webAfriqa / Library / Anthropology


webAfriqa Custom Search Engine

George Peter Murdock
1897 — 1985

Professor of Anthropology, Yale University

Africa. Its Peoples and Their Culture History

New York. McGraw-Hill. 1959. 456 p.


Previous Home Next

Explore also (a) the SemanticAfrica Peoples Vocabulary
(b) the Central Bantu mind-mapping diagram

Part Eight
Expansion of the Bantu
— 38 —
Central Bantu

When the ancestors of the Northwestern and Equatorial Bantu and of the Mongo peoples had occupied all their present territory, they reached the far limits of the tropical rainforest. Beyond them, to the east and south, the indigenous hunting population, Pygmies and Bushmen, offered no more of an obstacle to further expansion than they had in the past.
The geographical environment, however, was another matter, for this now changed to savanna, dry forest, and upland grasslands (see Map 4). Here the Malaysian food plants provided but an inadequate means of exploiting the new territory. The West African cereals, which would have proved suitable, had been lost in traversing the equatorial forest. It would be 1,000 years before the Portuguese introduced another adaptive complex, the American, and 500 years before the Bantu advance guard, reaching the east coast, could acquire rice and other crops from the Arabs. In the meantime, how could the Bantu who emerged from the rainforest manage to subsist?
The discussion of the Luba and their neighbors has already given a hint of the answer. The Bantu readopted the African cereals. But whence did they obtain them? A clue is provided by the nature of the crops themselves. One was eleusinc, an Ethiopian cu ltigen which never reached the Guinea coast and therefore could not have accompanied the Bantu on their migrations. A second was sorghum. The notable fact about this plant is that, wherever it occurs among the Bantu of eastern or southern Africa, it belongs to varieties developed in East Africa, notably durra and Kafir corn (see Snowden, 1936). Even today these have not spread to the western Sudan, so that again the Bantu could not have brought them with them across the Congo Basin. The source of borrowing, consequently, can only have been a people who cultivated African cereals of Ethiopian origin.
Around the eastern periphery of the rainforest in the early centuries of the Christian era archeological and other evidence reveals only one agricultural people. And they possessed these very plants, for they were the Cushites from the idamo country who had emigrated to Uganda and established there a relatively complex civilization (see Chapter 25). Having acquired in Uganda the African cereals in their Ethiopian forms, the Bantu could readily have passed these on from tribe to tribe completely around the edge of the tropical-forest zone until they reached the Atlantic coast near the mouth of the Congo River. With these crops, all groups could then have continued their expansion into the adjacent new geographical environments.
Several auxiliary sources of evidence confirm this interpretation. In the first place, Uganda lies closer to the original homeland of the Bantu than any other section of East Africa, and hence was presumably the first to be reached, especially since the Congo River system leads to the very gates of the lake region (see Map 2). Second, although most of the present Bantu inhabitants of East Africa possess matrilineal forms of social organization, the Interlacustrine group in Uganda and Ruanda-Urundi are patrilineal and thus reasonably to be derived from the adjacent Equatorial Bantu, who likewise have patrilineal social systems. Third, the Bantu who have emerged eastward from the tropical forest have largely abandoned the settlement pattern and house type prevalent there and have adopted dwellings of the cone-cylinder model characteristic of all sedentary Cushites. Finally, as already evidenced by the Luba and their neighbors, nearly the entire circumference of the equatorial forest, from Uganda around to the mouth of the Congo River, is characterized culturally by highly complex states reminiscent of those of the Sidamo peoples. Since, as we have seen, the Bantu inhabitants of the three rainforest provinces nowhere exhibit any consequential degree of political integration, they can scarcely have carried the pattern of large-scale despotic government with them into adjacent territory. It seems reasonable to assume, therefore, that their complex political institutions came from the same source as their cereal agriculture and that both cultural systems subsequently followed approximately the same course of diffusion. Later historical events of great consequence have markedly changed the culture of the Interlacustrine Bantu, making it necessary to defer considering them until Chapter 45. We shall therefore turn next to the broad belt of Central Bantu which lies immediately south of the rainforest and extends eastward to the shores of the Indian Ocean (see Map 11). Despite the high degree of uniformity in social organization, first noted by Richards (1950), certain local variants in this and other respects make it advisable to divide the peoples of this province into several clusters.

