previous chapter
6 How Ngoma Works The Social Reproduction of Health
next chapter

How Ngoma Works
The Social Reproduction of Health

Is it good for the children?
Rick Yoder, Planning Adviser, Swaziland Ministry of Health

Exploring the issue of efficacy continues in this chapter, in terms of the survivorship of at-risk sectors of society and the role of ngoma in the creation and maintenance of a social fabric that contributes to health. Whereas in the previous chapter I identified practitioners' and analysts' theories of how ngoma rituals are intended to work, here I shall examine the consequences of ngoma upon the "social reproduction" of health. Social reproduction as used here refers to the maintenance of a way of life and the commitment of resources to relationships, institutions, and support organizations that directly or indirectly maintain health.

Four widely different ngoma settings illustrate the social-reproduction-of-health model of analysis: the historic coastal Congo Lemba order in the context of the mercantile trade at the Atlantic coast; the Southern Savanna natality-enhancing ngoma orders of Zambia and southern Zaire; the township ngoma orders of Cape Town; the professionalized ngoma institutions of Dar es Salaam.

The questions that are at the heart of this inquiry are these: What type of social fabric is created, reinforced, or reproduced by a particular ngoma order? How does it address health issues and needs? Can this be assessed in terms or definitions that health planners or practitioners could use?


Health And Health Indicators

The analysis of ngoma ritual healing in terms of health requires both a distinctive perspective on definitions of health and a bridge to the definitions of health commonly used by demographers and health planners. Health is a universal human goal, like virtue or enough to eat, but it eludes definition except in terms of specific negative or positive criteria. Myriad definitions of health exist.[1] This is not the context in which to present an exhaustive review of them. For present purposes, we require, rather, a selection of those that seem to fit the phenomenon.

It has been observed that good health is not necessarily determined by medicine, especially not curative biomedicine, narrowly defined (Navarro 1974; World Health Organization 1978; Dawson 1979). Rather, a range of factors such as nutrition, housing, environmental quality, social order, and mechanisms for coping with stress are important (Gish 1979). Further, if improvement of health on a societal scale is to be attained, then health policies must be appropriate so as to avert health catastrophies or gradual declining levels of health.

One way to describe health, suitable for present purposes, is that it is embedded in a set of structured relationships, rights, and practices rooted in a worldview of values, truths, and ideals. In a stable situation this would include hygiene (although methods would vary), adequate and clean water, sufficient and clean food (although this would vary from nearly all-meat to all-vegetable), rules of social interaction and organization (varying enormously), and rituals spelling out the coherence and interconnectedness of things. As we will see, this understanding of health could be "paraphrased" in terms of the ecologist's concept of the "adaptive system," in which the human community obtains its food source through a given technology, a given social order, and a structured flow of energy from the environment to the human community and back into the environment (Janzen 1980:7–8).

More focused in scope and scale would be a description of health based on the society's self-conscious efforts to treat disease and restore health—that is, its medical or ritual systems. Here we would find at work "an explicit theory (or theories) of disease causation, a corresponding set of therapies, and a focus directing the scale of application of such theories and therapies. These aspects of the medical system will be evident in symptoms, complaints, and verbalized statements about illness and health, in the cultural premises used to evaluate the meaning of affliction, and in practices of heaters" (Janzen 1980:9).

Despite the appropriateness of such conceptualizations of health for


the purpose of understanding ritual healing, it is useful to begin this discussion with the types of health assessments that have generally been done with more precise measures of fertility, mortality, and morbidity—the demographer's stock in trade. We learn from a World Development Report issued by the World Bank (1985), that crude birth rates and crude death rates in Sub-Saharan Africa's "low income countries" and its "middle income countries" have declined from 1965 to 1983. For 1965, the birth rate in low-income countries was 48/1000 and the death rate 22/1000; for middle-income countries the birth rate was 50/1000, the death rate 22/1000. This produced a population gain per annum of about 2.7 percent. By 1983 both had declined, to 47/1000 births and 18/1000 deaths for low-income; 49/1000 births and 16/1000 deaths for middle-income—a population gain per annum of 3.3 percent. Thus, death rates declined faster than birth rates, leading to an overall growth in population increase rates.

A good portion of this decline in crude death rate came about through the decline in infant mortality rates. In 1965 IMR was 156/ 1000 in low-and middle-income countries in Sub-Saharan Africa. This was a great improvement over the 350/1000 IMR that had prevailed in the continent earlier in the century. In a few countries, such as Kenya, IMR had been brought down to under 100/1000 by the late 1980s. These "improving" health figures were, however, offset by 6 to 7 percent decline in food production, which translated in the lower-income countries into a nearly 10 percent decline in calorie intake (the middle-income countries saw a slight improvement in this category). Insofar as one can then extrapolate from these composite statistical figures, survivorship has increased, although real income and food production has decreased.

