Navigation – Plan du site
Études et essais

Dravidian Kinship Systems in Africa

Per Hage
p. 395-407

Résumés

Résumé
Alors que les systèmes de parenté dravidiens, fondés sur une règle de mariage bilatéral entre cousins croisés, sont généralement envisagés comme point de départ des théories universelles d’évolution de la parenté, on considère les systèmes iroquois, qui ne possèdent pas une telle règle, comme dérivés des systèmes dravidiens. Les systèmes dravidiens et iroquois n’ont pourtant pas la même répartition géographique: les premiers sont bien connus en Asie du Sud, en Australie et en Amérique, mais pas en Europe ni en Afrique; les seconds se retrouvent dans de nombreuses régions du monde, exceptée l’Asie du Sud. Cet article a pour objectif de décrire un système de parenté dravidien dans une société de langue bantou et de suggérer la présence, actuelle ou passée, de systèmes dravidiens en d’autres endroits d’Afrique.

Haut de page

Texte intégral

1Dravidianate (Dravidian- and Kariera-type) kinship systems, based on a rule of bilateral cross-cousin marriage, are usually taken as the starting point in universal theories of kinship evolution while Iroquois-type kinship systems, which lack such a rule, are treated as devolved versions of Dravidian systems (Needham 1967; Allen 1986, 1998; Kryukov 1998). Dravidianate and Iroquois kinship systems, however, have an uneven geographical distribution. The former are well known from South Asia, Australia and America (Godelier, Trautmann and Tjon Sie Fat 1998) but not from Africa and Europe while the latter are well known from many regions of the world but not from South Asia. Parkin (1998) was able to turn up a single Iroquois-like system among the Burushaski in northern Pakistan but no one has adduced a Dravidian kinship system from Africa. How should the Dravidian-Iroquois distribution be interpreted?

2In Trautmann’s view (2000, 2001), the unevenness of the Dravidian-Iroquois distribution argues against evolutionist explanations of kinship such as Allen’s tetradic theory (1986, 1998) in favor of historicist explanations such as Trautmann’s comparative study of Dravidian kinship systems in South India (1981). Tetradic theory holds that all attested kinship systems ultimately derive from a prescriptive truly elementary system based on bilateral cross-cousin marriage structured by cross-cutting exogamous descent moieties and endogamous alternate generation moieties similar to the Kariera four-section systems known from Australia (Radcliffe-Brown 1931). In tetradic theory, Dravidian systems are formerly Kariera-type systems that have lost their alternate generation equations and hence their sections (marriage classes). Allen (1989a) assumes a «patchy» retention of Dravidianate kinship systems from early – Upper Paleolithic – times while Trautmann emphasizes the historical distinctiveness of kinship systems in different world regions – Dravidian in South India and Iroquois in Africa.

3For Godelier, Trautmann & Tjon Sie Fat (1998) the Dravidian-Iroquois distribution raises theoretical and, in the first instance, empirical questions:

«[…] does [Parkin’s] discovery mean that many more Iroquois-like systems will be found in south Asia? If not, what does the scarcity or absence of Iroquois- (and Yafar- and Kuma-) like systems in South Asia mean? And how can we explain the situation in Africa, where Iroquois-like systems appear to be abundant and Dravidianate systems absent? These questions are difficult to answer. Any attempt to do so must, as a first step, undertake regional inventories, especially in the regions not covered in this [conference] most notably Africa and Europe».(Ibid.: 17)

  • 1 Bivar Segurado’s analysis (1989) of restricted exchange and preferential patrilateral marriage exch (...)

4In view of the conflation of Dravidian and Iroquois systems in the history of kinship studies (Trautmann & Barnes 1998), the African data clearly deserve a closer look. My purpose is to describe a Dravidian kinship system from a Bantu-speaking society and to suggest the presence or former presence of Dravidian-type kinship systems in other African societies. My analysis is based on Mitchell’s studies of the Yao (1951, 1956) which were published before the distinction between Dravidian and Iroquois kinship systems was recognized (Lounsbury 1964), formalized and generalized (Scheffler 1971; Trautmann 1981; Tjon Sie Fat 1998; Viveiros de Castro 1998)1.

Dravidian and Iroquois Kinship Systems

  • 2 Some kinship theorists treat Dravidian and Kariera as essentially the same, the essential feature b (...)