Kongo Cluster

A cluster of closely related peoples on the lower reaches of the Congo River and the adjacent coast to the north and south aroused the astonishment of the early Portuguese explorers because of the complexity of their political institutions and their social life.

  1. Kongo (Bacongo, Bakongo, Makongo), embracing the Bashikongo (Besikongo, Eshikongo, Exikongo, Mouchicongo), Hungu (Mahungo, Maungo), Mbamba (Bamba, Bambamba), Mbata (Bambata, Batta), Mpangu (Bampangu, Pango), Nzombo (Bazombo, Zombo), and Sorongo (Asolongo, Bashilongo, Basolongo, Misorongo, Mossilongi, Mouchilonge, Musarongo, Mushirongo, Musurongo). Ihle (1929) estimates that they numbered 2,500,000 in 1675.
  2. Kunyi (Bakougni, Bakunya), with the Yangela (Bayangela, Nyangela).
  3. Sundi (Basoundi, Basundi, Mayanga, Nsundi), with the Bembe (Babembe), Bwende (Babouendi, Babwende), Dondo (Badondo), and Kamba (Bakamba).
  4. Vili (Bafiote, Bavili, Chiloango, Fiore, Fjort, Loango, Mfioti), with the Kakongo (Cacongo, Kabinda, Kacongo, Makouango).
  5. Yombe (Bayombe, Majombe, Mayumbe). They numbered about 170,000 in 1933.

Kimbundu Cluster

This cluster of tribes, occupying west central Angola, is among the least adequately described in the entire African continent.

  1. Holo (Baholo, Hollo).
  2. Kimbundu (Ambundu, Mbundu), embracing the Luanda (Loanda), Mbaka (Ambaca, Ambaquista), Mbondo (Bondo), Ndongo (Ndonga, Ngola), Ndembu (Andembu, Atembu, Bandempo, Batembo, Demba, Dembu, Jindembu), and Tamba (Matamba).
  3. Kisama (Cisama, Kissama, Quissama). Their population was reduced by smallpox from over 30,000 around 1900 to about 10,000 in 1929.
  4. Lupolo (Libolo), with the Esela, Haku (Hako, Oako), Kipala (Cipala, Kibala), and Lemba (Malemba).
  5. Mbangala (Bangala, Bangela, Imbangala, Ngala), with the Shinje (Bashinshe, Chinge, Kasinji, Mashinge, Maxingi) .
  6. Sele (Basele, Esele, Selle), with the Kisanji (Cisanje, Kissandschi, Quissanje), Mbui (Amboim, Ambuin, Mbuiyi, Ombe), and Sumbe (Basumbe).

Kwango Cluster

This cluster embraces a number of well-described tribes in the drainage basin of the Kwango River, a southern affluent of the Congo.

  1. Kwese (Bakuese, Bakwese).
  2. Luwa (Balua, Baluwa, Lua), with the Nzofo and Sonde (Basonde).
  3. Mbala (Mmbala, Bambala), with the Huana (Baguana, Bahuana, Bauangana, Wengana) and Humbu (Bahumbu, Bavumbu, Bawumbu).
  4. Ngongo (Bangongo), with the Songo (Basongo).
  5. Pende (Bapende, Baphendi, Capende, Tupende). They number about 27,000.
  6. Suku (Bapindi, Basuku, Pindi, Pindji), with the Samba (Tsaam). They number about 80,000.
  7. Yaka (Aiacka, Bayaka, Bayakala, Djakka, Dsehagga, Giaca, Jaca, Majacalla, Mayacca, Muyaka, Myaka, Ngiaka, Vacca, Yagga), with the Lula (Balula).

Kasai Cluster

The tribes of thjs cluster lie northeast of the Kwango group on or near the Kasai tributary of the Congo River, and adjacent to the Mongo province.

