb, but conditioned over the years in the business of their operational use and non-use by ever more deeply situated action. I prefer to deal here with ‘commoner priests’, pictured as a set equivalent to ‘village trustees’, and with ‘hereditary priests’ as officially recognized doctors of medicine. The best documented hereditary priest was Kasitile the Rainmaker; the laying-on of hands would have been the passing on to him of the local set of rainstones, by a series of ‘hereditary’ steps with beginnings too early to chart. It would have been from the set of such career professionals as Kasitile, resulting (for a given generation in the Selya arena) that a ‘living Lwembe’ would have been co-opted, elevating the territorial claims of a rain maker to translocal scope and the fullest range of destructiveprotective faculties. In this model of Konde priesthoods, the ‘commoner’ is a person accustomed to decisive powers by a modest career as village head or his assistant, and ‘priest’ through the special wisdom of mystical concerns, earned taking charge in village and chiefly rituals, and living up as well as may be to the qualities ascribed to his other role. For he was expected to be not only the armed guardian of his village by day but its mystical guardian by night. Compared to a chief’s charisma, this is the routinized version of the community leader, acquired on the job not through chiefly clientship. At the same time, a senior generation of ‘great commoners’ will be serving him as priests/ counselors through his early years or decades. They will have chosen him for greatness in a village context during the formative stages following his Coming Out, and can expect tenure as his special patrons. He needs them, to begin with, for the protective medicines he wants for survival. Like the ‘old chief’ but on a humbler scale, the ‘old headmen’ straddle both sides of the Konde power structure, as leaders of men and ordained wizards. There are so many ‘priests’ in both societies because the ethnographic observers have found it proper to use that word in reference to the whole class of male elders whose explicit consent is required on an occasion requiring consensual action. The community concerned may range from a single village to a full chiefdom, and readers of the ethnographies will know how jealously a priest may react when he feels unjustly left out of the loop. It is perhaps no great mystery how the priestly biochemistry entailed in the ‘divine king’ obsequies was supposed to work. Details may remain inexplicit, but it is likely enough the perishing holder of such high office is considered to be living as long as his hair and nails are so perceived—we may be free to suppose the chrism made from 377

those peripheral somatic parts doesn’t normally require the acts of torture the European literature on the subject presumes. To be effective, only the rumor of such mysteries need be kept alive. The education of a priest has to have been mainly through apprenticeship and has to have achieved the standard of ‘cultural reproduction’—passing on a tradition intact. But it isn’t wise to assume a clinical sterility, no blarney at all, in this kind of training. A priest who says ‘we do thus and so’ has surely heard it earlier said to him. He has not necessarily done it ‘thus and so’ himself. What is required is only that he will reiterate this teaching in his own turn to another generation. I don’t reject the possibility that everything said to have been done in a secret seance was truly and often done. The Kinga priests may have caught and sacrificed a little boy in the bush for his blood every time they marched to Lubaga, and the Konde priests may have strangled many a chief who caught sick. But it isn’t necessarily so. What can’t be doubted is that it makes better theatre if it is dramatically true in the telling. It is the same, of course, with talk of witches around the evening fire. Mwaisumo, a clerk in the 1930s, made the endless possibilities of ‘witchcraft’ explicit to the Wilsons; and Ali, a young Christian, made the subjective experience clear: [Mwaisumo]: “Suppose I dream we are fighting, and that you knock me down, or that you come to beat me, then I will say to my friends that you bewitched me last night. Men talk with their friends and tell their dreams. If several people dream like this of one man, they say: “Why should we dream of him? He must be a witch.” [Ali]: “The pagans say that if I dream of flying that means that the witches came and tried to eat me. If I do not dream at all, but wake up in a sweat, then they say: “They came to bewitch you by night.” ...Again, if I dream that someone comes and holds me down...they say: “The witches came to kill you.” They say a man is a defender if he does not merely dream, but sees in his dream who it was that came. So with dreams of fighting, if I recognize that my opponent is so-and-so, then they say I have mystical power... We Christians have all these dreams but say they mean nothing and do not fear them. [B 226-7]

What is witchcraft? In the hands of a man playing doctor it is a tool for allowing a client to place blame without either of them bearing personal responsibility for doing so. Who are the doctors in a traditional Konde community? Any man or woman may claim the wisdom. Then who can you trust? When you are in trouble, you will be directed to a stranger in whom you may place your faith. If we’d ask which of these living men, the witch or the doctor, must be reckoned 378

spiritually closer to the invisible persona of divinity invoked in sacrifice, it would not be the one we are sometimes prepared to call ‘priest’. It is the one we know for inspiring fear. The prevalence of priests keeps pace with the prevalence of witches. Like law, the procedures for finding and blaming a witch are a department of the social structure responsive to trouble. Next to the turbulence of Konde life, the Kinga seem to have enjoyed a measure of seclusion in their settlements on the high mountain slopes. It allowed them especially in Ukwama, the capital domain of the (oldest) Central realm, a state-building opportunity consistent with a further stretch from the typical religious expression of animism toward ceremonialism. Lwembe was, after all, the divinity who had successfully been chased away. It is the local divinity, on whom an odour of witchraft waits, that one must quietly dread, while the distant can with care be magnified.

379

FEAR & BLAME, CHAPTER TEN

Implications About regional cultures This work has explored a particular congeries of social mechanisms which cooperated to produce the regional culture of the Sowetan peoples and, within it, the Malatan congeries. An elementary point is that the value of a ‘one people’ ethnographic focus can always be magnified by thoughtful comparison with the neighbours. The most fruitful questions may be about narrow differences. I have taken the position that wherever neighbour peoples have been well studied, as the Nyakyusa and Pangwa have, a fresh ethnography ought to deal thoroughly with the region itself. Another point is more pragmatic. Just as a thorough history of a town can put its facts better in context than a history of one of its families, so we should expect a profit—better depth and authenticity—from regional histories. This is not only because internal contrasts and homologies within the region always bear upon the facticity of one folk memory or the interpretation of one archaeological site. It is at least as important that new questions and altered perspectives arise in the broader study. It will have attracted more inquiring minds, differently schooled. Informants can be asked better questions. There are a few very good questions I should have asked Kinga elders, had I been thinking in regional terms. Though the language map suggest separate origins, it appears that a long history of contact among the several peoples of the Malatan cluster is evidenced in a common ‘culture of rules’. The prime example in the Sowetan region is a bilateral rule of marriage restriction. But even where basic rules differ, and general homologies can’t be expected, a regional assessment (say, of symbiotic relations among the neighbours: think of the Nilgiri Hills cluster in Tamil Nadu) would rarely be out of place. Supposing the ethnographer can afford the extra time and trouble, what militates most against a regional bias is an easy assumption that cultures (and regions) have hardedged borders. Historians know better. But we do have sound methodological reasons for insisting you must observe one people at a time, if only for focus. And reconstructing a people’s history, to be properly done, has to come after recognition of what is peculiar in its 380

culture and what regional. In the present study I’ve leaned on the concept of ‘deepstuff’ as a touchstone for recognizing cultural difference within a region. I mean to explain this way of thinking about emotional interaction, its roots and consequences.

How are taboos learnt? My position in brief is this. The survival of a people through history depends on an ability to fine-tune two systems of behaviour, which I’ll call routines and insights. Ethnographers have generally reported on established routines under the rubric of ‘social structure’, and insights under that of ‘culture’ or, quite often, ‘culture and personality’. Homo sapiens differs most sharply in behaviour from his philogenetic cousins in living within an institutionalizes social environment. The key human institutions, of which the most basic is the family, require enduring intimacy among members. Prolonged intimacy among individuals first becomes conceivable with the emergence of behavioral drills, established routines sufficient to maintain cooperation in the practical pursuits of life. These drills are not executed as ‘rules’ but are coded in terms of social danger. Our minds were first opened to this by Franz Steiner in the lectures published as the book Taboo. A reader might suppose a book with this title would be strictly about ‘the primitive world’. Steiner uses ethnographic studies to explore how danger signs make social control work. The principles apply widely to human society. ‡ How general is taboo? To begin with, the word does not in its general sense refer to a human institution, but is nevertheless about how mainstream human life is organized. The difference between a trained and an untrained dog can actually help us to see what training a child amounts to. The procedure shortcuts reason. It takes human institutions to do the training. It is also the case that it takes training to make or maintain a human institution. But there is no contradiction. The circularity is inherent in the phenomenon we call culture. The implications are quite basic. There are no autonomous human societies of hermits, hippies, or otherwise unshackled egos. Don’t bother looking for a society of always-equal voices if only for the fact that all the mainstream human routines are invested with complex language, conceptual and gestural, which some individuals will master better than others. New individuals must always be programmed to fit into a going human community complete with an extraordinarily fine-tuned social and emotional structure. To achieve a suitable level of autonomy within that structure, a self needs to 381

perfect a language of self-control, an intuitive routine completely private to each individual. Some are much better at social tactics and moral strategy than others. For the most part, social skills will never be learnt without mentors, and no two individuals ever have the same mentors, nor are any two mentors alike. So some voices have to be more equal than others. Politics is not about equalizing social training. That may be attempted some day in a Brave New World. For now, Bismarck’s ‘art of the possible’ is about managing the distribution of voice in society. But if this means language with all its complexities is coeval with social structure, it means in parallel that all the vehicles of selfexpression a particular language curries will have grown with and into the routines we think of as structural. And if the moral features of a social structure—any human institution—are as heavy as the practical, it seems the simplest social system must comprise a dense and intricate network of social contracts, all open to emotional bargaining. As it would have to appear to a newborn it seems to be a design for chaos not order. That is why socialization needs mentors, not just ‘role models’, and they must be found in every household and dormitory. I suggest what makes the system work as a newcomer’s maze is the level of insight present in the primary socializing groups. Most often, these are stable families. But in the Malatan region families aren’t as important as siblings and peers. Given that mutual trust is only produced and maintained in a close relationship by zealous emotional bargaining, what keeps people cool enough to live together is not a limited capacity for quarreling but the spontaneous cultivation of intelligent insight into the human condition in its local manifestation. This crucially comes out of quarreling. I’m not referring to philosophical dexterity or doctrine. I find it an obvious fact that humans of normal intelligence everywhere develop an existential awareness of the-other-in-terms-of-self. It develops primarily through the ordinary give-and-take of emotionally exciting interaction with others, especially others of approximately equal powers. It happens even in the scenario of ‘identification with the aggressor’, and it happens with ‘guilt trips’ spurred by self-blame. The insights are seldom in any sense codified, but are constantly brought into play by the mini-dramas of everyday life. The best evidence that this is a human norm—that socialization is the gradual incorporation of a new child as a valued member into a primary group of willing volunteers—is the literature on failed child rearing. It is understandably a literature mainly on modern urban environments. It has to be reckoned that during the long millennia of human prehistory most cultural groups were persistently successful in rearing children to 382

adulthood, who were able to continue the long dialectic of a people’s history without break. I set genocide aside as related to the phenomena of barbarism not systems of civility. I have discussed socialization and the human social capacity in other terms in this and earlier volumes on the Kinga and their neighbours. My principal concept has been ‘moral strategy’. An individual achieves a viable moral strategy by adopting a systematic but watchful approach to the task. The basis for and measure of success is the individual’s moral insight. The level of insight I have in mind is one which goes well beyond anything a person can quickly ‘explain’ to another: Compare the spontaneous response to a funny story with the lame response of listeners who don’t get the point— and the awkwardness of our attempts to show them what was funny and why. The difference is not one of degree but kind. Fortunately, the level of insight needed for getting through the day with relative strangers falls easily within the normal intellectual range of a child of ten. It develops from early years in relations with peers. Why it isn’t usually as important in relations with parents may be because parents feel obliged to instruct and explain and put themselves in a guileful parenting role in the relationship. This produces behaviour which may mask but won’t erase the child’s own social learning. Kinga children do very well with a peer education system, but theirs is not to be confused with playground or classroom learning organized from above. Continued development of insight through adulthood wants only a mature facility for error recognition, with help from friends. Kinga would be particularly puzzled to learn how often Western adults read books on how to do ordinary things with food, gardening, and personal care. Wisdom is a topic wise writers will avoid. Most of us learn, in the daily scrum, performance skills we couldn’t possibly pass on to others—couldn’t analyze and ‘explain’. The others learn, as we do, mainly from doing and being done to. It is a common flaw in our schooling-conscious Western perspective that mature insights can be intentionally ‘introjected’—that they are normally exogenous in that sense. My position is that ego is essentially on his own in learning about the core dangers of life in the primary domain of his maze. What is ‘social danger’ to the ethnographer has to be intuited as genuine alarm by the child and emerging adult. The reason that, against all biological probability, humans are remarkably good at living together even in large groups, is that they are all born into a quick-learning maze. With ordinary luck, there will be volunteer help. A prime condition for thriving in the maze is that socially situated action, though built necessarily on a scaffolding of 383

routine, be put and kept in motion by successful—insightful— emotional bargaining. In peer socialization the two ends of a conversation are reasonably balanced, if only because the green ones are so quick at catching up. The name we give to those scaffoldings of routine is ‘institution’. Social structure is a condition of the existence and reach of any human institution, but is the skeleton not the flesh of it from which we can discover its nature. What institutions are about is making community possible. They are set up against the great odds posed by human ingenuity. We are not and never were the automatons William Graham Sumner discussed a century ago in his essays on ‘primitive custom’. We are not born programmed in advance to build our own mazes, as socio-biologists seem sometimes to suppose. But in struggling against the layout of the maze we are allotted, we earn the insights which make it livable. I will now go back over this ground in somewhat more conventional language. To the extent an ethnographer’s task is portrayal, the idea of ‘culture’ is a better guide than ‘social structure’. How men and women get on with their fellows and with children—how they make it through the day—does lend itself to pragmatic (structural) analysis. There are daily, seasonal, and age-long agendas which must be served, and a structural mapping of the usual routines at these three time-levels is apt to explain much that a visitor to the ethnographer’s elected community would find curious. A good local guide doesn’t have to be a philosopher to explain in this pragmatic fashion what is going on before your eyes. In fieldwork you look especially for a range and variety of virtual experience, so as to generalize the pictures you have begun to form. The product of fieldwork is a big netfull of ‘inductive knowledge’. Unfortunately, your fieldnotes alone won’t help future generations know a lot about the bit of human history you have studied. But so far as you have presented a fair descriptive study of the social structure—how things are organized to make community possible—you will have offered valuable, fairly ‘hard’ knowledge about human history in general. An ethnographer may do well to settle for that. I think particularly of the estimable Fr. Hans Stirnimann. The idea of culture has a different provenance. Consider how prose and poetry, though they may importantly overlap, differ in what teachers used to call ‘inspiration’. The study of social structure has often gone under the name of ‘functionalism’ because of its adherence to pragmatic analysis, and in that form offers an alternative to conflict theories or neo-Marxist analysis of ancient or modern nation states. But while the idea of culture might help you formulate a difference between Italian and Spanish mating patterns, it offers weak tea to the grand theorist. For one thing, a culture 384

can’t be planned. It is always a result of what actually happened, sum total, not what anyone in authority or in any ‘planning’ agency meant to take place. Cultures are not conceived in heads but in what happens when heads meet, Origins obscure, future inscrutable, even in the here-and-now a culture is always too big with potentialities for its implications to fit in a human mind. In 1963 I traveled in Kenya en route back to Kampala from fieldwork in uKinga. Passing through Gikuyu country we saw women carrying heavy cans of water on their backs, using a headstrap. The strain on their faces, the rigid posture of the neck, the bumping of the hard-edged tank on the back I found shocking. The footfall was forced and ungainly, punishing even to watch. The task of fetching water Gikuyu style was evidently hellish and huge. I thought, ‘Every tiny aspect is graceless. Can they possibly not see it so? Would anyone not revile a world which had them fetching water at such cost?’ Nothing could be so graceful as a Kinga woman fetching water, the traditional great clay pot calmly balanced on her head. Even massive loads of firewood, heavier and clumsier than this water, move gracefully along the path, the women never so constrained by their work or obsessed with their mission they can’t sing greetings out to friends they pass in the fields. I asked my Kinga companion (Soda) what he would say about the Gikuyu when he got home. He said, ‘Whatever it would be, the women at home would not believe it.’ As a practical matter, it is not hard to explain why a woman would use a headstrap to carry such a heavy backburden. It was clear enough what she was doing and, in the structural context of Gikuyu, why. What was inexplicable was why any people would have constructed that context of coercive meaning in the first place. The idea of culture derives from quandaries of this kind, and can be a helpful guide in exploring them—this especially when it leads an observer to contextualize action usually apprehended in simple pragmatic terms. Still ‘culture’ is not the kind of term real pedants ought to use in teaching anthropology. The study of a culture demands an idiographic lens, whereas social structure prompts nomothetic thinking—attention to the possibility of general findings cutting right across cultural boundaries. I find the good ethnographer/historian needs both lenses. Stubbornly excluding one of them may be tactically useful but serves a strategy of denial. That is why I’ve dared discuss a regional ‘culture of rules’ and contrast it with the ‘deepstuff’ so privy to a closely integrated human community as to define its cultural genius. My ‘deepstuff’ refers to the peculiar moral slant on the human condition its people share and help each other cling to. The point to

385

notice about pragmatism is that it also comes in schools. Its premises are learned—dogmatic. The rules of a game must be set before the players enter— that is what defines the notion of games. For Bena, Pangwa, Kinga, and Konde, the politics of solidarity among men is a game played by friendship not by kinship rules. The four peoples contrast sharply in the way each stages its game. They are worlds apart in their ways of bending sex, dominance, and obligation. But they share a ‘culture of rules’ by which kinship is hinged upon friendship, not the reverse. However important kinship may seem to be in local reckoning, these are not ‘kinship societies’. Their political history as partners in a region—their ‘political archipelago’ from which Pangwa (along with Mawemba and Wanji and some others) are left out—is not properly brought into focus until the shared culture of rules is recognized. You have to understand that pivotal loyalties in this region are not ascribed but have to be won and steadily managed. It is only that the difference one might observe in the way this political game is played on the ground—the difference to usage in a typical kinship society— would be subtle. Kinga never told me their kinship organization was a hollow shell. The games of kinship and friendship may seem to be played for very similar stakes and employ quite comparable tactics. But in fact the difference in the two structural rules is radical enough to account for very different political histories. Looking (impressionistically) at ‘culture areas’ around the ethnographically reported world, I sense that the separate peoples of such regions quite often do show a set of basic organizing principles which all share. Think of the Melanesian Kula Ring. But there is good reason to suppose, all the same, that two peoples can sometimes be closely associated in space over centuries without developing more than trivial similarities. I would expect that some cases, carefully regarded, might show the same kind of regional structure I have described for the Sowetan peoples: a ‘culture of rules’ effecting similar infrastructural principles throughout the region, balanced by a sort of ethnic schizmogenesis at the level of deepstuff—making for significant (structural) differences in the way the main institutions of the region have severally matured. The major suggestion I have made in this volume on the Kinga and their neighbours is that every close-knit human community goes beyond ‘rules’ in its collective effort at achieving viability as a group. We may perhaps speak of an ‘institutional arena’ comprising all the emotional interaction entailed in making things work on the private, the domestic, and the public planes. Viewed analytically (and not historically), the culture of rules is prior to the culture of insight—I 386

mean that kind of insight into one’s fellow’s condition as well as one’s own, which is the only firm basis for enduring ego-concessions. Insight in this sense is not possible without interpersonal collisions of the sort which generate emotional heat. The whole process entailed in the evolution of a culture of insight favours localism because it depends on direct personal interaction. In the circumstances of a pedestrian civilization, this means that within any broad regional culture of rules there will normally be a plurality of more locally focused areas of cultural intensity reflecting the struggle for community through face work. I understand this two-tiered arrangement within the Sowetan region to be fairly typical for regions ethnographers have studied, because I suppose that a ‘culture of rules’ evolves over time through low-intensity contacts across what will become local (ethnic) borders. Start in ‘prepolitical’ times and the patterns of emotional intensity which differentiate one local settlement from another will be relatively subtle, if only because exogamy facilitates the flow of intermigration within a given muster of local groups. But as a consequence of political development and demographic merging on a broader plan, we should expect within a maturing political arena a heating up of the struggle for (mutual) ego-concession and a need for the more penetrating level of insight which in times of change makes for and stabilizes mutual compatibility. By ‘emotional interaction’ I mean face work among mutually significant others. Ego concession is always at issue. The terms for success in the consensual plotting of everyday ventures are set by the way (and to the extent that) ego claims can be seen to be honored, the burden of concession lightened, and the venture found to serve, in its outcome, the moral strategies of its several participants in a normatively balanced way. Strangers bring heavy hands—they would not cope. Being peculiarly creative of one culture, and coded perforce in a style intrinsic to it, just these or these events could not arise in a culturally other group. A Kinga youth in the cultural style of the 1960s, dropped into a Pangwa hamlet of the style Fr. Stirnimann was able to describe, would soon enough be able to cope in the public sphere but could be quite lost, emotionally, in the private. I think again of the Kinga lad, returning from another life far away, who lost his way with a woman and hung himself from a tree, dressed all in his cosmopolitan finery. There is much to be learned about these matters, especially about the way political growth patterns have varied among cultural regions around the world, by rethinking the ethnographic literature. Clues are scattered under such headings as tribe, culture area/ intensity, exchange and alliance, or the origin of the state, though a 387

writer’s viewpoint is not often regional. I find the terms I lay out here suitable at least for the Sowetan case. But what is distinctive of this presentation is the idea of ‘deepstuff’ as an integrative force. The connection I make to ‘personality theory’ is less important than reference to a concept of culture as social environment—context for the quotidian pursuit of happiness. I see the consistent effort to improve personal moral strategies as the principal contribution individuals make (whatever the flavour they give it) to their cultural lives. Social behaviorists will see a familiar irony in this view, which makes of minding one’s self interest an essentially constructive business.

