Africa: Journal of the International African Institute, Vol. 50, No. 3. (1980), pp. 305-310.

Africa, 50(3), 1980

THE DIFFUSION OF AGE-GROUP ORGANIZATION IN EAST
AFRICA:

A RECONSIDERATION

J. J. de Wolf

In 1962 LeVine and Sangree attempted to answer the question of why some Bantu-speaking peoples in Kenya and Tanzania had adopted the age-group organization and associated rituals of their non-Bantu neighbours, while other Bantu peoples in the same region had not. They examined the Tiriki, who adopted the age-set system of the Kalenjin-speaking Terik, and the Gusii, who are also Bantu and who were also in close contact with the Kalenjin-speaking Kipsigis, but who did not borrow their age-group organization . The authors suppose that the age-group organization of the Kalenjin and Maasai had military functions and that their Bantu neighbours whom they attacked and defeated may have decided to copy this institution, because they would have attributed Kalenjin and Maasai success in warfare to it (LeVine and Sangree 1962:97-8).

They see the problem primarily in terms of the requirements for survival in the face of inter-tribal warfare and suggest that three circumstances may explain the difference between Tiriki and Gusii. (1) The Gusii are much larger in size than the Tiriki, and also than the Kipsigis, while the Tiriki were a tiny group, even smaller than the Terik, when they adopted their age-set system¹. This meant that the Tiriki were in much greater need of allies than the Gusii. As they interacted on an individual basis with the Terik, they were also more likely to assimilate aspects of Kalenjin culture than the Gusii, who interacted with the Kipsigis as a group. (2) The Gusii were able to defeat or keep at a distance their major enemies, while for the Tiriki defence continued to be a serious problem, even after their alliance with the Terik. (3) The only allies available for the Tiriki were the Terik, who made initiation and age-group formation a condition of their alliance. In times of emergency the Gusii found allies among the Luo, who did not have age-groups or initiation ceremonies (LeVine and Sangree 1962:108).

LeVine and Sangree (1962:109) then formulate a more general hypothesis which is based on these considerations.

Given the general prowess of the Nilo-Hamites² and the ease with which age-groups can be borrowed, it seems likely than a Bantu group which did not have a clear advantage in terms of size, defensive capacity, or strong allies lacking age-groups, would tend to imitate the successful Nilo-Hamitic military organization and initiate far-reaching changes in their social structure.

They express the hope that further publications will contain enough historical material to test the validity of this hypothesis.

As new material has become available recently it is now possible to reconsider this solution to the problem of the diffusion of age-group organization. First I shall examine the comparison between Tiriki and Gusii in the light of the historical research

DIFFUSION OF AGE-GROUPS IN WEST AFRICA

307

of Ochieng' (1974) among the Gusii. Next I want to evaluate the supposed connection between age-set systems and military functions and success in warfare. Finally I shall suggest an alternative explanation to the adoption of Kalenjin age-set systems by Bantu tribes.

The Gusii were driven from the Kano plains near Kisumu, where they started their existence as a separate ethnic group, by the Luo who were always numerically and militarily superior to them (Ochieng* 1974:208). They found refuge in the Kericho area, where they held their own against Maasai from the Nyando valley (Ochieng* 1974:100-2). But when the Kipchoriat Maasai had been defeated by Kipsigis coming from the Rift Valley, Gusii were also driven away, with the exception of some groups who remained at Sotik and became Kipsigis (Ochieng* 1974:107-9). In the Trans-Mara triangle near Kilgoris the Gusii were attacked by the Isiria Maasai who defeated them heavily at the battle of Migori (Ochieng* 1974:111-2). When they tried to settle again near the borders with Luo, they met with strong opposition, while others were accepted as serfs by them. The latter revolted after the Maasai had been driven away by the Luo. They escaped with as much cattle as they could get hold of and set up their homes in Nyagoe Forest (Ochieng* 1974:141-8).

So it seems very doubtful whether the Gusii had less need for allies than the Tiriki and unlikely that the Luo can be called allies in times of emergency. The only time that Luo and Gusii combined forces against the Kipsigis was in 1891, but this was not a planned affair. The Kipsigis had become desperate for cattle after the rinderpest epidemic, so turned to raiding their neighbours. After doing much harm in northern Gusii country, they continued in great numbers into Luoland in order to replenish their stock. This gave the Gusii time to gather a large force from various sections of the country and ambush the returning Kipsigis as they were chased back by the Luo, in the famous battle of Saosao (Ochieng' 1974:128-32).