  1. Bunda (Ambundu, Ambuun, Babounda, Babunda, Bambunda, Mbuni).
  2. Dzing (Badinga, Badzing, Baringa, Dinga), with the Lori (Aiwer, Baloli, Balori), Ngoli (Angul, Bangodi, Bangoli, Bangulu), and Nzari (Bandzadi, Bandzala, Bandjadi, Banzari).
  3. Kuba (Bacouba, Bakuba, Tukubba), embracing the Binji (Babindji), Mbala (Bambala), Ngongo (Bangongo), and lesser subtribes. They number about 75,000.
  4. Lele (Bachilela, Bashilele, Bouxhilile), with the Wongo (Bagongo, Bakongo, Bawongo, Tukkongo). They number about 12,000.
  5. Sakata (Basaka, Basakaca), embracing the Baye (Babaye, Bai, Bobai, Tollo), Boma (Baboma, Babuma, Ouabouma, Uabuma, Wabuma), Dia (Badia, Bajia, Jia, Wadia, Wadja), and Lesa (Balesa, Lessa).
  6. Yanzi (Bajansi, Basehansi, Bayanchi, Bayanzi, Wachanzi), with the Amput (Bamputu)

Lunda Cluster

The tribes of this cluster live east of the Kimbundu and southeast of the Kwango groups. Most of them were subject to the Lunda empire of Mwata Yamwo (see Chapter 37).

  1. Chokwe (Ahioko, Aioko, Atsokwe, Bachoko, Badjok, Bakioko, Batchokwe, Batshoko, Chiboque, Kashioko, Katsokwe, Khioko, Kioko, Kioque, Makioko, Matchioko, Quioco, Shioko, Tsehiokwe, Tsiboko, Tsokwe, Tutsehokwe, Utshiokwe, Vichioko, Watsehokwe), with the Minungu (Tuminungu).
  2. Luchazi (Balochasi, Balojash, Kaluchazi, Lutsehase, Luxage, Malochazi, Makangala, Mulochazi, Valuchazi).
  3. Luimbe (Baluimbi, Loimbe, Lwimbi, Valuimbe), with the Mbande (Kimbande, Quimbande). They number about 40,000.
  4. Luvale (Aluena, Balovale, Kaluena, Lobare, Luena, Lovale, Lwena, Malobale, Muluena, Tulwena, Valwena).
  5. Mbunda (Ambunda, Mambunda).
  6. Mbwela (Amboella, Ambuela).
  7. Ndembu (Andembu, Bandempo, Southern Lunda).
  8. Songo (Basongo, Masongo).

Bemba Cluster

The tribes of this cluster live east of the Lunda group in Northern Rhodesia [now Republic of Zimbabwe] and the adjacent [former] Belgian Congo.

  1. Aushi (Avausi, Bahushi, Baoussi, Baushi, Umwausi, Ushi, Wouaousi, Wausi, Wauzhi), with the Chisinga (Wenachishinga), Kawendi (Kabende, Wenakawendi), Mukulu, and Ngumbu (Wenangumbu). They number about 100,000.
  2. Bemba (Awemba, Ayemba, Babemba, Muemba, Wawemba, Wemba). They number about 150,000.
  3. Bisa (Abisa, Aiza, Awisa, Babisa, Bawisa, Invisa, Moize, Moiza, Muiza, Wabisa, Wisa). They number about 50,000.
  4. Buye (Babudjue, Babui, Babuye, Bambuyu, Bubui, Luba-Hemba, Wabudjwe, Wabuyu, Waruwa), with the Bwari (Ouabouari, Ubwari, Wabwari), Goma (Wagoma), Hombo (Bango, Bangobango), Kalanga (Bakalanga), Kunda (Bakunda), and Lumbu (Balumbu).
  5. Holoholo (Baholoholo, Wahorohoro), also called Guha (Bagua, Baguha,Wuaguha, Waguha) and Tumbwe (Batumbwe, Watombwe).
  6. Kaonde (Bakahonde, Bakaonde), with the Sanga (Basanga). They number about 40,000.
  7. Lala (Balala, Bukanda), with the Ambo (Kambonsenga) and Luano. They number about 75,000.
  8. Lamba (Balamba, Walamba), with the Lima, Sewa, and Swaka. They number about 70,000.
  9. Luapula (Bena Kazembe), with the Londa (Alonda, Aronda, Balonda, Lunda,Warunda). They number about 100,000.
  10. Lungu (Alungu, Marungu, Walungu, Warungu), with the Mambwe (Amambwe).
  11. Senga (Asenga). These people, numbering about 25,000, are an offshoot of the Bisa and are strongly acculturated to the Tumbuka.
  12. Shila (Awasira, Bachila, Bashila, Messira, Wasira).
  13. Tabwa (Barabwa, Batambwa, ltawa, Watabwa), with the Bwile (Babwile, Swire).
  14. Unga (Bahonga, Bahunga, Baunga, Honja, Waunga). They number about 10,000

Maravi Cluster

The tribes of this cluster, who are known collectively as the Maravi, live mainly in Nyasaland (now Malawi) and on the lower Zambesi River.