Fear, anger, blame Concepts like fear, anger, and blame tend to be treated as known quantities—familiar stuff to anyone graduated from the sandbox. Most of us suppose we know the range of sentiments which make us human. The idea of culture is badly used when it seems to imply that fear, anger, and blame are not colours of the mind in all human communities. But it is another thing to imagine that the phenomenal manifestations of fear, anger, and blame everywhere share one same ‘human’ grammar. Only move from one household to another, one age-group to another in a single village, and you will feel a difference. Looking for contrast at the level of ethnic groups (or in time, generations) on a regional plan can mean climbing the ladder of abstraction a rung too far. It is helpful when the similarities are many, and shades of difference stand out. A ‘regional culture of rules’ has provided one major conceptual backdrop and one premise of homogeneity. Another such premise, far more general but no less pervasive, has been an Eastern Bantu version of the cosmological system we know as ‘animism’. The different ethnic communities I have discussed hold a culture of rules and a system of belief in common. Their cultural differences can be seen as variations on common themes. I have wanted to provide some measure of the way uses of passion have changed with the rise of Sowetan protostates. To do this I have been at pains to suggest how long a journey it has been from the Pangwa to Sanga or Konde kinds of world, using to this purpose the concrete cases—ethnographies—as exemplary. In effect, I make of them ideal types wherewith to plot a course of regional history. The two protostates I’ve examined in this volume have pushed against the limits of an animistic world in ways which illuminate its special qualities. In animisms, anger (vexation, frustra388

tion, panic, the rage of personal desolation) is socially transmuted into a legitimate grievance. The blame may fall on a ghost (ancestor, local spirit) or a ‘living ghost’ (witch, sorcerer)—in any case, a familiar. In acephalous societies, even when the culprit can be faced, and retributive action is wanted, the procedure may be unspectacular. The victim undertakes an errand not a crusade. The accused is apt to show surprise and do what seems to be required to cool the accuser out. But where chiefs have set themselves up as powerful protectors, they will want spectacle. You get the ordeal. You get diviners using special effects. You get the magnification of incidents into emergencies. In the Sanga case you could get summary execution. The contextual meanings of anger and blame are, we might say, politicized. The culture of anger in Kinga court culture has changed in pace with a shift from transactional sanctions to authoritative. What this means for blame is that its quality as a deeply personal transaction between private persons begins to take on a public burden. We have seen this on a fiercely local level in the cases of ostracism for witchcraft, which Monica Wilson reported in detail. But Tunginiye’s near-death experience was a deeply private event. It would not have been, even in its colonial circumstance, if Tunginiye had been a less private person. The Sanga courts as they are described for precolonial times must have cultivated a more extraverted, more collective lifestyle than I found half a century later. But the Sanga courts were balanced by a systematic circulation of men (in particular) between the bush life and the court. Close and lasting friendships were always more significant than mere ‘male bonding’. Once we understand that social coherence (ties among equal agents) must be ontologically prior to social integration (organization of dominance relations) we are in position to understand that the rise of a Sowetan political archipelago had to entail, within each ethnic site, some mechanism for the intensification of masculine solidarity. For Sangu and Hehe the evidence suggests a ‘military tyrant’ scenario. External threat leads to military mobilization. One political entrepreneur beats out his neighbours in establishing a rallying centre for ‘defensive conquest’. But this system of authority depends on war. Power at the centre is a function of pushing out a political horizon by transmuting the insecurity of life for fighting men into vicious self-assertion in a serially repeating struggle to quell and exploit (in local tongues, ‘eat’) an external enemy. Conquering and sacking is a hot-house scenario. In a longer time-scheme it can be no more (and often is less) than a bridge toward a self-stabilizing political system based on better-rooted forms of social coherence among men.

389

What I have taken to be distinctive of the Kinga and Konde protostates is the intensification of masculine solidarity through a politics of friendship. The two ethnic approaches are quite different—one can be seen as an inversion of the other. Nyakyusa keep boys and bachelor men ‘at home’ in a permanent men’s agevillage temporarily grouped closely around a perishable political centre until at last they are ‘sent out’—whole villages of them—to marry and fight and create new centres. Kinga do otherwise. Boys are at home in the political periphery. The bachelor youth seeks to the local court, where he associates with men gathered in from all about and is re-sophisticated. He forms in this gregarious context lifelong adult friendships some of which are eventually carried with him in his return to the political periphery as a married homesteader. Homophilia is expected of the Nyakyusa youth in early life but cooled down as an affective basis of friendship with maturity and marriage. For Kinga, open homophilic association is lifelong for both sexes. It is linked to a predominant preference for (heterosexual) monogamy and a matrifocal strategy of reproduction. The contrast to Nyakyusa polygyny and patrifocal parenting is striking. In the one case it is mother and daughter who garden together, joined only seasonally by a father with his circle of friends or a son with his. In the other case the main gardens are in the fullest sense the father’s and the devolved responsibility for them the sons’. Yet in each society homophilia among men is a politically sponsored basis for social cohesion, without which the system as a whole would lose rationale. It makes a nice paradox. On the one hand, solidarity within the bosom of the family is quite generally fostered by incest taboos, which damp petty jealousies and favouritism. On the other hand, sexual license is condoned to favour solidarity among (particularly) same-sex bachelors. In both of the Malatan protostate systems the antipolitan ethic blunts the edge of authority to a degree unthinkable in despotisms. Making a difference between social cohesion and social control as the predominant means for attaining social integration can help explain. Social cohesion has to be cultivated while control is, gracefully or otherwise, imposed. Social integration doesn’t just happen. It is not ‘natural’ to human populations but results from actions bearing political intent. Where we think of solidarity à la Durkheim we have in mind the kind of social cohesion which seems spontaneous—natural in that sense. When we speak of a community being deficient in social integration we mean that its several institutions are not nicely coordinate. The birth or rebirth of politics as a (coordinating) institution in its own right is prompted by the kind of social discord which arises whenever the institutional structure of a community needs patching. What we see in advanced protostate 390

systems is a base-level politics of settling trouble cases and of local issue-management, overlaid by a chiefly political theatre playing on themes of booty warfare and mystical power. For Kinga the political centre is the Sanga court, for Nyakyusa it is the person (persona) of the chief, wherever he may be resident. In each case there is a political team prepared to legitimate chiefly rulings and execute them at need. In each case also this team is party to every legitimate ruling, and in the scene of such decision boldly plays guardian to an effective antipolitan ethic. Without that, the tempered totalitarianism of the Malatan cluster would soon lose its temper. The ‘kinship society’ is ethnography’s prime example of political particularism and so of resistance to the politicization of authority. The Kinga evidence makes clear that the muting of kinship loyalties can permit system-change toward a rule of law, even from a start in the kind of ‘pedestrian particularism’ we glimpse in Fr. Stirnimann’s Pangwa. The particularism which local autonomy favours is that of the aggregative face-to-face community—Stirnimann’s settled community [ulutenana] not the ephemeral patriline [uluxolo] which Pangwa elders tend to see in their minds. The politics of seniority in a robustly patrilineal society is inherently pyramidal from top to base. Kinship loyalties are prescribed parameters not dependent variables. The antipolitan ethic is effectively neutered, since the allimportant loyalties are embedded in the internal structure of ‘family and clan’. This is to say that the structure of taboos which comprises the kinship system has paralyzed the politics of individual choice. This is a point which anticipates later argument, but for now the gist is that the kind and scope of political process we have seen in the Kinga and Konde reconstructions is enabled at the start by a relatively diminished ethic of kinship loyalty. Politics in amity-based communities, when it makes its appearance, quite naturally revolves about the allocation of loyalties. They are in a special way vendible. As the protostate process matures, the issue of authority itself becomes part of the problematic, and procedural ritual (as in any modern court) is required to anchor the authoritative legitimacy of rulings as coming from centre. What particularly separates this picture from that of a mature state is the plurality of centres to which a man (or through him a woman or child) may be attached. Each is autonomous in the sense and degree implicit in the reigning antipolitan ethic. A man is free to remove himself and his goods, and anticipate acceptance where he chooses to go. Nyakyusa, where a ‘political centre’ moves with the man who embodies it, always has something like a hundred such centres of varying importance, where Kinga have fixed court centres and attribute no political weight to a local leader without a well attended barracks of his own. Hovering 391

over all the chiefly histrionics and princely pretension there is the entrepreneur’s pragmatic awareness of a simple truth: if nothing succeeds like success, the obverse also holds. Maladroit leaders stand to lose following to rivals, each loss only further prejudicing the chances for recovery. The best established Sanga princes withdrew from the open arena into a sacerdotal role. But for an aging Nyakyusa prince there is some evidence that there could be no retreat even though his generation had given over cattle and marriage claims to the next. The game, as it seems in my reading, would only go on through clientships as long as the old prince could manage to remain at political centre. The divine king’s deathbed scenario is, in my reading, proof of the Nyakyusa priests’ conscious commitment to supplying political continuity in face of an awkward challenge. The Konde political system as it appears in the Selya area must be counted among a fair number of African protostates which disconfirm the strict equilibrium models of early structural analysis and point us to a more common-sense figure, the political entrepreneur. To be sure, he doesn’t run free of his structural bases, but he disports over time a great talent for bending things his way. ‡‡ Systems of fear are often made comprehensible when they are seen as semiotic devices for polarizing political thought. There is the in-group/out-group polarity, the deference and decency gradients, and (what is particularly political in its flexibility) there is voice. Available targets of antagonistic expression are graduated from most licit to least. Reputational attributes and achievements open doors or close them. A received wisdom determines whose voice will be heard at centre. Kinship systems initially define these matters in terms of gender, genealogical seniority, and distance. But any good description of kinship rules in action will have to be concerned as much with principles at variance with doctrine as with cases where kinship practice actually overrides the rules. Propinquity, as a practical measure of social distance, is most often noted as a basis for ignoring or annulling claims of faraway kin. Competition for women, conflict over debts and movable wealth, and other sources of disaffection or distrust are features of every political community, and they lead men (especially) to see strategic advantage in situational rather than system terms. Might always wants to make right. To counter this, politically developed kinship societies may use such devices as intermarriage treaties, recallable bridewealth debts, and threats of schism to keep local kin-groups inclined to good order. As a general rule, the political significance of clanship is parallel to that of chiefship in an amity-based system: there must 392

be a plurality of centres, with tolerance of mutual leverage, to deflect mutinous coalitions. What best distinguishes the Sowetan systems from kinship societies (as we generally conceive them) is the ease with which individual men can bring their grievances directly to centre. It is a kind of individualism which builds on a radically open domestic group and dispersal of the extended family. We have seen that this kind of balance is achieved in quite different ways by Kinga and Konde rules. But both societies depend utterly on a multicentred chiefly politics above the level of local amity groups. For Kinga the peculiar bush-court-bush circulation of men is a fundamental determinant of a man’s adult status and political enculturation. For Nyakyusa the peace within the age village of lifetime male peers would never be maintained without the broader Konde pax which allows a man free movement to a new chiefdom and new peermentors when that internal peace is broken. In both cases political alienation is relieved by a system which enjoins fear of a magnified ruler’s wrath even while offering the right of asylum in a neighbouring vicinage. In neither case is it easy to imagine that the system could persist in the absence of politically magnified seats of power. Still the system survives the incompetence of incumbent rulers: the Sanga prince may be cloistered away, and the Konde prince treated as a living relic, while their denizens in power see to keeping the peace under that royal aegis. To this point I have argued that fear, anger, and blame comprise an affect bundle energizing the power systems we find in the Malatan protostates. Most individual experience even in the most ‘political’ settings remains, as everywhere, at an apolitical level. It is existential dread—awareness of one’s mortality, fallibility, secret vulnerability—which is the extra-situational source of fear. Politics is the art of giving a realistic, situational focus to the fears all of us have, and have in common, but can’t easily share. Franklin Roosevelt’s cooling formula was ‘Ye have nothing to fear but fear itself’. Some of my school friends were sharp enough to answer with a chorus of ‘death and taxes’, but we were all soon enough into war. It proved a political product even worse than the taxes we would later learn to pay, if seldom without a muttered frenzy of fear, anger, and blame. The chiefly systems anthropologists study have learned to deal with extraprocessual episodes beginning in fear, and to gather round the cadres needed to handle them. Trouble cases require that blame be legitimately allocated and anger cooled. Maintaining the cadres seems to have called for taxes in meat or iron and, from time to time, death in war.

393

Blame, taboo, and the problem of order Let me, for my argument’s sake, take the position that religious premises often dominate a people’s sensitivity to moral issues. Religions of blame have different roots and represent different moral premises to religions of (guilt and) responsibility. Consider the political uses of moral indignation in a Sowetan polity, which I offer as an ideal-type case for animism. You may expect this sequence: (a) a chance but resonant incident of personal misfortune; (b) rumblings of indignation; (c) popular search for a target of blame; (d1) selection of an angry spirit being; or (d2) selection of a culpable individual; (e1) appropriate private sacrifice to a spirit being; or (e2) public presentation of a chosen culprit to authority; and where witchcraft was affirmed (f) authoritative confirmation or dismissal of the charge, which now presents itself as a popular cry for riddance. It is only in the last phase that political intervention is wanted. If the spectacular theatrics of a full-blown ordeal are employed it will be in the measure required for assuring an authoritative resolution. Only in this public manner can divination at once perfectly condemn the culprit and fortify the crown against backlash. But if the ordeal should decide against the accusation, the sequence has been terminated, the charge denied in accordance with infinite wisdom. The driving force which has brought the episode into the political arena has been a contagion of moral indignation. The turn to authority for judgement in itself confirms the standing of the ruler sought to. At the other pole of a linear continuum, which is to say in didactic monotheisms, you should expect scapegoating to be associated with a politics of responsibility not sublimative blame. Svend Ranulf’s Moral Indignation and Middle Class Psychology famously explained (as of the mid-1930s) the growing success of German ‘National Socialism’ as an end-run around Marxist rhetoric, mobilizing the petty bourgeoisie in an as-if proletarian style against a government held to have failed its responsibilities to taxpayers. More recently, some Islamic leaders have called for ‘revolutionary’ stampedes against dissenting or ‘non-Islamic’ visions of the good society. What is at issue is not ‘racism’ but responsibility for history. Islam teaches in no uncertain terms that its leaders must seize and hold responsibility for a global future. For those who take this teaching, politics has to be fundamentally about religion. This would make no more sense to the Kinga of 1900, for all the ritual associated with their princely courts, than would the fingerpointings of Marx or Ranulf. It is the intellectual context of Western civilization which makes it conceivable that political leaders can shape the world. Its major religions teach that Moses, Jesus, and 394

Mahomet did just that. But this, in the context of Bantu civilization, is a nonsense. The reason wouldn’t be that chiefs are not big enough. It is rather the animist’s premise that trouble arises in molecular form and must be handled in molecular form. Perhaps the best metaphor at this point would be the Kinga beer drinker’s ploy of letting his friend pour him beer instead of handing over the pot. The careless way would be to let him pour into your mouth. The Kinga way is to make a cup at your mouth with both hands, a thumb ever ready to flick away the tiny floater which will be a poison pill meant for you. In a Western reading this is bizarre: why would this gesture not instantly turn friends into enemies? But if you have to ask that, you have not understood. In 1930 the most respected Nyakyusa headman in a rabidly egalitarian society might have a reputation as a witch. A powerful chief could not rule without one. By 1960 one or two in every gathering of respectable Kinga men would be wearing the badge of witchcraft on his forehead. To my Kinga friends the greatest obscenity of the European rule was not the lash but the gaol. What could a man do there to re-establish himself? How could he even think about his life-choices? How was he expected to correct them in gaol? Men as they knew them lived in a world of autonomous individuals bound only by personal claims willingly granted those close. Rules were embodied in persons. We are in position now to define witchcraft as the suspension of the normative complex of taboos which establish amity among intimates. Witches are said to eat human flesh. To get this ritual food they are said to undertake killing a close relative or friend. All this is not to be dismissed simplistically as ‘folklore’ but should be addressed as veritable knowledge in one important sense. It constitutes an exposition in synoptic terms of the scope of taboos in intimate relations, and the lesson in their breach that it is the maintenance of taboos which make community life possible. It is important to be clear about the nature and scope of the taboos in question. They would be in place in any human community in the absence of witch belief but are made obvious by it. For the sake of more clarity about premises, I’ll briefly recapitulate the analytical position I have been taking with my review of the Malatan evidence. It is intimate relationships which make moral sanctions hold. This is the case because socialization is most apt to take hold in the context of supportable emotional interaction—episodes in which egos are at risk. It is generally conceded that individuals carry with them into new situations a host of insights previously gained, and that they predictably try to comprehend right behaviour in terms of what has been acceptable (or otherwise successful) elsewhere. We 395

often say this elliptically by reference to significant others ‘being present’ to ego when they are not demonstrably so. I find the most telling and enduring social learning, at least, does take place in contexts of emotional interaction. These contexts, I further suppose, will most probably be primary groups, in the inclusive sociological sense. The fact that Kinga of both genders were peer-socialized made me especially sensitive to evidence in Monica Wilson’s works that the same had been so in precolonial times for the Selya and like communities of Nyakyusa. Where peer socialization is strong, we should expect that the mapping of taboo sensitivities for individuals would reflect the fact. Authority exists only as a sort of free radical in peer groups. Closure in the style of familism, which can only divide peers between us and them, gives way with de-emphasis of kinship to non-prescriptive norms and conditional recruitment of new members. Marriages are sanctioned by family not peer wisdom. Incest taboos follow descent lines. The peer socialization house knows no internal sex taboos. But we have seen that for Nyakyusa men the taboos exposed by witchcraft revelation normally do fall within the intimate agevillage. So also for Kinga men, in later life it is the comrade of youthful times who is thought likely to carry black secrets. He may act through what we call poison or through less tangible means, but in either case if there is no political intercession, the drama of witchcraft has only episodic weight. The social danger presents itself as immediate and belongs, when it is over, to past time. It remains, no doubt, as current in the form of personal bemusement on the part of any principals who survive: personal enmity lives on. In the 1960s a witchcraft episode might remain as delicately scarified brows on men who had been cleansed by a Malawi prophet. But the badge was not read as a sign of lingering guilt. The nearest phenomenon in European law is the crime passionnel, considered to belong to a unique relationship, and not to be evidence of the perpetrator’s danger to others. The difference to be made for understanding Malatan witchcraft is that the deterrent force of taboo (upon whatever inimical behaviour) is to be taken as objectively real. Consequently, someone in breach of taboo doesn’t require ‘correctional’ intervention by society. Taboo is in this way an obverse of law. A man in breach who is fortunate enough to survive the episode will not learn to be careless of taboo but, from the grueling experience of social alienation, to fear it the more. This may illuminate some Western failures to understand Bantu ‘witchcraft’. The phenomenon is at home in a society of face396

to-face relationships where the prevailing expectation is that social danger is maximized in the breach of taboo. Political intervention in the Malatan milieu occurs when the sense of social danger is highest, and must aim to quell it. The means of riddance range from absolution to death, and respond less to the considered gravity of the offense than to the authority’s quick estimate of political points at risk. If it is not immediately apparent in what religions of responsibility differ from those of blame, it may be well to consider the halfway type, which is most conveniently called Olympian. ‘Fate’ gets most of the blame, and the rationale is that different gods play different favourites. The well-known examples are ancient Greek religion and Hinduism. Both are syncretic cosmologies embracing facets of monotheism and the animist’s sense for legions of spirit beings. If perhaps the scapegoats (expected to ‘take away the blame’ for trouble) are mainly four-legged, human sacrifice is certainly known at least to lore. The very choice in propitiation—the choice of which god to turn to—is in itself an act of divination by which a suppliant’s fate may be determined. Mortals have no prior knowledge of a god’s will but stand apart, usually looking up, in fear or reverence. Whereas in a skeletal model of monotheism the sole spirit-being crucially escapes all blame, in animism a spirit-being at blame is ex hypothesi the honest protagonist with claims to righteous indignation, and must be propitiated. But in the Olympian cosmos the severalty of godhead means that currying favour with one may only incur the anger of another. Gods can be blamed but propitiation is fallible. Mortals can’t live in expectation of an unbiased world. Neither need they carry self-blame (which is the responsibility of guilt) for their failures, unless in the one exceptional case Greek legend has bequeathed us: that is when a Hero, living oblivious of a great blundering breach of taboo in his youth, through dramatic incident discovers at last the shattering truth. Having broken all the sacred taboos of family life, he wanders homeless, an exile from the civilized world. Interestingly, it is in the two polar cases, didactic monotheism and animism, that the sought-to spirit-being can be unimpeachable. In the middle ground gods of both genders take on all the charms, the vacillating virtues, and most of the evil ways of humankind. They are put on display in statuary along with human heroes and heroines. Monotheisms by contrast put godhead out of reach. It must be dealt with through intermediaries. Islam allows no images. Christianity offers every sort of likeness of its intermediaries but seems to have shied away from the Old Testament God who walked among men and tutored them. Most animists are content to deal with 397

