University of North Carolina

Forte's account of ancestor worship and kinship among the Tallensi has recently been subjected to a variety of criticisms. Most of these have stemmed at least in part from misunderstandings of Fortes. More fundamentally, many have tended to stress cultural categories and terminology to the exclusion of sociological analysis. They have reduced the authority of ancestors to a static set of norms or a seemingly random collection of ritual observances. Some have even suggested that the distinctions between living and dead, agnatic and non-agnatic ancestors, are matters of indifference. The present article challenges these interpretations with a reconsideration of Forte's material in which the authority of ancestors is treated sociologically and seen to be the key to the working of the Tale kinship system.


In tradition Tale social thought, all authority is vested in ancestors, fathers of at least two generations removed, dead, and significant in being points of genealogical unification and differentiation. Though living persons do have authority in some matters, it is never absolute. It is either the authority of parents, which is part of the same pattern of relationship at that of ancestors, and always limited by the authority of ancestors, or it is authority given by assumed devolution from ancestors. The authority of living persons is partial and subject to challenge; that of ancestors is pervasive and absolute. Ancestral authority is the key to the working of the Tale kinship system and the reproductive capacity of the society. Forte's analysis, though complex and sometimes with central points left implicit, is sound. Forte's critics have consistently, sometimes seemingly wilfully, misunderstood him and have offered markedly inadequate alternative treatments. In particular, they have abandoned sociological analysis for a variety of wholly culturological approaches. In this article I draw on Fortes and attempt sociologically to explain the authority of ancestors.

The concept of authority

Authority, in the present usage, refers to the recognised a priori right to determine the nature and outcome of social situations. It is right which exists distinct from any might, neither contingent nor subject to challenge.

is right derived from creation, which thus takes precedence over the reasonings and machinations of things created, of living men. In short, the existence of he living is entirely contingent on the past actions of the ancestors, and thus by extension, the lives of the living are subject to the present will of the ancestors. But the authority of ancestors among the Tallensi is genealogically specific, just as the relations of living men are genealogically specified. The logic of ancestral authority is a neat and accurate projection of the jural capacities of the living as organised over varying social identities, individual or collective. It is no accident, thus, that the ancestors are dead. Because of their decease they are unable to take any direct action; their response is 'after the fact' and mediated by the living. This helps to ensure that the ancestors will combine the virtues f flexibility with a complete absence of error. Error can only be an inaccurate interpretation on the part of the living. It is difficult for ancestral authority to upset the social order, since it is a direct extension of that order. Power wielded by a living individual is always likely to upset social order since it is by definition at least partially free from the constraints of that order. For this reason only dead ancestors have complete authority.

Power may be as capricious as the actor who wields it and may shift quickly among actors. Authority, because it depends for its strength on common recognition, is essentially collective and consistent in nature (although the capricious vicissitudes of life may be explained through reference to a figure of authority). Authority is right, whether or not it is followed; power is force. People thus often propitiate authority after violating its dicta; the presence of authority is seen in these acts of respect as much as in strict obedience. The presence of power is seen entirely in its effectuality. The rule of law, and by extension the rule of social order, may be seen as matters of authority, or right, rather than power or individual intention. This view follows a long philosophical tradition, and was particularly prominent in nineteenth-century conservatism. It is also directly congruent with Forte's description of distinctions made by his Tale informants (cf. 1945: 116; 1949: 235). In order to believe in authority, one must believe in the existence of standards outside the province of specific human intentions, judgements and enforcements. Such an ideal standard may itself be evolving; respect for divine authority did not vanish from medieval Christianity as it became Aristotelian instead of Platonic. It is the task of mortal men to try to discover the will of gods and ancestors.

Durkheim, in a similar view, saw an impending crisis in the weakening of the influence of religion which attended the growth of egoism. The authority of religion is inextricably linked to its ability to 'socialise', that is, to determine collectively the individual existence of its members. This is precisely what ancestral authority is effective in doing among the Tallensi, for ancestral authority represents a stronger social organisation than any indirectly worshipped by the adherents of the major (textual) world religions. 'Religions', according to Durkheim, 'can socialise us only insofar as they refuse us the right of free examination' (1897:376). Authority requires unquestioning acceptance. Even a bureaucracy, with its operation supposedly based on the rationality of its procedures, must founder if authority within it is constantly questioned and changing (cf. Weber 1921:196).

