Chapter 3 (pt.4)
THE PRINCIPLE OF AGNATIC DESCENT
Fission and Fusion
In discussions of segmentary lineage systems, the terms fission and fusion have been used in two ways: First, to denote the continuous operation of identification by contraposition, in which agnatic groups are either linked together or separated according to the order of relationship which involves them in interaction. Secondly, these terms have been used to deal with the cumulative processes in time which produce new lineage segment identities, and incorporate new members and other demographic and political changes. This is the sense in which we sue the terms here, and it is largely the same as what Fortes implies in his usage of the term 'developmental cycle'. I wish to place less weight on both the cyclic aspects of these corporate changes, and on a view of them as intrinsic to the corporation itself, as the word developmental implies.
By fission I mean quite simply the process by which corporations break into two or more parts, each of the same order as the original corporation (this is the central distinction from the first usage noted above). We shall look at it both in terms of the continuous variables or processes and the discrete variables or events through which it takes place. As such, fission occurs much more frequently in the lowest orders of the lineage organization, where groups are both smaller and less stable. Fusion, on the other hand, is much more frequent in the higher order corporations. It refers to the process by which morphologically distinct units are merged without differentiation into a single unit. We shall view both processes as essentially linear in time, since they are 'cyclic' only with regard to abstract concepts such as ' the family', or ' the lineage', as opposed to actual families or lineages.
Fission: I have described the two sorts of phenomena which we must consider here by analogy to discrete vs. continuous variables. This is only a partially accurate analogy. It correctly emphasizes the distinction between the particular occurrences in which fission becomes apparent and gains 'jural' recognition and those longer term processes and pressures which underly these events. It does not, however, adequately allow for some of the conditions of fusion which do operate with temporal continuity, but which are constant categories rather than variables. Paternal and maternal identifications are among these. We shall include these 'constants' as part of our discussion of the processes, that is, the continuous aspect of fission. Here we shall treat social maturity and the growth of the domestic family, economic and demographic pressures, and matrifocal identities and matrilateral influences. I recognize that the 'continuity' of these variables is, of course, an epiphenomenon of weak measurement in one sense. But I would emphasize that this is not simply a matter of error, avoidable or unavoidable, but is phenomenologically important. Tallensi are neither better equipped nor more apt than we are to find any discrete and absolute units by which to demarcate the operation of these variables. However important social maturity or demographic pressure, say, may be in producing a particular fission, there is never a simple answer to the question of how much is enough. It is always the conjunction of such 'forces' or processes with events which determines the actual timetable, along with such important adjuncts as the identity of a new lineage head.
Social Maturity and the Domestic Family: Some of the most obvious underlying processes leading to lineage segmentation are, together, simply the human life cycle. There are two coordinate aspects to this which are crucial. In one we see the growth of individuals in social maturity, and in the other the growth and size and sub-divisions of the domestic family. It must be obvious that for any one individual or group to increase in size or maturity, others must eventually yield place. In other words, while a segment, for example, grows with the growth of its parts up to a point, after that (indeterminate) point, the original group must either be weakened or dissolved in favour of the sub-groups. Further, while there is nothing about the natural processes of the life cycle which limits either the production or growth of new units in relation to old, the Tale morphological structure does so. It limits the occupants of positions and the parties to social relations.
Thus social maturity is not simply a matter of age or behaviour, or even of property or interpersonal relations, though it is influenced by each of these. Most clearly, social maturity is a matter of formal relations of agnatic kinship, which can only in part be abrogated by such intervening factors as age, wealth or intelligence:
a man's social status, particularly in jural and ritual matters, but also in economic affairs, advances step by step with the dying off of his fathers. Strictly speaking, he is not altogether autonomous until he has no more fathers within the range of the inner lineage, but in practice he is free of effective paternal control when he has no more fathers in the nuclear lineage. (1949a, p. 1957)
That this definition of social maturity does indeed extend to economic affairs, especially in the case of heritable property, is shown in the matter of debt:
Strictly speaking a son cannot contract a debt involving heritable property such as cattle or large sums of money without his father's authorization. Even if the son is a man of substance, a wise creditor will insist that his father's (and also his brothers' of the same nuclear lineage) consent to the transaction be obtained first and made known of him. If this is not done the creditor's chances of recovery are small. (1949a, p. 209)
A man, then, only achieves an element of autonomy in relation to his father; he does not achieve any outright categorical autonomy independently of his progressive 'fatherlessness'.