Another objection to the hypothesis of LeVine and Sangree arises from the use which they make of contemporary population statistics for estimating the relative strength of the Gusii in the past (LeVine and Sangree 1962:105, 108). Moreover, as the Gusii were divided into territorially discrete clans which fought one another as well as outsiders, it seems wrong to base the supposed military superiority of the Gusii on the total size of their populations (LeVine and Sangree 1962:106). One would also have thought that at least some Gusii would have found the Kalenjin age-set system a good organization for warfare and raiding against other Gusii, if the argument of LeVine and Sangree were valid.

Even though the association between age-sets and military organization is by no means universal (Baxter and Almagor 1978:17), such peoples as the Nandi, Kipsigis and Maasai found age-sets useful as ready sources of military manpower (Huntingford 1953:32, 48, 119). But it may well be that the military advantages of age-group organization are limited to offensive warfare and raiding in enemy territory. Goldschmidt (1976:105) tell us that the Sebei, a Kalenjin group on the northern flank of Mount Elgon, had an age-set system similar to that of the other Kalenjin. Typically all boys who are circumcised during a certain period form one named age-set. There are eight or sometimes seven names which recur in a fixed order. Members-^f the same set often change their age-group together, e.g. when warriors become elders. But the Sebei age-sets are not, and have not been within living memory, a salient element in the organization of political, legal or military affairs. New age-sets did not have the

308

DIFFUSION OF AGE-GROUPS IN WEST AFRICA

duty of warfare because defensive warfare, which seems to have been the rule, required the full support of all able-bodied men (Goldschmidt 1976:21).

The Bantu-speaking Bukusu who live to the south of Mount Elgon have an age-set system which is very similar to that of the Sebei and which also seems to lack any clear political, legal, or military function (de Wolf 1977:131-3). That this was also the case in the past becomes clear when we examine their traditional history. Bukusu claim that the practise of initiation into age-sets through circumcision was started by their ancestor Mango at Mwihala, which is probably Chelelemuk in what is now North Teso Location. Mango was the leader of the clans who are said to have come from Samia and Nyole country (where circumcision is still not practised). They used to live near Tororo in Uganda, but then they were chased by the Teso and had to cross Malaba river into Kenya (Were 1967:171-3). This happened in the early nineteenth century according to Bukusu age-set chronology.

Another group of Bukusu lived in those day at Ebwayi, which is now known as Amukura is South Teso Location. They had come directly from Gisu country in Uganda. It is not unlikely that they were already used to circumcising boys but it is very probable that they had no age-set system, as the Gisu do not have one. In fact the Bongomek, a small Kalenjin group known as Ng'oma to the Bukusu, claim that when these Bukusu came to Amukura they were already living there. They urged the newcomers to get circumcised and become initiated into their age-group organization, that they might live together in peace and security (Were 1967:18). However this may be, the initiation of Mango was certainly done at the instigation of Kalenjin, even according to Bukusu tradition (Were 1967:172).

When the Bukusu were living in the present-day Teso Locations they started to intermarry with a group of Kalenjin at Mwalie whom they call Yumbu. From them they seem to have taken over initiation of girls involving genital mutilation, normally referred to as 'female circumcision' (Were 1967:173)³. Around 1925 Bukusu gave up this custom, responding readily to a government campaign against it (Bryk 1928:65). The Tachoni, who are sometimes considered to be the descendants of the Yumbu of Mwalie, were less inclined to do so. This may well be the reason that Wagner (1949:336) reports that no other Luyia group but the Tachoni adheres to this custom. Yet my informants were adamant that this custom also existed among the Bukusu before 1925, which is confirmed by Bryk (1928:33)⁴. The language of the Kalenjin Yumbu gradually gave way to that of the Bukusu⁵.