Authority gains its potency not by denying all practice of reason, but by insuring that all reasoning begins from certain premisses. The rationalisations of the scholastic philosophers are thus legendary, but they were not in any way in conflict with their belief in God or Christian mysteries. The scholastics who constructed elaborate 'proofs' of the existence of God did so not to convince themselves or sceptics, but in order to learn about the nature of God. They began with the assumption of God's existence and proceeded to consider the conditions of our knowledge of God. So it is with the worship of ancestors; the Tallensi need not constantly search for signs of the existence of ancestors, but rather for signs of the will of ancestors. In this connexion I suggest it is not fruitful to treat additional religion or authority as a set of rules or ideas prescribed with finality once and for all.

Weber was among the earlier and more influential scholars to describe traditional authority as a static catechism of 'norms', and thereby neglect the process of reasoning involved:

Domination that rests upon ... piety for what actually, allegedly, or presumably has always existed, will be called 'traditionalist authority'. Patriarchalism is by far the most important type of domination the legitimacy of which rests upon tradition ... It is a characteristic of patriarchal and patrimonial authority ... that the system of inviolable norms is considered sacred; an infraction of them would result in magical or religious evils. (1915:296)

I argue that the central characteristic of traditional authority is not the timeless existence of a set of norms to which all adhere but rather a variety of reasoning about issues which do not fit clearly into any set of explicit rules. Divination of the will of ancestors is used to adjudicate issues which are either novel or complex enough to go beyond everyday rules of behaviour, or are inherently insoluble in terms of everyday rules of thought. After the fact explanations of natural disasters in social or moral terms are of the latter variety; settlements of disrupted inheritance are of the former.

Weber developed his view of traditionalist authority in constant mental opposition to his notion of the rational-legal authority predominant in the modern West (Parsons (1947:51) suggests that the opposition perhaps impeded his judgement). Central to Weber's definition of rational-legal authority is its impersonality (see Weber 1925, Vol. I: 212-40 on the contrast between legal and traditional authority). In a rational organisation, such as a bureaucracy, the office, not the individual, gives authority to pronouncements.

Fortes points out that the authority of parents and ancestors is not as much a matter of their personal characteristics as of their purely formal status as parents of living offspring.

Ancestorhood is conferred on persons of the parental generation who have jural authority in living social relations, not on those who imprint their personalities on their offspring by virtue of their part in bringing them up. (Fortes 1965:130)

The relations of offspring to ancestors (as to parents) are obligatory, not simply a matter of the enforcement of the superordinate of the choice of the subordinate. In this sense, the Tale pattern of organisation may be seen to fall within Weber's essential definition of 'legal' authority. The idealisation of ancestors, based on their incorporeality, contributes further to his fit. Weber continued his discussion of traditional, especially patriarchal authority (quoted above) by suggesting that in it:

there is a realm of free arbitrariness and favor of the lord, who in principle judges only in terms of 'personal', not 'functional', relations. In this sense, traditionalist authority is irrational (1915:296).

In fact, divination of ancestral will pays close attention to functional issues in the society of the living. It must be obvious to all outside analysts that the expressed whims of the ancestors do not come from a 'realm of free arbitrariness.'

The morality of kinship

It is in something of a similar context that Bloch has emphasised Forte's arguments regarding the 'morality' of kinship and rebutted those which take kinship as instrumental and as primarily the object of manipulations based on immediate individual self-interest (1973). Bloch argues that in segmentary kinship-based societies, morality is primarily a matter of relationships. In this he echoes the common sociological view of morality held by Europe's nineteenth-century conservatives (see Nisbet 1966:107-73, and Calhoun 1980). He suggests that moral relationships are those which are not organised for specific immediate ends, but are long-term and binding as part of the social order. Specifically, their source is focused in the ancestors, not in individual decisions. Bloch sees the contrast of which de Tcqueville also made much (cf. 1840:232), between particularistic norms of relationship (which leave everyone to establish and govern their own relationships by negotiation) and moral norms and kinship regulation (which make relations a matter of social obligation for the individual, and, especially in close kinship, one to be assumed, not chosen). The latter orientation to relationships provides for greater social stability:

I have tried to show that it is only because of the fact that to the ancestors kinship is moral, that is non-specific and long term, that it produces an adaptability potential to long term social change. If more rational ties were used, i.e. ties which are the fruit of a process of maximization, they would be more efficient in the short term but more costly in the middle and long term (Bloch 1973:86).