5The study of Dravidian kinship systems has a long history. For present purposes, the major punctuating events are Rivers’articulation of the relation between Dravidian kinship terminology and cross-cousin marriage (1906, 1907, 1914), and Lounsbury’s distinction between Dravidian and Iroquois crossness (1964). Dravidian systems can also be distinguished from Kariera systems by the presence in the latter of alternate generation equations which imply marriage classes and «global forms» of society (Dumont 1970, 1975)2.

  • 3 The following abbreviations are used in this paper: P parent, F father, M mother, G sibling, B brot (...)

6Dravidian kinship terminologies are defined by two properties consistent with a rule of bilateral cross-cousin marriage (Godelier, Trautmann & Tjon Sie Fat 1998). The first property is a set of affinal-consanguineal equations including3:

  • in the +1 level:
    MB = FZH = EF ? F = FB = MZH
    FZ = MBW = EM ? M = MZ = FBW
  • in the 0 (ego’s) level:
    MBD = FZD = W = WZ = BW ? Z = FBD = MZD
    MBS = FZS = H = HB = ZH ? B = FBS = MZS
  • in the -1 level:
    os-GD = SW ? ss-GD = D
    os-GS = DH ? ss-GS = S

7The second property of Dravidian terminologies is a distinctive pattern of cross-parallel classification of remote relatives. In the case of second cousins the children of parents’opposite sex cross-cousins are parallel while the children of parents’same sex cross-cousins are cross. Intuitively, opposite sex cross-cousins are potential spouses and their children are therefore classified with ego’s siblings and parallel cousins while same sex cross-cousins are potential in-laws and their children are therefore classified with ego’s cross-cousins (Trautmann 1981; Godelier, Trautmann & Tjon Sie Fat 1998). In an Iroquois system the situation is just the reverse: the children of same sex cross-cousins are parallel while the children of opposite sex cross-cousins are cross. The Dravidian and Iroquois classification of second cousins is shown in table 1. Evolutionarily, an Iroquois kinship system can be regarded as a formerly Dravidian-type system which has lost its prescriptive kin term equations and simplified its cross-parallel classification by defining parallel and cross-cousins solely on the basis of same or opposite sex relations in the parental (+1) generation.

Table 1. Cross-parallel classification of second cousins in Dravidian and Iroquois systems

Table 1. Cross-parallel classification of second cousins in Dravidian and Iroquois systems

8In the +1 and -1 generational levels Dravidian and Iroquois crossness are defined as follows:

  • in the +1 level:
    FMBS, FFZS are parallel in Iroquois but cross in Dravidian.
    MFZD, MMBD are parallel in Iroquois but cross in Dravidian.
    MFZS, MMBS are cross in Iroquois but parallel in Dravidian.
    FFZD, FMBD are cross in Iroquois but parallel in Dravidian.
  • In the -1 level:
    FZSC, MBSC, (man speaking), FZDC, MBDC, (woman speaking), are parallel in Iroquois but cross in Dravidian.
    FZDC, MBDC, (man speaking), FZSC, MBSC, (woman speaking), are cross in Iroquois but parallel in Dravidian.

9Dravidian and Iroquois terminologies have often been conflated because they have the same cross-parallel classification of close relatives (Lounsbury 1964). In the case of first cousins, in both Dravidian and Iroquois G = FBC = MZC ? FZC = MBC. In Murdock’s classification (1967) all terminologies with this first cousin pattern are «Iroquois». As a result, many of the «Iroquois» systems in Murdock’s Ethnographic Atlas, for example, Yao, the African system examined in this paper, may in fact be Dravidian in type.

The Yao Kinship System

10The Yao are a Bantu-speaking group in Malawi (formerly Nyasaland). The population in 1945, exclusive of a large number in Mozambique, was 360000 (Murdock 1959). The Yao raised both subsistence and cash crops. Traditionally, men were often absent from Yao villages engaged in trading and slave-raiding expeditions. Descent was matrilineal and residence matrilocal (Mitchell 1951, 1956). Matrilineally related women and their female descendants were organized into «sororities» (asyme mbumbu) centered on an eldest brother who served as a «warden» for the group. Yao society was unstratified, organized into small autonomous villages headed by senior matrilineal kinsmen. Genealogical reckoning rarely exceeded six or seven generations in depth.

