In the first place, compound size in the homeland is much more sensitive to household growth than to household shrinkage. For instance, when a new wife enters the household, it is almost always possible to erect a new hut for her, but the hut will rarely be torn down when she leaves. A hut may be maintained as an extra bedroom, kitchen, or storage area, requiring only that the roof be rethatched periodically. Therefore, compound size correlates with household size only when households are at the peak expansion of their developmental cycle (Stone 1983). Small households might have anywhere between 0 and 6 extra rooms. Plains households spend much more of their developmental cycle in the stage of greatest expansion. They have an "impacted" developmental cycle, due to shortage of the land needed for household fission (Stone et al. 1984).
Before the days of migrant farming, there was a near-perfect correspondence between the settlement unit of the compound and the social unit of the household. The exceptions were few. Sometimes a household fragment (such as a wife and an adolescent son) would move into a recently inherited farm and begin to work it while continuing to participate in other ways in the household's affairs. Also, women who moved in with a wumulak (medical/religious practitioner) to give birth often remained there for several years, acting as partial members of both the wumulak's and their own households. Otherwise, the co-resident group on the farmstead was the social unit of production, and the great majority of farm labor was accounted for by household members working on the farmstead or its outlying fields.
This changed when households began sending some of their ranks to migrant farms to the south. In most cases, the migrant farmers were still very much members of the homeland household.
The Neighborhood
Groups of farmsteads form neighborhoods the Kofyar call toenglu. Lu is a house or home, and toeng is a tree or plant; the name connotes a discrete, organically connected group of residences. Each toenglu is a spatially contiguous, named area.
In the hills, toenglu are often separated by natural boundaries or by unfarmed areas. There is an average of 7 farmsteads per toenglu in the hills, with a range of 3--20. The residents of a toenglu, at least in the hills where Netting made detailed maps, tend to be partly localized lineages (Netting 1968:146--147).
The toenglu serve only infrequently and informally as a forum for mobilizing labor. The toenglu might be organized for a roof-thatching party, or a small toenglu might join forces with the neighboring toenglu; "occasionally a man may invite a particular neighborhood to work for him, but usually the invitation was an open one, and anyone in the village who is not otherwise occupied may come" (Netting 1968:138).
There are groupings that might be considered toenglu on the plains, but I believe they are too large to be the functional equivalent of hill toenglu, and I discuss them below.
The Village and Supravillage
The village is a fundamental unit of settlement in the Kofyar hills, although the language lacks a term for it. Villages are typically made up of 4--7 toenglu with 25--50 households. The farmsteads are usually adjoining, but some may be scattered, depending on local topography and population. Most hill villages are separated from their neighbors by natural features and are readily discernable on air photos, but some have shared boundaries.
The hill village was a military unit in that the village always fought together, and decisions about whether or not to wage war were made by the village male elders (daskagam). It was an economic unit in that labor for mar muos could be drawn from throughout the village. The hill village was a political unit, although villages were not all taxonomically equivalent; some had their own chief, but some recognized another village's chief.
On the plains, villages are contiguous areas of settlement, often lacking physical boundaries, with populations of 31--363. These plains villages are in turn clustered into larger units, a taxonomic level also lacking a name in the local language but which I will call the supravillage, having populations of 265--1503. The supravillages are Kwa, Miket, Lardang, Doka, Doemak, Kwang, and Kwalla. Fogol also might be considered a supravillage. It is now considered a section of Kwa, although it is much more populous than other Kwa villages and probably more populous than the supravillage of Kwang. It now pays its taxes through Miket and formerly paid through Kwang. It also appears to have joined the hill communities in battles (Netting 1974). Temple wrote in 1919 that "though now incorporated the Mikiet and Lardang were formerly independent, and the Mikiet dialect differs from that of the Mirriam" (p. 276).
Supravillages are partly the result of colonial sociosettlement categorizations. Like colonialists elsewhere, the British saw formalized political hierarchies as indispensable for taxing and governing indigenes, and they enhanced or imposed hierarchy within the Kofyar sociosettlement pattern. For instance, the supravillage of Lardang, comprising the hill community of Goetoeng and the plains communities of Kongde and Lumuat did not always function as a decision-making unit before the colonial period. When taxes were imposed by the fledgling colonial government in 1913, the plains portion of Lardang acquiesced, whereas Goetoeng refused (and suffered a vicious punitive patrol that took 13 lives).