Men do not live by plans alone, or, as Merton many years ago stressed, we need to recognise the importance - often positive - of unanticipated consequences of particular purposive acts for social organisation (1936). Bloch clearly shows the importance of this analytic separation of motive from effect; he introduces something of a canard, however, with his reference to 'adaptability potential to long term social change'. He makes an analogy to the principle of generalised adaptation in biological evolution, a 'maximisation of options' principle of survival, which does not follow from the case he has presented. He details an instance of the significance of kinship morality in maintaining long-term continuity not as an adaptation to long-term change, but rather to short-term fluctuation. Peasants keep their reciprocal obligations most up-to-date with distant of fictive kin because they can count on closer kin under any circumstances, even in a particularly desperate year. This is, in fact, an aspect of the lineage system which Fortes emphasises:

Observation of the lineage system in action suggested that its distinguishing characteristic, as a regulating factor in the social structure, was its tendency toward equilibrium. This operated in such a way as to leave room for continual internal adjustments in the social organisation without endangering its long-term stability (1945:x).

Bloch's analysis differs from that of Fortes in another significant way (which cannot be considered matter of error). He describes patterns of labour co-operation among the Merina in which active reciprocity is most important with distant or fictive kin, somewhat secondary relationships in terms of the formal system of descent-based bonds. This, he suggests, is because the nearer kinship ties are more insured by the moral aspect of the relationship so that:

'Real' kinsmen would always come, they said, 'artificial' kinsmen would only come if one kept up the typical kinship behaviour of repeated requests for help. If one did not do so these 'artificial' kinsmen would lapse (Bloch 1973:79).

Fortes, on the other hand, tells us that among the Tallensi,

The more distant a genealogical tie is, the more does it become a matter of moral and ritual, rather than a jural and economic relations (1949:18).

Bloch argues that his findings contradict Fortes and demand some sort of resolution. Other differences in the two social organisation make it difficult to settle the matter conclusively, but we may offer two additional considerations. First, the two statements are not necessarily as contradictory as they might seem, if one bears in mind Forte's distinction between the mutuality of close agnatic relations and the reciprocity of more distant, especially merely cognatic, ties. Secondly, in Forte's description of the Tallensi we find that socio-spatial fields exist in which geographically proximate kin are also genealogically near. There would seem to be relatively few situations in traditional Tallensi farming which distant kin would be appropriate or practical to call upon for assistance. We might expect to find something closer t the situation Bloch describes among migrants into urban areas. Among the Merina, a high rate of geographical mobility has meant that neighbours are not necessarily kinsmen, let alone close kinsmen, and made the employment of kinship fictions to describe and govern relationships widespread. Although Bloch emphasised the importance of this mobility in an earlier monograph (1971), in his (1973) comparison with Forte's analysis, he does not bring out the difference between the two cases.

Bloch's usage of the notion of artificial kin also raises some questions, which apply to the whole treatment of fictive kin in anthropological debate. Bloch, in common with most authors, describes artificial kinship as simply a matter of comparison with other kin categories (in general, not with particular relations; hence apparently the term artificial instead of fictive), seeing these latter as arranged on a single continuum of intensity. This treats relationships too exclusively as matters of inter-individual transactions. Bloch's fictive kinsmen are analogous to cognatic kin, as Fortes describes them, but not agnatic (Fortes 1949: 13-14). They have relationships with individuals, which may be personally established, rather than group relationships situated in a formal corporate organisation. They are relations thus of reciprocal, but not common, interest. Agnatic bonds are characterised by mutuality in part because agnates share responsibility and rights in inheritance, because together they form social units at higher levels of aggregation. Kin linked only by cognatic not agnatic bonds, have interests in each other's welfare solely on the basis of exchange, not at all on the basis of commonality. Artificial kin are similar to cognatic kin but for the fact that their relations are not formal. This is why keeping up the relationship is so important to them. The relationship among fictive or artificial kin is never implied by the rest of a set of common relations, as, for example, is the relationship among brothers. It is not the product of descent or marriage. It exists through its actualisation alone.

This is the key to much of the distinction between the Tallensi case Fortes describes and the Merina case Bloch describes. Among the Tallensi, distant cognatic kin become relevant only in situations where they are called upon to assist in travel or to make sacrifices to their agnatic ancestors who are their kinsmen's matrilineal ancestors. These distant cognatic kin seldom live nearby, and are therefore unlikely to be of everyday significance. Close cognatic kin, however, are the persons most frequently called upon to form work parties. This is because the need for labour is seen as a personal issue. Among the Merina, unrelated persons are likely to live in close proximity while close cognates and agnates may be quite distant. In any event, however, the fluctuations in demand for co-operative labour makes it practical to maintain fictive kin relations on an everyday basis. It is in this that Bloch describes an adaptation to long-term change. The system itself is geared to deal with short-term fluctuation, but has been able to adapt to long-term change by substituting artificial for cognatic kin (compare Jackson 1977).

