Robert E. Daniels
A paper presented at the 10th Annual Meeting of the Southern Anthropological Society,
April 3, 1975, Clearwater Beach, Florida

The Kipsigis are the largest of Kalenjin speaking groups of the Rift Valley Province of Kenya.1 At the turn of the century they probably numbered less than 100,000, with a settlement pattern of dispersed homesteads, an economy based on pastoralism and hoe agriculture, and an acephalous political organization involving a cyclical age-set system. Unlike most East African pastoralists, however the Kipsigis live in a highland environment and at the start of the colonial era they controlled an extensive area, only a fifth of which was used for permanent settlement (Pilgrim 1961:33). Thus, on the one hand they have experienced a transformation during the twentieth century similar to that of most highland groups, involving, among other things, a four- or five-fold increase in population due to the introduction of maize, plows, medicine, roads, etc., and resulting today in land enclosure, a peasant economy, and extensive involvement in the many institutions of the modern nation (Manners 1967). On the other hand, the Kipsigis have been able, through expansion into formerly open grazing land, to keep local densities in many areas close to the precolonial level of 150 to 200 persons to the square mile (Pilgrim 1961; Daniels 1970:25). Thus, they have been able to maintain, to a surprising degree, their pastoral base and the patterns of domestic organization derived from keeping cattle. 2


Most of us, I suspect, true to the old cliché, turn first to our own names when receiving a new telephone book or directory. I also suspect that a great many of us, like myself, have turned from time to time to that directory of societies, Murdock's Ethnographic Atlas to look up the way in which "our people" have been coded. I have repeatedly pondered the codings for the Kipsigis with a sense of uneasiness. This uneasiness does not stem from a feeling that the codes are wrong exactly, or that the coding categories per se are false, or from a rejection of the basic approach for which they were devised. The problem is that these codings, individually and in combination, fail to pin down or map certain key features of the social process. I would like to discuss this problem in terms of a standard topic: rules of residence.

Column 16 of the Ethnographic Atlas contains codings for "the prevailing profile of marital residence in the society" (Murdock 1967). The Kipsigis are coded "P Patrilocal, i.e. normal residence with or near the male patrilineal kinsman of the husband, cf. V Virilocal. Virilocal on further inspection, is defined as "equivalent to Patrilocal but confined to instances where the husband's patrikin are not aggregated in patrilocal and patrilineal kin groups." This suggests that the Kipsigis do have some sort of aggregation of patrilineal kin, but under Column 19, Community Organization, we find that they ate coded as living in "A Agamous communities without localized clans or any marked tendency toward either local endogamy or local exogamy." In other words, there appears to be some sort of local concentrations of particular patrilineal kinsmen but no aggregation of larger descent categories. Elsewhere Murdock notes that there are no 11neages among the Kipsigis. I have quarrel with any of this, though it does leave room for a great deal of ambiguity.

These codings are based on the primary ethnographies (Peristiany 1939a, 1939b, Huntingford 1953; see also Orchardson 1961) which do a decent job of presenting the normative rules of the indigenous system. But while descriptions on this level are highly informative (they are, after all, crystallizations of informants' explanations), the fact that they tend to be expressed in terms of individual decisions obscures fundamental patterns on the level of the community and wider social contexts. My argument is in two parts. First, a consideration of the norms or rules of family organization themselves reveal that the term partilocal (or any term of the same class) is inadequate to describe the "prevailing profile of marital residence." Second, the term is misleading in the sense that inspection of genealogical and census data from one area reveals patterns far different from those called to mind by the term patrilocal.


Pastoralism in East Africa means family herds, the association of specific people with specific animals. In these societies the definitions of rights and obligations in the idiom of kinship are, in large measure, a transform of the rights and obligations involved in the ownership and use of cattle. Since ownership comes first and foremost through inheritance, and pastoralism strongly implies patrilineality (Schneider 1964), continuity of the main line must be maintained in every possible case.

But while a family exists as a unit in relation to a set of cattle, it is critical to mote that the constraints of pastoralism do not require the physical presence of the owner of the herd, the geographical aggregation of the people sharing rights to the herd, or even the geographical aggregation or herding together of the cattle they share. Indeed, successful pastoralism requires a physical dispersal of both family members and family property. This while the Kipsigis strive to maintain continuity of family lines, the need not maintain continuity of specific localized domestic groups.