In spite of the adoption of age-sets and their merger with the Yumbu of Mwalie the Bukusu were driven away from the area now inhabited by the Kenya Teso and from other areas in Bungoma District. They even had to seek refuge with the Nyala, Marachi and Kabras (Were 1967:174-5). It was only after Mukite, son of Nameme, had reorganized the army and Bukusu had acquired the art of fighting by night from the Maasai and adopted bow-and-arrow warfare from the Gisu, that they were able to drive the Teso back and settle again in their present locations in Bungoma District (Were 1967:177-88). Of course, when the Bukusu adopted the Kalenjin age-set system they could not have foreseen that its military advantages, if any, would have been of no use in their wars with the Teso. Yet there are no traditions which suggest that the Kalenjin Ng'oma and Yumbu were known for their success in warfare and raiding which could have induced Bukusu to adopt their age-group organization for the reasons given by LeVine and Sangree.

If the thesis of LeVine and Sangree seems to be doubtful in the light of the evidence

DIFFUSION OF AGE-GROUPS IN WEST AFRICA

309

presented here, we must try to find an alternative solution to their initial problem of why some Bantu adopted a non-Bantu age-group organization but others did not. In western Kenya only two cases of this adoption have been reported. They are the Tiriki and Bukusu⁶. In both cases there is a strong evidence that their Kalenjin neighbours were anxious that they should join their age-sets through initiation. If this is so, it would be better to cast the problem differently and ask the question: why did some Kalenjin allow Bantu neighbours to become members of their age-group organizations?

The Terik are said to have been motivated by their desperate need for military allies. They offered asylum to wandering or refugee segments of Luyia lineages, on condition that the men became incorporated into the Terik warrior groups through initiation into Terik age-groups. These Luyia speaking members of the tribe became known as Tiriki (ba-Diliji). Later these Tiriki initiated new groups of Luyia immigrants until the Bantu speaking section came to outnumber the Kalenjin-speakers. The Terik were probably able to maintain a separate identity because differences of belief concerning female initiation prevented intermarriage between the two groups. One may note that the migration and subsequent alliance of Luyia segments with the Terik was just another variation on the continuing process of Luyia lineage segmentation. Instead of either setting themselves up independently or putting themselves under the protection of another lineage in a different Luyia group they decided to join a Kalenjin community (LeVine and Sangree 1962:102). As we have seen, the conditions under which the Bukusu adopted the Kalenjin age-set system were rather similar.

Generalizing, then, we can conclude that Kalenjin communities would be willing to allow Bantu to become members of their age-organization and thereby grant the right to live among them (i) if they felt that their numbers were too small to continue to live undisturbed in the territory which they occupied, and (ii) if the Bantu were willing to acknowledge at least the ritual superiority of the Kalenjin through submitting themselves to their initiation rituals and become members of their age-sets. Only refugees and migrants in search of land were likely to do so.

NOTES

¹    In 1955 Tiriki Location had 40,000 inhabitants, less than 10% of whom were Terik (Sangree
1966:6). In 1969 there were 471,000 Kipsigis and 702,000 Gusii (Kenya Population Census 1969).

²    The term Nilo-Hamites has become obsolete and has been replaced by the term Nilotes, on the basis
of linguistic evidence. The Maasai are no longer grouped together with the Kalenjin as Southern Nilo-
Hamites (Huntingford 1953), but are now placed in the same sub-group as the Teso-Karimojong cluster
(Sutton 1968:96-8,80-1).

³    Unfortunately this is not clear from Were's translation which rendered khu-lichana as 'to cut up and
eat each other' instead of'to circumcise women'.

⁴    Roscoe (1915:186) reports that the same custom was practised by the Gisu, which was denied by La
Fontaine (1954:41). This discrepancy can easily be explained as Roscoe's visit was limited to areas near the
border with Kenya, where the culture is in many respects similar to that of the Bukusu (Roscoe 1915:161).

⁵    Were (1967:181) omitted this information when he summarized the original manuscript on the
history of the Bukusu written by the elders of Kimilili Location in 1955 (Were 1967:166). Ehret (1971:47)
identifies these Yumbu as Kitoki, who lived south of Mount Elgon, influenced Bukusu culture and
disappeared. His assertion that they gave up female circumcision must be revised in the light of the
evidence presented here (Ehret 1976:12).

⁶    Five of the names of Kalenjin age-sets are the same as the names of Kuria generation-classes.
However, the two systems of age-grouping are based on different principles or recruitment and function
within their respective societies in different ways (Ruel 1962:35).

310

DIFFUSION OF AGE-GROUPS IN WEST AFRICA