Ancestral authority and lineage structure

The morality of kinship an the authority of ancestors are both ideational extensions of a social organisation which is based upon segmentary kinship. By emphasising these connexions, we are better able to understand the workings of the system of social organisation than by granting primacy to the cultural plane or trying, for example, to interpret Tale cognatic kinship in terms of a forced and unsociological analogy to the Kwaio, as did Keesing (1970). It is also important to realise the particular workings of the system of ancestral authority in sociological terms rather than assimilating it to the more general culturological category of the morality of kinship. Ancestors define social groups - groups of persons subject of their authority. It is of course true that ancestors can sometimes be held to have interfered in the lives of descendants outside of the corporate groupings they define, that is, in the line of their descendants who are only cognatic, not also agnatic. This is not a matter of authority. That is a matter of power (the real social power of living cognates) is indicated by the fact that the interventions of cognatic ancestors are seen as generally much more capricious than those of agnatic ancestors. Keesing misinterprets this set of issues in part because he has failed to read Fortes carefully. Erroneously, he says that

a Tale man, like a Kwaio man, does not sacrifice only to the lineage ancestors (as opposed to parents and grandparents) of his own lineage .... He sacrifices to these nonagnatic ancestors on many occasions (1970:766).

In fact, a Tale man can only indirectly approach nonagnatic ancestors, may only attend but not lead sacrifices at their lineage homes. Furthermore, the personal 'destiny' (yin) ancestors are the only major part of such sacrifices for which a man must seek the intervention of his non-agnatic kin. It is not at all clear, as Keesing asserts, that Fortes undervalued cognatic kinship. Rather, he emphasised the important distinction between agnates and kin who are mere cognates. For the Tallensi it always matters whether connexions are matrilateral or patrilineal; it is never a matter of indifference, as Keesing's definitions would have it (cf. 1970:768).

One of the most important points in Forte's analysis of the Tale ancestral system is that is is always a specific individual, more especially a genealogically specific individual, who becomes an ancestor. Fortes's emphasis on this has not kept others from missing the point. Kopytoff, for example, has suggested that ancestors are merely elders and we are ethnocentric to see them as anything else (1971). Kopytoff's argument has attracted credence out of proportion to its quality (cf. Sangree (1974), though note the qualifications which his comments on the complementarity of generations require; or Mendosa (1976),who calls for a synthesis of Kopytoff and Fortes without realising the fundamental nature of the contradiction). Fortes is quite clear on the matter:

When a particular - and it is always a particular person - is thus reinstated as an ancestor, it is, as I have argued, because he has living descendants of the right category. His reinstatement in this thus establishes his continued relevance for his society, not as a ghost, but as a regulative focus for the social relations an activities that pursuit as the deposit, so the speak, of his life and career (1965:129).

There is, for a corporate group of agnates, a single ancestor who is the salient point of collective reference. Similarly, there is a single ancestress, forming a matrilateral link, as represented in the copying of ancestral shrines. It is the apices of groupings in the kinship hierarchy which are crucial, then, in patrilineal descent among the Tallensi:

Each segment has its focus of unity, and an index of its corporate identity, in the ancestor by reference to whom it is differentiated from other segments of the same order in the hierarchically organised set of lineages (Fortes 1945:31).

Important female ancestors are patrilineally determined - the mother of apical male ancestors - and matrilateral. Matrilateral connexions are not of enduring significance.

The identity of the particular ancestor inheres in the social relations of his descendants. Relations with ancestors are primarily manifestations of the social system and not of particular physical relations with parents or ungrounded metaphysical beliefs:

Ancestorhood is conferred on persons of the parental generation who have jural authority in living social relations, not on those who imprint their personalities on their offspring by virtue of their part in bringing them up (Fortes 1969:130)

An ancestor has his authority not because of his personality, nor in general because of his individual career. Exceptions to this occur only inasmuch as men of social importance are more likely to have sons and keep them in a large unit of social solidarity until their deaths. Neither does an individual become an ancestor because of any particular authority or esteem among members of his own generation. He has authority because of his genealogically structured position with regard to the living:

Ancestor worship is a representation of extension of the authority component in the jural relations of successive generations (Fortes 1965:133).

This authority component is of course formal. Although the ancestor must be someone in particular, he does not behave as anyone in particular. He represents the abstracted principles of lineage structure, authority and values. In Kuper's words, 'the ancestors are the ideal not the actual personality' (1947:188; see also, Fortes 1949:235). Because of this idealisation of ancestors, that is, the independence (in principle) of their ancestral status from their lifetime careers, any man who

dies leaving a son ... becomes an ancestor of equal status with any other ancestor (Fortes 1965:133).