11According to Mitchell, the Yao had a strong tradition of cross-cousin marriage with some preference for the patrilateral cross-cousin. The Yao said that all first marriages of a man should be with a cross-cousin even though the frequency of such marriages was only 15 % (Mitchell 1956: 200). However, the divorce rate was high – 80 % of men and women over 40 had been divorced at least once – and it may be that the total frequency of cross-cousin marriage was higher. Marriages to other types of kin was 21 %. Such marriages are «loosely called cross-cousin marriages, and often what has been called a cross-cousin marriage turns out to be marriage to a cross-cousin’s daughter or something similar» (Ibid.). We note that some Dravidian societies in South India also have relatively low rates of cross-cousin marriage (Trautmann 1981: 218).

  • 4 Mitchell gives two terms for (man speaking) yB without explanation. He does not give a term for W o (...)

12Yao kin terms, as they occur in the course of Mitchell’s ethnographies (1951: 336; 1956: 136, 147, 175, 199-202), are collected in table 2. Mitchell’s kin term data are incomplete but sufficient for our purpose4.

Table 2. Yao kin terms (from Mitchell 1951, 1957)

Table 2. Yao kin terms (from Mitchell 1951, 1957)

13Mitchell showed that Yao kinship terminology assumes a rule of bilateral cross-cousin marriage. The affinal-consanguineal equations are MBW = EP = CE (akwegwe) and [man speaking] yB = cross-cousin’s H (mpwao) on Yao «reasoning that if a man does not marry his cross-cousin then his younger brother will» (Mitchell 1956: 199). According to Mitchell (1951: 336) the roots of the term akwelume (mother’s brother) «seem to be that of akwegwe (parent-in-law) and mlume male and not as Radcliffe-Brown [1924] pointed out for other African tribes composed of the roots for mother and male.» In Laumanns’comparative study of Bantu kin terms kwe is a Proto-Bantu (PB) root for affine (1941: 32). The PB roots for mother are ma and nina.

14Consistent with the assumption of cross-cousin marriage, Yao cross-parallel reckoning for remote relatives in the +1, 0, and -1 generational levels is Dravidian:

  • in the +1 level:
    MMBS, MFZS = F
    FMBD, FFZD = M
    MMBD, MFZD = FZ
    FMBS, FFZS = MB
  • in the 0 (ego’s level):
    MFZDC, MMBDC, FFZSC, FMBSC = FZC, MBC
    MFZSC, MMBSC, FFZDC, FMBDC = FBC, MZC
  • in the -1 level:

15MBDC, FZDC, (man speaking), MBSC, FZSC, (woman speaking) = C

16The Yao classification of third cousins is apparently Dravidian as well. Thus MMMBDDD is a cross-cousin (Mitchell 1956: 201) consistent with the classification of these relatives worked out by Tjon Sie Fat (1998: 69). (This is easily done using the standard tetradic diagram [Allen 1986].)

Discussion

Bantu parallels and variations

17Yao is not the only Bantu kinship system with Dravidian crossness. In a paper published in 1965 De Sousberghe anticipated recent research on cross-parallel variations in kinship terminologies by comparing the classification of second cousins in two Bantu systems, in Burundi and Rwanda as described by Delacauw (1936) and Kagame (1954). Burundi and Rwanda are neighboring, closely related Interlacustrine Bantu regions numbering around 2 000 000 speakers each (Murdock 1959). Societies in both regions were stratified into distinct classes or castes with agricultural, pastoral, hunting and commercial economies. Descent was patrilineal and residence patrilocal. Succession was by primogeniture. In both cases there was an expressed preference for bilateral cross-cousin marriage.

18In the Burundi kinship system the preference for cross-cousin marriage was expressed in a Dravidian classification of all second cousins as either marriageable «cousins» bavyarawe or unmarriageable «sisters» bashikwe. As explicated by De Sousberghe:

«[…] les enfants des cousins croisés de sexes opposés se retrouvent dans la même relation que les cousins parallèles: assimilés aux siblings, conjoints prohibés. Mais, d’après un autre principe de la coutume, les cousins parallèles, s’ils sont de sexe opposé, donnent naissance à des cousin croisés, tout comme les siblings de même père et mère» (De Sousberghe 1965: 398).