spirit beings as familiars. They are demanding, not ‘giving’ beings. In the (unseen) Sowetan pantheon there are no spirit beings who can’t be depicted as ancestral humans. The moral emphasis in animism is on exemplary behaviour toward those who deserve your trust. In ancestral religions the demand is focused in the taboo construct which comprises the moral code of seniority by which ego is expected to orient action with close kin. All of Eastern Bantu civilization reflects its development from a seed culture imbued with such a seniority code. But for Kinga, moral seniority is partially displaced from the kin toward the peer/age and political identity groups. Animist belief systems elsewhere range through quite various systems of moral orientation to village and political life, coded into their teachings about the supernatural world. What we know or can expect to know of Kinga or Konde ancestors—how their quality of life is conceived—is drastically less than we know of hero-gods and (for all the shelves of disinformation) omnipotents. The reason for that is the molecular or granular structure of animist worlds, which in the Sowetan protostates is only in small ways modified by the heroics of war and princely claims. The underworld extends under all localities, down to the smallest; and human knowledge of it is from local experience. I take this as the mark of animistic worldviews: the supernatural is an extension of the local and the worldly. For Kinga it has been a religion that migrates with little groups traveling without luggage, only their needed tools and skills, prepared to put down roots among strangers and lay claim to the land by burying their dead in it. Pangwa show us how the familiarity between living and dead becomes a tie to the land. I never heard talk from Kinga men of glimpsing ancestors walking above ground after dark, though I fancy that is still part of the talk with some of them—most of my friends there were not among the mystically inclined. But Lwembe, Mbasi, and other broadly ‘territorial’ spirits of the Valley expected prestations on a decidedly munificent scale, setting themselves apart from plebeian spirit practice. What generally keeps me from accepting these shrine-spirits as Olympian ‘gods’ in spite of the hero-tales about them, and in spite of the spectacles they stage? I think it is their never-ending claims on mortal help. With the power to devastate the land with pestilence, they are uniformly disinclined to use their tricks for the good of human kind. Think of ‘the cattle herdsman of Mbasi’ or think of the histrionics, beer, and passage of trade goods at Lwembe’s shrine when the Kinga put in their official visit. It is not that the human minds behind Mbasi or Lwembe do not think big, but even the wellestablished shrine, viewed as an institution, radiates a molecular style: one case, one deal at a time. Over and over again without a 398

noticeable cumulation in experience, the Sanga pageant would repeat itself, the bearing of gifts in fear and trembling re-enacted with fresh hardware, crisis management achieved once again with the stilling of Lwembe’s histrionic anger. As the ritual recapitulation of an ancient mythological charter, this was drama going beyond the minimal scope of animism in its political significance, but not in its theology. We can mislead ourselves when we call Lwembe or Mbasi a hero—Trickster is not quite right either. They are the heavies, the big men of the underworld, anti-heroes needing appeasement. A politics of fear is meant to concentrate the mind, not paralyze it. When a religion of blame is used politically in an ideal kinship society, it strengthens social solidarity in characteristic ways. At ground level there are the domestic establishments, each radiating its own ‘internal domain’ founded, as I have argued, in the taboo of internal violence. These domains are aligned in parallel not opposition with others to create a political segment (often ordered in the manner of clans); but the very solidarity of the segment always requires external opposition, and the result is a ‘balance through tension’ throughout the whole of a pyramidal social order. When the religion of blame is politically exploited in an amitybased society, the characteristic style of social solidarity is fundamentally different, because individual men (in particular) or domestic groups are in position to detach themselves from one segment and join another. This is particularly so in Kinga communities, where very few men expect a notable property inheritance from fathers or uncles. An ethnographic monitor in the Kinga or Konde worlds in 1880 would have been kept busy recording the daily moves of realignment especially among the many young bachelor adults who were scarcely to be seen at gatherings without their spears. The main spur to disalliance would be the aftermath of witchcraft blaming. Fast friends were likely to move together or in sympathy. Small, close peer groups would often represent the maximum of solidarity and group autonomy. Fear is an emotion fastened on the future. The violence of warfare, always present to Kinga or Konde as the nineteenth century progressed, was a constant spur to fear and to the kind of rattle-brained blame that wants to stomp out witches. But blame is deeply situational, and in this cannot nicely mix with an ethos of time-binding responsibility. The fear of witches in early-modern Europe was a signal of culture crisis in a world made coherent only by a difficult belief in divine providence. At least by the middle of the nineteenth century the Malatan region had developed a level of political density which made the old system of transactional sanctions obsolete. The maintenance of 399

translocal stability required a full-scaled shift toward authoritative sanctions—what ethnographers think of primarily as ‘chiefship’. But it needs be said that chiefs are not in place to ‘keep the peace’. The sense of good order they can maintain flows from their ability to protect and cultivate the solidarity of the village. The shaping force toward stable polities radiates from the internal domains to which we can metaphorically refer by this word ‘village’. We have to examine again the twin notions of taboo and law. The distinction I have been making parallels the distinction I make between society and culture. Law can be tempered, laws can be changed. Taboo is either changeless or it changes in spite of men’s best efforts—or more simply, it is no longer there. Culture is only predictable in the way narrative is.

What grounds social identity? Consider the quandary of the good citizen of a Nyakyusa peer village, accused of witchcraft, trying only to shrug it off with denial. If he does insist on refusing blame he is testing his friends. If he fails to deny he feels the double bind. We’ve seen that his best bet would have been not to deny but directly insist the case be taken to authority. But if he can’t believe his friends are serious in accusing or failing to defend him, these others will arraign him, thus revoking his best-bet option. Now his hesitant friends are left with Hobson’s choice. The accusation is driven (say) by the progressive illness of a neighbour’s child. If the panic is general enough, the authority must condemn or lose his own station in case the child dies. Why does this kind of thing ‘work’? Why has this people constructed this culture? this huis clos? I find the answer in the villagers’ need, in crisis, for reconfirmation of the intuitive moral code by which they are tacitly governed. I mean, of course, the taboo-complex on which mutual trust is founded in a village of unlocked houses, and the very ability of any single one to speak for any other. It is this moral code which makes community possible, makes the fiduciary nexus dense enough to keep men to their promises, to say, We know who and what you are. Community comes into existence gradually, as a group of autonomous individuals begins to understand that all of them share in a common fate. When one among them is found and confirmed in secretly despising them, riddance seems to be the one choice which stands to restore the security of life under a sacred code. Ancestors may be selfish but they aren’t freakish. In some other animist world, it may not be your angry yet reconcilable ancestors who plague you but quite another sort of spirit being; and the cure may not be magical placation but exorcism, a tense-making 400

kind of mystical surgery. In a highland Burmese village where animism lives on in the shelter of an ancient Buddhist establishment, pain and suffering may be caused by witches or ghosts or demons, or by powerful supernaturals called nats. Every village also has a few known and several more suspected witches, all female, who appear to be living ordinary lives though ‘known’ to have killed or otherwise bewitched someone. Compared to this up-country Burmese cosmos, the Malatan is uncomplicated. In Konde culture it makes little sense to blame a child’s illness on ancestors—how would they think to act this way? why cut off a nose to spite a face? The sole available alternative in this religion of blame is claiming witchcraft. Further, the outcome of crying witch and sending the culprit off in exile is in ‘system’ terms positive, whether the child recovers (because the witch was stopped) or dies (so the past can be put away). We have seen how this works. Like the spectacle of exorcism, that of witch-finding leaves a village cleansed of evil. Ex-witches around the Kinga fire are known from their brow-scars to be clean because recidivists will drop dead in the very act—that is the witch-doctor’s guarantee. If doing witchcraft is acting in secret in defiance of an unwritten, uninstitutionalized local code of righteousness, then revealing and shunning a witch is revealing and endorsing the code. It is through ritual theatre, mimes of curing or cleansing, that lessons of evil are learnt, and the moral magnitude of personal betrayal measured. Fireside tales alone may acquaint a child with ideas about taboo. Without real incident, the piercing confirmation of first-hand experience is left out. Soda believed in witch duels, complete with fireballs traveling over rooftops. Without doubting his belief I couldn’t share it, as the firsthand feel was missing. I think the same steps of learning apply to all the taboos on mistreatment of persons close to you. Until you see and feel it, in a spat, in a careless accident, knowing the rancour and injustice, you don’t really know that the source of normal restraint is not ‘common sense’ but taboo. Even then, of course, you are unlikely and perhaps unable to put the lesson in propositional form. Witchcraft at least in the Malatan cases is exceptional. Taboo is the norm, witchcraft its unwitnessed breach. Sanga courts used a dramatic form of the ordeal. It was long gone by my time there, but was described as a fleshed-out version of the familiar rubbing-board oracle. A priest rubbed a basin of hot coals on a suspect’s back while he was questioned. The incriminating signal seems to have been the operator’s inability to move the basin freely—it would suddenly adhere to the flesh. What followed the finding of culpability in a major case would be summary execution, now in grand spectacle. Social memory here becomes vivid. Marched to a high cliff, a man is 401

thrown to his death on rocks below and left as carrion. I assume the culprit who might earn this treatment was close to the court circle, where witchcraft would be tantamount to disloyalty—treason in presuming powers equal to the ruler’s. There is no indication the ordeal was at home in Kinga bush culture. Tunginiye thought elders would take up the case and make decisions. A serious outcome meant one or both principals would move away. In this apolitical context, the theatrics of face-to-face accusation and judging made spectacle enough to incite a move favouring solidarity under the code of amity. Taboo required no re-institutionalization in procedural law. Apart from witchcraft, so often invoked in a child’s death, the moral code of the inner structural domain can be breached in quarrels or in sexual affairs. The true home, as I think, of maximalclaim ties is the socializing domestic group. In Kinga culture this is pre-eminently the boys’ or girls’ house group. For the most part, the individual’s awareness of a ‘group code’ or ‘moral ground’ is kept alive by small incidents in which they may fare especially well or badly, and through which they gain insight into the moral strategies they have been following, seen as others see them. Traumatic experiences from time to time may cause a lurch in a young person’s moral behaviour, whether toward conformity or dissent. Movement of young people from one house-group to another, sometimes distant, is generally motivated by attraction to an individual or clique and is often the occasion of switching loyalties. Peer socialization tends in this way to produce a more open society than kinship organization will. But in either system, the story of a typical moral career for either gender is going to be a tale of constructive interference by others, jab after jab in defense or in emotional repudiation of the inner domain’s structural codes. As an intellectual experience, this give-and-take doesn’t necessarily add up at all, but as a continuing series of forced choices it accumulates. To put the point: In taboo is to be found the root of moral justice, and through claims to justice the grounding of social identity. If all of us were moral chameleons we could hardly expect much justice; and where justice can’t prevail you will find skinchangers everywhere. Human beings are everywhere socialized, if not always well, into what we call their social identities. These are selves they are disposed to defend and develop throughout their long moral careers. Subjectively they are selves we don’t feel trapped in, however clearly it may seem to others we are. They are one crucial element in the paradox of being human. The other element is, of course, the ego. I think that is most clearly seen when it is sifted

402

from the long-term observation of moral careers as they are everywhere pursued.

Least polities The culture of the Kinga is perhaps unique in its implications for understanding the role of taboo in shaping the structure of internal domains—those self-segregating regions of private sociality which centre in the houses where infants make their way to adulthood. I suggest this partly because, of societies which avoid serious dependence on the nuclear family, the Kinga people is the one I know best. But it is also the one society I know, where the structuring of sexual relations within the socializing groups is bent to accommodate same-sex attachments. Incest taboos are a particularly conscious and lively category in the thou-shalt-not codes which make family systems, in all their manifold permutations, possible. The main taboos (on harming intimates) apply as well to the conjugal bond as to kin-by-descent. It is only the incest taboo which stands apart. It is also subtly different to the harming taboos in referring to occasions of breach which do not ordinarily carry much meaning in the codes of childhood. To understand why incest taboos are so nearly universal is to understand how close they lie to the primordial concept of social identity. It is the idea of self-and-other in terms of right which makes it conceivable that I might successfully hold another person responsible for an act, and others so hold me. The powers of mutual moral suasion among cats, dogs, horses, or bonobos are limited to the now-situation. Consider though with what delicacy our taboo codes apply to some not others, some situations not all. Under religions of responsibility the moral code itself may demand a man kill his wife for adultery, or justify his beating her for insubordination. Corporal punishment is only seldom denied a parent, even in otherwise lawlumbered societies. The critical context is the situation wherein the harm may jeopardize the structure of intimacy itself. And that is to say, taboos are only as strong as the relationship they protect. Unlike law, they are self-stabilizing. Yet once broken they no longer apply—the bridge they sustained is fallen in. I have elsewhere called the maximal-claim tie limitless. It is so in the sense that should any limits be put upon it, it will break away. Think of the sanctions of treason or personal betrayal. But think also of the great stick game of Kinga boys. Your best friend might break your leg with a wildly milling stick, yet it is not he but you who will break the peace by seeking to blame. 403

Is it true there are specific provisions of moral restraint in dealing with others, which are equally written into all the main scriptural religions? I find it hard to conceive the notion could be either confirmed or denied without getting down to detail, as ‘moral laws’ are inevitably stated in the broad sort of terms no lawyer can respect. But I think the question may well be posed, whether all human moral experience does not have a centre in the restraints we may call ‘intimate taboos’. These are constraints on conflict and congress in the internal domain. They can’t be defined to suit a lawyer’s standard, because they are context-sensitive and enjoy distinct contexts from one culture area to another. They have to be intuitively grasped in each such context—they are taboos. The Ten Commandments of the Judaeo-Christian tradition illustrate. Incest is unmentioned because making that taboo an express teaching would only violate the very mysteries it is meant to hide. So also with the kind of sacred trust in one’s own security which is vested in the harm-taboos within an intimate socializing group. Preaching against parricide might only serve to make the act thinkable. The premise of moral commandments is a heedful not a heedless client, an already viable human society not a human menagerie. Essential intimate taboos are presumed. But if the attainment of a systematic political-jural-ritual system (Fortes’s external domain) depends on the presence of a systematic infrastructure (the internal domain comprising the domestic or hearthside groups which must see to the adequate socialization of the young and the maintenance of required supportive rules) the implication is not that taboo is the father of law and civilization. You can’t ‘ban incest’ until you already have close-up kinship—a full system of family identities. A ‘horror of patricide’ or a ‘guilt of fratricide or incest’ presupposes the prior sanctioning of such identities. In effect, positive law (and with it the possibility of political order) depends on and emerges with the negative structure of intimate taboos which foster educability and enable the nailing down of responsibility. Compare Aristotle’s insistence that the law has no power to command obedience except that of habit.‡‡ Premising the advent of Culture—with all that’s human packed and ready—on a mock-historical acceptance of incest taboos has ineptly been tried by such rhetoricians as Freud and Levi-Strauss. The link is just functional. What society can boast stable institutions of political, jural, or ritual character if it can’t boast a stable basis for fostering morally responsive adults? Taboo is as old as human history but didn’t beget it, as Solon didn’t beget the law. Taboo assumes law just as law assumes taboo. Taboos of the hearth are familiar the world around but everywhere are seen with a local 404

face. What is the same wherever the tabooing of incest is found, is that each new generation is born into a uniquely placed genealogical slot, through which an unambiguous social identity of its own can be conferred on each offspring. (I set aside twins. Some societies find them problematic as lacking a firm birth order and defying sure identification.) Everywhere, individuals must each be assigned an unambiguous role in the structure of rights and responsibilities which prevails. This is normally done by birth-placement in a leastpolity tagged by a marriage—a mooring in social space. ‡‡ As for the emergence of incest taboos, the possibility of ‘holding others responsible’ would have built steadily through the sifting of moral concerns into early-human discourse, species behaviour skewing slowly toward contingent structure. The varieties of ‘least polity’ in the world suggest many branchings and restarts. Perhaps what has made the incest taboos so interesting to theorists is the special vulnerability which marks them. Among the taboos comprising the moral code of an emotionally close socializing group, those relating to incest have the most ambiguous bearing on the group’s viability. Atop that, incest taboos are peculiarly prone to breakage. The combination of access and temptation might not overcome sibling inhibitions (if familiarity does at least breed diffidence), but others with family access (e.g. parents) may not be so armed. Sexual opportunities notoriously screw up our personal moral strategies. Erotic involvement and sexual predation get mixed. Forbidden danger is commonly found the most compelling. Then, granting that incest taboos are not always respected, why is it they so often are? I believe we should ask if they are not generally submerged or embedded in the transcendent moral ban on breaking trust with intimates. In a case of incest, dyadic withdrawal of a couple is being attempted in a teapot. Breaching one link breaks a circle of trust. What needs examination is the nature of trust itself as a catalyst or demiurge of intimacy and its uncodified rules. If the nagging question has not quite been answered, how incest taboos are engendered and fostered in one autonomous society after another, it wants explanation in the same way as the full set of taboos comprising a local moral code. This has been discussed. If it seems necessary to explain why the incest taboos are so seldom dropped out of the set, I find the answer fairly straightforward. To start with, the taboos of doing harm and those of incest are not coterminous when they are reckoned as reaching beyond the primary socializing group. A taboo of violence can stretch easily out to include friends of friends or kin of kin, though this may include eligible mates for ego. In some tightly exogamous kinship 405

systems, war regularly breaks out between intermarrying groups. When a man kills his sister’s husband we see legitimate and morally defensible harm indirectly done a woman within the man’s primary ‘incest circle’. The two sets of taboos are not confused the one with the other. Still, I think within the primary socializing process itself the two kinds of ban do reinforce each other. It is well to look at the relatively transparent case for microcultures lacking documentation or licensing laws. (a) If socialization is to be linked to the bequeathal of social identity, certain conditions must be met. And (b) if the parenting agents will be expected without duress to support and own moral responsibility for their offspring into adulthood, further conditions must be met. Finally, (c) if there will be a regular pattern of growth and decline of households over time, with carry-over of personal values (deepstuff) from the old to the new, still further conditions must be met. There must be system in the way internal domains relate with each other and institutions of the external domain. The system must buffer socializing groups as they move toward marrying out, merging or segmenting, and property exchanges; and the system must be popularly known as just. I take these points in order. (a) Social identity can’t be ambiguous. Folk must be able to hold individuals, including the minors, responsible for their acts. For this standard to be reached, every birth has to be amply witnessed, usually with both genetrix and pater unambiguously noted. Claims of social parenthood based on adoption or its equivalent must be readily ratified. The same applies to claims hinging on broken marriages or reconstituted households. But these conditions can’t be dependably maintained without uniform rules for identifying all the members of each socializing group and their interrelations. An incest taboo at once magically deflates the balloon of possibilities for ambiguity and guile, improves the probabilities for ‘cellular’ homogeneity and self-balancing scale among the socializing units, and often allows both lads and lasses to be properly socialized while contributing to their ‘family of orientation’—get socialized before starting to procreate and socialize a new generation. If the reader wants a picture of what happens when social identities become unmanageably chaotic, I suggest the dossiers of relief and pacification agencies in West and Central Africa for the early years of the third millennium a.d. The variables at play in the socialization process include trustworthiness and civility. These and social judgement are often considered as benchmarks for assessing ‘readiness for having children of your own’. All this is pretty well implied in the idea of a ‘family system’, but bears heavily upon the matter of bequeathing a

406

secure social identity for all newborns who will be taken up from the birthing mat and offered a chance of life. (b) Market forces do not see to it that children are well fed. The system which does do parenting is another one, arranged to put children in the position of wanted members in a solidary small group. Economic responsibility for raising offspring to the age of selfsupport devolves upon the parenting unit. Usually we call it ‘the domestic group’, but for careful analytical purposes I prefer ‘the least polity’. Motivationally, this group is set up to make a decent living for all its members, and at least to reproduce itself in a new generation. Except for special cases the pertinent group is some kind of ‘family’. No society this side of a Brave New World can prove itself viable without these ‘least polities’ which are dedicated to managing and handing-on the arts of reproduction and survival on a quotidian basis. Market forces do not imbue the very young with social and moral skills which lead into lives of greater than market value. It is the genius of the internal domain to produce individuals easily trained into the roles the external institutions require. Policing cases of child abuse can occur even in uKinga—thumbs show up when dikes leak. But there has to be more to the external domain than handling trouble cases and common defense. It has to be involved in keeping its ‘families’ viable—there is rarely any other recruiting ground than its own ‘internal domains’. We have seen how Pangwa elders, as a community force, have their way with the young. They do this mainly through powerful ‘ritual schools’ staffed by graduates of the same. Their ‘external domain’ has perhaps no houses of its own. But because the Pangwa system gives great scope to these distinctive teach-ins, we are struck by the intervention of the community in the intimate affairs of a private household. In fact, the practice is universal, only less obvious to us in cases where procedures are less dramatically intrusive. (c) Relations among ‘household groups’ comprise the problem to which the institutions of an external domain are primarily addressed. The simplest foragers, exemplified in the popular mind by the San (Bushmen) of southwestern Africa, are almost without such an external domain. But settled agriculturists, wherever land and easy mobility are scarce, do supply themselves with politicians, courts of judgement, and a host of ritual practitioners. These institutions are staffed by and for the householders of each community, and these householders must be able to command respect mutually. It is probably important to be clear that each household has ‘exit autonomy’ in such a setting, and with that, ‘voice’. These basic rights are present in more distinctly acephalous societies as well, but are less visible there—as the external domain begins to develop, proce407

dures become more visible for according voice or forcing exit. My point is that households in a subsistence-economic community are ‘members’ in a polity. One can think of the household as a ‘micropolity’ nested in a ‘local polity’ nested in an ‘ethnic polity’. Kinship societies are often formally presented in that way, where a term such as ‘lineage’ can substitute for the less specific ‘polity’. I find ‘least polity’ the better term than ‘domestic group’ to think with, where general statements are being made in a cross-cultural forum. Wherever a family system is in charge of the process by which citizens are recruited from eggs, the political responsibility for this process will belong to the ‘least polities’. It is their work, and doing it comprises their claim to deserving a fair morsel of autonomy.

