Introduction

That which was a stitch of pain,
has become the path to the priesthood.
Lemba song text,
Kongo society, 1910

An important feature of Sub-Saharan African religion and healing, historically and in the twentieth century, has been the interpretation of adversity, paradox, and change within the framework of specialized communities, cells, and networks. In Central Africa these communities have come to be called rituals or cults of affliction, defined by Victor Turner, a major author on the subject, as "the interpretation of misfortune in terms of domination by a specific non-human agent and the attempt to come to terms with the misfortune by having the afflicted individual, under the guidance of a 'doctor' of that mode, join the cult association venerating that specific agent" (Turner 1968:15–16). In some circles these communities are called "drums of affliction," reflecting the significance of their use of drumming and rhythmic song-dancing, and the colloquial designation in many societies of the region of the whole gamut of expressive dimensions by the term ngoma (drum). The drumming is considered to be the voice or influence of the ancestral shades or other spirits that visit the sufferer and offer the treatment.

This work is concerned with institutions carrying the designation ngoma and related terms. By entering African religious and therapeutic expression through its own language, we are identifying some important underlying, and possibly historic, commonalities and connections. We can also establish the basis for variants and transformations more intelligibly.

A number of modern scholars have looked at this institution in Cen-

tral and Southern Africa, although not always through the indigenously labeled categories. For example, Hans Cory, in the thirties, studied the constellation of ngoma groups among the Sukuma in colonial western Tanganyika and on the Islamized coast. His work for the British colonial government was concerned with the potential of these groups for social unrest. This work today provides a useful cross section of ethnographic and historic interest at one moment in time (Cory 1936).

The reference point of scholarship on African rituals or "drums of affliction" continues to be Victor Turner's work among the Ndembu of northern Zambia in the fifties; he introduced the term as a translation for the indigenous word and concept ngoma (Turner 1968, 1975). Turner's in-depth studies on several of the twenty-three Ndembu cults of affliction showed their inner workings and social contexts, intricate ritual symbolisms, therapeutic motivations, and societal support systems. At the same time, although he put forth the Ndembu as universal persons with believable aches, pains, and expressions, we now see that his account of them was largely ahistorical, localized in its coverage to the villages in which he did fieldwork, and presented in a largely static analysis characteristic of the prevailing structural-functionalist paradigm of the time. It was not clear in his work how widespread this genre of institution might be, nor whether it was particular to the Ndembu of Zambia on the Southern Savanna.

A variety of authors, researching and writing about the central and southern regions of the continent, described similar features in connection with the verbal cognate ngoma , but they usually did not make the connection between their own work and that of other scholars in other regions. In the era of structural-functionalism and colonial domination, the local "tribe" was the unit of study. Rarely were comparisons, or concerns for historical directions, articulated. However, useful work was accumulating which would make the task of historical comparison possible later on.

J. Clyde Mitchell (1956), a colleague of Turner's, followed the Beni-Ngoma movement into the migrant labor camps of the copperbelt. Terence Ranger found, in coastal and historic trade-route Tanzania, that the revivalist and dance dimensions of ngoma had followed the trade routes and population movements between early colonial settlements (Ranger 1975). Maria Lisa Swantz (1970, 1976, 1977a , 1977b , 1979) and Lloyd Swantz (1974) studied ngoma and related ritual heating on the Swahili coast in connection with social change and development. Wim. Van Binsbergen, Gwyn Prins, and Anita Spring

studied ngoma in Zambia. Spring's work added comparative ethnographic data from the "ngoma mode" of healing among Luvale women (1978, 1985). Prins and Van Binsbergen contributed to the history of western Zambian ngoma, the first to the cognitive framework of ngomalike therapeutic ritual (Prins 1979), the second to the linkage between numerous cults in the history of a region as an expression of differing modes of production and forces of historical change (Van Binsbergen 1977, 1981). Monica Wilson (1936), Harriet Ngubane (1981), and others studied the therapeutic ngoma settings in Southern Africa, where it was perceived as having largely to do with divination (especially among the Nguni-speaking societies).

Despite the value of these authors' writings on the subject of the cult of affliction, none has looked at the larger picture. They do not tell us how far-reaching the institution, as a culturally particular institution, might be. Luc de Heusch and Jan Vansina have been among the few to attempt broader surveys of possession cults in Central and Southern Africa. DeCraemer, with Vansina and Fox (1976), offered a summary profile of Central African religious movements, which they suggested were part of a cultural expression reaching back a millennium or more. But this article, suggestive in its general lines, did not provide a lexical or structural handle on how to study it further. De Heusch (1971) established a structuralist comparison of types of possession cults and relationships throughout the West African and Central African region that had far-reaching ramifications in such scholarship of the area. He emphasized, for example, the important contrast between possession cults, which entailed healing and exorcism, and cults that venerated shades and spirits, on the one hand, and cults that utilized mediumship for the interpretation of misfortune, on the other.