Morbidity patterns in the ngoma region, as measured by agencies such as the World Health Organization and standardized measures, indicate that although some of the major contagious diseases such as smallpox, cholera, and measles had been eliminated or brought under control, others such as diarrhea and respiratory infection continued to be the greatest killers of infants. Generally, the disease profile continued to be that of contagious diseases rather than of degenerative and other "diseases of civilization."

What do these figures mean in terms of our understanding of the backdrop of ritual healing, or the status of health in the region with which we are concerned? Does ngoma relate to these population measures?

There is a discussion that concerns our agenda, perhaps indirectly, having to do with whether Sub-Saharan Africa's health measures repre-


sent a conventional pattern of demographic transition, comparable to that which has been seen in Europe and Asia. There, following the decline of mortality and morbidity earlier in this century, fertility measures declined as well. In Africa, although mortality and morbidity have declined, fertility has probably increased. This pattern may be related to what ngoma does.

One set of analysts argues that the demographic profile of Sub-Saharan Africa is to be explained by the peripheralization of the subcontinent in the world capitalist system, and by capital's need for a large, cheap, surplus labor pool (Gregory and Piche 1982; Cordell, Gregory, and Piche 1992; Stock 1986). Another group (e.g., Lesthaege and Eelens 1985) evaluates what they call "demographic regimes," characteristic profiles, with an interest in the fate of socioeconomic and cultural mechanisms such as child spacing, household structures, and number and composition of family members available for childcare, which might affect overall demographic trends. They document the apparent collapse in recent decades in Africa of a number of these long-term historic structures and the consequent increase in fertility rates. A related perspective has suggested that much of Central and Southern Africa's social priorities are derived from a lineage base and a particular technology of hoe agriculture (Goody 1976; Caldwell, Caldwell, and Quiggin 1989) and that so long as these institutional structures prevail, health and fertility measures will be affected in distinctive ways.

The outcome of this debate on Africa's demographic transition will not be clear for some time. We know health has improved considerably in the past half-century, but there is still much room for improvement in diseases and conditions for which cures are available. We know fertility has not begun to decline, and in some areas it has probably increased. Clearly many of these factors affecting health status have to do with knowledge, access to resources, and social policies. Therefore, in our assessment of the role of ritual healing in affecting health status, we do well to keep "hard" demographic data in mind but to develop social theories of the ways health-related resources are generated and utilized. This is why it seems appropriate to bring the notion of "social reproduction" to bear upon the analysis of health and healing.

The Social Reproduction of Health

The definition of social reproduction with which this chapter began—the maintenance of a way of life and the commitment of resources


to relationships, institutions, and support organizations that directly or indirectly maintain health—was a generic statement derived from a number of writers whose work on this concept may be presented more fully here.

In assessing a "social reproduction" concept of health, it is important to identify the social units or sectors involved and to identify some indices of this process that are separate from biological reproduction and the reproduction of labor. Meillassoux's work offers one perspective on this type of analysis. He separates the "domestic community" as a social formation from both biological reproduction and the reproduction of labor for capitalist needs (1981). This analytical model distinguishes the energy or "social product" needed (a) to reconstitute productive adult producers, (b) to nurture future producers, that is, "not yet productive" children, and (c) to maintain the postproductive elderly and the sick (1981:51–57). The sum of these products offers an indication of what is required to socially reproduce the domestic community. Relative surpluses enhance and enrich the community; relative deficits erode it. Over several years one can in theory determine the level at which a community reproduces itself or falls below a minimal replacement level.

Other authors go beyond Meillassoux's approach to include more than material needs in the calculus of social reproduction. The distinctions between social, biological, and symbolic reproduction are further developed in Pierre Bourdieu's well-known study of the Kabyle of Algeria, published in his Outline of a Theory of Practice (1977). Here the patrilineage, composed of households headed by brothers, is the principal social institution. Various centrifugal forces are at work to bring disintegration to the lineage. Strategies to pull the lineage together, to maintain the family—socially, symbolically, and biologically—concentrate on the appropriate marriage. This is often a marriage of a man to his father's brother's daughter (bint amm ), so common in pastoral nomadic societies. Such a marriage serves materially to keep the herds and other aspects of the estate intact. However, Bourdieu points out that such a strategy involves far more than just economic management (1977:60). The ethos of honor attaches to the unity of the land, to equal status alliances, and the unity of the agnatic group, the prestige of the house. Bourdieu says it is impossible to separate ends and means in the collective matrimonial strategies. Every marriage tends to reproduce the conditions that have made it possible. "Matrimonial strategies, objectively directed towards the conservation or expansion of the material and symbolic capital jointly possessed by a more or less extended group,


belong to the system of reproduction strategies , defined as the sum total of the strategies through which individuals or groups reproduce the relations of production associated with a determinate mode of production by striving to reproduce or improve their position in the social structure" (1977:70).