Logics of the Least Polity Rule In kinship societies the external regulation of householder relations is largely invested in the running of an effective marriage system, able to replace old families with young ones. Bridewealth, for example, should be reckoned as an ‘external’ institution where its terms are imposed on individual households by courts of law, not freely chosen. No sort of marriage happens naturally: the key institution of the internal domain is imposed from without, just as the staffing of the external institutions is afforded by the ‘internal’ institutional sphere. The two spheres are coeval, having evolved in dialectic balance. In the Malatan region’s amity-based societies, families did the ‘breeding’ but peer households did most of the rearing to adulthood. This is rather common around the world where you find children housed apart from their parents. But in either case economic responsibility for children’s’ survival devolves on the independent households legitimated by a marriage. Politically, their usually ‘cellular’ occupation of social space warrants regarding household groups as making up an infrastructure, at least where institutions external to the household can be styled as superstructure. But what is a key to understanding Kinga and Nyakyusa organization is that the main socializing institution (the peer household) labours for the parental households and in each culture sees male offspring fed in ‘communal’ fashion, even while taking the food from household fires. When I say the Sanga court culture arranges the system of court-bush circulation I have described for the Kinga, I am showing how external-domain institutions, through vested authority, can validate the political-economic mandate of the community’s standard ‘least polity’ even as it takes form on the political margins of a court’s acknowledged sphere of influence. 408

For clarity, consider now the same segmentary ‘nesting’ structure familiar to students of lineage systems, as it will appear in its planar aspect. Look at the ‘system on the ground’ which results. Heads of household, like their children, deal more often directly with their neighbours than they deal with them ‘through channels’. Torts and witchcraft accusations occur on the ‘civil’ plane, and in such everyday neighbourly politics each voice is about equal, and each member’s interests are (as is not the case the world around) best defended by himself. The main guarantee for this condition is that each voice in this system of keeping order through kinship should be close to and inclined to take care of all his or her own folk, but disinclined to expand claims on the common good at the expense of neighbourly envy. Every society studied by ethnographers has its own ways of working this kind of system. Scale and detail in the use of kinship links are always relative to local norms. But the main way the neighbourly entente is engendered is by sanctioning an optimal model of the self-sufficient subsistencecum-socializing group—a made-to-standard least polity—to suit the setting. The combination of equal voice with exit-autonomy for the householders is the mechanism which serves this purpose of mutual adjustment in the Malatan communities under study. Now we have a political cluster of subsistence households, each at some particular stage of a domestic life-cycle which will produce well-fed nubile women and woman-ready men inclined eventually to set up as a fresh, unencumbered generation of autonomous micropolities fashioned on the model of the households which have reared them. In the kind of society in question, family life is valued as an end in itself. But in the Malatan region we have three quite different scenarios for ‘family life’. In uPangwa there is scarcely any ‘group’ to be found but the family itself. The functions which belong outside in the ‘external domain’ are pop-up functions brought at need right to the family household. In the planar map of Pangwa social structure there is no footprint for the external domain as a member structure and voice. In the Kinga case there are quite a few different kinds of group which can be reliably mapped. There is a central court and outlying isivaga barrack houses for men. There are boys-and-goats houses for every grouping of households into a hamlet. And there are the isaka bachelor girls’ houses. Of course on this map you might also want to indicate ties between particular households and particular fields, storage structures, and the like. But ask, What are the parts, the ‘member groups’ through which individuals are counted as members of the inclusive political community? The answer, which for other societies might be ‘the domestic groups’, has to be couched for Kinga in terms 409

of ‘multilocal households’ or ‘dispersed families’. A map of the planar constitution of the polity surrounding a central court has to be coded to reflect lines of right and responsibility. These lines pass through a household head to individual family members. The Kinga family system is not typical world-wide. Heterosexual motivation is bumped from first place, and facework in the parent-child relationship perhaps from second, in the moral strategies of either Kinga gender. What is most obviously different, structurally, is the spatial dispersion of the sleeping quarters for household members. Labels for the model I am developing might be ‘ground-level politics’ or ‘planar analysis’ or ‘least-level politics’. The short reference I’ll use is ‘Least Polity Rule’ or LPR. I want to show that incest taboos are essential to maintaining this kind of self-stabilizing planar structure. This will be so, if always with a difference, in any of the empirical culture settings where incest taboos are actively maintained. In complex industrial social systems the diagrammatic mapping might require a special mathematics, but the main principles can be stated on the Malatan evidence. It is axiomatic that (a) good neighbourly relations are optimized by applying equivalent rights and responsibilities to all households in a political neighbourhood; (b) the enabling conditions for doing this include substantial replication of the internal household structure across the field—what might be called family custom law. This is carried through so as to achieve a high level of spontaneous order—minimizing the need for external intervention to keep the peace. In modern terms, disparate family systems will aggravate neighbourly turbulence; so also a mixing of rowdies with innocents, conformists with free spirits. The governing principle to have in mind is that an effective political system will require the greater use of force and coercion, the less it can depend on spontaneous order within its member communities. One corollary is that a dual society, suddenly losing the political barriers which set the two cultures apart, will see the escalation of disorder and government coercion. Most of the civil violence (Africa, South Asia, Northern Ireland) in the closing decades of the twentieth century fit this diagnosis. Another corollary: a political island able to maximize internal stability through coercion will be in position to dispense with the Least Polity Rule. The historic example I’ll adduce is the court of an African despot. It is the LPR which produces a consistent domestic cycle at grass roots and brings it about (most often through the institution of marriage) that each generation sees socialization within its own set of internal domains. Societies which build families into corporate 410

groups have proven hard to govern, requiring quite special constitutional arrangements. The tendency to press for such time-binding domestic groups, at the expense of neighbourly collegium, is present in most Bantu lands. It is obvious in the Wilsons’ ethnography of Nyakyusa, where the full members of a village collegium are few and comprise a greedy gerontocratic elite. Young men are kept in bachelor status and young women hoarded as surplus wives of the elite. But eventually the elders’ big stem village is allowed to complete its cycle, and the bachelor men are sent out on the elders road to join the game of privatizing women and cattle and exiling their sons in turn to satellite villages. ‡‡ The court barracks in the Kinga case was a fully corporate group, always replacing staff exits with new entrants. But as part of the political island of the court village it didn’t displace the leastpolity rule in ‘private’ domestic politics. What this rule yields is a citizens’ collegium comprised of household heads—the male elders. We have to understand a ‘citizens collegium’ in light of the need for giving political voice directly to each member household and so indirectly to everyone. In the Sowetan region this ordinarily means giving equal political voice to all husbands. The expectation is obvious: the sanctioned household structure bears rights to all members through its member in the collegium. When we speak of ‘elders’ in the region we advert to this scheme, and it is appropriate to style the overall planar political institution an ‘elders collegium’. It is not inappropriate to ask what sort of political guarantee this is to women. The answer might be, better at least than the guarantee of personal influence her sister may have on cases and policies through an urban ballot box. Out in the country, virtually every woman will have personal leverage on at least one male with a citizen’s clear voice. Gemeinschaftlich conditions do make a difference. The political logic of the incest taboos has now been outlined. In order that something like ‘equal voice’ may obtain, the Least Polity Rule must hold, so that the inequalities which are generated by one generation are not geometrically multiplied in the next. The result of unequal voice is trouble, often followed by exit or other symptoms of a broken peace, always in principle perceived by the collegium as weakening the whole community’s position. ‡‡ The logic of a citizens collegium will be evident in a review of the Nyakyusa example. Their political-domestic system is aberrant enough to test the logic of my formulation. The Nyakyusa are obliged by their age-village system to stretch out the political tenure of a village to match the standard of all the villages of an established 411

chiefdom. All must join in the Coming Out. It takes place about thirty years from the time the village was granted autonomy with the last Coming Out. So that all these villages can cycle as one, a son or nephew may be coopted from bachelor life to replace an elder whose death has left an elite identity vacant. The custom underlines the importance of maintaining a peer village intact to the end, even if it means bringing in a half-grown ‘peer’. This stem village is no mere network of chums but an equal-voice peer polity. The rule is equality through (say) ‘matched domestic standing among male elders’. So does the logic not hold? There are no smaller polities with voice. The Nyakyusa case is one of ‘rampant polygyny’ at the expense of structural skew. The age-village system means building up an overblown village of elders—men—intent on a centralizing accumulation of the movable wealth of the larger community. This includes produce, women, cattle, and offspring. Within the inclusive village, everything is owned and managed by some male elder. The process is ultimately limited by the Coming Out, an institution which summarily gives voice to men of a new generation. The Least Polity Rule is hard to see until finally the generational basis of ownership and political responsibility is made visible, as the new generation begins with a sound basis for equal voice within the political space of a new set of political domains—commoner and chiefly villages. Nyakyusa incest taboos belong first to the prepolitical world of the tiny age-village, where near-peers pass on the rhetoric of adult men to their juniors, then the climber’s household of modest means where a man may, in effect, be trading a sister for a wife. Taboo is not really proper to the manipulative culture of the external (public talk) domain which is used to justify the exile of youths to the outskirts and to a cultural life on their own. The seriousness of the taboo is in fact confected with the genuine seductive interest of pubescent boys in the little ‘mothers’ who are the new wives of the village ‘fathers’. The result of a rhetoric of suspicion is that elder learns to stay out of younger’s private life and vice versa. Then the younger but maturing citizens, now in their own house-head roles, learn as elders did before them to concern themselves with ‘incest’, bending the concept to explain why boys of the next generation must now also be sent to live outside in wee villages of their own, only coming in to till the fields and take food, and after flirting a bit be chased home to own quarters. As the business of the elders is only in building up harems and cattle and protecting their holdings, it is by amassing wealth at a strong centre that they keep dominion over the young men, otherwise let go on free rein. The main aim is keeping the undivided 412

loyalty of a son by employing him in the gardens, with the herds, and in war-making, which is a high calling. There is in the age-village system an effective provision for neutralizing citizen envy: men of the bachelor generation are at once held out of the political loop and held to the game of peer rivalry. Since they run their own village, the game is a political education in any event. Meanwhile, by using lateral inheritance, the decay of the older generation is systematically delayed, and young man’s invidious sentiment for his father is deflected as ‘incestuous’ design on some one of the elder’s young brides. If this logic bends the word, it trades on the incest code in a politically strategic fashion. Ironically, the viability of the Nyakyusa system of delayed marriage and massive circulation of the population through a Coming Out seems to depend crucially on binding a son’s loyalty by stunting his sexuality. In the youthful are-village homosexuality is normal. But as it becomes a village of young warriors, an otherwise repressed heterosexuality gets full play in the dance, particularly at funerals abroad, where ‘incest’ has no bearing. Homosexuality is put down as a thing of immature boys. The prickly side of this is that all along the heterosexual interest has been teased but staved off by the broad ‘incestuous’ caricature of young men lusting after their father’s young wives. We have seen that all the elders are virtual ‘fathers’ and all the elders’ wives ‘mothers’ to all the lads, who are scarcely separated at mealtime according to ‘actual’ kin ties. Similar logics apply in societies which have stretched and magnified the kin-group conception to embrace the full roster of institutions belonging in the external domain. An ideal model of the segmentary lineage system, best known from Meyer Fortes’s monumental Tallensi study, illustrates. Here the senior voices belong to the ‘clan’—or it may be ‘maximal’ or ‘medial lineage’ heads, depending on the jurisdictional level you have under consideration— with the lead always devolving on the eldest capable male belonging in the fold. But power in the system is always devolving through mechanisms adequate to renewing it with ‘new blood’. Internal demographic shifts get to be recognized by semiotic alterations in the alignment of ancestral claims. Hiving off occurs when too few men of a given group (at whatever genealogical level) have the voice they seem to deserve. The result is admitting new voices into the power loop. In the very long run migratory pressures realize themselves in the gradual remapping of the Tallensi polity and the larger system of alliance and opposition of the fuller ethnographic region. Throughout the system structural change proceeds at all the genealogical levels, accommodating in unforeseen ways to a pyramid of political voices. A pruner of orchards will know the 413

formula: the life is in the roots. I have argued it is the salutary cover of moral taboo which keeps it there. Taboos have a way of cooling down quibbles before they escalate.

The management of intimacy Of the internal domain’s ‘intimacy taboos’ that on doing violence to the point of bodily harm is the most elastic. Children have fights which soon pass over. Corporal punishment if approved is limited to what ‘does no harm’. Many societies allow that a wife may leave a marriage to an abusive or otherwise malicious husband, and where that solution is not approved, continues abuse is nonetheless socially condemned as harmful to the social order, and usually hidden. But violence is less likely to be deliberate than seduction. Violent episodes within the inner circle of a family are often followed by a show of mending the peace. Fights are apt to leave two clear losers or, at best, one ambivalent winner. By contrast, voluntary incest is its own reward and self-motivating. Even involuntary incest can take on the charm of a bad habit. For the most part, in the same way that family members become inured to a certain level of conflict, they will become inured to a certain level of sexual tension and ‘learn to live within the law’. But a clear breach of code in the inner circle, on sex or bodily harm, is not readily made known beyond that circle. Taboo takes its power by shaming and secrecy. Few ‘moderns’ in the West even regard the keeping of peace within the home as a matter of taboo, and many adults have no word for ‘incest’ in their vocabulary. Is there a slang word for it in English? A single, well established ethnographic case, mentioned before, does seem to say that of the two sorts of intimate-group taboos it is the prohibition of incest which is the most sensitive to breach. In Evans-Pritchard’s reconstruction of the Zande kingly courts we find incestuous sibling and father-daughter relationships well tolerated. It is not accidental that there is also a rather high rate of fratricide, but neither is if surprising that such murders remain taboo. The internal structure of the ruling house is that of a despotic microcosm. I call it a ‘swollen’ internal domain because though it was based on only one man’s marriages—the king’s—the court with its harem and ‘family members’ was a major village of its own. ‡ My reading of the case is that free-love incestuous attachments were the price of preventing planar structuration of the court. This the king could hardly want. It would compartmentalize his court, limiting his access to members. It would prevent him keeping an eye on everything the junior royals were up to. Evans-Pritchard found sons were kept in keen mutual rivalry to the point of snuffing 414

out any move toward an entente against the throne. Separate ‘families within the family’ of the king could more easily build power structures of their own. As for the psychology of breaking taboo, there is the dramatic meaning of flouting convention: it makes of the royals one kind, able to see its interests solidly in opposition to the masses of commoner clansmen. Flaunting evil makes the king’s kind special and keeps royal aspirants to power from being trusted by commoners who might otherwise be teemed up in revolt. Most important: each Zande court faced a chronically turbulent military situation. Azande by the late nineteenth century was a rapidly expanding despotic system of principalities with rubber borders. ‡ What I think the case does show is that the near-universal practice of including incest taboos in the package of a people’s moral code hinges on the systematic importance of the citizens collegium. The chief executive officer of the Zande ruling clan did not want princely segments hiving off to build increasingly autonomous fiefdoms of their own. He established a communal organization at the court with no place in it for other households than his. His own security as a householder relied on the sheer size of his establishment and his personal power at its centre. The crown had been his not to inherit but to take. His means of communication were oral and pedestrian, limiting scale to what the tricks of first-person majesty can manage. In a region of kin-oriented polities, a conquest group may have to set itself up as a politically integral caste. Short of that, each political island will—as indeed was the case with the Sanga principalities in uKinga—come to depend on a positive policy of local fraternization between rulers and ruled. In the Zande case the grasp of each kingdom was meant to stretch over a congeries of unrelated and rivalrous commoner clans and their widespread settlements. Legitimacy through fraternization was no option. Where might must make right there is safety in numbers. The key to the Zande court’s inner structure was repealing the incest taboos and centralizing the political-jural-ritual institutional complex—the stuff of the ‘external domain’—in an expanded court always under the royal thumb. The fact that a community, however uniquely situated, can so readily slough off the incest taboo makes it clear such taboos must be socially enforced or cease to exist. Biological factors are of marginal importance. But wherever autonomous households are spread like tiles across the ground of an ethnic community, incest taboos will be retained. They are crucial to a system by which a self-stabilizing infrastructure can reproduce itself. Call this ground-zero politics.

415

I am now in position to review the lessons of Kinga sociology. The problem is to show how Sanga deviation from a standard-model family organization affects the form of incest taboos and illuminates their nature. The Kinga case may challenge the view that family life is the cradle of morality. Parenting is not central to socialization in Kinga tradition. All the same, manners and morals are surely no worse (and likely better) than standard. The case is not without general sociological significance. A point for educators to note is that ‘peer socialization’ in this case entails small groups ranging in age from about five to (say) maturity, with a break for each individual at puberty to move from the children’s to the youths’ sleeping mats. If all a boy’s peers were in his own narrow age range, or if the longterm continuity of the group were less ‘family-like’, the process and quality of the education would plainly be different. Even where no rules effectively segregate adult age-groups, individuals are crucially in close emotional interaction with an odd-lot of age peers. But in the Malatan regional culture of rules age segregation is as bold as we find anywhere. While this is most particularly true for males, praxis for women is clearly forced in the same direction. Age villages for Konde men and barracks life for Kinga have to be reckoned as ideal schools for establishing an open-door pattern of amity among peers. In both cases the political rules thus foster an internal domain of intimate interaction on a broadly localist plan, belittling family-centred values or, arguably, drawing on the ‘family’ code in a more nearly communal system of peer socialization. I have to this point been assuming that the following three major premises are valid. But each is challenged by Kinga data. Premise One. If the human way of begetting entailed no erotic attachment to a partner, the primary socializing group would not resemble our ‘nuclear family’. Premise Two. If descent were not the crucial determinant of social identity, the institution of marriage would not be a standard feature of all major civilizations. Premise Three. If primary socializing groups did not engender maximal-claim ties, the routines of community life as we know it would be unknown.