Ian Lewis, with a scholarly focus in the Horn of Africa, has offered important hypotheses on the nature of African cults and religions, first with his "peripheralization" or "deprivation" approach (1977), more recently with emphasis on the extent of "controlled" and "uncontrolled" power in society, and the relationship of witchcraft patterns to patterns of possession (1986). Most recently, DeMaret (1980, 1984) has used archaeological and linguistic findings to attempt an overview of Central and Southern African religious and social features. I will have occasion to come back to these authors and their work.

This brief review of some of the scholarship relating to healing in community settings in Central and Southern Africa suffices to demonstrate that the field is not well defined, nor is it clear where one begins.

Current scholarship tends to break down into a distinction between religion and healing, but this distinction is not so useful in the present setting. A fundamental ambiguity that will need to be worked out in this study is that between the indigenous categories and terms, on the one hand, and the analytical models we devise for such an institution as the rite, cult, or drum of affliction, on the other hand. The term ngoma has been identified as the indigenous word for an institution. And yet, in many regions it is not necessarily used, nor exclusively used, to describe collective rites of healing. In the course of this study, therefore, the layered ontology of the "unit of study" will need to be clarified and variations around themes explained. This will be done ethnographically or contextually, culture historically, ethnologically, and analytically, in sequential chapters.

My own concern for the understanding of the shape and character of African therapeutics began, like that of other scholars, with very local work—in Kongo society of coastal Zaire (1969, 1978a , 1978b )—and has gradually moved to increasingly expansive coverage of institutional arrangements (1979b ), therapeutic dynamics (1986, 1987), and historical processes (1983, 1985). Following intensive fieldwork in Kongo society on the "quest for therapy" and the structure of local institutions, I looked at a major historic cult, Lemba, which had emerged in the context of the coastal trade in the seventeenth century, and which had mediated the disintegrative mercantile forces of the overland caravan routes in that trade with the local lineage-based communities (1982). It became clear that local descriptions and explanations made little sense of the continuities and variations in Lemba. One had to take both the regional view of the cult phenomenon and a long-term historical perspective of the economic, political, and social climate to understand its emergence and duration.

After extensive reading in connection with my own local fieldwork and after historical study, it has become apparent that the cult of affliction and the ngoma designation of it is widespread throughout Central and Southern Africa, although there are many institutional and terminological variations. The scholarly task of the present moment, therefore, is to situate this work in wider regional, societal, and subcontinental context, and in the process to ask how widespread this institution might be, whether its many manifestations are transformations of an underlying common institution, why particular forms of it rise and decline, and how it relates in a dynamic relationship to other features of society and religion. A major lacuna in studying the wider phenome-

non of the cult or drum of affliction across its appearance in Central and Southern Africa has been the absence of a set of comparable studies. Scholars have either done local ethnographic studies with careful attention to the structure of customs and languages and have done little to seek broader generalizations, or they have attempted broader generalizations without careful attention to the cultural particulars.

In 1982–83 I undertook to remedy this situation for myself with an extensive field survey of ngoma manifestations in four settings of Sub-Saharan Africa where the literature suggested it occurred. I was especially interested in how ngoma impulses and organizations were represented in major urban settings. The sites I visited in this work were Kinshasa, a hub of Western Bantu societies, including Kongo, in the Zairian national capital; Dar es Salaam, where Eastern Bantu and national Tanzanian cultures come together, with a strong Islamic presence; the Mbabane-Manzini corridor in Swaziland, at the northern end of Nguni-speaking societies, in a strong traditional kingdom; and Cape Town, in whose black townships all Southern African traditions merge in the underside of a society torn by apartheid.

Why were urban settings selected, which tended to feature immigrants to cities and transplanted practitioners from home areas in the countryside? First, the urban capitals studied offered much more accessibility to regional traditions than single rural areas. Indeed, one could find all regional traditions represented in these cities. Further, it was easy to identify ongoing local scholarship on these traditions and to converse with scholars and practitioners about the unfolding direction of the therapies. Ongoing practice in the urban setting would demonstrate continuing life, although changing, of the institution. Finally, it was virtually impossible for me to do justice to the subcontinental survey short of visiting a selective set of points on the map, such as Kinshasa, Dar es Salaam, Mbabane-Manzini, and Cape Town. Of course, other capitals could have served equally well, including Harare in Zimbabwe or Lusaka in Zambia.

