Spirits, A Scholarly Bugaboo?

The majority of writers about cults of affliction have simply subordinated the phenomenon to another called "spirit possession." On the subject of spirits indwelling people, scholarship itself seems to have been obsessed with a bugaboo—a peculiar fascination, a fear, a concern—a word perhaps derived from the Western Bantu notion buka lubuka , to divine, to treat.^[1] So pervasive has been this fixation on spirit possession that it has apparently become a sui generis category of Western scholarship, one that has given rise to entire bibliographies on the subject (Crapanzano and Garrison 1977; Zaretsky and Shambaugh 1978). The Western bugaboo has certainly gotten in the way of clear understanding of the African institution.

A Part of the problem has been the use of the term shamanism , both in the use and in the breach. Although many authors on African spirit possession make a routine reference to the difficulty of using the term shamanism to refer to the African setting, the use of spirit possession tends to serve as a euphemism for shamanism. Authors such as I. Lewis (1986, especially his chapter "The Shaman's Career") and De Heusch (1971, especially his articles "Possession et chamanisme" and "La folie des dieux et la raison des hommes"), who do use African material in their general discussions on spirit possession, and use the term shaman , do not lay to rest the definitional issue. Other authors offer less charged descriptive distinctions in their writing on African spirits. Van Binsbergen and Schoffeleers (1985:39–40) distinguish between shamanistic and mediumistic divination. In the former, the diviner is said to go on a spiritual visionary quest from which he returns with his revelations; this is rare in Africa. In the latter, the diviner is locally considered to be entered or possessed by an external, invisible revelatory agent. This form is said to be prevalent in Africa.

A distinction between those societies that reflect possession belief and those that practice possession trance was introduced by Erika Bourguignon (1976:44–46) to account for another important dichotomy in relation to spirits. According to Bourguignon, the distinction between the two can be accounted for in terms of types of social organization. Less complex societies in which individuals take initiatives are less likely to utilize trance states than more complex societies in which religious expression is controlled or channeled.

Structural interpretations, such as these provided by De Heusch and Bourguignon, offer comparisons of differing valences of spirit presence

between entire societies. De Heusch's distinction between adoricisme and exorcisme emphasizes the alleged pattern that in the first case creates a permanent bond between human and spirit, and in the second leads to the cleansing, or casting out, of the possessing spirit. In the first case the sufferer becomes a vehicle, or medium, of the spirit; in the second the sufferer needs to be cured of possession (De Heusch 1971:235). The first is "good" possession; the second "bad."

In practice, what do these structural surveys of possession types reveal? How helpful are they? For De Heusch, the Kongo (in the Nkita possession) and the Thonga feature possession sickness and trance, thus calling for exorcism or healing. The Sukuma and Lovedu reflect sickness possession, from which therapy is the only solution. Mediumistic possession is found among the Sukuma, Kuba, Luba, and Nuba, thus approaching shamanism. But only the Vandau, a small group living among the Thonga, have what De Heusch calls "authentic shamanism" (1971:258–276).

This approach, which lumps entire societies into structural types, has been correctly criticized by a number of writers (e.g., I. Lewis 1986), who point out that these ideal types are in fact usually moments in individual "careers." The typologies are based on rather inadequate ethnographic information that generalizes to the entire society. They need to be given a more historical and contextual interpretation in order to explain why some sectors of a society, at particular junctures in history, are prone to possession.

Ian Lewis's earlier approach to this question, developed from his work among Somalian pastoralists in the Horn of Africa, has become widely known for its emphasis on the "marginal cult," which stands in contrast to the "central" or dominant cult in a society. Spirit possession, he argued earlier, was an expression of crisis, impasse, or confinement, and the only legitimate or permissible outlet by members of society—often women—in subordinate or marginal positions. Nature spirits, the source of their possession, were marginal in the cult of the dominant—that is, Islam, controlled by men.

Later (1986) Lewis elaborated on this hypothesis of the "epidemiology of possession" with a more dynamic model that stressed the individual course from "uncontrolled" to "controlled" mystical experience. In social sectors that were marginal, the career of possession led from membership in a cult of affliction by a peripheral spirit in the dominant pantheon to a kind of permanent accommodation with the mystical force. In social sectors that were more central to power in society,

possession led to control and even exorcism of the mystical force. If control, or channeling, of the force was the outcome, such holders of mystical power were generally thought to possess witchcraft—that is, mystical—power over others. This, of course, could lead to the use of this power in social control and government. For Lewis, the main difference between witchcraft (malefic mystical power) and possession had to do with whether it was handled "obliquely" in cult accommodation or "directly" through exorcism by the dominant religion or utilized in social control (1986:60).

De Heusch's and Lewis's approaches to possession in African religion offer a sketch of the work of a major school of writers who view it as a compensatory effort to come to terms with misfortune, suffering, and evil. There are many reasons why spirit possession might be a dominant hypothesis of difficult experience in Central and Southern Africa. In societies that are acephalous, either today or historically, individuals who claim original knowledge are vulnerable to envy and criticism. Thus, just as such individuals are often code-labeled as witches, in keeping with Lewis's approach, so cultural norms offset this by fostering as the source of all original knowledge and of change the realm of spirits, especially ancestral spirits or shades, the custodians of societal core values.

