2
Identifying Ngoma
Historical and Comparative Perspectives

The contemporary settings of ngoma-type cults of affliction in Central and Southern Africa, as seen in the previous chapter, may now be joined by historical and comparative perspectives of the entire region within which these cults appear. Evidence for ngoma's origin, spread, and distribution can be gleaned from a range of types of sources: linguistic evidence from a comparison of Bantu language cognates; evidence for the distribution of material culture artifacts of ngoma, mainly musical instruments utilized in healing rituals; evidence of political variables in the presence of distinctive alternative forms taken by ngoma. This historical evidence, joined with the contemporary profile, permits us to sketch the basis of a unique institutional profile for ngoma, one that has often been misrepresented by scholars.

Readers acquainted with African culture history will recognize the "ngoma" region as approximately that of the distribution of the Bantu languages, that is, with the distribution of the cognate ntu which gave the Bantu language family its name. By itself, the term ngoma and its distribution do not tell us much about the common, and varying, features of the institution. However, when this verbal cognate is associated with the cluster of other terms and features that commonly accompany it over a wider region, we begin to see the larger picture that relates linguistic evidence to religious, social, and therapeutic, or health-related, cultural phenomena. This approach, to which is sometimes added archaeological research, has found increasing application in Sub-

Saharan African historical research of the history and nature of domains such as livestock husbandry, agriculture, iron working, political systems, and titles.

Such an approach to scholarship raises questions about the formal relationship of language to other domains of culture and society. Do phonetic units, the cognates, consistently carry common referents to conceptual thought, symbols, and emotions? How consistent are linguistic referents to behavior, techniques, and material culture? How consistent are rates of change in the relationship between "words" and "things"? These are practical research issues that may require specialized methodologies in particular domains. They are also theoretical issues bearing on the claim that language does (or does not) convey meaning, does (or does not) carry structures homologous with institutional or behavioral patterns, or is (or is not) patterned by physiological determinants.

After presenting core therapeutic-related cognates in Bantu languages, and some of the issues pertaining to the basic research on Central African linguistic history, this chapter will consider similar evidence about the constellation of musical instruments utilized in the rituals of therapy, in an attempt to clarify, in a formal sense, the relationship of the instrument type named ngoma to the other aspects of the ritual and the wider institution. The chapter closes with an exclamation of social and cultural variables that have shaped the institution, and considers how, as a distinctive institution, it has been represented, or misrepresented, in social research.

The Bantu Conundrum

Societies across the middle of the continent, from Luanda and Libreville in the west to Dar es Salaam in the east, and from Cameroon to the Drakensberg range and the Cape in the south, share many words and grammatical features that have come to be called the Bantu family of languages. Designation of these languages as "Bantu" is to some extent arbitrary, the result of nineteenth-century European linguistic research, which recognized large regions in Southern and Central Africa whose languages shared cognates. The cognate ntu , meaning "person," (plural, bantu , people) was but one of hundreds that could have been utilized to describe the entire set. In twentieth-century linguistic and archaeological research, the extent of commonality and variation in this common linguistic base has become much clearer.

In the twentieth century the notion "Bantu" has also taken on a range of connotations, positive in one setting, negative in another, and both in other settings, depending on whose perspective is entertained. It has come to stand for a mode of thought, or ethnophilosophy, presumably based on indigenous ideas, an approach to the study of African thought that is roundly criticized by some. It has become a rubric of major historical research, especially in Equatorial Africa, where the Centre International des Civilisations Bantoues of Libreville, Gabon, conducts cross-disciplinary work and publishes the journal Muntu . In Southern Africa, however, the notion "Bantu" has taken on a negative connotation because of the South African government's reification of a stultified tribal and imposed interpretation of African culture, particularly as carried out in education for blacks. The simultaneous positive search for civilizational heritage that one sees in Equatorial Africa under the rubric of "Bantu," and the negative tribal connotation of "Bantu Education" in South Africa, contribute to the Bantu conundrum.

For present purposes it suffices to summarize the central findings of some of this historical research. A body of current scholarship (Bastin 1983; Bastin, Coupez, and de Halleux 1981; Meeussen 1967, 1980; DeMaret 1984; Heine 1984; Phillipson 1985; Vansina 1984; Hyman and Voorhoeve 1980; Ehret and Posnansky 1982; and Van Noten 1981) establishes the origin of Bantu languages in the eastern Nigerian and western Cameroonian area, in the early first, possibly the late second, millennium B.C. Linguistic classifications, based on methods of "least common" and "most common" (or shared) lexical and grammatical features, determine that these languages are genetically related to West African languages.^[1] "Bantu" is thus defined as a narrow language group—though spread across a vast subcontinent—in a wider set of interrelated language families that are sometimes referred to as "Bantoid" (Heine 1984) in the much more extensive "Niger-Congo" group (Greenberg 1955).

The same methodology—that is, genetic classification of least common and most common cognates or features in sets of languages within the family—establishes further that the Bantu languages had, by the first millennium B.C. , begun to spread southward through the forest zones and the Atlantic coast of Equatorial Africa, and eastward along the northern edge of the forest-savanna border into the Interlacustrine region (see fig. 7).

Thereafter, additional "nuclear zones" are posited, from which further dispersion occurred eastward and southward. One of these was the

[Full Size]

Figure 7. 
Major lines of Bantu migrations, according to recent scholarship,
and four sites of field research for this book: Kinshasa, representing
Western Bantu; Dar es Salaam, on the Indian Ocean coast, representing
both inland East African and Swahili coastal influences; Swaziland, the
northern Nguni-speaking setting; Cape Town, a cosmopolitan South
African setting.

"Congo nucleus" in the forest regions of Cameroon, Gabon, Congo, and Zaire (Heine, Hoff, Vossen 1977), also spoken of as "Western Bantu" (Vansina 1984, 1990). Another was "East Highland," from the lake regions (Heine, Hoff, Vossen 1977), also called "Eastern Bantu." Later, in the first millennium A.D. , the expansions continued into East Africa and to Southern Africa, and, with a mixture of Eastern and West-

ern, across the Southern Savanna to the southwest, in Zaire, Angola, and Namibia.

This demonstrated expansion of the Bantu languages raises a number of questions about the relationship of language to other facets of culture that are germane to our interest in therapeutics, particularly ngoma. What was the technological basis of these societies that permitted their expansion into territories occupied by the hunting-and-gathering populations preexistent in the continent? Was this Bantu expansion more on the order of a gradual technological and cultural transfer, community by community, or was it more like a migration? Were Bantu languages tied to a more intensive agriculture? If so, at what point did iron working become the basis of this agrarian technology? Further, what types of food crops were used? Given the contrasting environments into which the speakers of Bantu languages moved—from savanna into rain forest and back to savannas—what were the adaptive advantages that gave the Bantu speakers the resilience to replace other ways of life and other languages? Were the speakers of Bantu languages organized in any particular social structures? What were their beliefs? What were their assumptions and values about health and society, and their approaches to healing? Are these reflected consistently in the common cognates across the region? How did they utilize the varied environments and natural products to improve health? Were the ngoma rituals of any particular importance in the process of expansion? Finally, if we assume a common source for ngoma, which features have remained continuous and which have changed? Are these reflected in the common and varying cognates across the region?