The Warfare Alliance
 |
Netting described Kofyar militarism in some detail (Netting 1973, 1974, 1987). He portrayed the Kofyars' historic warfare alliances called sargwat ("shield arm"), each consisting of a group of villages that defended each other during hostilities (Netting 1974). The Doka, Plains Merniang, Plains Doemak,9 and Kwalla sargwat were based on the plains, whereas Jipal and Ganguk were made up of hill communities. The Gankogom sargwat was mainly in the hills but also included the plains subvillages of Lumuat, Kongde, and Kwanoeng (fig. 4.2).10
Sargwat are not formal entities. Although some were named, they have no officers, regular meetings, or social functions outside of the armed conflicts that had died out by the 1940s. Their membership was not even completely firm. They were often relative entities, activated by opposition, and not immune to the possibility of internal fission and raiding. As in segmentary lineage coalitions, there was no durable leadership. Thus, even though few young Kofyar adults today could tell you which sargwat their grandfathers fought with, these divisions do reflect social affiliation, or social propinquity, as will be apparent in the analysis of marriage patterns.
The Tribe
Colonial governments in general promote hierarchy in indigenous sociosettlement systems, including both formalization of informal or contested indigenous hierarchies and imposition of hierarchies where none existed before. This hierarchical imperative was especially important in the British colonization of sub-Saharan Africa, with its explicit reliance on indirect rule. Fried (1975) believed "tribes" to be a characteristic product of interaction with colonialists, as is well illustrated by the "mental map of . . . neatly bounded, homogeneous tribes" (Ambler 1988:32) that came to characterize British colonial policy in Africa. Colonial records attest to official satisfaction in cases where "tribal organization had been created" out of "a very disorganized state" (Chanock 1985:112; Berry 1993:28).
It is clear that the "tribe" was the operative unit of social affiliation and classification to colonial British administrators and scholars (Temple 1919), who divided the area shown in figure 4.2 into the tribes of Doemak ("Dimmuk" to the British), Merniang ("Mirriam"), Kwalla ("Kwolla"), Doka ("Jorto"), and Jipal (Ames 1934). The taxonomy was reified by taxation and by the instituting of customary courts, and sociosettlement units considered themselves subordinated to the units that collected their taxes:
"In Shendam, where a payment of train ('Zakka') is due from all farmers to their administrative superior, a lessor or claimant to fallow taken up gets half the amount due. (Incidentally this payment sheds an interesting sidelight on the traditional administrative organisation about which so much inconclusive argument goes on. Trace the 'Zakka' and you get the line of subordination.)" (Rowling 1946:29).
Although the colonial tribalization corresponded to indigenous social geography in the case of Jipal, Doka, and Kwalla, it caused problems in the heart of the homeland. In a reflection of the colonial administrators' bias toward the more accessible plains communities, hill villages were classified into tribes named for plains communities and forced to pay taxes through plains chiefs. The chief of Kofyar (who was symbolically the paramount chief) was therefore subordinated to a plains chief, and the Ganguk villages, labeled "hill Dimmuks" villages, were put under a Doemak chief who had held no sway in the hills before (although the administrators were aware of cultural differences between hill and plain and sometimes referred to the hill villagers as a "sub-tribe" [e.g., Josprof n.d.]).
Today, it is by these "tribes" that people from the area often self- identify when they are outside of the area, although the tribes are sociosettlement taxa that correspond poorly to both precolonial political organization and, as shown below, marriage geography. It is somewhat ironic that this level, which is so important by the criterion of self- identification, is in large part a colonial invention.
The sociosettlement unit of "the Kofyar" that I use in this book was defined by Netting on the basis of shared language, agricultural regime, social organization, origin myth, and various other cultural traits, contrasting with those of the neighboring Ron, Mwahavul, and Goemai (Netting 1968:35--43, 1974:146). For lack of a better term, I will call this the "inclusive tribe."11
FOOTNOTES
9. Plains Merniang refers to the sargwat comprising the Kwa-Kwang-Fogol area, and Merniang refers to the "tribe" comprising the Plains Merniang and Gankogom alliances. Similarly, Plains Doemak refers to the sargwat at the escarpment base, and Doemak refers to the "tribe" comprising the Plains Doemak and Ganguk alliances.
10. Gankogom means "the harmattan side," referring to the dust-laden winds that migrate annually from the Sahara, enveloping the area during the dry season. Ganguk means "the lee side."
11. Opinions vary on the propriety of the term tribe (Gandonu 1978; du Toit 1978; Sithole 1985), but the alternatives seem to pose more problems than they solve (e.g., Bromley 1978). I use the term here simply to label a group within which there is social propinquity relative to the surrounding populations. Patterns of social propinquity may be quite different from political organization and self-concept (Cornell 1988).