1. Introduction

Possession, Identity, and Power: Theoretical and Methodological Considerations

Women, migration, and power: these are the themes that frame this study of spirit possession and identity in northwest Madagascar. The setting is Ambanja, a booming migrant town in the heart of a prosperous plantation region called the Sambirano Valley. Here identity is shaped by polyculturalism and manipulated through religious experience. Healing rituals, involving possession by ancestral tromba spirits, provide an important arena in which to articulate the problems of urban life. In this latter respect this is a study in medical anthropology.

Identity in Ambanja is dynamic and multifaceted, defined in reference to a subjective conceptualization of self, an individual’s role as social actor, and cultural—or ethnic—origin. In this migrant town, these three levels of experience are bounded by competing statuses of insider and outsider. Insiders are, specifically, the indigenous Bemazava-Sakalava;^[1] outsiders are other Malagasy speakers who have come to this region as wage laborers searching for work. Ethnicity is the most important marker for defining identity and status in Ambanja. Although migrants form the majority of the valley’s population, the Sakalava comprise the largest single ethnic group.

These competing statuses define the powerful and the vulnerable, where local ancestors and access to land are pivotal. To some extent, length of stay and material wealth may affect one’s position in local arenas of influence, yet indigenous notions of power require that one be symbolically rooted to the land. The Sambirano Valley is the tanindrazan̂a (HP: tanindrazana), or “ancestral land” of the Sakalava. The significance of this concept is reflected in indigenous terms that are used to distinguish insiders from outsiders: Sakalava are the tera-tany (“the children” or “possessors of the soil”) or the original inhabitants of the Valley; while migrants are referred to as vahiny (“guests”). These symbolic distinctions also carry significant economic weight, ultimately affecting success or, even, survival. It is Sakalava tera-tany who most often control access to and hold rights over the use of the most important local resource: arable land.

Tromba, as the spirits of dead Sakalava royalty, define the key features of contemporary Sakalava identity.^[2] Amid rapid social change, they remain the guardians of local sacred space and they are significant historical actors for collective experience. Possession by the dead is deeply rooted in Sakalava history: among the earliest surviving written accounts is one recorded by Luis Mariano, a Portuguese Jesuit who described this phenomenon when he visted the island in 1616 (Lombard 1988: 23, who in turn cites A. and G. Grandidier, eds. 1903–1920, vol. 2: 251, 255). Spirit possession as an institution subsequently accompanied the rise of royal dynasties and their associated kingdoms in the sixteenth century. Because spirit possession is central to Sakalava culture, descriptions of this form of possession provide a means through which to trace the historical development of the Sakalava as a people, who today are organized into a chain of kingdoms along the west coast of Madagascar (Kent 1968; see also S. Ramamonjisoa 1984). In turn, since tromba spirits are known, historic personalities, spirit possession is an indigenous form of recorded history where Sakalava preserve knowledge of royal genealogies and, ultimately, of who they are more generally. In everyday terms, it is the tromba spirits that define who Sakalava are in contrast to the ever-increasing influx of migrants.

Tromba possession is a mainstay of everyday life in Ambanja. Tromba spirits address the living through mediums, the majority of whom are female. In precolonial times (prior to 1896) there was only a handful of mediums who served as counselors to members of the royal lineages. Within the last fifty years, however, there has been a virtual explosion in the incidence of tromba possession throughout Sakalava territory. Today, tromba spirits proliferate, especially in Ambanja. In this town of approximately 26,000,^[3] perhaps 60 percent of all women are possessed, of whom nearly 50 percent are non-Sakalava migrants. Tromba mediums are respected by both commoner Sakalava and non-Sakalava as powerful healers and as advisers on personal affairs. Through the assistance of tromba spirits and their mediums, insiders and outsiders may manipulate their statuses and their personal relationships in this town.

In Ambanja, tromba is perhaps the most significant local instititution for both Sakalava and non-Sakalava, and its importance is reflected in several ways. First, as a key aspect of indigenous religion, it is dynamic, constantly changing in form so as to remain a central defining principle of what it means to be Sakalava. In turn, as an essential Sakalava cultural institution, it also manages or regulates the incorporation of outsiders into the Sakalava community. An important feature of tromba in this polycultural town is that most non-Sakalava may, over time, actively participate in tromba, either as observers, mediums, or mediums’ clients. Tromba is the only local institution that enables migrants to permanently transform their identities. Through a complex fictive kinship system, non-Sakalava mediums are symbolically redefined as Sakalava. In this way they become enmeshed in local networks that increase their access to local institutions of power. Since tromba mediumship is primarily a female experience, tromba facilitates the incorporation of migrant women over men, and no comparable institution exists for men.