He may not be of equal structural significance, or be equally likely to be remembered, but he is an ancestor in an equal sense. He cannot be overruled, and is not subject to qualifications on his authority.

Ancestors, not elders

Kopytoff finds this description unsatisfying (at least in 1971; not so much so in his earlier work):

I shall ... try to show that by viewing what have been called African ancestor cults as part of the eldership complex, we can account more simply for many of Forte's generalisations and at the same time make redundant some of the problems he raises (1971: 129-30).

Kopytoff's argument is apparently based on the assumption that all Africans are the same, for he continually speaks of 'African ancestor cults' as a unit, while unthinkingly criticising Forte's analysis of the Tallensi on the basis of his own material on the Suku. He does this without even considering the relevance of the fact that Suku descent-reckoning is matrilineal and Tallensi patrilineal. He pays little heed generally to differences in social contexts or the changes wrought in the decades between Forte's fieldwork and his own.

Much of Kopytoff's article is occupied with an examination of various Bantu languages in search of words translatable as 'ancestor'. His search is part of an attempt to show that the term and its connoted distinction of living from dead are but ethnocentric impositions of Western anthropologists. Brain has convincingly challenged Kopytoff's assertions on this, suggesting that his claim 'that Bantu languages have no word for ancestral spirit is patently absurd', and indeed that the noun classes used are different from those for living persons (1973:126; see also Sangree 1974). Fortes also offers a term for ancestor and a distinct one for elder (yaab and kpeem), translations Kopytoff does not consider. Under Tale definition, an ancestor is any man who dies leaving a son, as we saw. By this criterion, there are ancestors in all societies; surely our questions concern their significance, or what variations there may be in ideas concerning them. Kopytoff is concerned to show us that:

Once we recognize that African 'ancestors' are above all elders and are to be understood in terms of the same category as living elders, we shall stop pursuing a multitude of problems of our own creation (1971:138).

He is wrong. The distinction between ancestors and elders is a useful one. Indeed, it distinguishes much in the Tale organisation of authority from that of the Suku.

The two peoples differ in a number of ways. The most relevant at present is that the latter abuse authority on relative age:

In short, to those on the outside, a lineage is represented by the eldest member present. Within the lineage, the lineage is represented to any one member by any older member present and, collectively, by all older members living and dead (Kopytoff 1971:133)

Kopytoff suggests that the distinction between living and dead is 'incidental and contextual' and that the Suku share this view (1971:133). Suku lineages do not seem to be internally differentiated and hierarchically organised with reference to a monistic system such as that of which ancestors are the focus for the Tallensi. Only at the highest levels, of clans, are Tale ancestors dealt with 'in general' (Fortes 1945:137). Surprisingly, Kopytoff asserts that:

The Suku pattern ... is congruent with most ethnographic descriptions of African 'ancestral cults' and of the role of elders (1971:134).

Certainly this is not true of Forte's writings on the Tallensi, despite the fact that it is these to which Kopytoff directs the most attention.

Residence in both the matrilineal Suku an the patrilineal Tallensi is primarily patrilocal. This means that a single system of ties of graduated intensity and density among the Tallensi is paralleled by two crossed dimensions of social relations among the Suku. The cognatic kinship relations which supplement and differentiate agnatic kin among the Tallensi appear as more diametrically opposed among the Suku. Fortes described ancestral authority as an extension of the authority component in relations between the generations. Suku lineage elders, clearly, are not the same as the generational authority figures confronting particular individuals where they live. Residential affiliation, but not the structuring of descent is, congruent with paternal authority among the Suku. Descent, residence and paternity are all part of the same monistic organisation of Tale society (see Jackson's 1977:134-5 argument that ancestor 'cults' are strong where lineage organisation controls local contiguity rather than being superseded by it). Tale society is thus the more logically consistent, in Sorokin's term (1957:20). In the phrasing Lévi-Strauss adopted to describe patterns of matrilateral cross-cousin marriage, the Tale system is harmonic, while the Suku system is no (1949:334). The former has strict genealogical determination of identity and authority relations which the latter does not, and which would be incongruent with the rest of Suku social organisation. Age, as a basis for authority, necessarily introduces a greater individualism into the organisation. Since all authority relations among the Suku are determined by relative age, they necessarily link individuals. They do not form the corporate units which Tale organisation of authority does. There are thus significant reasons, quite beyond anyone's ethnocentrism, for the difference in descriptions of authority systems, and for retaining the term 'ancestor', at least in the Tale case.