These issues of threshold levels of household and community maintenance are addressed in Colin Murray's work "The Work of Men, Women and the Ancestors: Social Reproduction in the Periphery of Southern Africa" (1979), in Lesotho communities that are deeply involved in oscillating labor migration to the mines, cities, and factories of South Africa. As has been abundantly documented, Lesotho has moved from being an autonomous agrarian society earlier in this century, which exported grain, to being a society now largely dependent on wage labor, and which imports most of its food. Disease levels have risen during this period; tuberculosis in particular constitutes a major endemic disease (Murray 1981). Aggregate demographic data for Lesotho are comparable to data elsewhere in the continent, with crude birth rates having remained constant at 42/1000 per annum, and death rates having declined from 18 to 15/1000 from 1965 to 1983, resulting in an increase in population growth over that time from 2.4 to 2.7 percent per annum (World Health Organization 1979; World Bank 1985).

Murray's careful anthropological fieldwork suggests that these aggregate demographic data mask the significant intra-community and household disparities, having to do with household makeup, related mortality and morbidity, and overall health rates. Up to 70 percent of the households are managed by women who are almost entirely dependent on their husbands' wage labor for survival. Only 6 percent of household income came from the sale of farm produce. At highest risk for disease were those families of single household heads, and children in families in which the spouse of the resident head did not provide a cash stipend (Murray 1979:346). In these homes infant mortality often reached 50 percent, far higher than the 120/1000 average. Although these "at risk" households reproduced the labor pool for the South African industries and contributed to population increase, they were not "socially reproducing themselves." In Meillassoux's equation of social reproduction, they plainly reflected a social product deficit.

In Murray's account, social mechanisms most often utilized in Lesotho society to achieve social reproduction were "inter-household income transfers," such as cash and in-kind remittances, bridewealth transfers, share-cropping arrangements, and other contractual and re-


ciprocal arrangements connected with agriculture; also, informal sector transfers such as beer brewing, petty trading, and concubinage, all of which maintained a wider scale of social relations than the household, and thus extended viable social support links for those in short-term or long-term need. In noting a further feature of Lesotho society that is directly related to the way ngoma reproduces society, Murray demonstrates how the exchanges and feasts of ancestor rituals, often in connection with leaving for, or returning from, migrant labor, play an important role in forging and renewing the alliances needed to survive the absence of the family head (Murray 1979:347). Although ngoma in a narrow sense is absent in historic Sotho-Tswana society, perhaps because of its centralized judicial institutions, the initiation case in chapter 1 is that of a Sotho man whose family is caught up in labor migration.

A further noteworthy author who has taken a view comparable to "social reproduction of health," without utilizing these exact words, is Steven Frankel, an M.D. -anthropologist who has worked among the Huli in New Guinea. Frankel's medical anthropological study of the Huli of New Guinea (1986) develops both the negative indices of health, the "absence of disease," and the positive concept of "social effectiveness." The elaboration of rhetorical skills, esoteric knowledge, ritual practices, and cosmetic decorations are considered essential in the ability to be effective in social exchanges. These are seen as prerequisites in soliciting others' generosity and ultimately enabling an individual to care for a family and to lead the community. For the Huli, and for Frankel, social effectiveness is seen as operating at a level to include not only the household but larger societal levels as well, and not merely the material basis of existence but the symbolic exchanges needed to extend public institutions.

In sum, the concept of social reproduction offers a model of how therapeutic rituals such as ngoma might prove efficacious in enhancing health. Can it be tested?

Profiles Of Ngoma Social Reproduction

The social networks and therapeutic cell communities formed through the long-term association of master-healer and novice, as well as the "lay" clients and the assemblies of people at ritual events, may well offer, in their ability to recreate society, the most pronounced characteristic


of ngoma therapy in achieving and maintaining health where it has collapsed. The structure of these emergent social forms shows the social reproduction around sufferers and former sufferers, healers and the healed. The four illustrations of this process that follow will focus, as noted earlier, on the type of social fabric that is created, reinforced, or reproduced by the particular ngoma ritual so as to more effectively address health needs.

Reconciling Lineage and Trade in Precolonial Kongo Society

In the setting of the ancient Lemba cult of affliction in Lower Zaire, the local lineage was challenged to come to grips with the mercantile wealth of the great trade with the coast without having its egalitarian ethos shattered. I have suggested, in chapter 1 of the present work and elsewhere (1982:70–79), that this was possible through the judicious creation of alliances between lineages, which forged links across the countryside along the trade routes, and through adequate exchange and distribution within these nodes of society, thereby assuring that Lemba members could safely travel from market to market, and to the coast with their caravans.

The demographic profile of coastal Kongo society during the seventeenth to early twentieth centuries, during which Lemba existed, is known to have suffered from an overall decline of about 50 percent of what it had been in the sixteenth century (Sautter 1966:2–71). An objective assessment of the factors involved in this would include not only the slave trade, which drew the best young adults out of the society, but the compounding influences of societal breakdown resulting from raids and epidemics.