These premises as they are framed reflect the ideal-typical society you might have in mind from a wide reading in world ethnography. The first premise looks virtually self-evident: if sex were no more significant for human relations than defecation, why would we have couples wanting to live together, share their privacy, pool their 416

means of subsistence, and attach themselves to offspring in the possessive and emotionally overloaded way they characteristically do? Yet for Kinga the bond between husband and wife was traditionally nothing like the happy Christian marriage I occasionally met in the Kingaland of the 1960s. The scene of cohabitation in the unconverted bush culture was a tiny grass hut with no fire, which a man would crawl into backward to avoid showing off his bottom. Under the Sanga dispensation, except when a woman was in search of a pregnancy, the man would prefer the barracks building where he slept with his fellows. Even for recent times, the self-sufficiency of the Kinga wife is famous. The attention she needs, so long as she has a pregnancy advanced or finds herself in the four-year-plus period of nursing, is strictly practical. The sexual relation does not noticeably magnify the emotional tie, and can’t compete at all with the nursing tie. We have seen that Pangwa traditionally extended nursing several more years and without pregnancy. It appears that erotic attachment, as distinct from episodic lust, may not be particularly important in shaping a human social system. In Kinga perspective, though, we must reckon that what does remain stable in family life is devotion between mother and child. The mother feeds sons and daughters and husband from her fire. The food is from her own gardens, where she works the year through. So we have a family system, but one very short on cohabitation. Even when it comes to commensality there is little. Boys come with their chums, all of them eating from many fires. Men sleeping in barracks have their food delivered. There is certainly a sense of mutual dependence on an economic plane. Every locality knows widows who live on short rations, but the worst-off are widows who fall ill, perhaps with young children. A husband or a caring brother in such a case would not leave the family bereft. But in the one actual case I knew it was a peer, an unrelated local woman, who came to aid the failing mother and afterward raised the child. The illness of a child is first tended by peers in their sleeping house. Only in severe illness is the news brought to a parent. The father might bundle up the child and carry it in haste over long mountain paths to a curer or (in modern times) a dispensary. But there is, oddly in an outsider’s view, no tradition of nursing a fever. Is peer socialization accountable? I can only speculate. In short, we do find that in the absence of erotic love the nuclear family loses much of its magnetic character and plays a reduced role in socialization. Premise One seems to be justified.

Premise Two links clear details of a person’s descent to an unambiguous social identity, and supposes this in turn accounts for 417

the institution of marriage as a basis for bestowing a distinct social identity on most children. Kinga practice neither confirms or weakens the premise. Descent is not by any means the prime basis of social identity for men, except in direct relation to sex taboos, which come to be reckoned at the time a man and woman feel ready to marry. If the elders can’t name a common great-grandparent, marriage and sexual relations are approved. But it is notable that even with the advent of writing Kinga seldom had a clear view of the two sets of grandparents and four sets of greats who must be identified. The problem is not made easier by the freedom with which individuals choose new names or go for periods under an almost prankish alias. In many or most local areas, there are only a handful of surnames, and these do not often reflect known kinship ties. In some areas everyone has the same surname and no one would suppose that indicated close kinship. The original birth name is discovered on the spot, when a sisterly attendant catches the mother’s first clear words and takes from them a distinctive expression suitable as a name. In short, descent matters but lies in the background of ordinary life, seldom coming to the foreground. Wherever society is on a favoured-name basis, no one uses any other to refer to a friend or acquaintance; very often I would find a well known man in one community instantly recognized by his idiosyncratic favoured name in another place, however far. Is the premise valid? The Kinga evidence suggests that descent would always matter and might be irreplaceable as a final basis for establishing identity. But the main implication of the evidence is that marriages may be expected of all men and women in reasonable health by a community which mainly uses the marriage to allocate special rights and responsibilities to men and women in connection to begetting and rearing children. In the Kinga instance the traditional dispensation assigned most of the task, by far, to the woman. But the importance of descent in this allocation can be seen in the scenario for the absent husband: if a husband (through his absence or otherwise) can’t service his wife when she is in need of a pregnancy, the responsibility devolves on the husband’s brother. On the other hand, if a widowed woman does not choose to be passed on to the husband’s brother, she may decline. Descent is recognized in the brother’s default position as heir to the social contract of marriage, but may be set aside if proper procedures are followed. In such detailed provisions of law we may read the degrees of freedom which pertain to each status assigned in the social structure. The social contract entails assigning the newborn a social identity important as a guide to the danger of incestuous marriage, which is burdened with social danger even if the parties are ignorant 418

of their kinship link. The point to have in mind is that incest in the jural sense is impossible without unambiguous data on descent, for each individual in the community. This is to say that a clearly reckoned descent structure is a necessary condition of an incest taboo. The taboo, then, is not (as Levi-Strauss and others had it) the genetic source of the family as a human institution. Premise Two is not invalidated by the Kinga evidence. But packaging marriage with descent reckoning, even if it could be empirically affirmed as a universal feature of cultures, leaves us the question, whether the incest taboos are to be reckoned as part of that package. Inconveniently, I don’t think so. Judging by their lasting quality, there was always a personal attachment in Kinga marriages. But as we’ll see, there were obstacles to close relations of father and son, despite the fact that Kinga men see themselves in patrilineal terms. From my observation I would have to say the father-son tie was rarely invested with the kind of positive affect expressed in mother-son and especially mother-daughter ties. For men, that kind of affection was displaced to the elder-younger brother tie. In sum, a sociology of marriage among Kinga would focus on the importance for social identity of common descent among near peers (sister-sister, brother-brother, and extensions to cousins), and would relate the marriage contract particularly to a woman’s desire to suckle children, have them about her, and know daughters in amity as they grow. It is not irrelevant that the traditional marriage in the bush culture was always done by the maiden giving herself to a man, knowing she could not then return to friends but must follow him. The idea of a parentally contracted marriage was alien to this, and generally despised by the young. I found that Kinga women married-in from uPangwa had enjoyed only a few maiden years after menarche. The gradually declining age of marriage for all girls in the 1960s suggested that a traditional girl who didn’t want eventually to bear children would have been hard to find. But given the high morale of bachelor girls it isn’t hard to understand their choosing to keep the bloom of youth while it lasted. The apparent suddenness of the maiden’s plunge into marriage (as seen by young men) bespeaks a deliberate choice to start a pregnancy. Again, the fact that descent and descendants are not all-important to Kinga men is part of the picture of delayed marriage for both genders. Another part of the picture is the absence of descent-related barriers to sex with same-sex peers, whether for boys or girls. Easy dalliance is a fair part of the charm of bachelor life for both genders. Setting Kinga custom against Pangwa, the meaning of youth is utterly different. Small Kinga boys and girls may, for convenience, live 419

together in a house with somewhat older peers; but well in anticipation of a boy’s puberty he will shift or be shifted to an all-male house. Heterosexual awakening is deliberately delayed. Adolescent girls enjoy caring for young girls, and don’t let them interfere with more ‘grown up’ matters at the other end of the hut—there are only the sounds in the dark, and until sleep the small ones will usually be busy with their own talk. I take it the scene is much the same in a boys’ house where both age-groups are present, except that older boys would be heavy-handed with discipline. Talk across the floor, when it is meant to be heard, helps the small ones grow up quickly. For one thing, the small ones will soon know that the only answer to ‘What was happening last night?’ would be ‘Nothing happened’.

Premise Three poses the familiar problematic of ‘structure and sentiment’. Put baldly: Are psyches caught in the institutional web of the social system, or is the system serving the psyches? But putting it thus baldly offers the easy answer, both-and. Most of us in the social sciences today are content with that; we differ with the extremists of either environmentalism or bio-determinism. But the problem ought to be carefully laid out where it concerns maximalclaim ties. These are the ties pre-eminently associated with close kinship, but also found in ‘blood-brotherhood’ (well known among Eastern Bantu) and other forms of close friendship, including the conjugal bond. We should be aware also that they give us the archetype on which politicians have constructed jihads, crusades, genocides, and suicide bombings. Technically, maximal-claim ties are ‘extreme’. Our problem is how certain structural ties come to be invested with so much affect that either party is prepared to risk life in order to save the tie. I put the matter that way because I am bound to see structure as prior to what I consider to be structurally stabilized affect. Premise Three proposes that the existence of maximal-claim ties deeply colours the routines of community life outside the circle of the domestic groups which give us the archetypal form of such ties. Kinga usage suggests that ties of amity may be invested with the ‘maximal-claim’ grade of affective commitment. But what we quickly see is that in either the Konde or Kinga amity systems, the infection of all ties of symbolic ‘brotherhood’ with such commitment would constitute an unbearable neurotic load for each individual. The structure of kinship limits the number of ties a man must be willing to die for. Even a man’s ego can only be so big. The structure of amity rather dilutes the affect which can be invested in ties between a crowd of ‘good neighbours’. This means only ‘true’ bonds of kinship can be ‘maximal-claim ties’. The telling case is the episode of the leopard’s tail: 420

The man A in a group trying to drive off a rogue leopard threatening their community lost his spear and saved himself by grasping the creature’s tail. His friends were unable to attack the leopard without stirring it up to the point of endangering the man behind it. They therefore kept its attention while sending for A’s brother. Only a brother, they tell, could take the responsibility of rescue with all its risks. A’s brother comes, and between the two of them and a little help from their friends they kill the leopard.

The first point to make is that amity has been shown weaker than kinship. But to follow that up, we must focus on the kind of bond which features in the tale—the fraternal or ‘BB’ tie. This is in a Kinga community the one peer bond which is seemingly overloaded with affect. Not only does an elder brother serve as mentor and model in what might elsewhere be a father’s place, but the sexual bond between brothers has a homely ritual place in bridging a young man’s orientation from homophilic to heterophilic relationships. The event is rather similar to the procedure at male puberty, when a boy is discovered ready to be formally shifted out of the prepubescent crowd, and is taken over by an elder peer (perhaps his brother) to join the crowd at the other end of the local boys’ house. I think it is crucially important to know that this special case (BB) is sexually toned and condoned. I am about to deal with the question, how this sexual bond can be condoned within the very heart of the family when we expect to see sex tabooed for all but the conjugal bond. That bond is licit because one partner is an outsider to the home kinship group. For now, the import I would focus on is that the BB tie is redoubled in strength when compared to the peer-amity tie alone. Talk about sex does tend to wander, and I think it best to be quite clear about the Kinga evidence we are addressing. First of all, I want to be sure that proper observation is not jeopardized by an implicit premise that only heterosexual intimacy is eligible for taboo as incestuous. An observer should withhold judgement on that. Kinga are not primarily heterosexual. Their ambisexuality is coloured by no doctrine that one orientation is morally inferior to the other. Due care on our part seems called for to avoid dealing with the microsociology of intimate life among Kinga on premises which don’t quite fit the case. I also want to be clear that the peculiar shape of the Kinga family system can best be understood in light of a disclosure model I will call ‘the absent male syndrome’. If you blank out the males, the structure of the Kinga family looks ‘normal’. That is, the family is centripetally organized with strong ties of loyalty among all its 421

members. Add the husband and you are essentially bringing in an absent male. Father-daughter relations are seldom intimate. The father is by objective standards a sometimes member. He is needed in the circle only for begetting, the rest of his contacts are tangential. Further take into this picture the sons and you are adding boys whose nature is conceived as untamed—‘wild’. They show up once a day at feeding time, and are polite, as if to deny their natural wildness. Another special feature of the family is that it is by Bantu standards small. The maximal size limit is four offspring, and extension (or reduplication) of a family through polygyny is almost reserved for lords and princes. The principal domicile for bush-traditional times is a woman’s hut where a husband stores his things. For court culture the domicile is more sturdy and contains the family fire, but commensality is not the rule. Pertinent family rules: Boys avoid all secondary intimacy with the father, such as taking his place at a fire or touching his clothing. Boys present this as a far more serious matter than avoiding virtual intimacy with the mother. Mother-daughter rules are not symmetrical, though ‘babying’ a girl, as with a boy, ends entirely with weaning and setting the child out with peers. Girls work closely and efficiently with their mothers or older sisters and are effervescent without being ‘wild’ in their amusements. For bachelors of either gender there are no set limits on same-sex intimacies, but ‘crushes’ in either group should not become ‘couples’. I have tended to discuss this as non-possessive bonding, and to distinguish it from the social contract expected with any heterosexual coupling. Elder siblings as a class are always in a socializing role vis-à-vis youngsters in the appropriate peer hut. This applies cross-gender only for avapapwa, the not-yet gendered child carried about by an own older sibling. At the point of marriage a younger sister may possibly join the household of her elder, as a second wife to a man inclined to take her. This is more frequent in fictive talk than in practice. The arrangement is supposed to be auspicious, and would be initiated by the elder sister. Cohabitation begins with a ménage à trois, a single household and marital bed. This parallels the case of the married elder brother and his open-bed policy toward the younger. Are brother-brother or sister-sister relations incestuous? The connecting structure is outside the parental household. The jury, unless it is the peer group itself, seems to be out. Sexual intimacy between sisters and between brothers is in fact condoned by the respective house rules under which they are socialized. Only the house of procreation itself is subject to the prohibition of ‘dangerous’ ties, and they are preeminently the ties between father and son. It is the existence of this prohibitory rule, 422

which we may go so far as to call a ‘father-son incest taboo’, which deserves special attention. What would happen if that prohibition were ignored and dropped? The structural separation of ‘family’ and ‘peer’ houses would collapse. The non-binding love relations of peers would be overlapped by the binding relation of lineal descent. It is the special feature of male descent lines which requires marriage bonds in the service of planar structuration. Filiation ties to a mother may be regarded as given by nature. To a father the tie is not such but needs be unambiguously asserted by social rules. Finally I want to suggest that the brother-sister sharing of a sleeping mat, which so astounded Soda (“Maybe his penis doesn’t work!”) falls into pattern if we recognize that many young men have not yet been and never may be wakened as heterosexuals. The girl, who was still playing in the daytime with a much younger brother, was not in the usual way sleeping with girls, as her location was isolated. If she had been an isaka member she would doubtless have been better socialized into the proprieties. But it shouldn’t have surprised me to find the ‘explosive’ taboos of incest so well hidden from the very young people they were conceived to protect from ritual danger. Fear and avoidance of incest is not any more instinct than homophobia. Human sexuality, though it needs be awakened, only by fair chance most often is. There is a crucial distinction to be made between ‘bonding’ and being structurally close. Most academic work on the prohibition of incest has tended to blur the point, and I find it best to start afresh with a more pointed application of terms. Involvements have to be set apart from structural ties, relationships, or bonds. Since a ‘bond’ may be either, for clarity in exposition I prefer to use other terms where the meaning may be mistaken. I can’t underwrite the use by biosociologists of ‘bonding’ (per apes) and ‘bonding’ (in the same breath and sense) in discussing transitive amity relations among men—the “any friend of Jim’s...” phenomenon. What that careless usage masks is the subtler phenomena I have called ‘involvements within the tie’. Such involvements may either strengthen or destroy the preexisting tie. Love within a marriage usually strengthens it. But love between brother and sister, when it is transgressive can be destructive. Incest as commonly conceived always is ‘involvement within a structural tie’. But African ethnography is full of cases which show the stinger may be ritually removed where there was no real tie ‘on the ground’. The clear conclusion must be that it is the mixing of two sorts of regnancy which is found dangerous. The maximal claim bond is a sleeper, not to be used wilfully. A parent or sibling can refuse help when it is no matter of need. These are socially sanctioned bonds

423

meant to survive hostile involvements. But they are unlikely to survive ego-involved distrust. They will be treated as contingent. Malinowski’s famous argument that more than one sexual involvement inside the (nuclear) family would overload the system— blow a fuse—was tempting but failed to explain incest taboos because it was too easy to suppose some folk would think that kind of overload fun. Still Malinowski was half there. You can see it when you ask what would become of the maximal-claim tie and its axiomatic ethics if the family system were tolerant of random involvements within the tight structure of rights and duties built around its hearth. I think the notion of an overload is right. But it is not a psychic, it is a structural overload we have to consider. The strength of family ties is made, not inborn. It has been a prime error on our part to suppose maximal-claim ties are not unnatural. The Least Polity Rule is the basis of ground-zero politics. Without the LPR a kinship organization would face a ‘system crash’ after the run of a few generations. There would be more individuals without than with secure kinship identities. The predictability of trust would vanish. I have argued that the LPR requires the presence of a full range of intimacy taboos, incest included. What I have now to explain is why families know all about that, when scholars apparently don’t. The reason is that planar structuration has to be continuous in time and social space, to survive from generation to generation. The ability of a family system to meet this requirement depends on making the mutual loyalty of its members absolute—fear of losing it all has to be embedded in the psyche and when it comes to surface has to breathe an atmosphere of trust. Sexual involvements make loyalty a problem. Parents know this. It comes with marriage. It runs in the deepstuff of every culture everywhere. Where family solidarity can’t survive contention, the family system won’t.‡‡ When I wrote (some decades ago) that ‘involvements within the tie’ can override structure, I didn’t have only sex in mind, or only the customary law of Malatan peoples. I thought my essay on incest would require a tome of Westermarckian proportions, examining the whole ethnographic roster for structural differences in the family systems of the world. That work seems less pressing now. The full review is still wanted, but I think it would have to be a study of the many ways ‘least polities’ can be organized to do their work of riding a culture through generations and millennia of change—not a quest for the answer to ‘the incest problem’. There is a balancing tension between the ‘internal’ and ‘external’ norms in all societies, from the widespread foragers of the deserts 424

in southern Africa to the once-majestic states of Ghana or India. The deepstuff that determines the cultural style of a human community is product of the socializing institutions of the inner domain of social relations. It is mostly socialization by contagion. Any system of ‘marrying out’ means that the principals in the socializing process will be a cross-section of the larger, in-marrying community. Institutions of the external domain which comprise the political, macro-economic, jural, and ritual or spectacular activities of the larger community owe their manners and logics to the socialization purveyed within the many-tiled world of the community’s internal domains. Tiling—planar structuration—is the border-to-border infrastructure. It is the prime source of that social stability which is vested in the rights of citizenship. The quickest way to see this is to ask how incest taboos might be related to exogamy. Since a rule of exogamy presupposes a society comprised of domestic groups, such a rule can only come into being where a homogeneous interdomestic structure already exists. In short, exogamy is an institution belonging to the external domain—a kind of law, not a taboo. But an incest taboo must already be in place before a society can envision the use of exogamy to extend the horizon of domestic compatibility on a plane which makes a sense of common citizenship possible. There is no Solon and no Napoleon to lay out the groundwork. The pattern must emerge locally, bit by bit, before it becomes general. We have seen it at the emergent stage in Fr. Stirnimann’s reconstruction of midnineteenth century uPangwa: the language of patriarchy was used in the elders’ narratives of life in the local settlement [lutanana] which in fact was far even from being straightforwardly patrilineal. If we must talk of function, incest taboos can be said to pay off by subordinating sexual interest to an ethic of practical cooperation in the inner domain. They can be seen to work by damping destructive involvements within the structural ties which are the manifestations of descent in a socializing group. But the reason they exist is not because families agree to value them for their payoff. The taboos exist because family members have grown up thinking incestuous sex is morally dangerous. This idea is widespread and is maintained by the system of planar structuration as it is set up in any given community. That is, membership in the structure is based on the ability of neighbours to sanction neighbours. This is not done by authorized power but ‘spontaneous reaction’—by the sanctioning power Durkheim called ‘mechanical’ and others call ‘diffuse’. In its potent form it is surging moral indignation. Thus it happens that when incest is kept secret it survives, but when it becomes known at all it becomes known generally and condemned as morally dangerous 425

by self-respecting citizens. Every human institution has an appropriate structure and limited freedom to alter it without default. Rogue institutions plague particularly modern societies. My reading of the ‘Mbasi case’ suggests premodern societies may also be prone. The stability of any social structure depends on control of deviance by whatever combination of sanctions fits the case. Planar structuration is not restricted to the macro form of ‘least polities’ but can be seen in micro forms: schoolyards come to be kept in tenuous peace. Writers, entertainers, marketers, industrial designers—all have to heed one another and keep up to changing standards. No social structure can be reduced to mere roles and rules. Fiduciary relations are always the basis of structure, and while they can be sanctioned under the guidance of rules, they can’t be created by sanctions or inspired by rules. Players can be too slippery to depend on. Ground-zero politics always apply. Without the incest taboos, strict descent groups could not form because the internal structure of the ‘least polity’ would drift away from the established ‘least polity rule’. This would in turn confuse the programme for producing clear-cut social identities for new community members. There are several considerations. One is inheritance. There must be unambiguous rules for inheritance of status, property, and children as well as the handling of debts and women. Another consideration is systematics. The way descent is reckoned and weight given to it has to be readily agreed to at need. Because rules have to be clear and uniform with local law, all this requires a standardized least-polity institution. Without it, each (extended) household group would come to contain persons of ambiguous lineage credentials. Even without property problems the flaws would inflate geometrically over the generations. Fr. Stirnimann’s Pangwa were gatherings of scattered households recolonizing their ancient lands and nursing into existence the kind of social contract which could put them on a viable political basis. Their exogamous rule employed the rhetoric of social danger which radiates from the incest taboo, though in the absence of usable data on descent, the rule had the practical effect only of giving the elders control of a connubial lottery. To make brides available, though, maidens had to be movable pieces on the game board: they should bring with them no ‘foreign entanglements’ to carry responsibility for, bring to a marriage no ambiguity of parentage. Ironically, very few clear genealogies could have existed among the remnant populations Fr. Stirnimann studied. The sense that identities all the same were intact was supplied by sponsorship, underwritten by the bridewealth contract. Ritual solemnities can propagate truth and trust as features of a ‘system in mind’. ‡‡ 426