Kopytoff argues at great length that the distinction between the living and the dead is ethnocentric. My own biases are perhaps more vivocentric, for I am convinced that the dead are quite different from the living, and nowhere more so than when we consider what action an ancestor (or elder) can take to impose his authority on his kinsmen. In any given case, a majority of one's ancestors are likely to be dead, particularly those beyond the range of grandparent. The Tallensi make a distinction at this level (incidentally congruent with English language usage). An individual seldom has concrete dealings with people beyond this range, dealings in which their personalities rather than their structural, genetic, or ideological significances are at issue. Among the Tallensi, it is stereo-typically at the level of grandparents that ancestors become the focus of new effective minimal lineages.

The fact that ancestors are dead makes it much easier to idealise them, to have them represent lineage values as opposed to personal interests or idiosyncrasies of judgement. Living persons have a problematic propensity for doing things, for acting on the bases of their own social and psychical personalities. These are always only partial representations of overall principles of social order. One of the most important characteristics of ancestors as authority figures, then, is their inactivity. Those things which they 'do' tend to appear to us as either chance occurrences or societal actions.

Ancestral authority in social practice.

Fortes describes what I have called societal actions in terms of consensus. This seems problematic, since however high consensus may be, in our frame of reference, it is not absolute in Taleland. Nor is consensus integral to the workings of ancestral authority, although it is of course advantageous, as in any system of authority. What is essential is acceptance of certain means or processes of judgement, notably divination. Access to control over such processes of judgement is quite stratified, both in terms of general social position and in terms of relational proximity to the principles in the case at issue. Enough people are generally involved to ensure that the decision is more or less in accord with widespread opinion. It need not be exactly in accord, however, since opinion is seldom completely consensual and usually is somewhat malleable. The process of judgement and the presentation and symbolic language of judgement as much as the specific outcome of divination are central to reinforcing the social order and ensuring acceptance of the decision.

Appeal to ancestors may be concerned with the post hoc explanation of some natural occurrence. Although not immediately apparent in such cases, it is still generally true that divination concerns social issues. Its findings almost always concern moral lapses of one sort or another, and morality is essentially social; it is concerned with relationships among people, their behaviour towards each other. The disease of a wife, for example, may be familiarly explained as due to her husband's failure to establish his own homestead, distinct from that of his proxy father. Divination is used to force the latter to be less tight-fisted (Fortes 1949:176). Alternatively, the ancestors may be called upon to resolve a dispute; to put it another way, disputants may carry out their arguments in the language of ancestral authority. Specific ancestors hold primary authority over groups larger than households (that is, larger than those headed by living parents). They are also the prime referents in the definition and differentiation of those corporate groups. The primary means of bringing the ancestral voice into the affairs of the living, and thus rendering an authoritative decision, is through divination.

Fortes does not tell us a great deal about Tale diviners, but it is clear that they are not the entrepreneurial specialists reported among a number of South African peoples (this reading is confirmed by personal communication with both Fortes and the more recent ethnographer, Hart, 1975). Most divining involves only local grouping and draws on local talent. Most men are qualified diviners, although only a minority are in active practice. It is not a full-time occupation:

Diviners are numerous. There may be as many as ten or twenty in a large settlement, but only one or two of them will earn more than a few pence a day by divining (Fortes 1945:10).

There is no indication that the particular identity of the diviner has any structural relevance to the cases on which he works.

When ancestors are invoked to bring about more complex social decisions, such as the settlement of long-standing disputes, two additional considerations apply. The first is that a divination is subject to considerable structuration, often with the conscious awareness of the participants and interested parties (Fortes 1949:99). The second is that a divination is not final, and may be ignored or questioned:

There are no jural sanctions compelling a man to abide by custom in cases of this kind; and as the Tallensi often say, men do not fear to defy even the ancestor spirits, when their property of power is at stake (Fortes 1945:249).

Although divinations may be restaged, claiming a failure of the mortal portion of the procedure, one may never 'go over the head' of an ancestor (as, for example, one may sometimes go over the head of living authorities to an ancestor). The ancestors stick together.