Lemba initiation rituals approached these conditions at several levels. One was the divination and treatment of specific symptoms and signs related to the fear of subordinates' envy. However, a more important criterion of the initiates' acceptance was their ability to pay for the final stages of the initiation rite, the "graduation," whether with their own or their lineage's patronage. Effectively, Lemba was a cult of affliction of the elite households of north bank Congo River society, in the face of the disintegrative forces of the mercantile trade with the coast.

The social structural particulars of this arrangement are well known. Of strategic significance in the whole Lemba scheme of social reproduction was the Lemba household, which brought together two types of


groups. One was the alliance that linked major landowning freeholder lineages in adjacent communities; the second was the alliance that bonded such freeholder lineages with client lineage fragments. In the first instance, the marriage tended to be of the patrilateral, cross-cousin type between lineages of equal status, between whom equal exchange marked successive marriages of this type. In the second setting, they were often, from the perspective of the marrying male in the dominant lineage, matrilateral cross-cousin marriages. This pattern expressed the unequal exchange between the two lineages, serving nevertheless to weld the community of several exogamous lineages together in a hierarchic local society. The Lemba marriages between freeholders assured a regional network for the trade and peaceful relations in a region where no historic state extended its hegemony. The Lemba marriages between unequal—master and client, or slave—lineages enhanced the local community by enlarging its population and political base. Indeed, the rhetoric of Lemba stated that the lineage in possession of Lemba "could not die out."

In these two ways Lemba helped to socially reproduce the society in the face of the centrifugal forces unleashed by the coastal trade from 1600 on, economic divisions within lineages, slave raids, feuds and wars, and epidemics. We have no way of knowing whether Lemba diminished the episodes of fear of envy by subordinates, either through protective medicine or through redistribution of goods and food. However, it is clear, and north Kongo informants stress, that Lemba was usually an important deterrent to local feuds and thus averted the bloodshed, loss of property through burned houses, and chaos that otherwise resulted from local wars. In this sense Lemba did have a measurable effect on the well-being of the region where it was implanted.

We have only glimpses of the numbers of individuals in a region involved in Lemba. In terms of the percentages of marriages that might have been "Lemba marriages," extrapolating from historical data in the village on which I have such records, it appears that less than 5 percent were involved. Yet as the elite, they were influential, and the impact of Lemba was considerable.

Saving Lives of Mothers and Infants on the Southern Savanna

Another area in which we may usefully examine the "social reproduction" hypothesis of the efficacy of ngoma rituals is in the enhance-


ment of conception and successful birthing and the survival of healthy children. Although ngoma-type intervention has been studied in the Western Bantu Nkita rite described earlier, and although Victor Turner studied certain ngoma rites among the Ndembu, these studies do not focus on epidemiological or demographic variables. However, we now know that the region from the Atlantic coast inland to western Kasai in Zaire, and the Southern Savanna to western Zambia, and northward to Congo and Gabon, constituted—and in some areas still constitutes–the widespread "infertility zone" (Gaisie 1989).

The work by Anita Spring (1978, 1985) and Veronique Goblet-Vanormelingen (1988) on this ngoma-style institution on the Southern Savanna of northwestern Zambia and southern Zaire, respectively, is exceptionally valuable for its attention to the demographic and epidemiologic indicators associated with the purported goals of ngoma. Spring has studied the Kula rite and related mahamba (generic shrines) among the Luvale; Goblet-Vanormelingen has studied an institution called Mbombo among the Mutombo Mukulu Luba of Shaba Province in Zaire. These are similar in their emphasis on reproductive difficulties in the cults of affliction of Wubwangu, Isoma, and aspects of Nkula which Victor Turner (1968) described among the Ndembu of Zambia.

The core features of ngoma identified in chapter 3 may all be found in the reproductive enhancement rites on the Southern Savanna. The mode of affliction is identified as spirit-originated threat to the newly conceived fetus. After an initial rite of entrance, the pregnant woman leaves her husband and enters seclusion in a special enclosure constructed in the homestead of her sponsoring healer-gynecologist-midwife, where she is taken care of with anti-abortive medication, special diet, and hygienic attention. Upon the successful birth of the child, in some variants, and as much later as the first steps of the child in others, seclusion ends with a second-stage ngoma graduation and final entry of the mother into the order. In Mbombo, as described by Goblet-Vanormelingen, the end of seclusion comes shortly after the birth of the child, at which time the mother and child are washed in a nearby river and presented to the father. Ngoma music accompanies this "coming out" of the new child with its mother.