19The Burundi classification as modeled by Delacauw (1936: 337) and slightly redrawn by De Sousberghe (1965: 402) is shown in fig. 1. The classification of third cousins as illustrated in fig. 2 from de Sousberghe is also Dravidian. Delacauw and de Sousberghe do not give any other kin terms.

Figure 1. Cross-cousins and their descendants in Burundi after P. Delacauw (1936): 46, 48, 52, 58, 60, 68, 72, 74, 84 are bashikwe “sisters” prohibited spouses marked by a line beneath the number; 44, 50, 54, 56, 62, 66, 70, 76, 80, 82 are bavyarawe “cousins” with whom marriage is permitted (from L. de Sousberghe 1965).

Figure 1. Cross-cousins and their descendants in Burundi after P. Delacauw (1936): 46, 48, 52, 58, 60, 68, 72, 74, 84 are bashikwe “sisters” prohibited spouses marked by a line beneath the number; 44, 50, 54, 56, 62, 66, 70, 76, 80, 82 are bavyarawe “cousins” with whom marriage is permitted (from L. de Sousberghe 1965).

Figure 2. First, second and third cross-cousins in Burundi (from L. de Sousberghe 1965)

Figure 2. First, second and third cross-cousins in Burundi (from L. de Sousberghe 1965)
  • 5 «Unless they are related to ego in some other ways as well» (Reay 1959: 65).

20Soon after Lounsbury (1964) discovered the Iroquois-Dravidian distinction, Scheffler (1971) noted a third variation of cross-parallel classification – the Kuma type named after a society in Highland New Guinea. In Kuma, «all descendants of all the cross-cousins are “cross-cousins”» (Reay 1959: 65)5. The Rwanda classification of cross-cousins is Kuma in type as expressed in the adage: «les cousins croisés et leurs descendants engendrent des époux l’un pour l’autre» (De Sousberghe 1965: 400). In contrast to the Burundi relation bavyarawe:

« […] la [Rwanda] relation babyara et conjoints possibles se transmet invariable entre les descendants du couple initial babyara (ceux-ci cousins croisés au sens strict que lui donnent les anthropologues); cette relation s’étend et s’établit entre eux tous à générations égales ou inégales et à quelque génération qu’ils appartiennent».(Ibid.: 397)

21In fig. 3 any man or woman A or a can marry any man or woman B or b. As in Kuma, intergenerational cross-cousin marriages are permitted. (In practice Rwanda men marry a same or lower generation cross-cousin [Ibid.: 400].)

Figure 3. Cross-cousin relations (ababyara) in Rwanda from L. de Sousberghe (1965: 399)

Figure 3. Cross-cousin relations (ababyara) in Rwanda from L. de Sousberghe (1965: 399)
  • 6 Tjon Sie Fat (in Godelier, Trautmann & Tjon Sie Fat 1998) has modeled all 16 logically possible var (...)

22Most recently, kinship theorists have distinguished five different types of crossness: Dravidian, Iroquois, Kuma, Yafar and Ngawbe (Godelier, Trautmann & Tjon Sie Fat 1998; Viveiros de Castro 1998)6. Presumably, some Bantu terminologies are actually Iroquois in type which means that at least three of the five recognized variants of cross-parallel classification are found in societies belonging to the Bantu language family. With close to 500 Bantu languages (Ruhlen 1987) it would not surprise to eventually find the remaining variants.

23The three known variants of Bantu cousin classification are shown in fig. 4 using a model from Viveiros de Castro (1998) based on Scheffler (1971). G+2, G+1, G0 refer to generational levels; 1 and 0 refer to relative sex, opposite or same respectively, of two siblings in G+2, of their children in G+1, and the crossness or marriageability of cousins in G0. The apparent cognate relation of Burundi and Rwanda cross-cousin terms (bavyarawa and ababyara) and the pattern of 1s and 0s in fig. 4 suggest an evolutionary sequence based on single binary steps: either Dravidian Kuma ? Iroquois consistent with tetradic theory, or Iroquois ? Kuma ? Dravidian. Determination of the actual sequence would depend on a historical linguistic reconstruction of Proto-Bantu cousin terms and their specifications.