At least in ideology, everything in Tale society is subject to the authority of ancestors. This is so because kinship is the dominant system of social organisation (see Fortes 1949:340), and ancestors are 'the main ideological bulwark of the kinship system' (Fortes 1945:33). Kinship relations cannot be reduced to the economic, religious or any other category of activities; in Forte's term, they are axiomatic (1945:249; and 1949:346). The notion is important, and its logical analogy revealing. Kinship relations do not exist for any teleological reason; their moral status is not demonstrable on any sort of evidence, but rather is to be accepted prior to and as the foundation for reasoning about evidence. Just as Kierkegaard positively and Hume negatively suggested that the only possible relations to revealed religion are belief or disbelief (i.e., not demonstration) so it is with kinship for the Tallensi. The logical status of kinship is primitive; from within the system it is not subject to criticism. Fortes is clearly not arguing that kinship exists because it is functional. He is arguing that it exists, and that certain functions follow on its operation. There is no logical way to say that kinship is functional for the Tale social system, since that system derives essentially from kinship. Of course, kinship may be of more or less significance for any person or persons among the Tallensi, and it may be judged good or bad in its particular effects. The standards of such judgement, however, must come from outside the Tale system.

The authority of ancestors is an internal part of the kinship system, not an external idea applied discretely to certain social relations. Authority among the living comes only by 'transmission and assumed devolution from ancestors' (Fortes 1961:187). This universal authority of ancestors appears in practice (at all but the highest level of clans) as the relationship of a particular ancestor to a given population of descendants.

The ancestors acknowledged in a given situation are primarily those who are exclusive to the worshipping group and therefore distinguish that group unequivocally from collateral and coordinate groups of a like sort, who have remoter ascendants in common with them, and worship jointly with them in situations of common concern (Fortes 1965:123).

All facets and all variations of life are subject to ritual sanction within the general framework f ancestral authority (Fortes 1945:144). A man may become the victim of his ancestors (a woman of hers and/or her husband's) for a wide range of faults not necessarily his own or under his control. The system is not 'beatable' as there is no way to live a life of sufficient saintliness to place one beyond the reach of the ancestors. Men are not able to judge mystical wrongs for themselves (Fortes 1959:35); these wrongs are thus not matters for human retribution. 'The community is neutral' (Fortes 1949:180), sometimes even sympathetic to those plagued by ancestrally decreed misfortunes. The ancestors represent much that is complex, even contradictory in social and natural existence; their will is thus not simple to understand.

The relations between men and their ancestors among the Tallensi are a never-ceasing struggle. Men try to coerce and placate their ancestors by means of sacrifices. But the ancestors are unpredictable (Fortes 1945:145).

Agnatic descent reveals the authority of ancestors, for agnatic descent is fundamental to the most enduring features of the social structure - the organisation of corporate groups. Continuity through time is expressed by the agnatic ancestors; nonagnatic ancestors express the contingent features of life. Thus the solidarity of a corporate group, and its distinction from other groups of the same order is expressed by reference to genealogically specific agnatic ancestors. When ancestors give voice to the distinction of social individuals, it is through a one-to-one relationship not strictly genealogically determined Fortes tells us that:

In the strictly personal affairs of the individual, his matrilateral ancestors have as decisive a role in his life as his patrilineal ancestors (1949:294).

In fact, they may have even more such importance, since the patrilineal ancestors seldom affect the individual in 'strictly personal affairs', being concerned with him as a member of the lineage corporation. A man's matrilateral ancestors, however, come from a variety of different lineages, and through them he has his own particular constellation of extraclan descent relations, as symbolised in their inclusion in his set of Destiny ancestors. Far from supporting Keesing's (1970) case for treating ancestors indiscriminately, the importance of matrilateral ancestors in this specific context points up the importance of the differentiation between agnatic and nonagnatic kin.

A man's Destiny ancestors are an ascending series of both patrilineal and matrilateral ancestors, usually dominated by the latter, and always unique. Women have no distinct set of Destiny ancestors, but fall first under the tutelage of their fathers' and then of their husbands' Destinies (Fortes 1959:25). A woman (like a man) does have a distinct Prenatal Destiny which is the result of a sort of positive or negative spiritual sponsorship based on the unborn infant's supposed choice (thus the name 'spoken Destiny'; Fortes 1949; 165). The Yin shrine is the shrine dedicated to man's Destiny ancestors, and as such its placement is a sign of his growing social individuality or autonomy (Fortes 1959:25). An offering to a Yin shrine is often eaten only by the owner and his wife and children, his kin by birth not being allowed to share in it, as they do in his other sacrifices (Fortes 1959:26). Thus the common language of ancestral authority is able to take account of the individuation as well as the unity of Tallensi. This is in no small part because ancestors are themselves always 'particular' and not an undifferentiated mass for the Tallensi (Fortes 1965:125). Individuation is, in any case, not a matter of extraordinary social differences or radical structural opposition, as in a modern, especially a class society. Destiny ancestors do not account for individual choices, and indeed - largely for economic reasons - the range of individual choice is much constricted by out standards. What such ancestors do account for its events which distinguish individuals on matters of common or shared preferences. Thus, as Fortes says with regard to Prenatal Destiny:

Proof that it is working itself out in an evil way is the victim's irremediable but involuntary failure to fulfil the roles and achieve the performance regarded as normal for his status in the social structure ... a failure in the relationship of belonging to society, which, for the Tallensi, means family lineage, and kin (1959: 41-2).