The cycle of the Kipsigis domestic group, the homestead, can be summarized as follows: almost all homesteads are founded by an adult man with his wife or wives and their children. Let us consider the case of a monogamous family first. As the sons of the “house“ come of age, each marries in turn, usually in his early 20s, and in almost every case of a first marriage is joined at his residence by the bride. One son in the family will remain permanently on his natal homestead, establishing his wife and children as a separate household there. All other sons are expected sooner or later to establish separate homesteads elsewhere. This can be in the same community, in nearby communities, or indeed anywhere else in Kipsigisland or even among the closely related Nandi to the north. Marriage involves considerable expense, including the transfer of about eight head of cattle (tug'ap koito) to the bride's family. Most families cannot afford to establish a son's separate homestead immediately after his marriage (which requires assigning at least a few other cows to support the new wife, and form the beginnings of her sons' future inheritance). Technically the formal division of the father's property does not occur until after his death, possibly long after since the older sons should wait until their younger brothers (including any sons of their mother's co-wives) are adults capable of protecting their own interests and agreeing to the division. However, when a son establishes a separate household he is capitalized in part with cattle from his father's herd, and depending on his distance from the father, exercises a fairly high degree of de facto control over the cattle which will eventually form the core of his de jure inheritance.

In keeping with the ethos of pastoralism, individual autonomy is highly valued (and closely guarded by such mechanisms as formality and distancing between father s and sons and generally between men of different generations [cf. LeVine 1973; Skinner 1961]). To be the fatherf's favorite who lives with him may mean a closer tutoring in the intricacies of managing the family property and being entrusted with the knowledge of the father's private cattle loans and outstanding debts. But it also means being held in a position of dependence often quite explicit and public while one's brothers and others of one's age are independent in their affairs. Men are thus commonly motivated toward leaving their father's shadow and establishing their own homesteads.

Although land enclosure and individual ownership of plots now add a new significant to patrilocal residence,3 almost all families still follow the general rule of dispersal and there is a persistent feeling that homesteads should not be subdivided among sons if at all avoidable.

There are thus several immediate problems with the term patrilineal. Perhaps most sons at marriage do bring their first wives to the grooms' natal homesteads. But Kipsigis families are relatively large, A sample of 58 sets of brothers contained 188 men, or 58 who should live at the parental homestead, and 130, 69%, who would be expected to move elsewhere. Thus, most men subsequently do not reside within or adjacent to their fathers' homesteads.

It often happens, of course, that a man die before all his sons are married. In that case it is decided among his sons which will remain at the natal homestead. In most cases men die before their wives (who are almost always younger), but widows are not required to remain at their husband's homestead if they prefer to move in with another son living elsewhere. Thus, there are many men living at their natal homesteads but not with their fathers or even necessarily their mothers.

The situation is greatly complicated by polygyny. I estimate the rate of polygyny for men to be approximately 42% (98 out of a sample of 231 men mentioned in genealogies). Second and subsequent marriages usually occur years after first marriages and are most likely to take place after sons have sons have founded permanent residence, most of which will not be at their natal homesteads. Furthermore, as noted by Peristiany (1939) and others, and as discussed in detail by LeVine (1962), the Kipsigis prefer to place wives on separate homesteads, possibly miles apart. My data indicate that about half of all polygynists succeed in establishing two or more homesteads. In other words, somewhere between 30 and 45% of all married women have co-wives living elsewhere. In such cases, one son of each wife will remain at his natal home while the rather is free to divide his time between homesteads according to his wishes and his personal understandings with each wife. Thus, in roughly a third of all homesteads the father if alive, is only a part-time resident. Are the favorite sons in such cases to be judged as living patrilocally?