Figure 4. Types of crossness in Dravidian, Kuma and Iroquois kinship systems (from Viveiros de Castros 1998)

Figure 4. Types of crossness in Dravidian, Kuma and Iroquois kinship systems (from Viveiros de Castros 1998)

Alternate generations

24A tetradic kinship system has alternate generation equations that merge certain relations from one even generation with certain relations from other even generations and correspondingly for odd generations, e. g. PP = CC, MB = ZC. In a Kariera system as described by Radcliffe-Brown (1931): FF = SC, MM = DC, FM = SC, MF = DC. In tetradic theory alternate generation equations may continue for some time after the disappearance of marriage classes. In an essay on the «assimilation of alternate generations» in self-reciprocal kin terms, marriage practices and cultural beliefs such as reincarnation, Allen (1989) cites Mauss’s suspicion, based on data from Burkina Faso, of the existence in Africa «of something resembling Australian marriage classes, or what is more or less the same thing, a quadripartite tribal organization (two moieties, each divided in two, no doubt by generation» (Mauss 1968-1969: 20-21; Allen 1989b: 53).

25Alternate generation equations occur in three contexts in Yao kinship terminology:

  1. a single term, akwegwe, is used for EP and CE;
  2. a woman is succeeded by her DD:«A woman’s daughter’s daughter inherits her name, her position and her property… When this has taken place the terms of address of her parental generation are now reversed. In other words her mother’s brother instead of calling her “sister’s child” (cipwa) calls her “mother” (amao). Sometimes the qualificative wamauja (“who has returned”) is used to signify the difference. Her mother’s mother’s husband, even before the succession has taken place calls her “my wife” and I have recorded one or two marriages between a man and his wife’s daughter’s daughter. A man’s daughter’s daughter must also perform some of the funeral rites for her grandfather after his wife has died».(Mitchell 1956: 175)
  3. When a man of a dominant lineage brings his wife to live in his own village (contravening the rule of matrilocal residence) the kin terms applied by members of the dominant lineage to members of the dependent lineage may have an alternating pattern. In the dependent lineage all members of the +1/-1 generations are «father» or «female father» and «child»; all members of ego’s generation are «cross-cousins», and all members of the +2/-2 generations are «grandparents» and «grandchildren». The members of the +2/-2 generations may be merged and called «cross-cousin» (Ibid.: 201-202).

26The identification of alternate generations is marked in various ways in many other African societies (Douglas 1952; De Heusch 1981).

***

27Contrary to common belief Dravidian kinship systems are not absent from Africa. They are found in Yao and probably in Burundi and may be present in other Bantu- or Niger-Kordofanian-speaking societies. Murdock’s Ethnographic Atlas (1967) contains a large number of Bantu societies with «Iroquois» kinship terminologies (59 % by Kuper’s 1982 count). These societies and others, not represented in the Ethnographic Atlas, should be studied from the structural and historical perspective of this paper.

28The unevenness of the Dravidian-Iroquois distribution is not a conclusive argument against universal theories of kinship evolution. For example, there are 20 million Mayan speakers but only one (very small) Mayan society, Northern Lacondon, is known to have a Dravidianate kinship system (Boremanse 1979). Early Mayan society, however, was Dravidian or Kariera in type (Eggan 1934; Hage 2003).

  • 7 Vansina’s reconstruction (1990) of Proto-Western Bantu culture, based on Guthrie (1969-1971), gives (...)
  • * C’est peu de temps après la composition typographique de son article que nous avons appris la mort (...)

29Bantu is a well-defined subgroup of the Niger-Kordofanian language family (Ruhlen 1987), with Proto-Bantu dating to sometime between 3 000 and 2 000 BC. The Proto-Bantu kinship system could be reconstructed using the comparative method of historical linguistics just as linguists have done for proto-kinship systems of similar antiquity, for example Proto-Algonquian (Hockett 1964) and Early Austronesian (Blust 1980) (both of which are prescriptive in type). Only when the Proto-Bantu kinship terminology is reconstructed will we know whether the Yao and Burundi systems are continuations of a Dravidianate system consistent with tetradic theory or endogenous developments contrary to tetradic theory7. It is suggestive that the Proto-Bantu term *dúmè means «male», «brother» and also «husband» and «maternal uncle» (Guthrie 1969-1971: 1182-1184). Sister’s daughter marriage is a common correlate of Dravidian kinship systems (Lévi-Strauss 1969; Good 1980; Trautmann 1981)*.