Other authors have questioned the wider applicability of Lewis's marginality model of possession cults. Linda Giles (1987) has studied the interface of ngoma and Islam on the Swahili coast of Tanzania, where there is a far greater interpenetration of the two than Lewis might allow. In any event, she suggests that not only are ngoma and related rituals practiced by Muslims but the Muslim waganga are often among the most devout adherents of Islam in Swahili society. A feminist scholar of Central African ritual healing, Anita Spring, who has studied ngoma rituals among Luvale women of western Zambia, offers that Lewis and other "marginalist" interpreters of cults of affliction have generally failed to consider the real presence of disease in connection with spirit possession and entrance to an ngoma order. Spring argues that the predominantly male scholars of ritual in Africa have imposed on African experience the nineteenth-century Western view of woman that accounts for women's mystical experience as their failure to cope with social conflicts (Spring 1978, 1985). I return in the next chapter to Spring's important study of the epidemiology of possession in relation to demographic and social profiles.

Whereas the former interpretations of African possession are charac-

terized by their reduction of the phenomenon to social or structural themes, other approaches take the explanation in the direction of psychological theories. Bührman and Gqomfa (1981–82) have studied Xhosa healers in the Transkei of South Africa and are persuaded that the amagqira of the Transkei, whose work in most respects resembles that of the Cape Town therapists presented in chapter 1, can best be explained by Jungian psychotherapeutic models, and that the "sickness" (twasa) that ngoma treats is largely psychopathology, with schizophrenialike symptoms of excessive dreaming (1981). The similarity of dreams and dream therapies, and the songs derived from dreams, to those of other cultural and civilizational settings offer Bührman what she suggests is convincing case material for a Jungian approach (1978).

Another type of psychotherapeutic interpretation is offered by psychologist Ellen Corin of the Western Bantu rite Zebola, reported in chapter 1 as an Equateur Province Zairian rite brought to Kinshasa. Corin relates her understanding of Zebola to the approach of the "Dakar group" working under Henri Collomb, with a generally psychoanalytic orientation. A major concern in African psychotherapy, suggests Corin, is the differentiation of the self within a tight kin setting. This leads frequently to a diagnosis of witchcraft, with the identification of a specific other as the source, such as a mother's brother in a matrilineage, or an uncle in a patrilineage. However, etiologies in African therapies are often presented in a chainlike sequence, such that culturally standardized etiologies such as witchcraft and spirit possession are invoked to explain quite a variety of particular signs and symptoms.

This is the case with Zebola, in which all cases are ultimately explained by possession of the spirit of Zebola, a type of upriver nature spirit. Yet, on studying the sign-symptom sets of particular members of Zebola in Kinshasa, Corin demonstrated that a majority were recommended for recruitment due to transgressions of social rules (28 percent) and interpersonal conflicts (55 percent), whereas witchcraft (8 percent), meeting an evil spirit (2 percent), and direct possession by the Zebola spirit (7 percent) were relatively insignificant immediate causes (1980:150).

Corin's method of relating the individual experience to the culturally standardized cause offers important correctives to the problems of the structural reifiers as well as the psychological reductionist explanations of spirit possession. Corin suggests, first, that there is a loose, accommodating, link between sign-symptom and etiology and, second, that the spirit possession nosology is, as I have suggested throughout this

work, a kind of cultural hypothesis about misfortune (Corin 1980:152; see also Crapanzano and Garrison 1977, whom she cites). The narrow psychopathological interpretation of ngoma offered by Bührman not only seems forced, it does not allow for the "recruitment to leadership" understanding of ngoma that appears in the indigenous model of the amagqira themselves.

Yet other approaches to spirit possession in Africa move away from the sociological and structural perspective entirely on the grounds that this fails to take into consideration the terms of the experience itself. These authors, such as Lambek (1981) and Comaroff (1981, 1986) work with an approach that is generally called "phenomenological" or "hermeneutical." Lambek's work on the island of Mayotte in the Indian Ocean between Mozambique and Madagascar especially typifies the approach. Rather than presenting spirit possession as a phenomenon to be explained in terms of society or its own cultural structure, this approach presents it as a text that needs to be "read" in its own terms and categories. As Lambek points out, on Mayotte possession is presented in terms of "curing," but it really has little to do with "disease." The spirits, who are for the most part spirits of the dead, are related to as separate, almost human, beings. Interaction with these spirits can be differentiated by distinctive codes of food and gesture exchange, incense, the topics of conversations, and the role of third parties (1981: 11). Although Lambek approaches the question of the "epidemiology of possession "—that is, the frequency of possession and social categories of the possessed, who are mainly women—he rejects any direct sociological inference to their roles as marginal to Islam or to their distinctive role in society. Rather, he insists that possession is a culturally autonomous domain that must be seen in its own terms and its own logic.

Csordas (1988) has recently provided a comparable picture of religious healing in a non-African setting, namely charismatic Christian healing in the United States. However, he goes beyond the mere phenomenological portrayal of healing to attempt to explain the criteria for "efficacy" in its outcome. With penetrating case study comparisons, he has identified the following criteria as important variables in predicting the outcome: the sufferer's prior disposition toward the treatment, whether positive or skeptical; the quality and character of the particular religious experience; the possibility in the mind of the sufferer as to the outcome; the occurrence of personal changes in incremental steps (Csordas 1988:138). Csordas's application of these criteria to charis-

matic Christian healing takes us directly to evidence for ngoma therapy's efficacy as established by healers and sufferers themselves. After reviewing the literature and my own fieldnotes on this point, I am amazed that so little is available. Also, those with personal involvement in ngoma and who are divined to be possessed do not speak with one voice. Still, it is important to discern an indigenous "theory of ngoma" that can be generalized from particular settings to the entire range of manifestations.