Where out-marriage applies to both genders, as with Kinga and Nyakyusa, no one keeps home. No one stays on to inherit. Another system than kinship is needed for establishing solidarity beyond the least polity’s spatial-temporal frame. For Kinga it is the movement from boys’ house and men’s barracks to marriage and life with a few chosen peers in a newly settled hamlet; or from girls’ house and bachelor women’s quarters to marriage away from home. For Nyakyusa the whole emphasis on solidarity is laid on the masculine ethos of the age-village. The critical variables in both cases are propinquity and group-sanctioned ties of (absolute) amity. I suggest calling the prescribed solidarity ‘absolute’ because I find it a simulacrum of the maximal-claim tie of close kinship, informed as it is by a stubborn premise of mutual loyalty grounded in trust. Where kinship is not kept close, amity must be. Some earliest peer ties, at least for men, are lifelong ‘on the ground’ in either setting. Is it that every political system, from the acephalous to the vastly complex, tends as a regular thing to expand its ‘citizenly’ infrastructure? So far as least polities find themselves in ‘presocial competition’ with neighbour polities of the same order, the point may have to be conceded. My emphasis on planar structuration was meant to recognize that. But in effect we are reduced to a simple formula: you don’t get a cellular structure by tiling social space with iffy, amoebic entities. Incest taboos let a community be comprised of cellular units vested in a self-leveling form of power base, and this in turn allows the community to take form as a stable polity. ‡‡ Because it is political, the linkage between the internal and external institutional spheres is inherently tractable. New generations work out their own logics. The power of old symbols may remain while their meaning is turned about. Chiefly authority had evolved in the Malatan cultures in ways we can never recover, as political densities deepened and external pressures impinged. What the Lwembe cult was in 1900 was likely as different a century before as a century later. That we can’t document the ubiquity of change in preliterate worlds on any shorter scale than the archaeological is no argument for denying it. Once this is rightly understood, we have an answer to the origin and nature of the special taboos of a household moral code. Taboos on harm, and on the de-structuring personal involvements I’ve discussed, come to us from no Freudian or LeviStraussian big bang but from continuous creation. That taboos can be learned through the subtlest of signs, yet hold a rambunctious young person from breach even where the downside is only foggily seen, seems to me a remarkable fact of human nature. Some rules, though sanctioned by taboo, notoriously don’t get properly learned, or always hold when they have been learned, because the fatal kink in 427

the act has to be sensed through insight in the private arena of emotional interaction. Yet as a species, the evidence is we are apt at observing scarcely-mentioned rules about the unmentionable. Each generation interprets in its own way, and does so with a show of acting on its own volition, however antique its rationale. The talent to gloss over the unmentionable is presumably wired in, as the gift of language learning is. Learning civility in childhood doesn’t need a genius for conformity. It wants such simple gifts as shame, pluck, and pride. It does want a household normally at peace and able to hold its own with neighbours, though. The big lesson seems to be that effective socialization requires a closely structured primary group in which everyone has his or her place and the rights which go with it; the pecking order is clear; and morale is eventually restored when trouble comes. Next door, whichever way you turn, are primary groups operating on the same plan. On a larger scale than the household the social space we are considering is called a village (Nyakyusa) or hamlet (Kinga) or deme (Pangwa). It is a local polity tiled with households and committed to treating them as groups with mutually equal rights and obligations. What particularly distinguished the two Malatan protostates were their socializing and ordering systems, mixing ‘family’ and ‘peer’ households. These systems turn out to work well in both protostates, though organized in quite different ways. Peer groups have the usual taboos on harming members, but sanction their breach with less than parental authority. Being gender-segregated, they teach the doctrine of incest’s dangers in a somewhat academic or categorical way. In this they contrast with the Magoma community located partly in Konde and partly in Kinga country—Magoma have virtually unisex dormitories and early marriage. Magoma boys and girls know explicitly who can do what with whom. When I discussed these matters with Magoma elders, the local boys’ crowd would be right there ready to consult—it was a switch on the Kinga scene, with its structured-in generation gap. As for the peculiarity of the Kinga system, it is most notable that the family itself is not charged with the socializing project after weaning. This means there is no solidary family household which is home to the brothers (and the ‘BB’ tie) jointly. That is, BB is not so much a ‘family’ tie as a ‘peer group’ tie, and is not as such a maximal-claim tie. Peer group structure negatively sanctions ‘possessive’ involvements. Where structural equality is built in, ‘possessive’ means ‘dominating’. Kinga fathers are not hesitant to dominate their sons. But because both brothers emotionally avoid their father, they live in a social space apart from that of paternal 428

domination, and can be unaffectedly non-possessive in their mutual dealings. Consanguine solidarity in a Kinga family remains tied to sex avoidance in the form of bilateral exogamy. Boys’ and girls’ houses, being the near-total institutions they are, manage socialization outside the family. Solidarity there without sex avoidance is therefore appropriate. We have seen that it applies to siblings of either gender. Sex within a close kinship tie for such sibling pairs is comparable to sex within connubium: a tie of amity is enjoined, and it therefore lacks the perishability of affective ‘involvements’ and transcends them. We have to deal with caritas not with eros, structure not involvement. The norms of the ikivaga and isaka have no bearing on the fastening of birth identities. As if to say as much boys often change their names and girls always use riddling pseudonyms in telling tales on their friends or on themselves. ‡‡‡ Once more compare taboo and law. Blame and responsibility are the two main instruments useful in applying sanctions to support an internalized moral code. The one is most intimately associated with taboo, the other with law. The reason is that blame works at an intuitive level, while responsibility depends on powerful forms of socialization going well beyond notions bred in the bones of youth. Law as the key to a just system of government is no more than Hope’s Agenda if there are no heavies to enforce it. Blame is the key to justice known in Malatan socializing circles, which everywhere gather around the sundry fires which welcome them. Blame is a way of meeting trauma, handling fear, passing the business of redemption over to others. It is used by the doyens of the external domain, priest and prince, as their business, in which they may take pride and profit. For blame is a wild card until it is called. The act of naming and blaming a witch becomes a cry de profundis, an abomination of evil invoking the taboo of betrayal within a circle of trust which has been crossed. Where an ancestor is blamed, the circle remains a private one, the grievance is just, only the plaintiff’s own moral strategies want mending. Otherwise the spectre of witchcraft is raised up, and moral indignation requires a traitor be named in open forum. Always with the breach of taboo the air is made heavy with the presence of Fate. Comforting doubt is swept aside. For Kinga and Konde the idea of treachery within an inner circle, since it is a formula for sanctioning by riddance, matches a localist organization of life. Witchcraft ‘on the ground’ is usually settled by exile, sometimes more than half voluntary, and cured by time with a morsel of good sense.

429

Through time and history translocal organization, as it grows, will move the sanctioning procedures by the zigzag of trial and confirmation toward the idea of crime against society—a cataloguing of culpable acts, with moral sanctions meant to measure and punish infraction. The word ‘punish’ in the common tongue may still carry its latin meaning of doing vengeance, but it is hard to show that Western procedural law in practice goes farther than fixing responsibility and matching it with pragmatic measures of dissuasion. Taboo is submerged with the sanctions of constituted authority, even dismissed as insubstantial. It doesn’t match the new ‘rational’ politics of procedural intervention. In yesterday’s terms, we are looking at the long transformation ‘from sacred to secular’. Solemn rituals no longer celebrate and ‘teach’ the taboos by which a people must live. The spirit of ‘substantive law’ grows with the practicality of enforcement. But a ‘least poity’ that depends on enforcement is doomed. Taboo remains to guard that aspect of human cultures which must be submerged to persist. The spirit of this ‘custom law’ is not pragmatic, not even corrective unless in the absolute sense that fundamentalism has to be.

430

431

FEAR & BLAME, SOURCE NOTES

Source Notes

Preface 4 A Kinga priest calls Nyakyusa priests ‘all witches’: Monica Wilson (1959:39). 7 General statements I make about the Nyakyusa are mainly based on Monica Wilson’s works; Stirnimann is my sole sound source for the Pangwa. Where chapter and verse are required to identify documentary sources, I place a reference symbol (‡) at paragraph’s end. For uninterrupted reading these may be ignored. I use the same symbol doubled (‡‡) to indicate comments in elaboration of my argument or to offer a further interpretive point. This is to unburden the text for readers following the main line of argument.

Chapter One 9 M. Wilson 1958 12 Violence at Nyakyusa funerals: M. Wilson (1957: 21-25). Nyakyusa political identities: P. Weber (1998: 70-2, 242). 14 Intra-regional comparisons: Eggan (1954). 19 P. Weber on Nyakyusa sectionalism: (1998: 70-2).

Chapter Two 21 The major source for the culture of amity among Bena is the Cullwicks’ monograph (1935) on the riverine group, displaced by late 19thC. wars. Other learned studies were done at survey range, not close up. Njombe Boma archives are concerned with social and economic mapping and offer little in the way of a cultural portrait. I have opted to write about the riverine group as ‘Bena’ on the assump-

431

tion that the Culwicks’ warm portrait is not deeply misleading for other Bena groups. 23 My source for these dates is Shorter (1972: 155-62). 24-5 Political history of the Hehe: Alison Redmayne (1968). 28-9 Aidan Southall’s ‘segmentary state’: (1954, 1988). 29 The relationship people-priest-ruler is triangular and doesn’t sandwich the priest in the way a puppet chief was sandwiched in colonial days. It is not irrelevant that priests come in choirs not singly. 32 For the source on ‘old rainmakers’: Iliffe (1979: 121). 38-9 For the Coming Out tradition the prime source we have from the ethnographers is Godfrey Wilson (1951: 280). For the developed view that the Nyakyusa Coming Out is to be seen as a handing over of power from one generation to the next, and implicitly lilnked to ‘divine kingship’ see Monica Wilson (1959a: 49-57 & 1959b) and other indexed references in her several volumes. In chapters which follow I shall be rethinking the institution, taking up her account in some detail, as well as some published critiques. Just here I am concerned to set the terms of a puzzle, not to solve it. My position will be that power is devolved but scarcely given over. I think the root misunderstanding derives from overestimating the corporateness of a Nyakyusa chiefdom before the pax. For detailed source references I suggest looking at the later notes, particularly for Chapter Nine. 39 The history of the Kyelelo the Cruel and his wars is reviewed in Four Realms. 41 ‘Skin changing’ by heirs, adopting the paternal identity and personage, was common, and among Nyakyusa did not even require dropping one’s own signature and status in another village community. An important consideration is that the heir would seldom be of the deceased’s home village, where the inherited identity must of course remain. The duality of kinship vs. friendship is showing here. 42 On the nature of ‘external’ relations it may be noted that ‘in Nyakyusa thought the back is associated with sexual activity’ (Monica Wilson 1977: 101). 62 For details on the Kyungu throne I have followed Godfrey Wilson (1939). The pondering of motive is my own. 65 For a review of witchcraft cases in Selya district: Monica Wilson (1951: Select Document 8).

432

Chapter Three 72 For ‘level one’ conditions in the region: Park (1988a: 130). 75 For the contact side of Hehe history: Iliffe (1979: 107-16). 76 For the histrionics of Kinga warfare: Park (1994). For Bena ‘sporting’ notions about war: Culwick & Culwick (1935: 27). 81 Since Stirnimann’s two monographs (1976, 1979) on the Pangwa are not translated but feature largely in this work, I offer my renderings of important passages at length in later chapters. Gulliver’s work is Neighbours and Networks. Because I use the Pangwa and Ndendeuli cases to suggest a way of life more or less congruent with the acephalous style of the region antecedent to the rise of protostate politics, a warning is wanted. Just as are the Pangwa Stirnimann reconstructs for us, the Ndendeuli of Gulliver’s book are veterans of aborted history. Pangwa are ‘returned refugees’ from displacement. They had practised a slow-shifting agriculture in the mountain slopes south of the Kinga. But their country had been razed by Ngoni bands, and the earlier community traditions Stirnimann recorded were but partially re-established on their return to uPangwa. This ended an extended generation of ‘captivity’ in a neighbouring but foreign country. In another part of that same country, Ndendeuli were not so much displaced as forcibly assimilated by the Ngoni who settled in their region. Some were complicit in Ngoni raiding during the later nineteenth century, and almost all were converted to Islam. That the majority held on to their ethnic identity is clear on the record. With the coming of the colonial peace after the Maji Maji catastrophe, Pangwa began moving north to their old country, and Ndendeuli moved in pioneering fashion eastward into unpopulated miombo woodlands, where they eventually achieved ‘tribal’ status on their own from the British. In citing their case I am insisting that because of their movement away from towns and authority into a seclusive life in scattered, shifting hamlets, the Ndendeuli socioeconomic adjustment speaks to a more ancient past than the Pangwa. Particular sources are detailed in later Source Notes and References. 83 These field records of Godfrey Wilson are cited by Monica Wilson (1959a: 140-41) 84-5 The citations are from Monica Wilson. Hiding by night to foil a witch:(1959a: 125). Moving away to cool out a witch hunt: (1951: 102-3).

433

85 On the ‘constitutional’ character of chiefship: Godfrey Wilson (1937: 22). But I find charismatic heroics were no less important than such ‘committee work’ before the pax. 85-6 Charsley (1969) wants to call Konde hereditary rulers by the title ‘prince’ in lieu of the Wilsons’ ‘chief’. His thought is that these are men given access to power as a (conditional) birthright but must be cast (in my terms) as ‘political entrepreneurs’. They become powerful in proportion as they take up the challenge of the Coming Out. With Kasitile, I want to note further that they exercise a freedom from mystical fear no villager (not even the head) enjoys. This said, it remains true as Charsley (1969: 69-70) emphasizes that the political standing of a chief was always precarious and wanted astute handling. Again, the corporate standing of a chiefdom before the pax needs be doubted. In my eclectic usage, a man may be a ‘prince’ so far as he embodies a birthright to rule, and a ‘chief’ so far as he has a political establishment under control. 86-7 The commoner senior priest: Monica Wilson (1959a: 52). Shall we think of Mwambuputa (a.k.a. Mwafungo) as a demagog? The ‘extended case’ pieces together evidence from his various appearances in Monica Wilson’s corpus. 87-8 The cited cases of Nsekela & Mwambuputa: Monica Wilson (1951: 114-15 & 230). 88-9 Telling circumstances of Mwambuputa’s return: Monica Wilson (1951: 188-90). 89-90 See the same source for this political wrangle. While Monica Wilson thinks Mwaipopo is motivated by fear of witchcraft (in form of ‘the breath of men’) a more economical hypothesis would be appeasement politics. 90 Things said at Mwambuputa’s installation ceremony: Monica Wilson (1951: 272-4). 91 Mwambuputa now as the redoubtably established politician: Monica Wilson (1951: 111-12). 92-3 For Mwambuputa’s elderly wisdom on the role of repression in witchcraft: Monica Wilson (1951: 100).

434

Chapter Four 98 Main source for Burmese nats: Melford Spiro (1967). 103 The regional political culture is discussed against a putative time-scheme in Park (1988). 105 Godfrey Wilson’s field records and interview with Kasitile are cited from Monica Wilson (1959a: 115-17). 107 The functional link between chiefdom and ‘redistribution’ was the theme of Elman Service (1971). I discuss the logics in Four Realms , the second section of Chapter Two. 115 Monica Wilson gives us ‘the angry parent who rouses the shades to action by muttering over the fire’. She holds ‘the genius of Nyakyusa religion is an awareness of the corrupting power of anger in the heart’ (1959a: 161). 116 The association of youthful ‘great expectations’ with cattle-fancying in East Africa is well documented. Where you find one of the two you find the other. But goats are made of less inspiring stuff. The difference is best explained as the effect of distinct natural symbols: Mary Douglas (1973). 118 Neighbours more troublesome than shades: (1957: 221). 127 (1959a: 155). Probably a local equivalent to this mystical ‘rhetoric of witchcraft’ haunts and informs chiefly redistributional politics everywhere. Simply to dismiss Wilson’s ‘balance of power’ as a witless borrowing from yesterday’s political theory raises witlessness to the power of two. 132 The cheating of a descent line in the Western realm is recounted in Four Realms in the final section of Chapter Five. 133 Whereas Mahanzi informants led Peter Weber to see the female spirit medium as a Kinga institution, I know of only one such figure (Hikadiseku) in the social memory. Her attachment to the ruling house at Ukwama was an entrepreneurial achievement of her own, fed with gratuitous flattery and ‘percentage’ payoffs in meat to the crown. In the far more entrepreneurial Rift valley, the prevalence of private operators in the ritual theatre was more substantial. In part this can be attributed to the easy cultural commerce of Konde southward, with stateless Central African peoples. Some German missionary records confirm the interest of chiefly forces in Selya in suppressing such mystical entrepreneurship. Their trouble in doing so is prime evidence for my position that state-building had got farther

435

in the less turbulent mountain slopes of uKinga. On the basis of paxperiod politics in the region, that notion would not have arisen. 138 I will later argue that belief in the malevolent power of intimates operates as a negative sanction of violence within an inner circle. It is thus part of the taboo of personal betrayal. 139 The dark mood of Ngonde ghosts: Godfrey Wilson (1939: 32). The Kasitile tale runs through Monica Wilson’s monographs, in particular Communal Rituals . Follow its index entries under his name. 142 Tunginiye’s bewitching was recorded for me by him in his Swahili notebook. My translation [III ST 36ff] .

Chapter Five 181 The linguistic evidence (Nurse 1988) calls for a time scheme of close to two millennia to account for the pattern of differentiation in situ , and shows two culturally discrete groups dividing the Sowetan region. This implies the long-recognized political affinities which cross the divide (the great escarpment which separates the Kinga from the Konde world) must have developed in recent centuries. On linguistic evidence, Nurse’s ‘Southern Highlands’ cluster (including Hehe, Bena, Sangu, Pangwa, and Kinga) has had a history distinct from that of the Konde cluster for well over a millennium. The time-depth of oral tradition is slight by comparison, at least until a particular myth can be shown to have held a hidden clue. On the other hand, the combination of lexical and phonological evidence does let us infer the degree of ethnic isolation as between two related tongues with measurable confidence. I am thus indebted to Nurse’s work for showing me that the institutional connexions between Kinga and the Konde peoples must be judged as coeval with the rise of the Sowetan political archipelago. For one thing, this does breathe a bit of life into Kinga/Konde tales about the ‘coming of princely rule’ to the Rift valley. The princes are said to have descended from Kinga/Mahanzi as long ago as a.d. 1600. My argument assumes the absence of any serious flow of influence eastward or westward prior to the (militarism of the) protostate era. Still, I take it as probable that a mix of amity with kinship, as well as standing bilateral rules of ‘incest avoidance’, prevailed among the acephalous local groups on both sides of the divide in early times. Partly, this is because I see the earliest Bantu settlements in the region as cumulative not massive. Early Selya like early Kinga communities would have been assimilating and acculturating migratory drift from whatever direction through their centuries of

436

gradually developing intensive agriculture. In the regional culture as we still know it, the scion of a first settler, the clearer of the land, is always imagined to ‘own’ it. Newcomers and their heirs are beholden. (See Park 1988 & Peter Weber 1998: ch. 8) My thinking owes a rather different debt to the recent fieldwork of the ethnohistorian Peter Weber (1998). His lively presentation of the diversity of cults and the rich load of local attachments to the past, all still to be met in the Selya of the 1990s, would banish any doubts about the energy resources of Nyakyusa culture and its incorporative flair. Considering the revival and survival of tradition after the constitutional changes of colonialism and Tanzanian independence, I suggest we have to deal with what Spicer (1971) called a ‘persistent cultural identity system’. The persistent high morale of Kinga young men and women in the 1960s impressed me in like fashion. As I turn at this point in my argument to consider some of Monica Wilson’s critics, I’ll mention here my own reservations about her work. I think she didn’t scrutinize the gloss of ‘corporateness’ which the Nyakyusa nation had been given under British indirect rule. Her chiefs are not rivals forever out to trump each other in an open political arena. An anthropologist’s mandate then was to observe and describe, not to try reconstructing the unprotected tradition of former times. The premise was, once you have recorded the tradition you can add a chapter on ‘changes under the pax’. I went into the field myself in 1961 under the same mandate. It was understood that you can ‘collect’ history orally taken, but you knew as a good scientist you could not bring it to life. If I were asked to state the genius of Nyakyusa statecraft, I would have to say it was an ingrained virtuosity in political entrepreneurship. Interpreting their national vitality and collective weight in ‘tribal’ or ‘predatory’ terms is off the mark. But taking account of the pre-pax conditions, under which a people of such and such tradition must have coped with, is a step woard the ultimate scientific desideratum, which is to add this people’s story to the shelf of human history. As for the Konde story: the Coming Out speaks clearly of the lively regenerative faculties inherent to their culture, which set their historic pattern sharply apart from Sangu or Hehe expansionism. 181-2 The main source for Charsley is Princes (1969). 183 Following is Fülleborn’s comment on the major uprising of Nyakyusa against the Germans, which occurred at Mwaya in 1897 and reflected for him ‘the carefree character of the people’.