It is now argued, on the basis of recent archaeological research in Cameroon and Gabon (Van Noten 1981; Vansina 1984, 1990) that the Western Bantu expansion along the Atlantic coast and into the rain forest was a gradual expansion of stone-tool utilizing cultivators of West African forest-related cultigens, that is, trees such as the oil palm and root crops such as yams. The Eastern Bantu cultigens, which presumably brought the West African grassland crops to the Southern Savanna, and later to Southern Africa, included millet, sorghum, and cowpeas.

The appearance of iron working in this technological-cultural setting is now believed to have occurred about 750 to 500 B.C. , the date of early iron-working finds in Nigeria. Related iron-working sites, or evidence of iron, have been found recently in Gabon and Cameroon. The

expansion of the eastern lake Bantu may have occurred as early as 750 to 500 B.C. with the use of iron smelting and smithing. This accelerated the ability of Bantu speakers to dominate the landscape, to grow crops, to hunt, and in general to improve their adaptive advantage. It is thought also to have contributed later to the ability to form centralized states. The earlier hypotheses that Bantu expansion might have been due to iron and selected introductions of crops from Asia (Murdock 1959) have not been borne out in recent archaeological and linguistic research. For one thing, the archaeological finds of iron are simply too late to support such a hypothesis. Also, the language cognates that pertain to iron working are too disparate to lend credence to a hypothesis of dispersion from a common point.

The role of cattle and other livestock in the Bantu expansion has been studied at some length because of the absence of livestock in the Western forest region (due to tsetse fly infestation) and the extensive use of livestock in East African and Southern African Bantu-speaking societies. Ehret, studying the character of language cognates associated with livestock, demonstrates that this part of an Eastern Bantu complex was introduced from Central Sudanic peoples to the lake region by 500 B.C. (Ehret 1973–74), from where it spread southward.

The Bantu expansion must not be construed to have been a migratory spread of a biological or "racial" group, except in selected dimensions. Genetic markers of the northeast Bantu region north of the lakes resemble markers of populations in the origin area in the Cameroon and eastern Nigerian region (Hiernaux 1968). Otherwise, genetic studies have revealed diverse populations, in many instances reflecting "pre-Bantu," but speaking homogeneous Bantu languages.

Many of the recent specialized linguistic studies in the Bantu zone have had their basis in the massive lifetime work of Malcolm Guthrie, published in the four-volume work Comparative Bantu (1967–71), which maps the distribution of several thousand cognates. Guthrie's compilations are not exhaustive. The lexica are built around key word sets determined by linguists to represent basic cultural domains; they are then expanded into more culturally particular domains. Further, Guthrie's lexical reconstructions are of necessity limited to those Bantu languages for which dictionaries and vocabularies were available in the 1950s and 1960s when he did his work. Guthrie interpreted the evidence, based on degrees of common lexical stock, to argue for a Western and an Eastern Bantu divergence from a central point on the Southern Savanna. Although this hypothesis has been discredited by

subsequent research, his compendium, and subsequent work to build up the basic project,^[2] provides a growing foundation for new studies that test additional hypotheses. It is Guthrie's lexicon, with a few supplemental sources, that permits us to demonstrate the character of a proto-Bantu level of therapeutic-oriented verbal cognates and derivative secondary, Western and Eastern, subsets of cognates, concepts, and practices, and the place of ngoma in this set.

The Lexicon Of a Classical Sub-Saharan Therapeutics

The terms that pertain to health and healing in Guthrie's lexicon, and which occur in the entire region of Bantu languages, are assumed to have been part of the earliest—thus, proto—level of the Bantu-speaking societies several thousand years ago. Other cognates that appear only in Western or Eastern Bantu respectively, are either derived from the earlier terms as variants, were independently created, or were introduced from the outside. The following set of cognate terms reveals symptoms, etiologies, healer roles, medicines, and ritual activities with the end-goal of health (asterisks and numbers indicate Guthrie's "Comparative Series" (c.s.) of cognate reconstructions based on a comparison of modern semantic variations; see appendices A and B for the approximate distribution):

Proto-Bantu :
wound, boil, sore (1608 ^* -pútá -, 1609 ^* -púté )
to become ill, illness, to suffer (677 ^* -dúád -, 678 ^* -dúádì -, 679 ^* -dúáídé -)
to bewitch, curse, witchcraft, power of words (644 ^* -dòg -, 645 ^* -dògá -, 646 ^* -dògí , 647 ^* -dògò )
medicine man, medicine (786 ^* -gàngà , ngàngà , 787 ^* -ganga )
medicine, tree (1730 ^* -ti-)
medicine, consecrated charm (1534 ^* -píngú -); curse (1557 ^* -píng -, 1558 ^* -píngö -)
drum, drumming, dance (844 ^* -gòmà -, 1401 ^* -gòmà -)
to become cold, cool down, become cured, well (1564 ^* -pód -, 1565 ^* -pód -)
interdiction, prohibition, to abstain (826 ^* -gìdò -, 822 ^* -gìd -, ^* -gìdú -, gìda ) (also Huygens 1987:59–70; Obenga 1985:209–211)

Western Bantu :
disease (okon ) (Obenga 1985:196)
ghost (50 ^* -bàndà -); healer, medicine man (51 ^* -mbàndà -)
to cure (195 ^* -búk -); to divine or cure by divining (196 ^* -búk -)
white clay, kaolin (1474 ^* -pémbà -, 1477 ^* -pémbé -)
consecrated medicine, charm, fetish (1072 ^* -kítì -); spirit (1073 ^* -kítì , kícì )
to protect with medicine (990 ^* -kàg -); leaf (1021 ^* -káyí )
to bleed by cupping horn (439 ^* -cumik -)

Eastern Bantu :
sore (656 ^* -donda -)
witch, witchcraft (240 ^* -cábi -); ordeal, poison (1884 ^* -yábi -)
to practice medicine, divine (471 ^* -dagud -)
spirit, spirit of dead (619 ^* -dímu -)
consecrated medicine, fetish, charm (293 ^* -càngó -)
cupping horn (412 ^* -cúkù -, 440 ^* -cúmo -); to bleed by cupping (700 ^* -dumik -)

Recognition of sickness is signaled very widely by the cognate term dúád , or dúádì , whose verb form luala in KiKongo, halwa in KiNdembu (Turner 1967), or umkhuhlane in SiSwati (Makhubu 1978:61), expresses the existential quality of suffering, injury, or misfortune. Across the entire region this is differentiated from physical injury or sore, for which pútá or púté is used. In northwestern Bantu, okon expresses this (Obenga 1985:196), or in Eastern Bantu, the term donda .