Haut de page

Bibliographie

Allen, N. J., 1986 «Tetradic Theory: An Approach to Kinship», Journal of the Anthropological Society of Oxford 17: 87-109.
—, 1989a «The Evolution of Kinship Terminologies», Lingua 77: 173-185.
—, 1989b «Assimilation of Alternate Generations», Journal of the Anthropological Society of Oxford 20: 45-55.
—, 1998 «The Prehistory of Dravidian-Type Terminologies», in Maurice Godelier, Thomas Trautmann & Franklin Tjon Sie Fat, eds, Transformations of Kinship. Washington DC, Smithsonian Institution Press: 314-331.

Bivar Segurado, Joaquim, 1989 «L’Emprise de l’échange restreint en Afrique centrale matrilinéaire», L’Homme 29: 44-75.

Blust, Robert, 1980 «Early Austronesian Social Organization», Current Anthropology 21: 205-247.

Boremanse, Didier, 1979 «L’alliance prescriptive et la nomenclature de parenté des Lacandon septentrionaux», Journal de la Société des Américanistes 66: 265-283.

De Heusch, Luc, 1981 Why Marry Her? Society and Symbolic Structures. Cambridge, Cambridge University Press.

Delacauw, A., 1936 «Droit coutumier des Barundi», Congo 17: 332-357, 481-522.

De Sousberghe, L., 1965 «Cousins croisés et descendants: les systèmes du Rwanda et du Burundi comparés à ceux du Bas-Congo», Africa 35: 396-421.

Douglas, Mary, 1952 «Alternate Generations Among the Lele of the Kasai, South West Congo», Africa 22: 59-65.

Dumont, Louis, 1970 «The Kariera Kinship Vocabulary: An Analysis», in Jean Pouillon & Pierre Maranda, eds, Échanges et Communications, Mélanges offerts à Claude Lévi-Strauss. Paris-La Haye, Mouton: 276-286.
—, 1975 Dravidian et Kariera: l’alliance de mariage dans l’Inde du Sud et en Australie. Paris-La Haye, Mouton.

Eggan, Fred, 1934 «The Maya Kinship System and Cross-Cousin Marriage», American Anthropologist 36: 188-202.

Godelier, Maurice, Thomas Trautmann & Franklin Tjon Sie Fat, eds, 1998 Transformations of Kinship. Washington DC, Smithsonian Institution Press.

Good, Anthony, 1980 «Elder Sister’s Daughter Marriage in South Asia», Journal of Anthropological Research 36: 474-500.

Guthrie, Malcolm, 1969-1971 Comparative Bantu: An Introduction to the Comparative Linguistics and Prehistory of the Bantu Languages. Farnborough, Gregg Publishers, 4 vol.

Hage, Per, 2003 «The Ancient Maya Kinship System», Journal of Anthropological Research 59: 5-21.

Hage, Per & Frank Harary, 2002 «Qué es un hipercubo? Un codigo binario para relaciones de parentesco», in J. G. Mendiata & S. Schmidt, eds, Analisis de redes: applicaciones en ciencas sociales. México, Instituto de Investigaciones en Matematicas Applicadas y en Sistemas de la Universidad Nacional Autónoma de México: 15-22.

Hockett, Charles, 1964 «The Proto-Algonquian Kinship System», in W. H. Goodenough, ed., Explorations in Cultural Anthropology: Essays in Honour of George Peter Murdock. New York, McGraw-Hill Books: 239-257.

Kagame, Alexis, 1954 Les Organisations socio-familiales de l’ancien Rwanda. Bruxelles.

Kryukov, M. V., 1998 «The Synchro-Diachronic Method and the Multidirectionality of Kinship Transformations», in Maurice Godelier, Thomas Trautmann & Franklin Tjon Sie Fat, eds, Transformations of Kinship…: 294-331.

Kuper, Adam, 1982 Wives for Cattle: Bridewealth and Marriage in Southern Africa. London-Boston, Routledge & Kegan Paul.