Both Spring and Goblet-Vanormelingen are interested in the "efficacy" of these birth-enhancing procedures. Spring criticizes Turner for paying little attention to this question and for assuming that Isoma (for miscarriages, abortions, stillbirths, barrenness, menstrual disorders, illness of infants), Wubwangu (for twin pregnancies, infant disorders, barrenness, miscarriages, menstrual disorders) and Nkula (for


menstrual disorders, barrenness, miscarriages, ill health of infant) were primarily social, symbolic, and religious in nature. She is interested, first of all, in the "epidemiology of ritual participation" and in whether the seclusion procedures enhance survivorship. One important predisposing issue to be taken into account in the Luvale region of Zambia is that although the continuity of the lineage and family is highly valued, infertility and subfecundity of women is common and survivorship of infants is low. About 20 percent of Luvale women in Spring's sample of forty-five-year-old women had been barren during their adult lives. A high percentage were subfecund, that is, had gone five years without a live birth during reproductive years. Further, infant mortality was about 150/1000. This resulted in a completed family size on average per woman of only 2.05 children for the society, which is barely a replacement level (Spring 1978:175–176). The Luvale were understandably concerned with fertility enhancement.

Reasons for the high infertility included high levels of genital and urinary-tract disease, much barrenness, and fetal wastage. These are in part reflected in the high level of abdominal pains, dysmenorrhea, and fevers probably caused by bilharziasis, gonorrhea, and nonspecific bacterial infection, in addition to malaria, hookworm, and amebiasis, which are endemic (1978:176).

The isolation therapy at the time of conception is part of a more general cultural strategy of the Luvale to improve the chances of offspring. Childless women are first of all given treatment for barrenness. If they become pregnant, they receive the performances of several mahamba cults at a number of possible points in their reproductive years. If a pregnant woman has a miscarriage or delivers a stillborn child, she is a likely candidate for the seclusion ritual. If she has menstrual dysfunctions, she will receive herbal treatments, which, if unsuccessful, will be followed by further ritual treatment. If a woman's child becomes sickly, both mother and child will be secluded. If a woman's small child dies, she will receive seclusion. If a woman becomes ill with problems unrelated directly to childbirth, she will be a candidate for the seclusion ritual. "Possession" cult initiation may thus occur at any one of a number of points in the course of a woman's reproductive years. The etiology of spirit or shade involvement is usually made official by the diviner. Rarely does it involve trance, although the ngoma-type song-dance, a variety of particular medicinal and technical treatments, food prohibitions and special diets, the "white" symbolism of seclusion, and the two-stage passage are integral features of the rites.

Fully half of the women in Luvale society are initiated to one or


another manifestation of reproductive cults by the time they reach the end of their childbearing years. In Turner's Ndembu sample, women's reproductive issues were by far the most frequent encounters of individual Ndembu with the ngoma system. In his sample from two areas, Turner noted that nineteen of twenty-six women had gone through Nkula; twelve of twenty-four through Isoma and Wubwangu each (1968:303). Goblet-Vanormelingen, whose work is still in process, gives no statistical information of this type. Spring emphasizes that these reproduction enhancement rituals are thus responses to both generalized physical ill health and particular concerns of Southern Savanna families for effective biological and social reproduction.

How effective are the rituals and their related interventions? Goblet-Vanormelingen judges as "truly beneficial" the following aspects of Mbombo ritual: the continuous assistance of the healer and the woman's husband, creating an encouraging atmosphere for her; some rules of behavior, particularly the necessity of living away from the stresses of family life, the dangers of exposure to contagious diseases and work-related infections, and having complete rest; and, after childbirth, seclusion to reinforce the mother-child bond. More dangerous to the health of mother and child are certain practices in preparation for birth and delivery, which, because they are nonsterile, may increase the risk of infections or tetanus, and certain food prohibitions that appear to restrict intake of nutritious food. On balance, believes Goblet-Vanor-melingen, the Mbombo rituals appear to enhance the chances of wellbeing of both mother and child.

Spring's analysis of Luvale womens' reproduction rituals, although it does not lend itself to a precise statistical evaluation, does suggest that the isolation therapy and the maintenance of networks of ngoma orders probably improve the health of mothers and children, that is, increase survivorship. Spring's evidence suggests that most women who are divined to require ngoma rituals do enter them, and in their later years these women become the doctors of these cults. This means that because they are at risk, survival ratios of the infants of those women who are members of, involved in, and leaders in the ngoma orders will be lower than among nonmembers. Survivorship must be counted in terms of pregnancies saved that might otherwise have been lost and surviving infants who might otherwise have been lost.