The comparative survey emphasized eight points regarding the therapeutic dimension of cults of affliction: (1) the names of principle rites, their regions of origin, and terminologies; (2) modes of affliction (following Turner: symptomatological signs) and etiologies (spirits, social forces); (3) the characteristic therapy of a rite; (4) the social scale of the affliction (whether individual, group, or combination); (5) the socio-cultural context—class and status, ethnic group, gender—of the afflicted and of the healer; (6) characteristic devices and musical instru-

ments, dances, and songs of the rites; (7) profile of individual(s) in charge of the therapeutic rite—family, diviner, other specialists, association members; (8) perceptible changes in the last decade. These queries provided the underlying thrust of the investigation and were answered in each of the four regions, insofar as possible.

Although this work will address a particular type of institution in concrete historical settings and is in many ways simply a straightforward attempt to understand and to portray this institution, the framework of the inquiry is intended to be universally applicable. In other words, there is a theoretical subagenda to this work, for which ngoma is the case study. I present this agenda in the form of three issues—health, healing, and efficacy. They must be approached, theoretically, in this order. This order may seem reversed to some; however, it stems from a growing concern in medical anthropology that this field is not effectively applied to health issues (Harwood 1987:4). I contend this is the case because of a lack of concern for the ways in which healing, or medicine, affects health, that is, the subject of efficacy, or how the therapy "works."

Increasingly, in social science and medical writing, definitions of health provide the point of departure for the analysis and action of specific interventions. This gets us squarely into the debate on conceptions of health, which authors approach from a variety of viewpoints. Most of the time we use the negative "absence of disease" definition of health, or the demographer's profile of mortality, natality, and morbidity. However, definitions of health may also be philosophical (e.g., Boorse 1977), ecological (Dubos 1968), political economic (Doyal 1979; Savage 1979; Morsy 1981), sociological normativist (Parsons 1951; Freidson 1971; Zola 1966), or ritualistic, discursive, and interpretative. Exploration of definitions of health suitable for the analysis of African ngoma therapy will be addressed in some length in the final chapters of this work.

The application of a concept of "social reproduction" seems particularly suitable here. Given the widespread network relationship building that goes on in ngoma, "health" may be seen as a society or social unit's ability to regenerate itself (i.e., socially reproduce). This approach is inspired by the work in Southern Africa of Colin Murray on labor migration in Lesotho and the outflow of labor capital, resulting in a crisis of social reproduction (1979, 1981). Other authors who have developed a social reproduction analysis include Pierre Bourdieu (1977). This approach overcomes the chronic problem in classical medical anthropol-

ogy and other disciplines of not being able to completely explain the deterioration of health in a society or a sector of society and the way in which members of society cope with this situation. The perspective of health as social reproduction will set the stage for an analysis of the collective therapies of Central and Southern Africa.

Establishing the character of the conscious therapeutic intervention as the basis for the comparative study of medical systems and traditions is the second major theoretical issue in this work. What will be the framework with which to analyze, in common terms, varied phenomena? What are the criteria of the "common," the "comparable"? Are they that which is labeled in indigenous practice and parlance? Or do they have to do with behaviors? In the case of Central and Southern Africa, do the common cognates of Bantu languages play a major role in determining what is the core of the historic and contemporary therapeutic system? Then there are the "institutional" questions, having to do with the primacy of the individual versus the collective, or societal. Central and Southern African therapies such as ngoma are so different, and differ in so many ways, from Western therapy, that we must first ask how the boundaries of researchable reality are to be drawn to identify this as medicine, or as healing, in order for it to have anything in common with the institutions the Western industrial world identifies by these terms.

Criteria of efficacy in therapy will need to be formulated, both in terms of specific therapies and interventions found in ngoma and in terms of the more general question of whether, and how, they may contribute to health. Both individual (psychological, symbolic, pharmacological, musical) as well as social mechanisms (entering and extending a network, creating support groups and redistributive chains, social competence) need to be studied as therapeutic mechanisms that may have generalizable qualities. Many of these measures enhance the ability of individuals and societies to contain trauma and to deal appropriately with difficulties, thereby contributing to social reproduction in the marginalized, alienated, or stressed sectors of a society, which ngoma therapeutics appears to address.