There are even further complexities. As mentioned above, the Kipsigis are concerned with the continuation of the male line in each house. This leads to leviratic marriage and woman marriage, both of which are common occurrences, as well as other remote alternatives. With the levirate, the widow remains married to her dead husband and bears him further children if possible. Thus, out of a sample of 58 men, 19, or a third, did not have living fathers by the time of their own initiations at adolescence. Are we to classify a married couple as living patrilocally of the husband lives on the parental homestead but his father is dead? Or if his father died years before his birth and no male kin have evert resided on the homestead during the young man's life? Can a man be said to live patrilocally if his pater is a woman? In sum, a consideration of the rules of domestic group composition make it clear that the label patrilocal, while the most readily perceived choice, cannot in fact be accurate in more than a minority of cases. But let us not conclude that the rest is chaos or simply sweep the majority of cases into a meaningless box labelled 'neolocal'.


Where do most sons go? Clearly into other communities. Here the peculiarities of the community I studied, Kapsuswek, offer an opportunity to answer some of the questions involved.

Kapsuswek is one primary community, or kokwet in an area known as Itembe. Itembe was briefly alienated from Kipsigis possession early in the twentieth century and granted to a British settler. Many Kipsigis continued to live in the area as "squatters"f on his estate. After his death, sometime before World War II, the area officially reverted to the status of Crown Land, but was treated by the Kipsigis like surrounding areas as open for communal grazing and homesteading by the expanding population. Finally, in the mid-1950s, the colonial government moved to regularize the settlement of Itembe, un part because of the rapid spread of land enclosure throughout the neighboring areas of the Native Reserve (Kenya Government 1956). An unrecorded number of "illegal squatters" were removed, and 311 plots of thirty acres each were surveyed, Forty-five percent were given to people already resident in Itembe, the rest were settled by people selected from surrounding areas. Thus, there was a large influx of people in the space of a couple years. But the situation does not appear to be all that different from the rapid increase in unofficial homesteading taking place during the same general period through the surrounding areas.

More significantly, as far as I was able to determine, the settlers were free to choose among many available plots, and the choices which men made were considered to have been a function of traditional Kipsigis values and not a function of external governmental decisions.

What happened? Most of the settlers chose farms next to relatives, in a few cases next to direct patrilineal kin (a brother or half-brother), occasionally classificatory kinsmen i.e. members of the same descent category but persons for whom genealogical connects and rights to cattle ownership are irrelevant), and in many cases next to direct or indirect affines. Thus, there were cases of a man settling next to his father-in-law, next to a sister and her husband, next to a brother's wife's brother, and so forth. Among the Kipsigis there is a warm relationship between men married to sisters.4 In Kapsuswek there was one case of three full sisters whose husbands settled on adjacent farms, and several cases of neighbors who were indirectly lemenye through relatives in other communities. This is considered a natural choice by the Kipsigis, but not what most readers would expect in a society described as “patrilocal”. In other words, men moving into new communities generally seek land adjacent or near non-agnatic kin. It profits a man and his sons to be related in many ways to many people in many places in order to maintain their flexibility and options in the face of pressures for dispersing from the natal homestead.

This also shed light on cases which I had originally thought to be anomalies: cases in which a man in community A marries a second or third wife in community B and established a separate homestead for her in B, her natal community.

But rather than risk over interpretation of fragmentary information, I would like to return to one last set of more solid data, the pattern of marriages involving Kapsuswek since settlement in 1956 through 1972. So far marital residence has been discussed in terms of sons. If we consider these marriages in terms of the women involved, an interesting pattern emerges. In all there had been 41 marriages involving member of the community during these 16 years. Fourteen of these marriages, or roughly a third, were between two families in the community. Another twelve involved daughters of community households who married out and fifteen involved young women who married into the community. Significantly, in all cases the marriage distance was no more than a few miles, or roughly an hour's walk between the bride's and groom's homes.

Table 1
The Movement of Brides in 41 Marriages
Involving Kapsuswek Community 1956 - 1972

12 married
OUT OF Kapsuswek
14 married
WITHIN Kapsuswek
15 married
INTO Kapsuswek

The residence pattern which emerges can be summarized as follows: from the man's point of view one's agnates are dispersed rather widely, and in the local community and surrounding neighborhoods one is involved in a complex and dense network of basically affinal relationships of several types. From the woman's point of view, one's close female consanguineals (mother, married sisters, married daughters) are dispersed within a much narrower range. Thus, women have more immediate relatives close at hand than do men. I believe this is a critical feature of Kipsigis society, and one which is obscured, I feel, by descriptive terms such as patrilocal or virilocal.