Anita Spring focuses her discussion of efficacy—although she does not use the term—around the structure and the strength of the network of women active in reproduction enhancement, and what this does for their self-image, social role, and the structure of the community. The


sequence of being a sufferer-novice, an apprentice, and finally a cured-doctor results in a cooperative system of social relations permitting women to gain, and perpetuate through practice and teaching, the knowledge of how to deal with women's reproductive concerns. As in other ngoma contexts, adversity is turned into strength, anxiety into specialized knowledge, suffering into healing. The institutional framework brings spirit possession, as an ideology, into the set of etiological beliefs, although the type of knowledge needed to deal with infertility, threatened miscarriage, stillbirths, sick children, and the rest, is highly practical, what in the West we would call empirical and rational knowledge, rather than a trance state.

Not all ngoma attention to children is as salutary as that of the Southern Savanna. On two occasions in the 1982–83 survey I witnessed practices that I thought might lead to the child's death, rather than to its recovery. Both had to do with divinatory possessions. The first occurred in Kinshasa as Mama Kishi Nzembela, the Luba Bilumbu medium, allowed her apprentice medium to flail about with her child pressed between her legs, shouting that it was evil and that she had an evil spirit. Obviously the young woman was unhappy, and in a reckless abandon she took her misery out on her child. The child escaped that episode unscathed, but the mother nearly smashed its tender skull on the cement floor, with which it came into very close proximity on several occasions. The other time was during a divining session in Swaziland when the most elaborate and high-priced mediumship divination was done to determine the cause of the sickness of a small child. Granted, in both cases, the infants were born to women who were not married, or whose marriages had ended. Thus, ancestral and spirit displeasure as cause of a small child's respiratory infection was a type of diagnosis that should have been abandoned in favor of adequate clinical and parental care. Although these two mediumship divinations may have enhanced family solidarity, they did not contribute to the wellbeing of the infants involved.

Rick Yoder's concern, "Is it good for the children?" was not an idle one with regard to ngoma rites.

Regional Networks of the Isangoma/ Amagqira of Southern Africa

The demographic profile of South Africa reflects both third world and first world conditions. White society has low mortality and fertility, on par with Western Europe, and a morbidity profile in keeping with


an industrialized, affluent society. The leading causes of death are degenerative and cardiovascular diseases. By contrast, the society of the black townships reflects conditions of high mortality and fertility; the leading morbidity causes are infectious diseases and accidents. Among adults, tuberculosis is a serious problem. Among children, respiratory and intestinal infections and contagious diseases are the leading causes of death. South African homelands, whose health statistics are often not published in official records because some of them are "independent," have the highest infant mortality rates in Africa, near 300/1000.

Ngoma is well-represented in the black townships of the Western Cape, as noted in earlier chapters. However, its relationship to the foregoing health indicators is not well understood. Little work has been done on the subject. The composition of ngoma cells is not as strict or well-defined as in Central Africa. Ngoma activity, as noted, exists in a single inter-ethnic mode, rather than being thematically specialized and ethnically distinctive. Cells may be organized as informal friendship alliances between healers (sangoma, amagqira). One igqira with whom I discussed the composition of her group of collaborating healers, suggested that she met with "her friends," whom she invited to her events with about a week's advance notice, on the basis of friendship and compatibility, that is, that they did not drink excessively. It was along such lines that information would be passed and mutual help and gifts exchanged over divination, counseling, and healing-initiating.

In five related events I observed in Cape Town, a loosely linked "star" network pattern emerged, suggestive of overlapping sets of fully initiated sangoma or amagqira (see fig. 13). To the skeletal network of Western Cape diviner-healers, one must add an array of, on average, ten novices allied with each healer, plus the novices' families and friends. For the five amagqira with whom I was able to discuss this, each had nine, twenty, ten, ten, and six apprentices respectively. These novices regularly attend purification events on the occasion of a death in the family or kindred of any one of them and also of their master-healer; they attend all events put on by their master-healer, such as initiations and graduations. The novices' own families are to some extent involved, if not directly in the events, then indirectly in the benefits of the regular food distributions made at the time of goat, sheep, and cattle sacrifices and feasts. The attendance at one typical event, the initiation mentioned in figure 13, included six master healers (three Xhosa and one Zulu woman, two Zulu men), two female Xhosa senior novices, six Xhosa


Figure 13. 
Pattern of association of individual healers and their novices in five
events in Western Cape ngoma networks, late 1982. Capital letters refer to
senior healers present at particular events; clusters of enclosed  x 's refer to
these healers' novices in therapy/training with them.

female novices and one male Xhosa novice, a Sotho male initiate, and seven unidentified additional novices. Several dozen observers were on hand, including the initiate's family and neighbors.

The strength and depth of this type of organization in Cape Town township society is not easy to generalize without adequate survey information. However, a rough estimate may be projected from numbers of master healers in the two streets in Guguleto and Nyanga where ngoma participation was assessed. One street had four full igqira residents; the other had two, plus an ixwele herbalist. I was told that this was a common degree of representation for the townships. This average of 3.5 amagqira per street, however shaky its statistical significance, times the 153 equal-length streets and houses in two townships, would suggest that there are 535 fully qualified sangoma/amagqira in Guguleto and Nyanga. This figure, doubled to include the townships Langa and Crossroads, would yield an estimate of over a thousand healers of the ngoma


type in the townships of Cape Town. If the Western Cape Regional Authority's figure of 200,000 African inhabitants is used, there would be approximately one ngoma healer for 200 inhabitants.