437

When [Captain] von Elpons and I a few weeks after the battle were visiting Mwaya and they were informed that peace now reigned, their response was cheerful and jocular, they showed us their wounds, talked laughingly about the massive effect of the breech-loader, telling von Elpons quite naively that they had wanted to take him alive, so as to undress him and put Nyakyusa-style brass rings around him, as he had teased some of them about their nakedness. They behaved themselves overall as harmlessly cheerful as though they had been in league with us against a third party, even though almost all of them had lost family and friends to our soldiers’ bullets, and the respected chief Mwakalinga had fallen in the combat. [1906: 292, my translation]

185 Citation from Charsley (1974: 422). 186 Citation from Monica Wilson (1951: 80). 187 Citation from Abraham Kaplan (1964: 335). 190-1 Fülleborn: (1902: 8) as cited in Weber (1998: 169n) 191 Monica Wilson 1959a: 83. For the autocthone’s social identity see Peter Weber (1998: ch. 8). 191-2 The distinction of structure ‘on the ground’ from ‘in the head’ is at home as ‘the system on the ground’ and the ‘system of ideas’ in Edmund Leach (1964: xiii). The original 1954 printing did not contain this illuminating introductory note. For my take see ‘Rethinking Highland Burma’ in Hard Cases. 196 Nyakyusa insistence on ‘father-in-law avoidance’ for women offers a nice example of the nature of taboo. In the traditional setting the taboo worked an important structural insulation, avoiding occasions for violence between father/uncle and son, something a law as such could not have done. But taboos work precisely because they can’t be rationally questioned. The subject is revisited in Chapter Seven. 204 Arena politics is learnedly discussed by Marc Swartz (1968: 8-18). What matters for understanding the turbulent Nyakyusa scene is the agonistic interplay of tangential interests, set in motion as claim meets counter-claim outside the controlled political space of local (village-within-chiefdom) authorities. In a pedestrian community any ‘planning in foresight’ is likely to have been done under four eyes, not more. Confrontations can be easy to evade. Much depends on chance meetings and propinquity. Variant versions of the relevant facts are legion. The pragmatism a player goes in with is frequently lost in the outcome. The flow of events is not aimless but stochastic to the point that aims are smothered in their very abun-

438

dance, or lost. Finality is more often expressed than experienced. By comparison, Kinga politics is distinctly consensual. 208 Apart from sources already cited, there is a useful account of traditional Hehe political organization in Brown & Hutt (1935: 2082). The best source I know for background on the Sangu nation is Aylward Shorter’s Chiefship (1972: 227-263). An analytical account for the region as a whole is to be found in Park (1988: 156-196).

Chapter Six 210 The citation is from Weichert’s memoirs (1928: 59-60) and is given in my translation. I believe he wants praise among his fellows for deciding not to overlook the obvious. 211 See PeterWeber (1998: 206-223). 212-13 Godfrey Lienhardt (1961: 30-33) introduced the ‘free divinity’ to distinguish proprietary spirits (e.g. ‘clan spirits’ or as here, particular ‘ancestral chiefs’) from spirits with a universal clientele. 213 Weber’s Map III: (1998: 108). 215 Monica Wilson’s rejoinder to critics: (1975). 216 The major sources for Marcia Wright’s thesis of ‘rival cults’ are her book German Missions (1971) and an article, ‘Nyakyusa Cults’ (1972). Sources cited on shrines: (1972: 164) & (1971: 55). 216-17 Sources on ‘cults’: (1971: 54-5). See also ‘Nyakyusa cults and politics’ (1972) throughout. How misleading the rhetorical effect of ‘cult’ can be is evident when Peter Weber (1988: 257) convinces himself that the itinerant herdsman’s borrowed cattle and wives comprise a ‘cult of afflictioin’. The matter can best be discussed when the regional cum political context of ‘cult centres’ has been made clear. For George Simmel: (1955:111). 217 Source for the citation: Wright (1972: 167). A different narrative version of the same episode: Charsley (1969): 14). 218 Marcia Wright: (1972: 167). 266 The word ‘cult’ can hardly be banished despite its toogenerous connotations. But it remains a term which needs grounding in context. Is one single agency entailed or more? How are clients recruited and their affiliation nurtured ? Is there a track record of historical importance? So far as the revistionists’ use of ‘cult’ may seem derivative of ethnographic work in Zambia, Malawi, and

439

Zimbabwe the problem becomes one of subtle shifts in social and cultural context. Where ‘kinship is politics’ and cults have no relation to political alignment, conceiving of a cult’s clientage in terms of congregational rivalries might not badly miscue us even if it is off the mark. Allegiance, which normally goes with such terms, doesn’t apply in animisms, where individual clients shop around for cures, and collective (‘territorial’) tribulations are associated with ungenerous divine anger. If an African ‘cult of affliction’ were in question (as Peter Weber supposes in the Mbasi incidents) we would have to see some appropriate ‘sisterly’ signs. Paying clients are not cult members, nor are boughten wives. In the Ngoni conquest region of eastsern Zambia, Rau (1979) shows that the Chisumphi cults were informed by a local history of government by pop-up leaders, ‘owners of the land’ whose status was sacerdotally sanctioned. This has its parallels in the peripheral communities affected by the Konde protostate process, but the political situation is quite other. Schoffeleers (1979a) shows that in Malawi cults related to those of eastern Zambia operated in very loose relation to more properly ‘political’ structures. The extreme localism of the ‘owner of the land’ ideology was obviously anathema to the chiefly system developing in uNyakyusa. My position is that a political system built on two tiers of executive authority is scarcely possible without its reducing the pre-political institutions of government to marginal status. We have seen that the secular authority system controls access to ‘witch finding’ in the Selya generation of elders studied by the Wilsons. If people don’t seek elsewhere for protection from witchcraft, the secular system of authority seems to be intact. I have described the Sanga system as dualism in a reasonably steady state. Konde turbulence suggests pluralism, with secular affairs fairly well in hand where ceremonial chiefship is strongly developed, but a decentralized and spontaneous organization of divination and remedy would have been kept alive by the kind of ecological disasters which followed one another during the early mission years. What we have to explore is the degree to which the pluralistic ideological bent of the Nyakyusa made itself felt through the traditionally given 30-year centripetal-centrifugal periodicity of chiefly politics there. My arithmetic has suggested that a portion of those who ‘come out’ would normally get lost to the chiefdom of orientation. We have no careful studies of the ‘independent’ small ‘chiefdoms’ in the colonial registers. In some degree, at least, they must parallel the persistent bush settlements in the periphery of a Sanga court. By looking at this as exploring a Konde ‘bush culture’ surviving in the wake of a chiefly ‘court culture’ I’m aware I risk imposing Kinga structural

440

form on a neighbour people. But a case lies for this: the proto-state process of its very nature seems to draw energy from this duality, and ought not to be conceived as merely ‘transitional’ to the formation of a ‘genuine’ form of the state. This notion bears in turn on the question of time-depth for the two protostates, favouring the structural credibility (excepting particulars) of long kinglists, with their suggestion of a system-maintenance ethic. The political balance in these two protostates is not predicated on the triumphalist ethos of the Sangu or Hehe despotisms. An antipolitan ethos takes its place, favouring conditional loyalty and exploiting the politics of (prescriptive) male peer-bonding. 230 Citation from Monica Wilson (1959a: 157). 230-1 Hübner as cited by Fülleborn (1906-317-18). My translation. 232 Citation from Monica Wilson(1959a: 159). 233 Peter Weber, p.c. 02/99. Second citation: (1998: 221). My translation. 237-38 A major source for ‘territorial cults’ south of the Corridor region is Schoffeleers (ed.) 1978. 240-41 Mpezeni’s movements: W. E. Rau (1979:131). Mbelwa’s kingdom: Margaret Read (1956: 37-44). The following citation from the same work, pg. 163. Citation from Rau: (1979: 135). 243 Monica Wilson (1959a: 164). 2444 Joseph Campbell (1968). I don’t find a place in the Malatan cosmos for the exiled hero-divinity when he would return. The nearest place is the ghostly stand, which no one can envy. There is noprovision for free-living sky gods, only earth and its underworld. Ghosts rise up nightly from the ground in uPangwa but exercise no dominion there. Old men in uNyakyusa feel they are often in touch with male ancestors. Prophets act as information brokers, but it seems the news is always bad. 245-46 The narrative Peter Weber got for ‘Swebe’ is no match (1998: 100, 108, 111). But the linkage between hero-name and heromyth in the Konde region is notoriously contingent. 249 For the mythological charter coded into chiefly genealogies see Monica Wilson (1959a: 7-16) and for her fuller account of Kikungubeja talking of a Kinga ‘charter’ see the same source, pp. 8-9.

441

252 The pertinent material from mission records lis to be found in Charsley (1969: 12-15). 253 My general discussion of turbulence: (1974a: 62-74). Banyankole succession wars: Oberg (1940: 157-61). 255 I reason thus: If Kasitile and his priestly colleagues are in possession of rainstones, they can’t be see as structural opponents of the chiefs they advise. The powers of nature are known to underline their words. If this ‘chiefly’ stand of priests complains that commoner priests (as ‘retired’ village headmen) do not concede rank, the complaint of a ‘dirt farmer’ cited earlier (pg. 191) shows him critical of the chiefly system and the institutions supporting it. He wants to identify with ancient ‘owners of the soil’. The irony is that by semantically aligning himself in opposition to chiefly priests he has separated himself from the very agencies which continue the ritual practice of ancient ‘owners of the land’. He can feel damned both ways. This doesn’t show there is no ‘structural resentment’ in the Nyakyusa community. Both Peter Weber and Monica Wilson have found it alive and well right through the twentieth century. Our judgement has to be that the pluralistic structure of the larger Konde community diffuses and displaces resentment, much in the manner we know in more complex (urban) societies. 260-61 Contingent vs. axiomatic personal ties: Park (1974a: 203-19). 262 The reference to Ruth Benedict’s Patterns of Culture (1934) is meant to remind that her holistic Nietshean distinctions don’t match a ‘pluralistic’ culture. Of the three Malatan communities only the Pangwa are not pluralistic. Is it not increasing complexities of the numinous kind which give rise to a politics of fear? Monica Wilson on handling anger: (1957: 227).

Chapter Seven 264-5 Meyer Fortes (1958) was careful to show that the same human experience (for example, passage from one life stage to another) usually has context in both internal and external institutional domains. From the viewpoint of individual development, life within the family of orientation proceeds by a widening out of personal involvement outside that in-group circle. Much can happen to the psyche.

442

272 For Balkan familism see Sanders (1949) and J. K. Campbell (1964). For ‘amoral familism’ see Banfield (1958). 274 Michael Polanyi (1951: 157). Marshall Sahlins (1976). 275 The information available to us on the domestic life of the Avongara courts among the Zande is brought together by the Zande corpus of Evans-Pritchard. On the matter of enjoining rather than prohibiting incest the key source is ‘The Zande State’ (1965: 102-32). I take up the structural (political) puzzle in some detail in the final chapter. It is important that we have to deal here with a ‘composite’ or ‘conglomerate’ assortment of clans and ethnic communities—a vast league of conquest states ruled by an elite (the Avongara) within a clan-within-a-clan. So we have cultural pluralism as a setting, and a culture of inbreeding in the segregated ruling courts. This is not standard stuff for ethnography, either as to scale or complexity. But it is wise, since the culture context will be germane, to mention here the further evidence Evans-Pritchard presents on the unbridled fascination Zande commoners everywhere will show for fearless sexual exploits. Here is an apposite (and unusually modest) example, offered by an unnamed Zande informant on his schooling in the matter of incest: When a boy reaches puberty he may take his sister and with her build their little hut near his mother’s home and go into it with his sister and lay her down and get on top of her—and they copulate. His father then begins to keep a watch on them to catch them at this and seizes him and gives him a good hiding and asks him what he means by going after his sister, she is his sister, has he seen people going to bed with their sisters? Then he is afraid. He keeps a look-out for his father and when his father is away he again takes his sister and they go and hide in the bush to copulate...So people say that children are like dogs, for a boy will go after his own sister. After they have been stupid for a time, when they grow up they get a sense of shame...Since I have grown up, should I see my sister’s nakedness I would compensate her [with a gift] for that is just children’s play, I have long forgotten it; since I learnt sense I left off childish things; what I used to do with my sister when we were small we have given up. [Evans-Pritchard 1974: 107-8]

281-82 It was Franz Steiner (1956: 21) who noticed that “taboo is an element of all those situations in which attitudes to values are expressed in terms of danger behaviour.” Evans-Pritchard’s Zande informant (cited just above) is not atypical in having ‘missed’ the dangermessages (if any) in early years. We see him adopting danger strategies in adult life—a little ritual act of expiation as a ‘cooling out’ response to her nakedness. It is when the sense of danger becomes ego-involved that ‘taboo’ has made its appearance. In societies which make much of sexual

443

modesty in early years the sex-avoidance response is more likely to be triggered earlier and is more easily extended beyond ‘incest’ to cross-sex peers generally. Kinga allow the condition of danger-avoidance to stay in place into adulthood. 282 Oversocialized people are, of course, fictions of the would-be scientific observer. As a general fault in social anthropology it must be said that every model of a social system comprises such a fiction. I wish it to be clear that I find Stirnimann’s ‘fictions’ as close as any ethnographer’s to a ‘system on the ground’. This is because of Pangwa ritualism, which depends on popular not professional theatre and extends right into the bosom of the family, allowing less latitude in the pursuit of individual moral strategies. What at first seems implausible is the model itself. We are attuned to a more ‘pragmatic’ style in the description of a social system. Stirnimann’s Pangwa model is counter-pragmatic. It seems to demand ‘model behaviour’, as though the ‘system on the ground’ could be a clone of the ‘system in the mind’. What becomes apparent as you read is that in this acephalous society popular ritual serves the functions of social control we usually conceive as ‘political’. Latitude is not disallowed but significantly controlled by taboo. 290 Murdock (1949: 62-4). 296 Legitimizing identity by putting a ‘genuine’ stamp on what was at best ‘virtual kinship’ illustrates the way the ‘system in mind’ can be affected by praxis. It is not that mental modeling picks up the system ‘on the ground’. Rather say the mind sees ‘how things ought to be’ and how, by the simple pragmatic move of ‘labeling’, things can be made right. The full importance to acephalous societies of such ‘creative kinship’ is explored in Gulliver’s close study of Ndendeuli moral strategies, Neighbours and Networks (1971). In the two Malatan protostates the problem of using residential freedom as a way of escaping a destructive personal involvement is facilitated by the importance of common political loyalties. Praxis under the antipolitan ethic is that a man moves from one military camp to another. Amity (prescriptive male bonding) is the basis of this social contract. But for Nyakyusa the ancillary ploy of ‘creative kinship’ is obviated by a dependence (for inheritance) on actual kinship, systematically maintained. 297 Rite of installation for a new household: Stirnimann [II 217-19]. 298-9 On ‘group-wife adultery’ see Goody (1956). 300 Structure and sentiment are intricately braided-in to give firm structure to the ‘internal domain’. This should be true of any stable human community—it is a condition of such stability (as Aristotle had it) that patterns of motivation should match the patterns of behaviour expected of good citizens. Nyakyusa avoidance drills use the phenomenon of social danger to define boundaries and embed the individual identities those

444

boundaries imply. In this case, festive license at funerals also serves to confirm and give form to a young man’s kinship identity. Avoidance taboos create the possibility of license, in the same sense that any systematic restraints create the valued possibility of freedom from them. Egoinvolvement in the acting out of kinship roles is governed both by social danger and by the kind of merrymaking which courts it. For Nyakyusa men kinship identities are far more important than for Kinga men, yet they are marginal to daily experience in the peer village. Accordingly, Nyakyusa kinship rituals have a more theatrical or ceremonial style, imparting structural weight to kinship identity—the political counterbalance in the men’s culture to the peer-village experience. Any kinship system adapted to rampant polygyny must cope with the fact that father-son relations are diluted by the sheer numbers of a man’s sons. Kinga male moral strategies are not often built around unilineal kinship—there is no solidary kin group beyond the family and rather limited solidarity within it. Yet I will insist that individual identity, as it is reflected in maximal-claim networks and the depth of mutual trust they hold ready , is as much a matter of kinship for Kinga as for Konde youths. It is in the number of individuals in a given youth’s kin network that the two cultures most clearly contrast. But the importance of peers in the socializatioin process of both genders in uKinga certainly means that amity—sentiment and structure braided together— will prevail there as the basis of political trust. For uNyakyusa it appears that amity is more deeply laced with rivalry. 305-6 Monica Wilson (1977: 86, 196n.). 310 For the distinction of the kinds of structural tie (axiomatic and contingent) see The Idea of Social Structure (1974a: 204). In other terms, axiomatic ties are governed by taboo (incest avoidance, age respect, loyalty, trust) where contractual ties are governed by mutually accepted terms of agreement. Axiomatic ties are those considered immune to final broach, contingent ties are those considered sensitive to the degree and mood of reciprocal affective involvement. The structural weight of any tie is indexed by the danger behaviour it commonly evokes.

Chapter Eight 319 Compare Henry A. Murray (1938: 164-73) on ‘the need for achievement’ as one of the phenomena he calls ‘manifest needs’ with the appearance of the concept in David McClelland’s Achieving Society (1961‘). 320 Murray (1938: 45). 321 For dating the past I follow my SUGIA article (1988).

445

328-29 For the involuted, formal structuring of the internal domain in traditional Pangwa communities see especially Stirnimann (1979: 64-5 & 108-113). 334 Stirnimann (1979: 109-10). 344 Cf. ‘Bridewealth revisited’ (Park 1994a). The social system prevailing in 1960 had a novel structure ‘on the ground’ based on the articulation of bridewealth debts and credits which managed to keep most of the money from migrant labour at home, managed by elders. It was a money game whose structural meaning was nowhere represented by the ‘system in mind’. It is hardly enough to know that a son must ‘fear his father’ when you don’t know how the structural weight of the tie is produced in action. This ‘fear’ is articulated in the most common form of danger behaviour, accommodation. 348-49 For the uses of ‘proximal distance’ in Pangwa fatherson relations see Stirnimann [II 80].

Chapter Nine 360 Citation from Monica Wilson (1950: 129). 362 Ordeals are ‘domestic’ only in the political sense. Where decisions are keyed to the outcome of mortal combat between heroes, as in medieval Europe or the nineteenth-century Kinga-Konde war pattern, they are ‘interdomestic’. That is, the rules of the game have to be the same on both sides. The reason Kinga called Kyelelo ‘the Cruel’ though he was a very popular leader, is that his wars were not done on a game-board but in earnest. The ordeal has a bad reputation in the folk-ethnography of Western Academe. Seen as a format for dealing with trouble in a controlled form of public theatre, it falls in line with other forms of divination as the use of ‘indirect action’ or ‘procedural intervention’ to avoid physical combat as the basis for moral decision. I have elsewhere shown the relevance to ethnography of Ortega’s distinction (1932) between direct and indirect action. Indirect action is ‘lawyerly’ (‘formal’, ‘controlled’, ‘rational’) where direct action is ‘passionate’ (‘impulsive’, ‘rash’, ‘non-rational’). For the sublimation of danger behaviour in a theatrical war pattern common to Kinga and Konde traditions, see Park (1994b). 367-68 Southall (1988). 369 On Nyakyusa chiefs see Monica Wilson (B 23-4) & Godfrey Wilson (A 286).

446

375 The two sources on Nyakyusa chiefhood and priesthood are Monica Wilson (B 27 & B 25). 376 On the Konde chief’s preoccupation with external affairs see Godfrey Wilson (A 288-89).