Space does not permit a discussion of the many implications of this aspect of Kipsigis organization. Suffice it to say there that this pattern makes a lot of sense in a society that must balance a major agricultural contribution from women's gardens in a fertile highland zone with the pastoral virtues of aggressive male dominance.


Anthropologists have been right, I believe, in seeing the patterns of residence as being fundamental to an understanding of the social processes in the communities they study. I hope that I have demonstrated, in this one case, that our traditional terminology is insufficient for the task, and that the approach I have suggested a look at the differential dispersal and aggregation of both male and female kin, may be one possible way in which to grasp the underlying processes that precipitate residence patterns.

  • 1 Research was conducted among the Kipsigis from August, 1965 to March 1968, and during June, 1972. Research was supported by a grant from the National Science Foundation and by the Child Development Research Unit of Nairobi, funded by the Carnegie Corporation of New York.
  • 2 The area in which this research was focused, grasslands on the edge of more fertile hills, supported 2.6 cattle per person in 1968. Evans-Pritchard reckoned the Nuer, whose herds had not fully recovered from rinderpest at the time of his study, owned slightly more than one head of cattle per person.
  • 3 Pilgrim (1961) suggests that before land enclosure the last-born son most frequently stayed with the parents but since land enclosure, eldest sons have tended to stay, forcing younger brothers to see land elsewhere. I have not been able to detect such trends in my data. In any case, there is no recognized rule of primogeniture or ultimogeniture.
  • 4 Men married to sisters address each other as lemenye. The relationship is marked by a sense of relaxed equality unlike regular affinal relationships which are asymmetrical (kapyugoi, a man of the house to which betrothal was made, i.e. wife-giver, and sandaniin, the suitor, the wife-receiver). The Kipsigis say that two men who are lemenye are allies because they have struggled to pay marriage cattle to the same father-in-law (and do not have cattle at issue between each other). By classificatory extension, the term can be used between any two men who have agnatic kin married to women of the same third descent category (oret/ortinwek). The analytical distinction between direct and indirect lemenye relationships is mine, and is not made by the Kipsigis in everyday behavior.
Daniels. R. E.
1970 By Rites a Man: A Study of the Societal and Individual Foundations of Tribal Identity among the Kipsigis of Kenya. Unpublished Ph.D. Dissertation: University of Chcago.
Evans-Pritchatrd, E. P.
1940 The Nuer. London: Oxford University Presss.
Huntingford, G. W. B.
1953 The Southern Nilo-Hamites. London: International African Institute
Kenya Government
1956 African Land Development in Kenya. Nairobi: Ministry of Agriculture, Animal Husbandry, and Water Resources
LeVine, R. A.
1962 Witchcraft and Co-wife Proximity in Southwestern Kenya. Ethnology 1:39-45.
1973 Patterns of Personality in Africa. Ethos 1:123-152
Manners, R. A.
1967 The Kipsigis of Kenya: Culture Change in a "Model" East African Tribe. In Contemporary Change in Traditional Societies, vol. 1, Introduction and African tribes. J. H. Steward, ed. Urbana: University of Illinois Press.
Murdock, G. P.
1967 Ethnographic Atlas: a Summary. Ethnology 6, no. 2.
Orchardson, I. Q.
1961 The Kipsigis. Nairobi: The Eagle Press. (abridged, edited, and partially re-written by A. T. Matson)
Peristiany, J. G.
1939a La Vie et la droit Coutimier des Kipsigis. Paris: Les Editions Domat-Montchrestien; 1st edition
1939b The Social Institutions of the Kipsigis. London: George Routledge & Sons, Ltd.
Pilgrim, J. W.
1961 The Social and Economic Consequences of Land Enclosure in the Kipsigis Reserve. Kampala: East Afrian Institute for Socail Research. Unpublished manuscript.
Schneider, H. K.
1964 A Model of African Indigenous Economy and Society. Comparative Studies in Society and History 7:37-55.
Skinner, E. P.
1961 Intergenerational Conflict among the Mossi: Father and Son. Journal of Conflict Resolution 5:55-60.