The social relations of father and son, at every stage of the latter's development, are always a compromise between the mutual dependence of father and son and the limited [in-] dependence enjoyed by the latter. (1949a, p. 172)
The son's independence from his father is in the first instance a function of his matrilateral ties, for of necessity these are different than those of his father. Later, this difference will be supplemented by different affinal ties when the son marries. The ties through his mother are of particular importance, however, as symbolized by the fact that a boy is given his start in animal husbandry not by his father, but by his matrilateral kinsmen (1949a, p. 194). This differentiation is further the primary distinction between brothers, as we shall consider shortly.
Matters such as social maturity and matricentral identification are of primary significance at the lowest levels of lineage organization; the relations of patrilineally defined corporate groups governing the segmentation process at higher levels (although the same language may be used by largely metaphorical extension to describe both levels of events). The forum in which these matters are most prominently at issue is the domestic family. As Fortes describes it:
The characteristic domestic family of the Tallensi is the agnatic joint family, and other forms of domestic group appear as phases in the growth cycle of a joint family . . . A joint family consists of two or more close agnates, members of a single effective minimal lineage in all but exceptional cases, and often comprising all the members of a particular effective minimal lineage, together with their wives, children, and other dependents, if any, occupying a single homestead. (1949a, pp. 64-5)
Although this unit has a strong ideology of solidarity and unity, it also has working within it equally strong tendencies toward fission. The former of course dominate at the outset, the latter toward the end. Frequently, however, and adjacent such as the death of the head of the family is needed to produce a formal split into new segments. Before that time, however, it is common for one or more of a set of brothers to hive off from the domestic family to set up independent residential (although not genealogically, that is jurally and ritually, independent) units.
The first step in this process of differentiation is the construction of separate quarters for the wife of a junior member of the compound, the primary indication that he is establishing his own household, though still within his father or older brother's compound and domestic family (Fortes and Fortes, 1936, p. 239). The next major step involves the cutting of an additional gate into the compound, to accommodate the increasing independence of one of the junior men, usually an ortho-cousin or nephew (classificatory brother or son) of the head of the household, or the first-born son of the head (op. cit.). Closer agnates and other sons are less likely to have separate gates. In particular, two full brothers (soog) frequently continue to farm together and live in a single gate homestead, perhaps even all their lives. Half brothers are likely to do so only until their sons are old enough for productive labour. "In time, however, the combined effects of economic needs, structural cleavages, and personal tensions lead to a further stage of separation." (1949a, p. 68) This is the 'semi-detached' stage of separate gates.
This is followed by the stage of complete spatial segregation, which occurs occasionally between the families of full brothers, often between those of half-brothers, and almost always between those of cousins (1949a, p. 69). These separate homesteads will usually be very close - within a couple of hundred yards of each other, but may be much more distant, if one family head goes to ask land from a cognate in another area. This is the stage Fortes refers to as the 'expanded family', a close, usually local group which is residentially though not yet jurally or ritually segmented. It is in the gradual creation of this group from the narrower early domestic family - the signs of which we have just noted - that the processes of social maturation have their primary effect. They are accompanied by such pressures as the growing size of the families involved, each of which is likely to desire independence most especially when it is viable as a more separate economic group. Although Tallensi explain this in terms of desiring to provide 'better' for their families, it seems unlikely that there is any actual increase in per person commodities of a more material sort than independence. Nonetheless, the petty - or major - jealousies of co-wives in a large household are a strong precipitating factor in the desire of their husbands to 'go out on their own' (1949a, p. 134), and these are often centered on common food supplies (op. cit., pp. 129-30). These quarrels tend to follow the lines of structural cleavage within the domestic unit (p. 228) and serve to emphasize the importance of the growth of individual domestic families in the process of segmentation at the lowest levels. The splitting up of the residential unit provides the basic economic independence which is necessary to make formal genealogical segmentation viable, and to sustain the various political claims which it may accompany. Before we consider the incidents in which these claims are put forward an put to test, let us look at the influence of matrifocal identifications and economic and demographic pressures on the process of segmentation.
Matrifocal Identification and Matrilateral Kinship: First off, we may emphasize that these two categories, though closely related, are not identical. By the first I refer to the identification of the son with his mother within the patrilineal family, particularly as opposed to his brothers by the same father, and different mothers. By the second, I refer to the particular ties which he has with his mother's agnatic family, which though they do not directly differentiate him in nay formal sense from his brothers, do give him important social and economic independence. What the distinction in terms emphasizes, then, is the importance of both kinds of ties, within the lineage and out of it, which are not patrilineally determined.