Health is identified by numerous metaphors, including "balance" (lunga , in both Zulu and Kongo), "purity" (veedila in Kongo), and "coolness," whose most widespread cognate is pód , "to become cool, or cool down" or "to become well, healthy," in contrast to the heat of disease or witchcraft. This is related in some Western Bantu languages such as Kongo to the action of the cupping born (mpodi ), one of the most widespread therapeutic and religious features of Bantu-speaking African societies, which "sucks out" through the cupping horn (hola mpoka ) the impurity. Regional terms for the cupping horn are cúkù (Eastern Bantu), cúmo (Western Bantu); the action of bleeding by cupping, with scarification or scratching the skin followed by sucking, is cumik (Western) and dumik (Eastern).

In a recent thesis by Huygens (1987), as well as the work by Obenga

(1985), the concept of interdiction, prohibition, or abstinence has also been related to the Bantu-African sense of health and illness. The cognate gìdò, gìda, gìdú (in its nominative form, e.g., ngili in Teke, cijila in Luba) or gìd (the verbal form) emerges in the context of ritual activities, initiations, or naming, but also with practical circumstances having to do with abstinence from specified foods in connection with an affliction or ailment (Huygens 1987:59–70; Obenga 1985:210–211). Not only are sickness and misfortune held to be the result of transgression of the prohibition; health can be maintained or controlled by keeping the interdiction.

The widespread dichotomy that distinguishes misfortunes or afflictions stemming from "natural" or God-given causes from those stemming from human involvement is recognizable throughout the region, although there is no proto-Bantu term for God, suggesting that terms for God emerged later. In the Western region the most widespread term for God is Nzambi or Nyambi (925 jambé , 1917 yambé or nyambé ). In Eastern Bantu Mulungu (715 dungu ) is common, and in the Southern region Mukulu or Umkbulane (Zulu), utilizing the very widespread term kulu for aged one, or kula for growing up. Everywhere these High God terms are associated with misfortunes that are "in the order of things," thus "of God."

By contrast, unnatural or inauspicious misfortunes are said to be caused by "people": in Kongo, [kimbevo ] kia muntu ; in Zulu, [ukufa ] kwa bantu . The cognate stem used here—ntu —is of course that from which the name bantu was drawn by nineteenth-century linguists. The association of the basic term for human being with a major cause of affliction touches on a fundamental feature of etiological and therapeutic concepts couched in Bantu cognates.

The most common and widespread action terms that spell out human-caused misfortunes are the cognate stem verbs dòg , dògá , and nouns dògí or dògò . These terms describe the use of powerful words and the intentions behind them—whether good or evil. In Kongo in the west loka is the use of powerful words in oath and curse, and kindoki their imputed use, which is believed to cause sickness and misfortune. In a review of lok in Central African life, Luc de Heusch (1971) identifies the ways in which "good" and "bad" uses of spells or powerful words are used to reinforce particular dimensions of social structure. In Zulu, in the south, thak , the verb, and ubuthakathi , refer to the process or state of sorcery and witchcraft used by many scholars who

have encountered some derivation of the cognate dòg across Central and Southern Africa have masked the originality of the verbal concept, and the recognition that words, anger, and all other expressions of the social setting can affect health and illness. As words and their associated thoughts can afflict, so words and the thoughts of kin and affliction-peers can heal. Theophile Obenga, in his discussion of Bantu-African therapeutic practices among the Duala of Cameroon and the Mbochi of North Congo, stresses that the pronunciations made in the esa rite of the Duala and the ndoo rite of the Mbochi become a therapeutic force in their own right, simultaneously effective at the mental and the physical level (1985:204).^[3] In any event, the reality of kin and extra-kin communities in ngoma demonstrates the dynamic of the cognate dòg at work.

Other notions reflect further dimensions of human causation of disease and misfortune. Such a notion is reflected in the term cábi (or tsawi ), found in Ekoi and Tiv, as well as in Eastern Bantu languages, meaning the substance of witchcraft power in a person, and yábi , in Eastern Bantu, for the poison ordeal that identifies this power. Similar to this is the notion kundu in Western Bantu to denote the congealed power of witchcraft in the body of the perpetrator.

Ancestors represent an extension of the human community as a major cause of misfortune and cure in African society. Very widely, definitely in Eastern Bantu, ancestors or spirits of the dead are referred to by the common term, dimu , which is probably proto-Bantu. In Western Bantu, the generic term kulu , elder (in Zulu, unkulunkulu ), is also used to refer to ancestor or shade, and as with dimu , to the power that may cause both life, if channeled properly, and misfortune, if the community is not duly cognizant of the ancestral shades. A major dimension of ngoma is, of course, coming to terms with these "living dead" in relation to the fortune and misfortune of the sufferer and his or her community.

Moving from the etiological and worldview terms to those defining therapy, we may note that tí —tree, stick, or medicinal plant—is found through the entire region and may therefore be assumed to be proto-Bantu. In some settings, for example among the Hemba of Eastern Zaire, buti carries further connotations of consecrated medicine and of sorcery and witchcraft (Blakely and Blakely 1986). Kàg or káyí , herb, a Western Bantu variant, is the generic term for leaf, or for medicinal plants, as well as tobacco, an American import.

The well-known color triad of white, red, and black is represented

everywhere in the therapeutics of Bantu-speaking Africa and can probably be classed with a proto-Bantu ritual scheme, although none of the terms for color are pervasive throughout the region. One example of a widespread ceremonial color term is the Western Bantu pémbà , or pémbé , white clay or river chalk, denotative of "clarity" or "purity," an attribute of the ancestors. White clay, red ochre, and black charcoal, or other referents of these properties, are commonly used as basic ingredients of health care. Whiteness defines the status of the novice who, from the time of initial entry until graduation, is "in the white." This ritual symbolism is extended to the arena of the sick role, as ngoma well illustrates.

The notions of tí, káyí , and, pémbà are usually not, per se, charged with overtones of power. When combined with other substances into compounds and given the "interpretation" of a spoken phrase, they become so. Such compounds, especially when spoken or sung over, become powerful medicines invoking the attributes of ancestors and spirits. Guthrie believes that pengo is the proto-Bantu term for this function. Píngö and phungu in the West, and càngó in the east and south of the Bantu-speaking realm, are its derivative regional variants. The power of medicine at the level of the compound, spoken or sung over, is evident in its ambivalence. Thus píng or píngö refer not just to medicines but also to cursing in Western Bantu. Similarly, the Western Bantu notion of kítì or kícì refers both to the possessing spirit of an affliction, as well as to the medicine used to deal with the affliction. In the easternmost appearance of this cognate among the Lubaized Nsengo of Tanzania (Waite 1987), it refers only to spirits. In the Western region, as in Kongo, it carries mainly the connotation of the consecrated packaged medicine compound. Local variations represent both or a range of variants between the two extremes here. Ngoma in the kícì zone is often situated within this context as the therapeutic performance of the medicine. A more localized term among the Western Lunda peoples of the Southern Savanna (e.g., Yaka, Suku, Tshokwe, and Pende) for the category of medicine as spirit-induced compound or shrine is hamba . In East Africa, the Arabic-derived term dawa carries this same connotation.