In order to accomplish the basic ethnographic-historical task of presenting ngoma and to open the theoretical discussions raised above, this book has the following structure. Chapter 1, "Settings and Samples," is a straightforward comparative study of four regional settings: Western Bantu, as found in Kinshasa, Zaire; eastern Africa, as found in Dar es Salaam, Tanzania; southern Africa, focusing on Mbabane in Swazi-

land, which is one of the North Nguni-speaking societies; and the townships of Cape Town, South Africa, predominately Xhosa, or South Nguni, but also a cosmopolitan synthesis of all of Southern Africa. This survey is largely a presentation of my field research of 1982–83, and thus it has all the strengths and weaknesses of a single scholar's work: limiting, in that it is only one individual traveling vast distances; enhancing, in that a trained eye can see much and make connections that a casual observer misses. In the western setting (Kinshasa, Zaire) I concentrate on the particular cults of affliction called Lemba and Nkita, of Lower Congo origin; Zebola, of the Equator origin; and Bilumbu, of Luba, or Kasai origin. Most of these are couched within the lineage setting or are designed to buttress the lineage. In East Africa (Dar es Salaam, Tanzania), because of the early work of Hans Cory on historic Sukuma ritual organizations, it is possible to offer a profile of both western Tanzanian ngoma and coastal Swahili, Islamized society, and ngoma expressions. From Southern Africa (Mbabane, Swaziland, and Cape Town, South Africa) come some of my best full accounts of ngoma, partly because of fieldwork luck and also because the institution may be less specialized there and may represent a more generic manifestation.

Chapter 2, "Identifying Ngoma: Historical and Comparative Perspectives," raises the possibility that ngoma is indeed a classical manifestation of Central and Southern African ritual. This chapter situates the book's subject in the context of research on the origins and dispersions of Bantu languages and cultures and the distributions of cognate lexica for ngoma and other African therapeutic-religious institutions. What is the evidence for culturally homogeneous domains beyond the linguistic tags? How do we account for the immense variations around the linguistic commonalities in this vast subcontinental region?

Chapter 3, "Core Features in Ngoma Therapy," develops a description of the main characteristics in ngoma therapy underlying the myriad manifestations of the institution throughout the region. These features include a phased rite of passage in which the sufferer, following identification of a sponsoring healer, moves gradually through the therapeutic initiation to membership in the order; a similar pattern of defining and interpreting misfortune through the invocation of, and possession by, ancestor shades, nature spirits, and other spirits; a common symbolism defining the status of the sufferer-novice moving through "the white," the ritual status of being "in process"; the role of sacrifice and exchange; and the empowerment of the novice through the transformation of the self and the composition and use of medicinal substances.

Perhaps the most important core feature, however, is the subject to which the next chapter is devoted.

Chapter 4, "Doing Ngoma: The Texture of Personal Transformation," moves beyond the behavioral and symbolic features of therapeutic initiation to the conscious, verbal dimension found in the ngoma sessions. A single session is described and analyzed in depth. It provides the basis for a wider comparison with other examples. The centrality of song to ngoma becomes apparent here. The variations in communicative structure of ngoma provide important clues to the understanding of the institution.

Chapter 5, "How Ngoma Works: Of Codes and Consciousness," proceeds with a presentation of the indigenous theory of this form of healing. From there it moves to the application of several academic analytic evaluations of ngoma, including the role of metaphor shaping, of consensus, and of the range of manipulations that shape affect of sufferer and therapists alike.

Chapter 6, "The Social Reproduction of Health," looks at ngoma from the standpoint of its contribution to society's fabric, attempting to answer the question of ngoma's contribution to health as understood in today's world. This chapter, of necessity, opens with a discussion of various health definitions, to determine which set might be appropriate for an understanding of ngoma's contribution to health in a contemporary context.

A project such as this is at once audacious and precarious. It is an attempt to demonstrate something that has not heretofore been known, essentially a mapping out of the core feature of a classic civilizational healing system in Central and Southern Africa, or at least a major feature of it. It is precarious because the assumptions that must be made in attempting this are not well validated. Working with linguistic reconstructions and variations around core behaviors often leads to interpretations of local evidence collected by others. One scholar, looking "over the shoulders" of others, is bound to be wrong some of the time in others' ethnographic backyards. To make matters more complicated, my ethnographic "home territory," Lower Zaire, in the Western Bantuspeaking region, fits the generalizations on ngoma least well, in some respects.

However, it will have been worth the risk if the end result, if only through criticism, provides the stimulation of new ideas and better research, especially that which goes beyond the confines of tribe and territory in Africa.