  1. Chewa (Achewa, Ancheya, Cewa, Masheba, Sheva), with the kindred Chipeta (Achipeta). They number about 650,000.
  2. Chuabo (Achwabo, Atxuabo, Chwampo), with the Podzo.
  3. Kunda (Achikunda, Chikunda), with the Nyungwe (Wanyongwe). They number about 75,000 in Nyasaland alone.
  4. Nsenga (Senga). They number about 50,000.
  5. Nyanja (Anyanja, Wanyanja), with the Manganja (Maganja, Waganga). They number in excess of 300,000.
  6. Nyasa (Anyassa, Wanyassa).
  7. Sena (Asena, Wasena).
  8. Tumbuka (Barumbuka. Matumboka, Watumbuka), with the Fuliwa, Fungwe (Wafungwe), Henga (Bahenga, Wahenga), Hewe (Rewa), Kamanga (Wakamanga), Kandawire, Nthali, Phoka (Poka), Sisya (Siska), Tonga (Baronga, Matoka), Wenya, and Yombe. They number about 120,000.
  9. Zimba (Azimba, Bazimba, Ouazimba, Wasimba. Wazimba).

Yao Cluster

The tribes of this cluster live mainly in northern Mozambique and adjacent coastal Tanganyika, although a few spill over into southern Nyasaland.

  1. Lomwe (Acilowe. Alomwe, Anguru, Nguru, Walomwe, Wanguru), with the Mihavani. They number about 400,000 exclusive of those in Portuguese territory.
  2. Makonde (Wakonde, Wamakonde), with the Matambwe (Wamatambwe) and Mawia (Mabiha, Mavia, Wawiha, Wamavia, Wawia). Those in British territory, a small fraction of the total, number about 150,000.
  3. Makua (Makoa, Makoane, Wakua, Wamakua). They number about 850,000.
  4. Mwera (Wamuera), with the Kiturika (Kirusika).
  5. Ngindo (Wangindo), with the Chobo (Choo), Hamba, lkemba, Tdonde (Wandonde), and Ndwewe.
  6. Yao (chawa, Adjao, Adsawa, Adsoa, Ajawa, Ayo, Hian, Mudsau, Mujano, Mujoa, Myao, Veiao, Wahaiao, Waiyau, Wayao). Those in British territory, a small fraction of the total, number about 360,000.