However, to estimate the full extent of ngoma networks in the populace, this figure must be multiplied by ten, for the average number of each master healer's apprentice-novices. Assuming five to seven individuals in each household, this would indicate ngoma involvement, as master healer or apprentice, in one in four households.

Such brandishing of statistical information is admittedly of dubious value. The claim that ngoma cells and networks socially reproduce relations needs to be given context not only in the ngoma setting but also in the domestic setting out of which the individuals originate. Evidence for this is again somewhat anecdotal and case-study specific. A review of one igqira's roster of current (1982) apprentice-novices, their family settings and domestic relations, and work records showed the following. Novice one, a married woman with five children in the house, had entered ngoma seven years earlier. With a secondary school degree, she worked for her husband's boss at a construction company. She was a "five to" (midnight), that is, near completion, and was assembling goods for her graduation. Novice two, also married to a man with "good work," and with children in the home, worked as a cleaning woman. She had begun seven years earlier and was gathering resources for her graduation. Novice three, married to a man with steady work in a blanket factory, also worked as a cleaning woman. Her children were small when she entered ngoma nine years earlier. She was trying to collect goods for her graduation. Novice four had children at home and was married to a taxi-bus owner who employed other drivers. She did not work outside the home. She had begun ngoma nine years earlier, having been a Zionist, and was a "five-to" assembling goods for her graduation. These four women, with seven to nine years experience as novices, were the mainstays in ngoma sessions under the direction of their master healer.

A second group of four novices in this cell seemed to have greater difficulty moving "through the white." They were all single mothers working outside the home. Novice five worked as a domestic and had three children. She had begun ngoma nine years earlier but had only progressed to the status of junior novice. Novice six, a single mother of one daughter, also worked as a domestic. She had problems with alcoholism and was not progressing well in her therapeutic initiation. Novices seven and eight were a mother-daughter set, living together


without husbands. The daughter was a schoolteacher. Both were junior novices.

Two new apprentices were both parents in relatively stable marital relationships with small children at home. Novice nine, a member of the Bantu Presbyterian Church, was married to a man with a good job. Novice ten was a married man with children and stable employment.

Few generalizations may be drawn from this set of ten novices in one healer's group of apprentice-novices. One common factor is that they enter ngoma as young or middle-aged adults, and most require from five to ten years to move through the novitiate. But little is known about the "epidemiology" of ngoma involvement in South Africa.

Janet Mills, however, has sought to identify factors within the household that might be associated with the appeal to ngoma (1982). Although the explicit reason commonly given for entering an ngoma cell is the "call" (twasa) from ancestors, she has investigated the possible correlation between this support seeking and active tuberculosis cases in the household. She found tuberculosis to be slightly more frequent in ngoma-related than in non-ngoma households. Although no conscious or explicit linkage is made in ngoma participants' explanations between tuberculosis and the call to join ngoma, the possible basis for such a link may be hypothesized. It is unlikely that ngoma healing is directly beneficial for tuberculosis. Rather, it may be that just as tuberculosis—endemic in its latent state—has erupted in active infection, the energies of productive individuals have been so strained as to lead to support seeking in an ngoma or similar network. Tuberculosis is endemic in South African blacks, and a variety of stressors, such as inadequate nutrition, poverty, lack of adequate shelter, can bring on an active episode. Also, prolonged rest along with medication is required to recover from an active episode. Participation of the care providers—that is, spouse—in an ngoma cell or network would provide the requisite support, additional contacts, and sources of aid needed to deal with the long-term crisis of tuberculosis in the family.

Although one could figure the nutritional intake offered to participants in the ngoma network events as a possible point of departure for the analysis of its existence, more significant may well be the social investment in ngoma therapeutic structures. If one in four households is involved, this network covers the entire urban society, as well as connecting it to other cities and rural areas.

This significance of social reproduction, which I argue is also health-building efficacy, would seem to make an impact at three levels: the


nucleus of the master with apprentices; the broad network spanning the whole society; and those households that are connected with ngoma. During the years of apprenticeship-initiation-therapy, the master heater serves as role model, counselor, therapist, guide; in return, the master may expect services and some goods from the novices. Although the novices or their family pay their masters a goodly sum of money for the sacrifice animals and for their own costume, they stand to benefit in the reshaping of their lives, in sorting out problems, finding contacts to jobs, and referrals of all sorts. Intense resocialization occurs within the group of novice-peers. The possibilities go beyond this, to the more public network into which the master healer is the "hub" of the wheel, radiating out along "spokes" to many households, and ultimately, to the entire society. A case can be argued that twasa is not madness, but the call to social reproduction. It is, as Harriet Ngubane has suggested (1981), a pan-societal network extending across Southern Africa, and, as we now know, well beyond.