In the same way as the formulation and identification of the sickness or its agent may generate the treatment mode, through control of that agent, so often the therapeutic technique and the specialist are described by the same term. Thus, very widely in the Bantu-speaking area tonal or contextual emphases separate the cognate stem ganga , medicine,

from gàngà , doctor, according to Guthrie. Whereas the former emphasis is sporadic, the latter is nearly universal in the region, ranging from ngàngà in Western Kongo, to mganga in East Africa, and inyanga among the Zulu. Regional terms for specialized types of practitioner include, in Western Bantu, buki , combining curing and divination/ diagnosis; banda and mbanda in Western Bantu, the first meaning spirit of affliction, the second healer; dagud in the northeast, meaning to practice medicine or to divine; and lumbu or bilumbu in the central Southern Savanna region and igqira in the south. These general terms for doctor or healer are given more specificity through combination with another term. Thus, the Kongo nganga nkisi deals with consecrated medicines (minkisi ); nganga lunga is the orthopedist; nganga mpodi the cupping horn expert, nkisi mpodi , the cupping horn, and so on.

Divination is a nearly universal technique in these societies, although the particular character of the technique varies immensely. The basic diagnostic question for which an answer is sought is that of whether a misfortune or affliction is due to natural—God-ordained—causes or to human forces, or is related to the misuse of medicines or inadequate control of spirit forces. Because of its integral place in the interpretation of human experience in societies of the area, divination methods have shifted frequently to respond to the forces of social change. Unlike the pervasiveness of the term that expresses the hypothesis of human-derived misfortune (dòg ), there seems to be no common or even regional term for the interpretation of misfortune in particular cases (unless it is the cognate gàngà ). In recent decades there has been a widespread trend for mechanistic techniques such as the Ngombo basket of the Southern Savanna and west coast, and pengula bone-throwing in the southern region, to be replaced by inspirational diviners who may be possessed by a range of spirits: nature, ancestral, Christian, Islamic. The urgency to respond to the basic metaphysical and social questions about the causes of misfortune has maintained the diagnostic and therapeutic system of Central and Southern Africa amid major changes in the past century. The ideational superstructure, more than particular techniques, has remained highly resilient.

Ngòmà is a final central cognate of the medicine of Bantu-speaking Africa which arises from and draws on all the foregoing features. So far this work has translated it as "cult of affliction," or "drum of affliction," since in many regions ngòmà refers to drum, the instrument. However, this connection between the term and the wider phenomenon

is far from being a neat single-stranded relationship. At the core of the "Bantu conundrum" is the fact that although there are some widespread referents of the terms we have seen here, few if any are present everywhere as a fully consistent complex of different elements. The musical instruments that accompany ngoma and other healing rites offer intriguing clues from a nonverbal domain which will support preliminary hypotheses to the questions we seek to answer.

Instruments Of Ritual Healing As A Nonverbal Cognate Set

We may think of the issues to be addressed here as a puzzle with several interlocking pieces. If the verbal cognate ngoma is nearly pervasive in Bantu languages (with the exception of zones B and C, and in the Khoisan south; see maps, appendix B), how do we account for the limited distribution, within that cultural and linguistic space, of ritual therapeutic institutions named ngoma ? Related to this, how do we account for the even less widely distributed occurrence of the drum type ngoma within that space? The distribution of the set made up of word, behavior, and object will offer clues as to the origin and character of the institution.

There is more to the puzzle. Although the type of music that accompanies therapeutic activities in Central and Southern Africa varies a great deal, well beyond the core area in which ngoma drums are used, the association of musical instrument types with healing is not random. Instrument types seem to be defined by regional sets or traditions, but in such a way as to dissuade any zealous reductionist of an inherent relationship between the ngoma drum type and the therapeutic rituals.

A first step beyond the lexicon, for the purpose of testing the extent to which behavioral or nonverbal culture may align with verbal cognates having to do with healing, takes us into the realm of the material culture of ritual therapy, namely the musical instruments, their names, and the constellations in which they are combined in ritual performance with singers. This evidence, added to the lexical evidence offered above, strengthens the inferences that may be drawn about the origin, history, and character of the institution. As we progress in this analysis of a complex institution, we will examine the relationship of these instruments to voice, song text, rhythm, trance-possession, and the social makeup of therapeutic communication within a sociopolitical context. (appendix C offers a distillation of findings on the formal composition

of instruments, singers, and the overall makeup of the ngoma-type groups, from a range of sources based on observation and the literature.) The research to date is of mixed quality. In addition to my own observations, and the recordings and publications of specialized scholars on particular peoples—for example, John Blacking (1973, 1985) on the Venda, Paul Berliner (1981) among the Shona—a major systematic compendium of great usefulness is the survey of music in Zairian healing rites by researchers Arnaut, Biolo, Esole, Gansemans, Kishilo, Malutshi, and Querson of the National Museum of Zaire in the 1970s and early 1980s (see appendix C). This work is backed up by another, earlier, set of data found in the work of Olga Boone, whose 1951 classic Les tambours du Congo belge et du Ruanda-Urundi identified drum types and their names across Central Africa.

In the belt across the middle of the continent, from Kongo to Swahili, where Guthrie finds the greatest convergence of common Bantu terms, ngoma refers primarily to the elongated wooden drum with a single membrane attached at one end with pegs. Boone noted that this major drum type was distributed along an east-west line roughly at the Southern Savanna/forest border. North of this line was a region of "mixed" drum types, with the pegged ngoma type interspersed with a type that uses cords to fasten the membrane to the body of the drum. Drums whose membranes were attached with cord or string were rarely called ngoma .

In the region of ngoma rituals, ngoma the instrument usually is also identified as a dance drum and a sacred medicine drum. It may also be a drum of state. In societies where trance-possession and therapeutic cults are present, ngoma more than any other drum is used in this therapeutic setting, to the accompaniment of shakers and singing. To the north of the region where this set of practices prevails, stringed and wind instruments are more common in healing rites; ngoma drums are absent.

Therapeutic rituals in the rain forest of Equateur Province of northern Zaire generally demonstrate the typical call-and-response pattern of musical interaction found elsewhere: a sufferer and healer, and a "choir" made up of sets of additional individuals on either or both sides, with the accompaniment of hand clapping, rasps, rattles, whistles, bells, stringed zithers or harps, horns, gongs, and kettle or slit-gong drums, as well as occasionally the xylophone (see appendix C). Spirits that are invoked in these rites are often ancestors ("Elima," "Balimo,"

"Malimu," consonant with the cognate dímu ) or nature or wild spirits ("Wetshi," "Nzondo").