Among the eastern Central Bantu, particularly those of the Maravi cluster, reside a number of intrusive peoples known collectively as the Ngoni. They represent the descendants of military bands of Nguni who fled Natal in the 1820s to escape Zulu hegemony, ranged northward as far as Lake Victoria raiding for cattle and women, and eventually settled down in the vicinity of Lake Nyasa. Here they married women of the Central Bantu tribes and have become to a considerable extent acculturated to the latter. They will be considered along with their South African kinsmen in Chapter 50.
The economy of the Central Bantu rests primarily upon swidden agriculture. Their original dependence upon the Malaysian food plants survives today only in the Kongo cluster. Elsewhere bananas and yams appear only sporadically, and taro disappears altogether. Eleusine, millet, and sorghum, the African cereals which supplanted them, still retain a prominent position among the tribes of the Bemba, Kasai, Lunda, and Yao clusters; and lesser crops of the Sudanic complex, notably cow peas, earth peas, gourds, and sesame, occur here and there. The plants that dominate the agriculture of the province today, however, are those of American origin, all introduced subsequent to 1500. Manioc constitutes the staple in the Kasai and Kwango clusters, and maize in the Maravi cluster, whereas both share the top position in the Kimbundu and Lunda clusters and are even important in the Bemba, Kongo, and Yao clusters. Among other American plants, peanuts and sweet potatoes are especially widely cultivated. Rice, acquired from the Arabs, also enjoys a modest standing in the eastern part of the province.
The Central Bantu augment their starchy vegetable diet with animal proteins in various ways. They do considerable hunting. They fish to an even greater extent, and the Luimbe, Songo, Unga, and some Shila depend more on fishing than on agriculture. A few tribes in the west even practice cannibalism; the Dzing, Mbangala, Sele, and Yanzi do so extensively, the Bunda, Mbala, Pende, and Suku poradically. Animal husbandry, however, provides little meat and no milk. Although all group keep dogs and chickens, most of them a few goats as well, and some of them even pigs and sheep, the overwhelming majority of the tribes of the province have no cattle at all. In the west, the Holo of the Kimbundu cluster keep them in considerable number today, but only the southernmost Chokwe and the westernmost Mbwela, who adjoin the pastoral Southwestern Bantu, had them in precontact times. In the east, an occasional chief among the Holoholo, Luapula, Lungu, and Nyasa maintains a small herd as a matter of prestige, but the striking ignorance of herding and milking techniques demonstrate that cattle are a rarity. Only the Tumbuka in the extreme north of the Maravi cluster seem to have kept them in appreciable numbers prior to the Ngoni invasions of the nineteenth century. Even the intrusive Ngoni, though strongly pastoral when they arrived, have in several instances lost their cattle today. There exists, in short, not a shred of justification for including the eastern Central Bantu in an alleged “East African cattle area.”
Intergroup trade plays a prominent role in the economic life of the peoples in the four westernmost clusters. More than half the tribes in this region maintain regular markets and possess conventional media of exchange, such as cowrie shells, salt, raffia cloth, brass rods, and iron bars. The peoples of the four easternmost cluster, on the other hand, conduct relatively little trade and are characterized, indeed, by a striking poverty in material possessions.
The province reveals some variation in settlement patterns. A few tribes in the extreme north, e.g., the Dzing, Holoholo, Kuba, Sakata, Sundi, and Yombe, align their dwellings in two rows along either side of a single street, as do the Bantu of the adjacent rainforest. The members of the Kwango cluster and some of their immediate neighbors live in settlements of dispersed homesteads or small hamlets, and this pattern recurs in the extreme cast among the tribes of the Yao cluster. Elsewhere, for the most part, the Central Bantu occupy compact and frequently stockaded villages, often with the dwellings arranged around a central plaza. From the western Kaonde to the Atlantic coast the prevailing house type is a rectangular hut with a thatched gable roof and walls of palm leaves, grass, wood, or wattle and daub. The Bunda, Chokwe, Luimbe, Mbala, Ndembu, and Pende, however, prefer square houses with pyramidal roofs, and beehive dwellings occur among the Kisama. From the eastern Kaonde to the Indian Ocean, all groups have exchanged the rectangular house type of the tropical forest for the old Sudanic hut with a conical thatched roof and cylindrical walls of wattle and daub, and this same structure recurs in the extreme southwest among the Mbangala and Mbwela.
The inhabitants of the province reveal a somewhat surprising complexity in social stratification. Besides the universality of slavery, usually hereditary, and the widespread practice of enslavement for debt and for serious crimes, most groups differentiate a privileged nobility from ordinary freemen, and a number also recognize a special class of royalty. The sources specifically deny such privileged classes only for the Kisama, Ngongo, akata, and Yanzi in the west and for the peoples of the Maravi and Yao clusters in the east. Age-grades and secret societies are rare. Most tribes of the five western clusters, with the noteworthy exception of the Kuba, practice circumcision, but the eastern societies do not, and genital mutilations for females are nowhere reported.