Laumanns, Grete, 1941 Verwandtschaftsnamen und Verwandschaftsordnungen im Bantugebiet. Lippstadt, C. Laumanns.

Lévi-Strauss, Claude, 1969 The Elementary Structures of Kinship. Boston, Beacon Press.

Lounsbury, FLoyd G., 1964 «The Structural Analysis of Kinship Semantics», in H. G. Lunt, ed., Proceedings of the Ninth International Congress of Linguists. La Haye, Mouton: 1073-1093.

Mauss, Marcel, 1968-1969 Œuvres. Paris, 3 vol.

Mitchell, James Clyde, 1951 «The Yao of Southern Nyasaland», in Elizabeth Colson & Max Gluckman, eds, Seven Tribes of British Central Africa. London, Oxford University Press.
—, 1956 The Yao Village: A Study in the Social Structure of Malawian People. Manchester, Manchester University Press.

Murdock, George Peter, 1959 Africa: Its People and Their Culture History. New York, McGraw-Hill.
—, 1967 Ethnographic Atlas. Pittsburgh, University of Pittsburgh Press.

Needham, Rodney, 1967 «Terminology and Alliance: II Mapuche; Conclusions», Sociologus 17: 39-54.

Parkin, Robert, 1998 «Dravidian and Iroquois in South Asia», in Marcel Godelier, Thomas Trautmann & Franklin Tjon Sie Fat, eds, Transformations of Kinship…: 252-270.

Radcliffe-Brown, Alfred Reginald, 1924 «The Mother’s Brother in South Africa», South African Journal of Science 22: 542-555.
—, 1931 The Social Organization of the Australian Tribes. Sydney («Oceania Monograph» 1).
—, 1953 «Dravidian Kinship Terminology», Man 53: 169.

Reay, Marie, 1959 The Kuma: Freedom and Conformity in the New Guinea Highlands. Carlton, Melbourne University Press.

Rivers, William H. R., 1906 The Todas. London.
—, 1907 «The Marriage of Cousins in India», Journal of the Royal Asiatic Society: 611-640.
—, 1914 Kinship and Social Organization. London, Constable & Co.

Ruhlen, Merritt, 1987 A Guide to the World’s Languages. Stanford, Stanford University Press.

Scheffler, Harold, 1971 «Dravidian-Iroquois: The Melanesian Evidence», in Lester Hiatt & Chandra Jayawardena, eds, Anthropology in Oceania: Essays Presented to Ian Hogbin. Sydney, Angus & Robertson: 231-254.

Tjon Sie Fat, Franklin, 1998 «On the Formal Analysis of “Dravidian” “Iroquois” and “Generational” Varieties as Nearly Associative Combinations», in Marcel Godelier, Thomas Trautmann & Franklin Tjon Sie Fat, eds, Transformations of Kinship…: 59-93.

Trautmann, Thomas, 1981 Dravidian Kinship. Cambridge, Cambridge University Press.
—, 2000 «India and the Study of Kinship Terminologies», L’Homme 154-155: 559-571.
—, 2001 «The Whole History of Kinship in Three Chapters: Before Morgan, Morgan and After Morgan», Anthropological Theory 2: 268-287.

Trautmann, Thomas & Robert H. Barnes, 1998 «“Dravidian” “Iroquois” and “Crow-Omaha” in North American Perspective», in Maurice Godelier, Thomas Trautmann & Franklin Tjon Sie Fat, eds, Transformations of Kinship…: 27-58.

Vansina, Jan, 1990 Paths in the Rainforest: Toward a History of Political Tradition in Equatorial Africa. Madison, University of Wisconsin Press.

Viveiros de Castro, Eduardo, 1998 «Dravidian and Related Kinship Systems», in Maurice Godelier, Thomas Trautmann & Franklin Tjon Sie Fat, eds, Transformations of Kinship…: 332-385.

I wish to thank N. J. Allen for helpful comments on this paper. Thanks as always to Ursula Hanly for the figures.

Haut de page

Notes

1 Bivar Segurado’s analysis (1989) of restricted exchange and preferential patrilateral marriage exchange in Yao does not pose the Dravidian-Iroquois question.