All of the Equateur examples in the Zaire survey are taken from north of the line, established in Boone's Tambours (1951) study, beyond which no lanced-skin drums are said to have been made and used. In other words, the northern forest picture of therapeutics from the Zaire survey demonstrates that healing rituals in Central Africa occur widely without the characteristic ngoma drum. Although Guthrie includes these regions in his Bantu language area, they are conspicuous for the absence of ngoma as a verbal cognate, a pattern that is also true of the eastern Zaire Warega rite "Butii."

By contrast, the Southern Savanna, East African, Southeast, and Southern African examples of therapeutic rituals demonstrate the near pervasive presence of the ngoma-type lanced-skin drum in the performance of the ngoma rite. As one moves south and east, ngoma drums are the rule until one reaches the North Nguni beyond the Zambezi river. They are present in Venda ngoma dza vadzimu rites, in Swazi and Shangan rituals, and in some area royal settings. But among the southern Nguni peoples (Zulu, Xhosa) they are absent from both royal and cult settings. Here, ngoma refers neither to the drums used (cowhide stretched over sticks or oil drums), nor to the dancing, but exclusively to the singing, divining, and the designation of those who do these things. Thus, the Zulu isangoma diviner is literally "one who does ngoma"—that is, sings the songs. Among Xhosa, much influenced linguistically by Khoisan, the role term for the ngoma-singer becomes igqira ; divination is handled not with bones but through contemplation. Group and network support plays a more important place than individualized divining in the work of ngoma. There is thus a host of regional and societal variations around which the notion "cult of affliction" or "drum of affliction" must be analyzed.

Exceptions to this pattern are coastal Swahili, the Shona setting of Zimbabwe, and the western Kongo setting. Generally, in the examples we have from the vast region mentioned, sufferer(s) and healer(s) either constitute or are joined by a choir and other instruments, such as shakers or rattles, gongs, and hand clapping.

In the Shona region the drum is replaced by another instrument, the large gourd-resonating hand piano, mbira , usually played in an orchestra of a dozen or so members, in performances called bira . The Kongo region reveals a mixed picture, insofar as musical instrumenta-

tion of healing rites is concerned, consonant with Boone's determination (1951) that western Zaire was a region of "mixed" drum types. Kongo therapeutic rites utilize a mixture of horns, single and double gongs, whistles, rattles, and a range of drums (including ngoma to drum up major nkisi , "ritual medicine" see appendix C). The Swahili coast features the "pure" ngoma types of healing rites, invariably from the interior, although the Islamic-influenced rites utilize small double-membrane drums and shakers.

This instrument survey, suggestive of type clusters around the widespread ngoma region, belies the impressive musical consistency across the subcontinent in terms of a few features I shall take up later in more detail, but which need to be pointed out here. Throughout the rites cited, the musical scores offer a pervasive use of call and response between the single "soloist"—the sufferer, novice—and the "choir"—the local cell or a group of "significant others."

The patterning of the instruments in therapeutic rites, against this backdrop of the call and response and song-dance, suggests that there are regionally, or culturally, specific constellations of instruments. We may think of these in terms of the common-sense designations of instrument types—harp, zither, whistle, horn, drum—or in terms of the formal designations of musicologists (Marcuse, in Merriam 1977:250): the idiophone , an "instrument that yields a sound by its own substance, being stiff and elastic enough to vibrate without requiring a stretched membrane or string" (e.g., xylophone, mbira, sansa, likembe, rattles, bells, gongs, slit drums); membrophone , "any instrument in which sound is produced by vibration of a stretched membrane, brought about by striking, friction, or sound waves" (e.g., drums); aerophone , "any music instrument in which tone is generated by means of air set in vibration" (1977:252) (e.g., horns, flutes, panpipes, and ocarinas); and chordophone , "any instrument having strings as tone-producing elements, the pitch of the instrument being dependent on the strings" (e.g., harp, zither).

The chordophones—the stringed instruments—although they are present throughout, are used in healing rites only in the northern forest region. Most common in healing rites throughout the Central and Southern African region, is the idiophone, that is, the shaker, gong, xylophone, slit-gong drum, and the thumb or hand piano. Second most common in healing rites of the entire region is the membrophone, the single or double membrane drum. The areophone seems to be more common in the northern forest region than in the Southern Savanna and other southern regions.

In terms of the hypothesis announced earlier concerning the relationship of instrument type to the therapeutic rite, great variation is apparent. Nevertheless, the variation is patterned. It is not merely a reflection of the general stock of musical instruments used in the regional culture. Stringed instruments and horns are widespread but are not frequently used in therapeutic rites outside the forest region. Drums, readily available in forest societies, are not frequently used for healing there. On the Southern Savanna, drums are the primary instrument of healing. The pervasive African hand-piano—nsanza, mbira —is used in major healing rites only among the Shona group of Zimbabwe.

A second conclusion, announced earlier, follows from this finding on the pattern of distribution of musical instruments utilized in therapeutic rites. Thus we must be suspicious of claims that explain a specific pattern of therapy, or possession ritual, in terms of the effects of a particular type of instrument, such as the drum. This is particularly the case given the background of common musical style features such as call and response and polyrhythm, and of the choral nature of African therapeutic song-dance.

One final piece of the puzzle, of those with which we began this section, remains to be put in place. What, then, explains the distribution of the ngoma-style drum and the use of this name for the therapeutic and celebrative song-dance? As we have seen, this distribution is roughly outlined on the north by an arc running southward of the Congo/Zaire River, then northeastward from Lake Victoria across southern Kenya. It extends from the Atlantic to the Indian oceans, and southward to the boundary of Nguni-speakers and the Shona. Plotted on Guthrie's language map, this corresponds approximately to the F, H, K, L, M, N, P, and parts of the S zones (see appendix B). It is absent from the C and most of D zones in the north, present only along the coast of A, and sparse in the R zone; absent in the large Khoisan zone, as well as in part of the S zone, the south Nguni.

A tentative interpretation of this distribution of the associated drum type and therapeutic rite would point to its presence, as a cognate, in early or proto-Bantu, with some kind of amplification in the Eastern Bantu migrations from the lake region, westward across the savanna to the Atlantic coast, southward across the Zambezi and the Limpopo, and eastward to the Indian Ocean coast. The early Western Bantu pattern of ngoma drums is not clear at this point; the identification of both the drum type and the ritual is extremely diverse and needs further study. It is clear, however, that over much of the western Congo basin, as in Kongo society, there is an overlay or melding of presumably East-

ern and Western Bantu elements. It is probable that ngoma techniques and material culture—the ngoma drum—complemented or incorporated the eastward spread of distinctly Western Bantu cultural elements such as the nkisi (known from the Kongo coast to the Luba-ized Nsenga in Malawi and far western Tanzania). The coastal Atlantic rites in Gabon and Cameroon, where the pegged-membrane drum is present, are suggestive of Central African rites. Whether these are due to proto-Bantu or more recent Eastern Bantu impulses is unclear. Our Eastern Bantu origins hypothesis for ngoma must remain suggestive for the present.