Chapter Ten 381 Franz Steiner (1956). 392 I have ‘put to proof’ this model of political life as a competitive game in a rethinking of Edmund Leach’s Highland Burma. He deals with an unusually pluralistic society by inventing an ‘oscillating equilibrium’ to account for what I class as turbulence. Social turbulence is, by definition, always between the poles of structure and chaos. Action is so far improvisational that it can’t be dealt with in the conventional terms of structural analysis. In the things that count, individuals in the most ‘traditional’ societies have choice. The weak forces which ultimately determine the course of social change may be as minuscule as the choice between biting the bullet and slacking off. My publications on the themes turbulence and political entrepreneurship are out of print and available only in electronic form. The books are, respectively, Idea and Hard Cases . The out-of-print version of the latter was fancifully called The Flying Armchair —see (1990a). 404 The structural rule Aristotle offered was that the citizen should be moulded to suit the form of government under which he lives. He taught that ‘customary laws’ (taboos) have more weight—concern more important matters—than written (positive) laws; and where a strong ruler might have to protect us from the law, taboo could be needed to protect us from him. ( Politics , Book III). 404-5 On the ‘origin of the incest taboos’ I say only that it is one feature of the evolution of a domestic infrastructure based on the self-reliance of small groups articulated about the business of marriage and family. What is peculiarly human can be compared, of course, to the nesting pattern in flocking birds—the devolution to small-group self-reliance in the breeding season. Crows, so famous for their flocking, actually support themselves in small foraging groups. The special feature of human populations is the long and arduous work of getting offspring safely off, each to a nest of their own. Ethnographers find that populations organized as interdomestic networks (almost no division of labour above the household level) can spend the better part of a typical lifespan in re-nesting the offspring of a

447

marriage. If this work is to be accomplished predictably by self-reliant primary groups, and in such fashion that offspring always have a unique and legitimate social identity with the rights and responsibilities adhering thereto, the non-contractual mechanisms for accomplishing this work must comprise the ground-zero structure of the society in question. As I suggested in Chapter Seven (pg. 275 & source note) the absence of incest taboos in a Zande royal court reflect the simple political truth that the King does not cherish self-reliance in his hundreds of offspring. 410-11 There are two points of special concern in thinking about ‘corporate’ varieties of the Least Polity Rule. One is that, though the inclusive ‘least polity’ is represented in the larger community’s policy decisions by only one voice, there is likely to be internal consensus behind that voice when it will speak with assurance. This is the principle of the consensual filter operating in face-to-face thinking. The other point of concern is allocating political voice to the noncorporate segments (families) of a corporate bloc. The family cells of an Iban longhouse are perishable but may achieve considerable lineal continuity. The longhouse is, as a fixed local polity, corporate. Nyakyusa chiefdoms, on the other hand, are not corporate, though again they may contrive to last under one chiefly identity for several generations. At the ‘least polity’ level, in the Nyakyusa scheme voice is tied to marriage—and marriage to voice. Delayed marriage means delayed voice and prolonged dependency. As for a consensual filter, rampant polygyny cuts most young women out of the political loop entirely, and male bachelor status, where there is scant facework between sons and fathers, leaves the unmarried individual outside the consensual fold. This is manifest in the exclusion of sons from inheritance where a father’s brother survives. Women are formally outside the active political fold in Kinga communities, but the highest throne has been held by a woman, and a woman like Hikadiseku had much influence on the High Prince. For Kinga the two gender worlds are far better segregated than in uNyakyusa. Having voice within a household is obviously gender-specific in both cases. But Kinga marital households are so fragmented as to leave women to cope with women, men with men. Women have by right full voice in agricultural matters, and a status system of their own. It is only in holding the hearth-position in the least polity that they play a prime role in ground-zero politics. 411 It is crucially important for reading the Least Polity model, to see that relaxing moral prejudice against incest leads inexorably to crumbling the borders which separate one least polity from its neigh-

448

bours. Evans-Pritchard’s Zande (commoner) informant on the topic makes clear that ‘incest discipline’ within a family may be lax. (See pg. 275 and source note.) But the wisdom in anthropology has tended to play on the internal domain as a microcosm, looking for rifts within the microcosm as the ‘bad news’ which might account for the taboo. In that case, Zande children could regularly copulate in secret, and only a cranky father would be giving them trouble. The Zande case suggests the real problem is more likely to be the continuation of an incestuous involvement after an offspring has been re-nested elsewhere. If we can’t explain the maintenance of the taboo as a crucial feature of the social system, we can’t explain the (continued) existence of the taboo. It is most cleanly seen as a necessary if not sufficient feature of a generic kinship system. Without the incest taboo, a least polity becomes a rogue, unable to pass on to a new generation the unambiguous social identities which are the result of keeping social and genetic (natural) models of parenthood congruent. 413-14 The main Tallensi volumes: Fortes (1945 & 1949). 414-15 See Evans-Pritchard (1965 & 1974) and earlier source notes on the Azande (for pp. 275, 404-5, 411). 424 In my own terms, the stability of least polities everywhre is formed around the internal production of durable maximal-claim ties. Sexual involvements entail voluntarism to the point of forcing a definition of the tie as contingent. If children are drawn into this frame while they are in the socialization process, they will not mature ‘true’—the maximal-claim ties willl not be embedded. The least polity structure has failed. 426 It is the taboo which must be intact, not the maiden. She will have had her hymen cut before her first menstruation, ‘to make the blood flow better’ (Stirnimann 1979: 165-66). If the taboo were not intact, it is the social identity of the bride which would be would be questionable, leaving distribution of the bridewealth contestable, and longterm certainties about debts and more general moral obligations in jeopardy. Actual incest might pass, if it did not become ‘jurally relevant’ through disclosure and the full scenario of moral indignation. While many societies are slow to trigger sanctions, I take the sense from Fr. Stirnimann that Pangwa were not lax about sexual conduct. 427 Presocial competition is a concept of human ecology. Planar structuration (institution of a Least Polity Rule appropriate to context) offers institutional control of presocial competition and in doing so commonly hides it. The result, in a macroscopic view, is an appearance of ‘spontaneous order’. As an ecological construct,

449

presocial competition refers to the situation of the many claiming parallel access to a limited pool of social goods—the situation at ground zero in any society the world around. Least polities are the smallest groups effectively organized to reproduce their format in succeeding generations. In the Kinga case they are ‘hearth centred’ rather than ‘domestic’ groups, as the latter term denotes a group with common domicile. For each cooking hearth in daily use there is a primary group, organized as a nuclear family, with primary claim to the product of that hearth and the primary responsibility for providing it. The law by which conduct within the family is governed is customary law or taboo, and custom governs conduct within the same-sex sleeping houses they share with neighbour families. What gives prime political importance to ‘family’ before ‘peer group’ in this circumstance is the responsible role a family takes in making a living for its membership. In this all-important respect, the pivotal role in the traditional Kinga system was certainly the wife-and-mother’s, though her political voice (be it in choosing garden sites or in ‘troubles’) was always mediated by a man. It was the devotion of women to their gardens and their fires which made a firm centre for the Kinga ‘least polity’ despite its dispersion in social space.

450

F EA R

& BLAME, REFERENCES

References Allport (Gordon W) 1955: Becoming. Yale University Press. Allport (Gordon W.) 1960, 1964: Personality and Social Encounter. Boston: Beacon Press. Benedict (Ruth) 1934: Patterns of Culture. New York: Houghton Mifflin. BMB = Berliner Missions-berichte [serial publication of Berlin Mission]. Bohannan (Laura) 1958: Political aspects of Tiv social organization, in John Middleton & David Tait (eds.) Tribes without Rulers (pp.33-66). London: Routledge & Kegan Paul. Brock (Beverley [Gartrell]) 1966: The Nyiha of Mbozi, Tanzania Notes & Records 65:1-30. Brown (G.Gordon) 1934: Hehe cross-cousin marriage, in E.E.Evans-Pritchard (ed.) Essays Presented to C.G.Seligman. London: Kegan Paul, Trench, Trubner. Brown (G.Gordon) & Hutt (Bruce) 1935: Anthropology in Action. London: Humphrey Milford. Charsley (S.R.) 1969: The Princes of Nyakyusa. Nairobi: East African Publishing House. Collingwood (R. G.) 1946: The Idea of History. London: Oxford University Press. Coopersmith (Stanley) 1967: The Antecedents of Self-esteem. San Francisco: W.H.Freeman. Culwick (A.T. & G.M.) 1935: Ubena of the Rivers. London: George Allen & Unwin. Dentan (Robert Knox) 1968: The Semai. New York: Holt, Rinehart & Winston. Douglas (Mary) 1973: Natural Symbols (2nd ed.). London: Barrie & Jenkins. Edgerton (Robert B.) 1971: The Individual in Cultural Adaptation. University of California Press. Eggan (Fred) 1954: Social anthropology and the method of controlled comparison, American Anthropologist (n.s.) 56: 743-763. Eggan (Fred) 1955: Social anthropology: methods and results. In his (edited) volume,Social Anthropology of North American Tribes (pp.485-551). University of Chicago Press.

443

Elton (J.F.) 1879: Travels and Researches among the Lakes and Mountains of Eastern and Central Africa. London: John Murray. Emeneau (Murray B.) 1971: Toda Songs. Oxford: Clarendon Press. Evans-Pritchard (E.E.) 1940: The Nuer. Oxford University Press. Fortes (Meyer) 1958: Introduction to Jack Goody (ed.) The Developmental Cycle of Domestic Groups (pp.1-14). Cambridge University Press. Fortune (Reo) 1932: The Sorcerers of Dobu. New York: Dutton. Foster (George M.) 1965: Peasant society and the image of the Limited Good, American Anthropologist 67:2:293-315. Foster (George M.) 1972: The anatomy of envy, Current Anthropology 13:2:165-202. Freud (Sigmund) 1918: Totem and Taboo. New York: Moffat, Yard. Freud (Sigmund) 1930: Civilization and its Discontents. London: Cape & Smith. Fried (Morton H.) 1967: The Evolution of Political Society. New York: Random House. Fülleborn (Friedrich) 1906: Das Deutsche Njassa- und Ruwuma-gebiet, Land und Leute. Berlin: Dietrich Riemer. Gartrell (Beverley [Brock]) 1979: Is ethnography possible? A critique of African Odyssey, Journal of Anthropological Research 35:4:426-446. Geertz (Clifford) 1973: The Interpretation of Culture. New York: Basic Books. Goffman (Erving) 1959: The Presentation of Self in Everyday Life. Garden City, New York: Doubleday. Goffman (Erving) 1974: Frame Analysis. Harvard University Press. Goode (William J.) 1960: A Theory of Role Strain, American Sociological Review 25:483-96. Gulliver (P.H.) 1971: Neighbours and Networks. University of California Press. Hallpike (C.R.) 1972: The Konso of Ethiopia. Oxford: Clarendon Press. Harwood (Alan) 1970: Witchcraft, Sorcery, and Social Categories among the Safwa. Oxford University Press. Hayek (Friedrich) 1952: The Sensory Order. University of Chicago Press. Heine (Bernd) 1985: The Mountain People: some notes on the Ik of northeastern Uganda, Africa 55:1:3-16. Hofstadter (Douglas R.): Gödel, Escher, Bach. New York: Vintage Books. Iliffe (John) 1979: A Modern History of Tanganyika. Cambridge University Press. Jaeger (Werner) 1945: Paideia: the Ideals of Greek Culture (I). Tr. G. Highet. New York: Oxford University Press.

444

Kalinga (Owen J. M.) 1985: Colonial rule, missionaries and ethnicity in the North Nyasa district, 1891-1938, African Studies Review 28:1:57-72. Kaplan (Abraham) 1964: The Conduct of Inquiry. San Francisco: Chandler. Kardiner (Abram) 1939: The Individual and his Society. Columbia University Press. Kroeber (A.L.) 1947 (1939): Cultural and Natural Areas of Native North America. University of California Press. Kroeber (A.L.) 1963: An Anthropologist Looks at History. University of California Press. Lacey (W.K.) 1968: The Family in Classical Greece. London: Thames & Hudson. Lasch (Christopher) 1979: The Culture of Narcissism. New York: Norton. Leach (Edmund) 1964: Political Systems of Highland Burma. Boston: Beacon Press. Leach (Edmund) 1958: Magical hair. J. of the Royal Anthropological Institute 88: 147-64. Lewis (I.M.) 1971: Ecstatic Religion. Penguin Books. Lewis (I. M.) 1986: Religion in Context: Cults and Charisma. Cambridge University Press. Lewis (Oscar) 1951: Life in a Mexican Village: Tepoztlán Revisited. Urbana: University of Illinois Press. Louch (A. R.) 1966: Explanation and Human Action. University of California Press. MacKenzie (D.R.) 1925: The Spirit-ridden Konde. London: Seeley, Service. Maine, Henry Sumner 1931 (1861): Ancient Law. London: Oxford University Press. Malinowski (Bronislaw) 1922: Argonauts of the Western Pacific. London: George Routledge. Malinowski (Bronislaw) 1926: Crime and Custom in Savage Society. London: Routledge and Kegan Paul. Malinowski (Bronislaw) 1927: Sex and Repression in Savage Society. London: Routledge and Kegan Paul. Malinowski (Bronislaw) 1929: The Sexual Life of Savages in North-western Melanesia. London: George Routledge. Malinowski (Bronislaw) 1935: Coral Gardens and their Magic. London: Allen & Unwin. Maslow (A.H.) 1968: Motivation and Personality. New York: Harper & Bros. Mead (George H.) 1934: Mind, Self, and Society. University of Chicago Press. Merensky (A.) 1894: Deutsche Arbeit am Njassa. Berlin. Berliner Evangelische Missionsgesellschaft. Mills (C. Wright) 1940: Situated actions and vocabularies of motive. American Sociological Review 5: 904-13.

445

Mitchell (J. Clyde) 1956: The Yao Village. Manchester University Press. Moffett (J.P.) ed. 1958: Handbook of Tanganyika. Dar es Salaam: Government Printer. Murdock (George Peter) 1981: Atlas of World Cultures. University of Pittsburgh Press. Murray (Henry A.) 1938: Explorations in Personality. Oxford University Press. Nadel (S.F.) 1957: The Theory of Social Structure. London: Cohen & West. Nigmann (E.) 1908: Die Wahehe. Berlin: Mittler. Nurse (Derek) 1988: The diachronic background to the language communities of southwestern Tanzania, Sprache und Geschichte in Afrika 9: 15-115. Nurse (Derek) & George Park 1988: Regional anthropology and historical linguistics, Sprache und Geschichte in Afrika 9: 7-14. Oliver (Douglas) 1955: A Solomon Island Society. Cambridge: Harvard University Press. Oliver (Roland) 1963: Discernible developments in the interior, in Roland Oliver & Gervase Mathew (eds.) History of East Africa (I). Oxford: Clarendon Press. Orans (Martin) 1965:The Santal. Detroit: Wayne State University Press. Park (George) 1966: Kinga priests: the politics of pestilence, in Marc J. Swartz et al. (eds.) Political Anthropology (pp. 229-238). Chicago: Aldine. Park (George) 1972: The motivational interpretation of institutions, Canadian Review of Sociology and Anthropology 9:2:134-49. Park (George) 1974a: The Idea of Social Structure. Garden City, New York: Doubleday. Park (George) 1974b: Longterm change in Kinga culture: arts, paper presented to Canadian Association for African Studies, Halifax. Park (George) 1988: Evolution of a regional culture in East Africa, Sprache und Geschichte in Afrika 9: 117-204. Park (George) 1990a: The Flying Armchair: Anthropology to Think With. St. John's: Prosit 1990. Park (George) (1990b): Making sense of religion by direct observation: an application of frame analysis, in Stephen Riggins (ed.) Beyond Goffman (pp. 235276). Berlin: Mouton de Gruyter. Park (George) 1994a: Bridewealth revisited, in Stephen Riggins (ed.) The Socialness of Things (pp. 253-292). Berlin: Mouton de Gruyter. Park (George) 1994b: Peace and power in an African proto-state, in Leslie Sponsel and Thomas Gregor (eds.) The Anthropology of Peace and Non-violence. Boulder: Lynne Rienner.

446

Parsons (Talcott) 1957: Malinowski and the theory of social systems, in R. Firth (ed.) Man and Culture (pp. 53-70). London: Routledge & Kegan Paul. Polanyi (Michael) 1951: The Logic of Liberty. University of Chicago Press. Polanyi (Michael) 1958: Personal Knowledge. London: Routledge & Kegan Paul. Radcliffe-Brown (A.R.) 1922/1948: The Andaman Islanders. Glencoe: The Free Press. Radcliffe-Brown (A.R.) 1940: On social structure, J. of the Royal Anthropological Institute 52. Rau (W.E.) 1979: Chewa religion and the Ngoni conquest. In Schoffeleers (1979: 131-146) Raum (O.F.) 1965: German East Africa: changes in African tribal life under German administration, 1892-1914, in Vincent Harlow et al. (eds.) History of East Africa (II) (pp.163-207). Oxford: Clarendon Press. Redfield (Robert) 1930: Tepoztlán: a Mexican Village. University of Chicago Press. Redfield (Robert) 1956: Peasant Society and Culture. University of Chicago Press. Redfield (Robert) 1962: Human Nature and the Study of Society. (M. P. Redfield, ed.). University of Chicago Press. Redmayne (Alison) 1968: The Hehe, in Andrew Roberts (ed.) Tanzania before 1900. Nairobi: East African Publishing House. Redmayne (Alison) 1983: Research on customary law in German East Africa, Journal of African Law 27:1:22-41. Reichard (P.) 1892: Deutsch-ostafrika. Das Land und seine Bewohner. Leipzig. Richards (Audrey I.) 1951. The Bemba of Northeastern Rhodesia. In Elizabeth Colson and Max Gluckman (eds.) Seven Tribes of British Central Africa (pp. 164-193). London: Oxford University Press. Richards (Audrey I.) 1956: Chisungu. New York: Grove Press. Rogers (Carl R.) 1961: On Becoming a Person. Boston: Houghton Mifflin. Rougemont (Denis de) 1957: Love in the Western World. Garden City, New York: Doubleday. Schoffeleers (j.M., ed.) 1979: Guardians of the Land. Gwelo: Mambo Press. Schoffeleers (J. M.) 1979a: The Chisuphi and Mbona cults in Malawi. In J.M.Schoffeleers (1979: 147-186). Sevice (Elman R.). 1971: Primitive Social Organization (2nd Ed). New York: Random House. Slater (Mariam K.) 1976: African Odyssey. Indiana University Press. Slater (Philip E.) 1968: The Glory of Hera. Boston: Beacon Press.

447

Sorokin (Pitirim A.) 1947: Society, Culture, and Personality. New York: Harper & Bros. Southall (Aidan W.) 1954: Alur Society. Cambridge: W. Heffer. Southall (Aidan W.) 1976: Nuer and Dinka are people: ecology, ethnicity, and logical possibility, Man 11:4:463-491. Southall (Aidan W.) 1988: The segmentary state in Africa and Asia. Comparative Studies in Society and History 30:1: 52-82. Spicer (Edward H.) 1971. Persistent cultural systems. In Science 174: 4011: 795800. Spiro (Melford) 1967. Burmese Supernaturalism. Englewood Cliffs:Prentice-Hall. Stirnimann (Hans) 1976: Existenzgrundlagen und traditionnelles Handwerk der Pangwa von SW.-TANSANIA. Freiburg: Universit tsverlag. Stirnimann (Hans) 1979: Die Pangwa von SW.-Tansania. Freiburg: Universitätsverlag. Sumner (William Graham) 1906: Folkways. Boston: Ginn and Company. Swartz (Marc J.) 1968: Introduction. In M. J. Swartz (ed.) Local -level Politics (pp. 146). Chicago: Aldine. Turnbull (Colin M.) 1973: The Mountain People. London: Jonathan Cape. Turner (Victor W.) 1956: Schism and Continuity in an AfricanSociety. Manchester University Press. Turner, (Victor W.) 1967: The Forest of Symbols. Cornell Unbiversity Press. Turner (Victor W.) 1969: The Ritual Process. Chicago: Aldine. ‡Weber (Peter ) 1998. Ritual und Identität. Hamburg: Lit. Weiner (Annette B.) 1988: The Trobrianders of Papua New Guinea. New York: Holt, Rinehart & Winston. White (Robert W.) 1948: The Abnormal Personality. New York: Ronald Press. Willis (Roy) 1981: A State in the Making. Indiana University Press. Wilson (Godfrey) 1939: The Constitution of Ngonde. Manchester University Press. Wilson (Godfrey) 1936: An Introduction to Nyakyusa Society, Bantu Studies 10. [Reprinted as: The Nyakyusa of South-western Tanganyika, in Elizabeth Colson & Max Gluckman (eds. 1951) Seven Tribes of British Central Africa (pp. 253-291). Oxford University Press.] Wilson (Godfrey) 1937: Introduction to Nyakyusa Law. Africa 10: 1: 16-36). Wilson (Monica) 1950: Nyakyusa Kinship, in A. R. Radcliffe-Brown & Daryll Forde (eds.) African Systems of Kinship and Marriage. Oxford University Press. Wilson (Monica) 1951: Good Company. Oxford University Press.

448

Wilson (Monica) 1957: Rituals of Kinship among the Nyakyusa. Oxford University Press. Wilson (Monica) 1958: Peoples of the Nyasa-tanganyika Corridor. University of Cape Town. Wilson (Monica) 1959a: Communal Rituals of the Nyakyusa. Oxford University Press. Wilson (Monica) 1959b: Divine Kings and the Breath of Men. Cambridge University Press. Wilson (Monica) 1975. Letter. Africa 45, 2: 202-5. Wilson (Monica) 1977: For Men and Elders. New York: Africana. Wolff (R.) 1905: Grammatik der Kinga-sprache. Berlin: Georg Reimer. Wright (Marcia) 1971: German Missions in Tanzania. Oxford: Clarendon Press. Wright (Marcia) 1972: Nyakyusa cults and politics in the later nineteenth century, in T.O.Ranger and I.Kimambo (eds.) The Historical Study of African Religion. London: Heinemann. Young (Michael W.) ed. 1976: The Ethnography of Malinowski. London: Routledge & Kegan Paul.

449