Each settlement has a hereditary headman and an advisory council of elders. The Makonde, Mwera, and Ngindo in the extreme northeast recognize no higher political authority, but all other tribes have at least paramount chiefs over districts, although in some instances these are small and the chief's functions are more religious than regulative. Genuinely complex states occur with a frequency probably greater than anywhere else in Africa south of the Sahara-Sudan fringe. In addition to powerful tribal states among the Bemba, Chokwe, Kimbundu, Kuba, Luapula, Vili, Yaka, and formerly also the Buye and Tumbuka, the province had two empires of even greater magnitude at the time of first European contact. We have already described the Lunda empire of Mwata Yamwo, which during the seventeenth century reduced to vassalage the Chokwe, Luchazi, Luvale, Mbangala, Ndembu, and Yaka tribes (Chapter 37). Resisting the temptation to summarize the extremely interesting political systems of the Bemba and Kuba, we shall take the organization of the Kongo empire as our sole example. Since, however, it was destroyed by the Yaka about 1569, a few details missing in the early sources must be filled in from accounts of the adjacent and very similar Vili state of Loango.
The Kongo king maintained an elaborate court at the capital town of San Salvador, where he was surrounded by numerous slaves, pages, and personal attendants; a harem of wives; and such special officials as a chief priest and a royal executioner. The Queen-Mother held a respected position, though one appreciably less prominent than in the Bemba, Chokwe, Kuba, and Vili states. The ruler was an absolute monarch with a ritual relationship to the land, and no one might observe him eating or drinking, on pain of death. The symbols of his high office included a throne with ivory carvings, a white cap, and a zebra's tail. He exercised supreme judicial power and reserved for himself the exclusive right to inflict the death penalty.
A hierarchical administrative organization, with six great provinces divided into districts with a number of villages in each, assured the maintenance of order and a regular flow of tribute in cowrie-shell money, livestock, and agricultural produce. The hereditary headman in each community had the responsibility of making these collections once a year and transmitting them to the capital. Other sources of revenue included tolls from trade, fines paid in laves for breaches of court etiquette, a royal monopoly on productive cowrie fisheries, and the presentation for the king of all pelts of particular animal and all fish of certain favored species. The king appointed all provincial governors and district chiefs, usually from among his own relatives. Each maintained a local capital and court in his own administrative territory, and a courier system provided communication between them and San Salvador. Each territorial official also maintained a residence at the national capital. Here the provincial governors formed an advisory council of state and also exercised specialized individual functions. One, for example, served as commander in chief of the army, a second as minister of commerce and markets, a third as supreme ruler during the interregnum following the death of a king.
When the ruler died, the sacred fire which had been maintained throughout his reign was extinguished, as were hearth fires throughout the kingdom, and all work ceased. His corpse was mummified by smoking, wrapped in many cloths, and interred with human sacrifices. Meanwhile the council of state, acting as an electoral college, chose his successor from among the male members of the royal matrilineage. He was then installed with much pomp, including a new-fire ceremony.
The Kisama in the west and the Ngindo and Tumbuka in the east have social systems characterized by patrilineal descent and patrilocal residence. In the last two cases, and very likely in the first as well, these represent recent shifts from the matrilineal types of social organization which prevail everywhere else in the Central Bantu province. Since the latter reveal two somewhat divergent patterns, we shall deal with them separately. The first and most basic one occurs among the Buye, Holoholo, and Shila tribes on the northwestern edge of the Bemba cluster; throughout the Kasai, Jongo, Kwango, and Lunda clusters to the west thereof; and also, so far as the scanty available information can tell us, in the Kimbundu cluster in Angola. When we have described this pattern, we shall note the departures from it which occur in the three easternmost clusters.
The regulations governing marriage among the western Central Bantu usually prescribe both local and lineage exogamy. They prevent unions with a parallel cousin, and most tribes of the Masai cluster also prohibit cross-cousin marriage. The latter is preferred, however, among the Bunda, the northwestern tribes of the Bemba cluster, and throughout the Kongo, Kwango, and Lunda clusters, although the Pende and Vili permit such unions only with a father's sister's daughter. On the basis of incomplete evidence, all groups prescribe the levirate, and all except the Kimbundu allow the sororate, for the remarriage of widowed persons.
A primary marriage always entails a bride-price. Although usually substantial in amount, this consists only of a small or token payment among the Pende and most tribes of the Lunda cluster. The Kuba allow both alternatives. If the payment is large, the wife's family must replace her if she dies, and she is obligated to reside with her husband and to marry his brother on his death; with only a token payment, however, these obligations do not prevail, and the couple reside alternately with the relatives of each for periods of five or six years.
Polygyny is preferential, general, and, except among the Bunda, exclusively nonsororal. The first wife enjoys a special status, but each co-wife has a separate hut, and the husband spends a prescribed period with each in rotation. The household unit normally assumes the form of an independent polygynous family, an extended family organization being reported only for the Sundi.
The Lele, one of the few tribes of the area possessing organized agegrades, exhibit an interesting form of polyandry. A girl abducted from another group becomes for six months the “wife of the village” and has sex relations with any or all of its male members. Thereafter she is reserved as the exclusive common wife of the unmarried men of the junior age-set. As these obtain wives of their own, they lose their communal rights. When only one man remains, he and the girl become husband and wife in a permanent monogamous union. Similar practices prevail among the adjacent Bunda, Dzing, and Pende.