Social And Political Variables Of A Complex Institution

Formal one-dimensional cultural historical indices such as verbal cognates and the distribution of material culture have set the broad historic boundaries of ngoma. We must now look within the region and its societies to further identify the subject at hand and to establish the hallmarks for its presence within this broad context. The next two chapters on "core features" and "doing ngoma" will further identify behavioral and normative correlates of ngoma, as used in its broader meaning as has emerged in the foregoing pages, and as the set we have been describing as a "cult of affliction." However, first we can identify some of the broad-stroke social and political corollaries of those settings in which ngoma is present and in which it is absent. Second, within the region and societies where this set occurs, we can begin to look for the reasons for the rise and decline of particular manifestations of ngoma the institution, and why it is segmented into many specialized groups in one setting and homogeneous or unitary in other settings. Third, given that ngoma combines features that are normally differentiated in Western institutions and in Western scholarship, what is an appropriate understanding of the cult of affliction and its functions as an institution?

Cults Of Affliction In Centralized And Segmentary Societies

Many of the societies of the subcontinent have been lineage-based agrarian communities, practicing some hunting, and in regions where the sleeping sickness-carrying tsetse fly is absent, livestock tending.

Especially in coastal regions, commercial cities have emerged, linking the continent to overseas mercantile centers. The region includes Southern Savanna matrilineal societies such as the Kongo, Lunda, Cokwe, Kimbundu, and Bemba of Zaire, Angola, Zambia, and Malawi; patrilineal societies such as the Luba, Lozi, Nyamwezi, and others of the central region, and in the southern region, the Nguni-speaking societies of the Zulu, Swazi, and Xhosa; and nearby, the Shona, Sotho, and Tswana, to name a few. The region has seen the emergence of numerous precolonial states and empires, including the cluster of Luba, Lunda, Kimbundu, and Cokwe states; on the western coast, the Kongo, Loango, Kakongo, and Ngoyo states; the states of the eastern lakes, Busoga and Buganda, and eastward, Nyamwezi; in the Zimbabwe region, the historic state of Monamotapa; more recently, in the early nineteenth century, the Zulu empire and the Tswana chiefdoms, and the Sotho kingdom in the Southern Africa area, associated with the great disturbances known as Mfecane.

Ngoma-type cults of affliction have related dynamically to these states. They have either been brought under the tutelage of government and served the purposes of, and the legitimation for, sovereign power, or they have preserved and perpetuated segments of society not directly related to the state. In the absence of the state, they have provided a format for the perpetuation of social segments, particularly those marginalized or afflicted, such as women, the handicapped, those struck with misfortune in economy-related tasks such as hunting, women's reproductive capacity, or commerce. In some settings, the model of the cult has provided the basis for normative social authority, defining and organizing economic activity, social organization, and more esoteric religious and artistic activities.

In colonial and postcolonial Africa, the logic of the use of affliction and adversity for the organization of social reproduction has contributed to the perpetuation, even the proliferation, of cults of affliction, often in a way that has baffled governmental authorities and outside observers. Cults have arisen in connection with epidemics, migration and trade routes, shifts in modes of production, and in response to changes in social organization and the deterioration of juridical institutions. Colonialism itself undoubtedly generated many of the cults of affliction that appeared in the twentieth century. Postindependence conditions have continued to provide grist for the mill of cult formation.

The picture of cults of affliction within, or in relation to, centralized historic states contrasts markedly with that in the decentralized societ-

ies. Under the shadow of the state they are less influential, or entirely absent, or transformed into the rituals of statecraft. Instructive is their apparent absence in the Tswana chiefdoms, where strong historic chief-ship has provided social continuity, a format for the juridical process, and some means of material support to marginalized and needy people. By contrast, in neighboring Nguni societies, they have thrived alongside or under the tutelage of chiefs and kings.

In other contexts cults are known to have provided the impetus for the emergence of centralized polities, as in the case of the Bunzi shrine of coastal Kongo. Elsewhere, cults have emerged in the wake of historic states, picking up the aura of royal authority, the trappings of sovereignty, and transforming them into a continuing source of mystical power. A prime example is the cult of Ryangombe and the BuCwezi of the lakes region of eastern Central Africa, whose spirits are said to be the royal dynasties of the ancient Cwezi kingdom (Berger 1981). BuCwezi is today found in Tanzania's major cities. The same model has been reported in Mayotte, off the coast of East Africa, where possession spirits are the Saklava kings of Madagascar (Lambek 1981:152).

The dynamic relationship of cults to centralized polities has been accompanied by changes in the way spirits and shades are focused in consciousness and ritual. As the scale and function of a cult expands, narrowly defined ancestor shades may give way to nature, alien, or hero spirits. In a few instances, centralized shrine cults have persisted over centuries, defining primary values and social patterns for generations of adepts. The Bunzi shrine of coastal Kongo, Mbona of Malawi, and Korekore and Chikunda in Zimbabwe are well-studied examples that continue into the present. Some authors have made a distinction between these centralized "regional" cults and topically focused cults of affliction (Werbner 1977). But the orders, taken in their entirety, suggest more of a continuum along several axes: centralized to segmentary, inclusive to specialized, controlled by state sovereignty to independent (or even opposed to state sovereignty). Cults have crystallized opposition to states, both in precolonial, colonial, and to a lesser degree, postcolonial settings. Thus, the Cwezi cult channeled opposition to hierarchized structures in the Interlacustrine state of Rwanda (Berger 1981). Cult leaders organized opposition to Rhodesian labor recruitment practices in the early twentieth century and inspired early strikes in the mines (Van Onselen 1976). In the Zimbabwean war of independence, mediums played a role of legitimating the claims to land by the elders, and the aspirations of the guerrilla fighters, although the par-

ticulars have only begun to be studied carefully (Fry 1976; Ranger 1985:187–216). The role of ngoma networks in popular resistance in South Africa's townships is not yet known to scholars, but it may be substantial.

In the twentieth century, cults of affliction have tended to be short-term movements of panacea (DeCraemer, Vansina, and Fox 1976), often born in desperation. They have provided expression to the pains and social problems of wide segments of the populace. There has been a great deal of interpenetration between the cults and independent Christian churches, and with Islamic orders. New permanent cults have arisen around characteristic ills such as the isolated nuclear household in the urban setting; epidemic diseases such as tuberculosis, and getting by with the chronic problems related to it; the divination of social problems such as unemployment in a proletarian setting; how to succeed in business and how to retain a job; how to protect wealth once it is acquired. Many cults focus on the alienation and entrapment so common in the African urban setting.

In the urban centers of Zaire, Tanzania, Swaziland, and South Africa the historic cults as well as new adaptations are represented by part-time and full-time healers and priests and their adepts. In most instances the ethnic communities of the rural hinterlands have brought their religious institutions with them to the city, where they have undergone shifts of function and signification.