A certain extent of individual variation is thus directly accounted for in the very terms of authority which generally work to ensure solidarity among those with common interest. In such a monistic system, sub-groupings may serve to provide for incorporation into larger groups, not necessarily to pull the larger apart, as the common sociological generalisation has it (cf. Merton 1957:287, or, to the contrary, Calhoun in press).

The system of ancestral authority operates with a conception of almost entirely socialised man. Its stress is first and foremost on the solidarity of lineage segments, secondarily on differentiation within such corporate groups, and thirdly on individuation within an accepted range of social personalities. Failure to fit into society is synonymous with stepping outside the pale of ancestral authority. In an early work, Gluckman at once stressed this, and reduced ancestral authority to a mere cultural epiphenomenon of social relations:

The ancestral cult is a mechanism by which kinship bonds are affirmed ... and the hierarchy of society expressed. In this the ancestral cult is, like much ritual, a form of mnemonic, legally prescribed actions which vividly express social relationships (1937:129).

Although Fortes has emphasised that this is more than just a mnemonic (1959:19), he is clearly in accord.

There is the jural component of status in one's lineage and locus in the web of kinship, acquired by birth, through one's parents, and forming an element in the continuity of the social structure through time. The ritual imprimatur for this is the conception of the ancestors as sovereign and eternal, mirroring the total system of kinship and descent which is keen as an everlasting and fixed framework for the individual's social existence (1959:40).

The authority of ancestors is a good deal more than just a mnemonic, and more even than just a ritual imprimatur. It is a way of reasoning about social relations, and as such inextricably tied up in the social practice which constitutes those relations. Fortes need not be read entirely in the unfortunate rhetoric of 'structures' and 'principle' for which Sahlins faults him (Sahlins 1976: 4-18). He also has an account of ancestral authority directly in terms of social relationships.

Unlike the system of authority by age, which Kopytoff finds to be the norm of Suku organisation (1971, see esp. example: 132), ancestral authority leaves no living person with the authority to act in isolation on behalf of a corporation. Divination and other modes of relating to ancestors are means of representing a collectivity, and often a drawing into action. The heads of segments may have more voice than others in the affairs of the higher order lineages, but beyond the effective minimal lineage, no living person holds final authority: the lineage ancestor has the 'last word' (Fortes 1959:33). This means that only an essentially public process can arrive at an authoritative decision. Even were each headman able to represent his segment absolutely and without challenge , this would limit the centralisation of power to a sort of oligarchy. Formally, at least, every headman of any given level has equal access to the salient ancestor of the level above, the level of their commonality.

The continuity of this system across levels (that is, essentially, through the statically viewed generations) is also important. Fortes has characterised ancestor worship as 'in essence the ritualisation of filial piety' (1959:18). It is clear that this is not epiphenomenal to material social practice, for, among other things, it means that the criteria of political enfranchisement are integral with the process of decision-making itself. A man seriously violating the norm of filial piety would not only

lay himself open to immediate punishment by the ancestors; he would, indeed, be unable to participate in the life of the community since he could not act for himself in ritual or jural affairs (Fortes 1949:218)

A man can be a member of the ritual and jural communities (which are the same) only through his descent. In other words, he can only have a practical voice in the running of society in and through his relations to his ancestors. The ancestors, as we have seen, are primarily concerned with his relations to their other descendants - in other words, with perpetuating the social order.


The authority of ancestors is part of a system of social practice which works to ensure moral relationships among the members of society predominantly organised through the structuration of kinship. Ancestors are qualitatively different in death from the more idiosyncratic existences they had earlier in life. Agnatic ancestors provide the central scheme of reference and reasoning for the reproduction of Tale society. Through their authority they give sanction to the solidarity of lineage segments. Nonagnatic ancestors provide ritual voice and sanction for the continuing dynamic of individuation within the highly sociated overall organisation. All this is done through the representation of specific moral relationships among the living in terms of common moral subjection to specific ancestors.


Go to the main Ancestors in Africa page Go to the main Experience Rich Anthropology page - online Go to the CSAC Anthropology Pages - online