With remarkable unanimity the western tribes of the province conform to a single basic pattern of social organization, characterized by matrilineal descent, inheritance, and succession; avunculocal residence; noncorporate matrisibs (usually exogamous and often totemic); and exogamous matrilineages, each localized in a village as an avunculocal clancommunity. Only for the Kimbundu cluster does some uncertainty exist because of inadequate ethnographic coverage; the available sources clearly attest matrilineal descent, inheritance, and succession but do not specify whether residence follows the avunculocal or patrilocal rule.
Different tribes naturally reveal minor variations. Thus boys leave their paternal homes to join their maternal uncles in another village at the age of five or six in the Lunda cluster, at ten or twelve in the Kongo cluster, but often not until their marriage or the death of their father in the Kasai and Kwango clusters. In a few tribes a minority of men even reside permanently in the villages of their fathers, a tendency which has been increasing in recent years. Perhaps in consequence of this, the clan or localized lineage comprises only a segment of a village in some societies. Different groups also vary in regard to whether a younger brother or the eldest son of the eldest sister takes precedence in inheritance and succession. Some variation likewise occurs in kinship terminology. Cousin terms of the Iroquois type prevail throughout the Lunda cluster and in many tribes of the Kasai, Kwango, and northwest Bemba clusters as well, but the Pende and apparently the Kongo kinship systems reveal features of the Crow pattern.
That so clear an over-all picture of a rather exceptional type of social structure emerges from the decriptive sources must be taken as a tribute to the competence of the Belgian and other ethnographers who have worked in the area. Particularly significant, however, is the fact that this, the largest of all Bantu provinces and in most respects the most conservative, provides full confirmation of the reconstruction made in Chapter 13 of the original form of social organization among the closest linguistic relatives of the Bantu on the Nigerian plateau.
Social organization in the eastern part of the province reveals an interesting shift from that prevailing in the west. Though descent, inheritance, and succession conform ro the matrilineal rule, only the Buye, Holoholo, and Shila, as we have seen, still exhibit the presumptive ancestral pattern characterized by a substantial bride-price, avunculocal residence, local exogamy, and avuncuclans. All groups, to be sure, have exogamous matrisibs, typically divided into lineages and often aggregated into phratries on the basis of the complementary or antipathetic characteristics of their totems, but the principles governing the localization of kin groups diverge markedly from those prevailing farther west.
The source of these differences appears to lie in the relative poverty of the people in movable and heritable wealth. The eastern tribes lack the iron money and quantities of trade goods held by their western cousins. Of the Bemba, for example, it is reported that the only important physical object which a man can leave to his heir is his bow. In the absence of movable property of consequence, inheritance tends to assume the form of positional succession, according to which a man's nonmaterial acquisitions—such as his name, his social status, and his guardian spirittend to descend as a unit, along with his widows and his few material possessions, to his preferred heir, i.e., his younger brother or perhaps his sister's son.
In the absence of livestock, native money, or other movable property in substantial amounts, the bride-price dwindles to a small and purely symbolic gift or disappears entirely, and is replaced as a consideration by bride-service. Throughout the area, consequently, a man must pay for his wife by working for her parents over a period of at least one year and usually several. This renders residence initially matrilocal, instead of being avunculocal from the beginning as among the western Central Bantu.
When he has served his period of bride-service, a man can usually remove with his wife and family to either (1) his father's settlement in patrilocal residence, (2) a new village which he founds himself in neolocal residence, or (3) his mother's original home in avunculocal residence. British ethnographers, through their habitual use of the one ambiguous term “virilocal” for these three very different residence patterns, too often present a confused picture of the resulting structural relationships. It seems reasonably clear, nevertheless, that removal after bride-service entails avunculocal residence far more often than it does either of the alternatives. Removal is by no means inevitable, however, for a man may always choose to remain in permanent matrilocal residence after he has fulfilled his service obligation, and the widespread preference for sororal polygyny provides a motive for such a decision. Among the Yao and neighboring groups, moreover, removal depends upon an opportunity for positional succession to a c hiefship or the headship of an extended family. Ultimate residence is thus variable—often matrilocal, often avuncular, sometimes patrilocal or neolocal.
The early German ethnographers called attention to a structural point which has since been overlooked, namely, the effect of cross-cousin marriage. Preferential marriage between cross-cousins, coupled with Iroquois cousin terminology, prevails throughout the area, and a number of tribes, e.g., the Chewa, Kaonde, Lamba, and Makonde, are specifically reported to favor unions with a matrilateral over a patrilateral cross-cousin, i.e., with a mother's brother's daughter rather than a father's sister's daughter. Whenever a man marries his mother's brother's daughter and resides matrilocally, he lives, of course, with his maternal uncle. The preferential rule of marriage thus equates matrilocal with avunculocal residence, and in all such cases removal naturally becomes unnecessary. Ultimate residence in the area is thus resolvable into combinations of four distinct patterns:

  1. Permanent matrilocal residence which is at the same time avunculocal
  2. Matrilocal but nonavunculocal residence persisting beyond bride-service
  3. Avunculocal and nonmatrilocal residence resulting from removal
  4. Patrilocal or neolocal residence after removal

The first three patterns are common; the fourth, exceptional. Regional differences within the area appear to reflect the varying incidence of patterns two and three in different tribal clusters. Statistically or descriptively, matrilocal residence probably preponderates, at least if one overlooks the avunculocality implicit in preferential cross-cousin marriages. Historically, however, the social systems of the area appear to be derived, not from an original matrilocal form of the marrilineate, bur from an avunculocal form in which an initial period of matrilocal residence has been extended to the point where, in the extreme case, avunculocality survives only in cases of positional succession. Such a development, though unusual, is by no means unprecedented. Murphy (1956), for example, has documented a case where a patrilineal and formerly patrilocal society has become matrilocal, except for local headmen, through a progressive extension of the period of bride-service.
The complexity of residence practices has in general prevented the development of localized clans of any of the tandard types. Instead, kin ties are utilized in unusual combinations to produce local groups. The Yao, for example, add to a matrilocal extended family, or localized minor matrilineage, the family of the cnior male of the lineage, who lives, as it were, in sororilocal residence with his married sisters and their families. Among the Bemba, a local group evolves out of shifting ties to an influential man, beginning usually with his sons and younger brothers and their families, who are later replaced by his sisters and their families (and perhaps an affine or two), and ultimately by his sisters' sons and their families.
The area also provides an actual case of transition from matrilineal to patrilineal descent, documented for three succes ive stages. The Tumbuka, prior to 1780, resembled most of the neighboring tribes in adhering to matrilineal descent, inheritance, and succession, in requiring matrilocal bride-service rather than a bride-price, and in permitting an ultimate shift to avunculocal residence.
Between 1780 and 1800 some ivory traders from the east conquered the Tumbuka and established the Kamanga kingdom. Under the influence of these patrilineal invaders, bride-service was reduced to a nominal period, the token bridal gift was increased ro enable the husband to remove his wife almost immediately to his own village, and rights to inherit and succeed, after younger brothers had had their turn, were transferred from sisters' sons to own sons.
Around 1855 a new invasion, this time by the Ngoni, overthrew the Kamanga dynasty and subjugated the Tumbuka to new rulers with even stronger patrilineal institutions. In consequence of their influence the Tumbuka abandoned even nominal bride-service, adopted the fullfledged South African bride-price, or lobola, substituted the eldest son for the younger brother as the preferred heir and successor, and transformed what had originally been matrisibs into indubitable exogamous patrisibs. The process of change here attested for the Tumbuka has doubtless been recapitulated in Tanganyika among peoples for whom we unfortunately lack comparable documentation.

Selected Bibliography