Professional and State Control of Ngoma on the Swahili Coast

The final example of the social reproduction of health through ngoma is from coastal Tanzania and shows yet another variation of the therapeutic focus and organizational structure of the institution. Among Zaramo and Zigua peoples of the Swahili coast, the coastal healers have utilized their ngoma networks to create a centralized institution, the Shirika la Madawa ya Kiasili, with officers, books, a treasury, and a "director," that is, a representative of government to be a liaison with appropriate ministries. Local ngoma performances draw, as everywhere in ngoma, a shifting set of senior participants with their apprentices and novices. At another level, there are individual waganga who belong to patrilineages, which in the Swahili coastal setting have often passed their therapeutic skills from generation to generation.

The Shirika had about five hundred members in 1983, with strongest representation in Dar es Salaam, Tanga, Bagamoyo, and Morogoro, and in the new capital, Dodoma. They were trying to create branches in other localities. Tanzania had earlier experimented with a national organization of healers, but this was abolished when it became too politicized, that is, fractured and powerful. The links between the Shirika and the government revealed some of the same forces of economic


interest, of political strength, and of party and government control that had surfaced in the national organization. Why was the state interested in ngoma?

The Shirika had official recognition through the Ministry of Culture, which generally handled licenses for ngoma entertainment groups and conducted research on song-dance. Why, though, should therapeutic dancing be controlled by the Ministry of Culture and not the Ministry of Health? The power of ngoma as a resource, especially one so large and well organized, was indicative of its symbolic power in society at large. The role of a party-appointed "director" with affiliation to the national political party hinted at some of the potential tensions between the organization and the state's interest. The director in effect controlled the recruitment of new ngoma cells to the Shirika, thereby exercising restrictions on its overall influence. This would avert what had occurred in the case of a cattle-rustling ngoma that had arisen in response to the need to locate stolen livestock. The government had heard about it only after it was fully constituted and then belatedly tried to gain its allegiance. Somewhat similarly, the ngoma for entertainment organized by the National Service indicated the effort to connect the state to the powerful symbolism of socially focused song-dance.

These details of the relationship between state and ngoma illustrate a significant general principal. A resource such as ngoma, which may arise in response to a need, and which symbolically, socially, and materially reproduces itself, by that very fact attracts the state, which seeks to co-opt power and legitimacy unto itself.

The resources to be drawn from the Shirika—an umbrella organization of therapeutic ngoma—were also apparent to the individuals and families who made their living from these performances. Unlike the populistic networks of reproduction-enhancing ngoma of the Southern Savanna or the township ngoma of the Western Cape, here by no means everyone who was brought into the initial stage of treatment followed through with the full initiation. In fact, according to the records and testimony of the Shirika leaders, only about 3 to 4 percent of the sufferers who entered ngoma dispensaries for treatment were fully initiated. This is corroborated by research of a decade earlier, in which only three of sixty waganga had entered their healing profession through spirit calling (L. Swantz 1974:203).

In Dar es Salaam, the recognition of healers and the opportunity for them to organize into associations has led to the strengthening of their control of resources (Unschuld 1975). On the one hand, this has led to


full-time practice, the utilization of therapeutics as income, and to the control of the therapeutic and symbolic resources, in this case divination, the diagnosis of spirit possession, and the performance of authorized therapeutic song-dance. On the other hand, this has led to the restriction of access to the role of the ngoma healer and to the consolidation of membership in the Shirika.

The effect of professionalization of health care elsewhere has been the codification of methods and the regulation of access to the ranks of those who practice (Last and Chavunduka 1986).


This chapter has offered an approach to the study of the relationship between social organization and the allocation of resources, and to their impact on health in the context of ngoma-type healing. Social reproduction theory, as put forward by Meillassoux, Bourdieu, Murray, and Frankel offers tools for a more rigorous analysis of the manner in which society itself structures the resources of health.

Unfortunately, very little research has been conducted to actually test propositions about the efficacy of ngoma-type therapy in terms of measurable health indicators. In the cases we have reviewed here, only the work by Anita Spring and Veronique Goblet-Vanormelingen on fertility-enhancing rituals of the Southern Savanna approaches the question in such a way as to offer clear comparative results. Ngoma structured care and isolation from the stresses of household duty appear to make a difference in survival of at-risk pregnancies.

A retrospective hypothesis for the controlled study of health as social reproduction would need to provide the following minimal information. What is the nature and extent of social support and its allocation to health-related arenas in the household, the extra-household networks, and society at large? Are there measurable differential effects upon survival of at-risk segments of society or the improvement of perceived health?


previous chapter
6 How Ngoma Works The Social Reproduction of Health
next chapter