Unitary And Diverse Manifestations

A further issue in considering independent variables surrounding the origin, persistence, and change in ngoma has to do with its alternative unitary and diverse manifestation across the region where it is found. This contrast is most marked in comparing the central region of the continent with the southern region. Across the mid-continent, from the Congo coast, across the Southern savanna, to the Tanzanian highlands and the coastal region, ngoma-type institutions are usually represented in multiples. Turner's work (1968) among the Ndembu, a society incorporated on the periphery of the Lunda empire, counted twenty-three ngoma orders; Cory's writing on the Sukuma of the Lake Victoria shores in western Tanganyika enumerated about twenty-five ngoma orders. Some of these pertained to women's reproductive disorders or child rearing. Others had to do with men's problems, either in produc-

tive work or in social roles. Several had to do with societal dangers, either from the natural world (e.g., poisonous snakes) or from spiritual threats (witches) or alien spirits. Others could be seen as ceremonial leadership organizations that consolidated responsibilities such as witch finding or the sponsorship of periodic rituals. In the southern region, particularly in Nguni-speaking societies of Mozambique, Zimbabwe, South Africa, Swaziland, and Lesotho, ngoma is mainly presented as a homogeneous type of institution, devoted to the recognition of ancestors and addressing general human problems.

It is difficult to interpret this contrasting configuration in terms of an independent variable, either in the past or in the present. It is tempting to look at this contrast in terms of a kind of Durkheimian or Spencerian social structural proliferation or specialization that occurs in most societies with advancing time. The older societies in the Bantu expansion, notably of the central area, would have shown greater institutional diversity because of their greater historical depth in that setting. Whereas the Nguni in the south, exemplifying the end result of migrations and frontier-type settings, would have retained a less differentiated type of society. This perspective might then reveal something about the original role of ngoma in the "Bantu frontier," perhaps in the need to consolidate authority and to come to terms with threats and contradictions of various kinds. However, there is no way to test such a theory or hypothesis, much less determine which are the independent and dependent variables, until far better historical understanding is available.

There are indications in specific settings of trends in cults of affliction toward greater proliferation, or toward greater homogeneity, which may offer a less grandiose approach to the issue. The salient independent variable here seems to be political and social consolidation. In the seventeenth-to nineteenth-century coastal Congo setting, in which the Lemba cult emerged, there was a proliferation of nkisi medicines, charms, and ngoma-type orders, especially along the coast as the coastal trade eroded political states such as Loango, Kakongo, and Ngoyo, as well as the Kongo kingdom, and undermined the juridical functions that these states were able to fulfill. The decline of the states may be correlated directly with the increase of charms, medicines, and cults, including Lemba. However, within Lemba—the major regional ngoma-cult organization, which reflected trade, alliance building, and healing—there emerged a consolidation of some of the diverse sub-charms and functions. For example, the coastal midwifery order Pfemba, the preva-

lent way of dealing with women's reproductive issues, was co-opted by Lemba (Janzen 1982:56). Its representation in the Lemba order and nkobe basket of medicines came to be known as "Pfemba-Lemba" (Janzen 1982:253–254). In other regions Lemba appears to have incorporated, or aligned with, other distinctive ritual functions and medicines. We may project this procedure to its logical extension and imagine that multiple ritual functions might be similarly absorbed within a single institution, leading to greater functional homogeneity. At the extreme, this might have led to complete integration of ritual functions within the state or some other absolutistic type of institution. Or, it could, as in the case of coastal Kongo, indicate that where the centralized state had collapsed, ngoma-type orders took up some of the functions of state, such as conflict resolution, social control of threat, and the channeling of useful knowledge as applied to problem solving.

Another example of movement toward ritual consolidation from the contemporary ngoma picture comes from Tanzania. There, a modern state-sanctioned organization of ngoma healers, the Shirika la Madawa described in the previous chapter, controls the resource of ngoma recruitment. Indeed, this control of accesss to the role of ngoma healer and membership in the association is sufficiently restrictive that one of the major legitimating criteria of admission, namely certified possession by a sheitani spirit, occurs in only four out of a hundred individuals treated by the ngoma group. But the Shirika's ability to control diversity is offset by competition from the many other ngoma orders in Dar es Salaam.

It is difficult to formulate a strict calculus of the myriad range of transformations ngoma may undergo across the region where it has been reported. A few generalizations are possible. Ngoma appears to fade away where there is a strong central authority with a highly developed judicial tradition (e.g., Tswana). It seems to proliferate on the social and geographical margins of large empires (e.g., Ndembu in the Lunda empire; BaCwezi in Ryangombe) or as a mechanism for the consolidation of authority in the interstices of society where misfortune lurks (e.g., Bilumbu in Luba society). It proliferates where misfortune is rampant and where social chaos prevails (e.g., early colonial resistance, postwar Zimbabwe, South African urban townships). In the wake of the demise of centralized states, it may take on the functions of the state (e.g., Lemba in coastal Kongo).

Do the constant features through all these transformations represent an institution? If so, how can that be characterized?

Scholarly Blinders And The Ontology Of A Unique Institution

Scholars, administrators, policymakers, and therapists have predictably come up with varying opinions on how to characterize the ngoma-type cult of affliction. As a final task in this chapter on identifying ngoma, I wish to argue for the proposition that it is a unique institution.

Definitions of institution abound, but they reflect a common understanding of how society is put together and functions. Durkheim suggested (in Parsons 1949:407) that "[a] body of rules governing action in pursuit of immediate ends insofar as they exercise moral authority derivable from a common value system may be called social institutions." A falsifiable proof of this definition held, said Durkheim, that "the means to these ends may vary, but the rules reflect the common values. If they are lost sight of, the result is a breakdown of control, and anomie." In another tradition, M. G. Smith noted (1974:212) that "whether culture is restrictively defined as the symbols, norms, values, and ideational systems of a given population, or more inclusively as their standardized and transmitted patterns of thought and action, all institutional organization has a cultural coefficient, since each institution involves collective norms, ideas, and symbols as well as standardized modes of procedure."

These general theoretical comments about the broad basis of institutions—norms, common beliefs, ends met by a range of means—would certainly be appropriate to describe what ngoma does in Central and Southern African society. The problem, of course, is that by Western institutional and scholarly standards some of the examples of ngoma are strange indeed. They have heretofore been put into rubrics of either Western institutions or have been allowed to languish in ethnographies as local culture, for example, Ndembu religion, Zulu diviners, and Kongo fertility magic.

The interpretation of African cults of affliction is analogous to the study of some other domains in anthropological research in that scholars have been faced with the need to bridge the indigenous concept with the analytical notion. Sometimes scholarship has come down on the side of the former, as in totemism, taboo, or shamanism; other times it has come down on the latter, which is frequently a reflection of a Western institutional category.