2 Some kinship theorists treat Dravidian and Kariera as essentially the same, the essential feature being the absence of terms for relatives by marriage (Radcliffe-Brown 1953). With one major…/…
[Suite de la note 2] exception (Scheffler 1971), most kinship specialists are in agreement about the association of Dravidian kinship terminology with cross-cousin marriage.

3 The following abbreviations are used in this paper: P parent, F father, M mother, G sibling, B brother, Z sister, C child, S son, D daughter, E spouse, H husband, W wife, ss same sex as ego, os opposite sex from ego, man speaking, woman speaking, + and - for ascending and for descending generations, 0 for ego’s generation.

4 Mitchell gives two terms for (man speaking) yB without explanation. He does not give a term for W or a term for (woman speaking) BC. ZD may also be called amao and ZDD cemwali as described below. Women call all Bs acimweni «eB».

5 «Unless they are related to ego in some other ways as well» (Reay 1959: 65).

6 Tjon Sie Fat (in Godelier, Trautmann & Tjon Sie Fat 1998) has modeled all 16 logically possible variants of crossness as a hypercube whose nodes represent types of crossness and whose edges represent evolutionary paths between them. Dravidian and Iroquois are maximally far from each other with Kuma occupying an intermediary position. For a presentation of the hypercube as a structural model in kinship analysis see Hage & Harary (2002).

7 Vansina’s reconstruction (1990) of Proto-Western Bantu culture, based on Guthrie (1969-1971), gives only the most basic meanings of a limited set of kin terms.

* C’est peu de temps après la composition typographique de son article que nous avons appris la mort de Per Hage; il ne put ni le parcourir ni éventuellement le corriger. Son collègue et ami, Bojka Milicic, accepta de relire les épreuves et, à notre demande, rédigea le texte que nous publions plus loin en hommage à Per Hage, évoquant, avec émotion et admiration, ce que furent sa vie, sa carrière et ses travaux. Nous tenons à le remercier, ainsi que l’assistante de Per Hage, Ursula Hanly. Ndlr.

Haut de page

Table des illustrations

Titre Table 1. Cross-parallel classification of second cousins in Dravidian and Iroquois systems
URL http://lhomme.revues.org/docannexe/image/21745/img-1.jpg
Fichier image/jpeg, 42k
Titre Table 2. Yao kin terms (from Mitchell 1951, 1957)
URL http://lhomme.revues.org/docannexe/image/21745/img-2.jpg
Fichier image/jpeg, 119k
Titre Figure 1. Cross-cousins and their descendants in Burundi after P. Delacauw (1936): 46, 48, 52, 58, 60, 68, 72, 74, 84 are bashikwe “sisters” prohibited spouses marked by a line beneath the number; 44, 50, 54, 56, 62, 66, 70, 76, 80, 82 are bavyarawe “cousins” with whom marriage is permitted (from L. de Sousberghe 1965).
URL http://lhomme.revues.org/docannexe/image/21745/img-3.jpg
Fichier image/jpeg, 88k
Titre Figure 2. First, second and third cross-cousins in Burundi (from L. de Sousberghe 1965)
URL http://lhomme.revues.org/docannexe/image/21745/img-4.jpg
Fichier image/jpeg, 12k
Titre Figure 3. Cross-cousin relations (ababyara) in Rwanda from L. de Sousberghe (1965: 399)
URL http://lhomme.revues.org/docannexe/image/21745/img-5.jpg
Fichier image/jpeg, 30k
Titre Figure 4. Types of crossness in Dravidian, Kuma and Iroquois kinship systems (from Viveiros de Castros 1998)
URL http://lhomme.revues.org/docannexe/image/21745/img-6.jpg
Fichier image/jpeg, 42k
Haut de page

Pour citer cet article

Référence papier

Per Hage, « Dravidian Kinship Systems in Africa », L’Homme, 177-178 | 2006, 395-407.

Référence électronique

Per Hage, « Dravidian Kinship Systems in Africa », L’Homme [En ligne], 177-178 | 2006, mis en ligne le 01 janvier 2008, consulté le 28 avril 2017. URL : http://lhomme.revues.org/21745 ; DOI : 10.4000/lhomme.21745

Haut de page

Auteur

Per Hage

University of UtahSalt Lake City (États-Unis)

Haut de page

Droits d’auteur

© École des hautes études en sciences sociales

Haut de page
  • Revues.org