The basic significance of matrifocal identification is indicated by Fortes:
This unity of the joint family is summed up in the concept of the patricentral 'house' (yir). But it is a unity superimposed upon a strict cellular differentiation into matricentral 'rooms' (dug2t), each centered upon a wife. (1945, p. 198)
Thus:
The patricentral unit is segmented into incipient matri-lateral segments. (1945, p. 199)
A stratification by generation provides for a number of patri-segments, these are in turn divided into matri-segments. The formula works to distinguish groups of quasi-groups in any everyday usage, as well as to delineate the lines of eventual segmentation into new corporate groups:
Patrilateral origin and matrilateral origin work in polar opposition to each other. An agnatic group united by their patrilateral origin can be subdivided in accordance with their matrilateral origin; and vice versa, an agnatic group united by reference to their matrilateral origin can be subdivided by closer patrilateral origin. (1945, p. 200)
Actually, it seems dubious usage to call this a 'polar opposition'; rather we may simply say that the two principles work in alternation as reference moves up or down the lineage order. Though it is possible, it is in practice unusual for a matrilaterally unified group to differentiate itself by closer patrilateral origin. The only likely case would seem to be that of the incipient patri-segments of a very large effective minimal lineage, for in no smaller group would there be different fathers and a common mother, and in no larger group would identification not be a matter necessarily of 'an ancestor and his mother' taken as a unit. As we have noted, when a corporate group splits up, it does so along the lines of its matri-segments. There is thus a symbolic connection to be made between matrilateral origin and fission.
Thus it is that reference to a founding ancestor of a lineage denotes its unity, while reference to the various progenitrices of its component groups identifies it as a collection of separate segments (1945, pp. 210-2). This is reflected also in the symbolism of the bogar, where each component segment may make reference at a shrine to their ancestor and his mother, while at the higher levels the matrilateral tie is no longer remembered, and the 'mother's bogar' becomes a 'father's bogar', with reference to the founding ancestor of the lineage only (p. 202; actually Fortes is a bit vague and unclear on this process; see discussion below, in the last part of this chapter). We have already noted how this formal distinction by matrilateral focus, both symbolically and morphologically, is furthered by the material contributions which a boy/man's matrilateral kin make to his economic independence (cf. also 1949a, p. 194).
We shall see in the next chapter in more detail how even in cases where segmentation has not yet taken place, the relations between half- and full brothers by the same father differ. They are predominantly the same in pattern, but, markedly at times, variant in intensity (1949a, p. 257). It is only through socialization into such a variance in feeling (which Fortes suggest quite regularly takes place), perhaps amplified by quarrels between mothers, that matri-focal differentiation of identity becomes a pressure for segmentation, or, in the case of full-brothers against it. In general, this identification only maps out the lines of such segmentation. It is the matrilteral kin which may directly influence the progress, as well as the direction of differentiation. In any case, it is worth emphasizing that matrilineality (that is, either matrifocal identification or matrilateral kinship) operates only in inter-personal relations and as a factor of cleavage within a patrilineal group. It does not form corporate groups on its own any more than agnatic descent can be furthered without women. We may add that this latter fact, which necessitates the introduction of strangers into the lineage, and the identification of women with processes of fission and disunity in the lineage are very likely behind the threat or ambivalence of a number of symbolic references to women. These range from the casual but prevalent suggestion that women are fickle (1949a, p. 85; quite likely true, given their conflicting loyalties) through the association of women with witchcraft (1949a, pp. 34-5; which crops up primarily in quarrels between wives) and the potion the blood-guilt passes through the uterine line (1949a, pp. 35-37), to the ascription of arbitrary cruelty to female ancestors despite the kindness and devotion of living mothers (1949a, p. 235).
The segmentation of the lower order corporations takes place between brothers, then, and largely on lines of differing matrilateral origin. We shall look later at some of the implications this has for understanding cognatic ties. Here let us look at certain of the pressures which underly the segmentation process.
nheritance - The Extent of Agnatic Bonds - Exogamy - Representation and Collective Responsibility - Cooperative Action - Fission and Fusion - Fission - Social Maturity and the Domestic Family - Matrifocal Identification and Matrilateral Kinship - Economic and Demographic Pressures - Events - Constructing the Corporations - The Salience of Ancestors