The debate about kinship in Western anthropology is instructive here. David Schneider's Critique of the Study of Kinship took issue with

the pervasive assumption by generations of anthropologists since Morgan of the universality of the family and kinship. This he explained, not so much by faulting Louis Henry Morgan for finding the family and kinship among the Iroquois, but by faulting anthropology at large for adopting Western notions of basic institutions and imposing them upon societies of the world. The "big four" institutions of Western society are, for Schneider: kinship (the family), the economy (business), politics (the state), and religion (the church).

Western social science, including anthropology, has extended this quartet as analytical categories onto other societies, much to the detriment of insights to be gained. Societies in which these institutions or attributes seem to be combined differently, or are only partially represented, are held to be "undifferentiated," and therefore "more primitive," or somehow disorganized or muddled.

One of the major challenges, then, in presenting ngoma has been to transcend Western institutional categorization. The difficulty of Westerners, and of Western-trained Africans, in accepting ngoma or the cult of affliction as a valid institution in its own right, has been instructive in this regard. In many African settings the colonial legacy of Western institutional structures clashes markedly with the African institution.

Ngoma in Tanzania, where there has been a commitment to build on African foundations, illustrates the point. Officially, African medicine and its institutions are recognized. Research units devoted to the subject have been sponsored. However, the research effort and the statement of the reality of therapeutic ngoma are initiated from the specialized basis of Western institutional categories. Thus, the Traditional Medicine Research Unit at the National Hospital is charged with examining the botanical and chemical character of medicines used in ngoma and other types of indigenous medicine and with creating a program for primary health within the framework of indigenous healing. The Music section of the Ministry of Culture is charged with researching the dance and song basis of ngoma, as well as sponsoring dance competitions of current ngoma groups and licensing entertainment ngoma. The Ministry also sponsors the national dance troupe and allied ngoma groups. The political party of Tanzania has de facto liaisons to ngoma and large healer's associations. Tanzanian bureaucracy thus sections ngoma into distinctive categories consonant with Western rules of social order.

In my earlier work on Lemba (Janzen 1982) I observed the struggle of publishers and reviewers to come to terms with this dilemma of the integrity of the institution versus the categories of Western scholar-

ship. The publisher, in filling out the Library of Congress Catalogue Data page, described Lemba as a "cult," thus a subset of religion. Several reviewers tried to escape the straitjacket of Western institutional typologies but succeeded only partially in doing so, coming up with hyphenated categorical types. One reviewer, after a page of discussion, noted that "to define it simply [Lemba was] a cult and a social institution that controlled trade, markets, and processes of exchange" (Mudimbe 1986). Another, picking up on my vocabulary, called Lemba a "therapeutic-alliance-trading institution" (Feldman 1988).

These reviewers appreciated the unique institutional profile better than another who spoke of the book's having offered "a new framework for thinking about a little understood region of Africa and for analyzing the relations between the political, kinship, religions, and economic aspects of social structure" (Riesman 1985), thereby virtually sectioning the subject through the Western institutional categories.

Another reviewer (Stuart 1986), whose summary of the work is a model of succinct interpretation, offered this explanation of Lemba's mix of trade, alliances, and therapy: "The ngoma , or the 'drum of affliction' became the cultural symbol of a therapeutic society ... which evolved to deal with the social stress and cultural change created by Europe's growing commercial influence." He concluded with this insight: "[T]herapeutics may be the metaphor serving to facilitate consolidation of substantial resources, material and human, and to aid long-term reordering of institutions of redress, economic redistribution, and ideological change."

We are thus confronted in Lemba, as in many other variants of the ngoma profile across Central and Southern Africa, with a constellation of practices and perspectives that are unique and yield to understanding only with some critical analysis. This realization highlights the centrality of the need to deal carefully with the homology between language, behavior, and institutions, which is closely related to the first activity of divining, science, religion, and a host of other human enterprises, to wit, naming the phenomenon.

To understand this better, it is instructive to look at reviews of the Lemba book, particularly one that discussed the question of naming ngoma. This reviewer (Stevens 1984:29–31) thought the study had been done a disservice by allowing "the term 'drum of affliction' to stand in the subtitle." It was not that "drum of affliction "—derived from ngoma—might be an erroneous label for Lemba but that only a handful

of specialists like Victor Turner know the term refers to Ndembu rituals. It has little recognition value as a more widespread type of phenomenon, he argued. Yet if none of the Western institutional labels are appropriate, and an expert's English term for the indigenous term in one African society is not appropriate, then by which term do we describe, or understand, the phenomenon if we wish to avoid having it locked into one or more versions of, or a hyphenated version of, the Western institutional grid? Beginning with Turner's local Ndembu work, Stevens reasons out an approximation of ngoma in the broader sense, freed of its Western institutional categorical boxes.

[Turner's] use of the alternate Ndembu meaning, "drum," is significant to an anthropological investigation of symbol and meaning in African cultures, and it will serve students and collectors of African art well to consider for a moment the possibility that a drum, as an object, may be meaningful only as a construction of materials, barely even as a musical instrument; its cultural meaning is revealed only through the total socio-religious context of its use. In such a context, then, "drum" is symbol: it is drumming, and it is collective sentiment, catharsis, transcendence—indeed, the whole of the ritual process. The ritual process is social effort.

Finally Stevens comes to his "discovery point":

In this sense, then, the phrase "drum of affliction" is justified as referring to a type of ritual; and we can understand why Bantu-speakers may use the term ngoma to refer even to a ritual in which drums are not used. "Drum of affliction" is a ritual with a therapeutic aim, the exorcising of some malign agency, but as both Turner and Janzen make clear, "drum" (ngoma ) means the aims, activities, actors, and institutions, and the network of symbols by which they are linked and united, that constitute the ritual process. (Stevens 1984:29–30)

Stevens is correct that ngoma , however we wish to gloss this term in English or another analytical language, refers over a wide area of Central and Southern Africa to a cluster of recurring processes and perspectives having to do with the interpretation of misfortune, usually manifested by disease or disease symptoms that are imputed to spirits or ancestors, and the rites to bring the thus "afflicted" into a supportive network with others similarly afflicted and to treat them by empowering them to deal effectively with the adversity. The particular source of the adversity, whether it is the impact of foreign trade, twinning, snake-bite, or lineage segmentation, is secondary to the fact of its definition as the phenomenon of adversity.

Conclusion

Ngoma, then, is a composite, historically unique institution widespread throughout Central and Southern Africa, with many local and regional variations. Its identity as an institution and as a behavioral process, often with name recognition, should satisfy scholarship. That it has taken so long for scholarship to catch up with indigenous usage is tribute to the tendency of scholarship, and administration, to categorize in its own, often local, terms the phenomenon before it.

This chapter's goal of "identifying ngoma" has been external and formal, based on examining comparative and historical distributions of "words, acts, and things" and how they vary in relation to one another. Its apparent central purpose as an institution is to respond to the need for order, meaning, and control in the face of misfortune and affliction as defined by a core proto-Bantu cognate, dòg : that just as words and intentions by others can afflict, so they can heal.