Tromba possession as ritual has a performative dimension (G. Lewis 1980: 22ff) where the actions, gestures, words, and knowledge communicated during large-scale possession ceremonies and smaller, private healing sessions reveal indigenous conceptions of well-being. For Ambanja’s residents, well-being does not hinge simply on one’s physical state; it is also mediated by one’s status as tera-tany or vahiny. Tromba mediums, as the embodiment of royal ancestral power, assist others in their attempts to cope with the problems they encounter while living in this urban community, one that is dominated by a plantation economy. Thus, embedded in the symbolic order of tromba are critiques of community life and its tensions, the meaning of work, the local political economy, and the dynamics of local power relations over time.

Critical Approaches to the Study of Affliction

The politics of identity in this town is multifaceted, requiring an integrative theoretical approach. In order to comprehend the many levels of meaning and experience inherent to life in Ambanja, this study will draw from theory in medical anthropology, migration studies, and the study of ritual. A historical dimension is also necessary if we are to comprehend the meaning of power at this particular point in time.

Health and Critiques of Capitalism

This study is framed by theory generated within the field of critical medical anthropology. In many ways this is highly contested terrain: various authors argue that since the term critical is (neo-)Marxist in origin, it is this theoretical stance that should be the primary (or even exclusive) orientation. As a result, there are writers who, assuming the more conservative position, advocate a return to the older label, “political economy of health.” Still others strive for a more loosely defined or eclectic approach that draws from interpretative, phenomenological, and epistemological frames of reference.^[4] (For more details on this debate see, for example, Baer 1982, 1986; Baer et al. 1986; Doyal 1979; Elling 1981; Frankenberg 1988; Morgan 1987, 1990; Morsy 1979; Navarro 1976a, Scheper-Hughes 1990; Singer 1896.)

This present study is situated primarily within the latter (more eclectic) camp; nevertheless, it seeks to draw from the strengths of each. A key assumption that runs throughout this study from Madagascar is that, first, inequality and power are often significant (or critical) factors for understanding health and well-being. This is true not only within highly complex and stratified Western societies, but also in smaller (and, in this case, urban) communities in the Third World, where colonial policies and relations have either introduced new forms of inequality and stratification or exacerbated older ones.

Second, literature written in the genre of the political economy of health assumes the pervasiveness of biomedical (or what I will refer to as clinical) medicine worldwide and focuses on the advantages as well as the disadvantages of biomedicine; nevertheless, this is not the most appropriate focus in Ambanja. Tromba mediums and other indigenous healers are preferred by many local inhabitants over biomedically trained practitioners because they have a clearer understanding of indigenous conceptions of illness and disorder. In terms of daily practice, biomedicine is not hegemonic or heterodox in nature (Baer et al. 1986: 95–98) but instead is perceived as ineffective and is less valued than indigenous forms of healing. Unlike tromba mediums, for example, clinicians do not comprehend or embrace their patients’ beliefs about the relationship between the cosmos and the social world. In addition, it is not medicine per se that must be the primary focus of the research but, rather, affliction, as a broadly defined category of human experience. Thus, in this study from Madagascar, it is not solely disease or sickness that will be investigated, but individual and collective problems that are social, economic, and political in nature.

A third assumption is that the human body provides a rich terrain upon which these problems may be played out or expressed. This becomes apparent during healing sessions and other ritual settings. The body may also provide an appropriate medium for transformations in identity, through dress as well as action. This may occur during the pageantry of possession ceremonies, but these in turn may have long-lasting effects on a medium’s life and sense of self. Finally, comprehensive critical interpretation requires the investigation of linkages between political-economic, symbolic, and historical elements.

If we begin with critical approaches to affliction, Morgan’s (1987: 132) definition of the political economy of health proves useful:

…a macroanalytic, critical, and historical perspective for analyzing disease distribution and health services under a variety of economic systems, with particular emphasis on the effects of stratified social, political, and economic relations within the world economic system.

Morgan advocates that we analyze medicine (and, by extension, medical systems more generally) in reference to competing modes of production; take into account the significance of social class; give greater attention to history, including colonial as well as precolonial periods; and explore capitalism’s effect on the declines as well as the improvement of health status (1987: 146).

Although a Marxist orientation offers valuable tools for understanding the effects of capitalism on the local political economy, and on indigenous social relations and healing practices, this definition requires a few adjustments. First, a purely macroanalytic stance (not necessarily something that Morgan herself would advocate) overlooks key aspects of everyday life which are so much a part of anthropological analysis. Thus, for the purpose of this study, the microlevel will be the primary unit of analysis. Following Morsy’s approach (1978; 1991: 205ff), the phenomenon of spirit possession will be analyzed against historically based, macrolevel developments.

Second, as Morgan notes, too often the political economy of health relies heavily on world systems and dependency theory, yet such an orientation on its own is outdated. In turn, it is also too restrictive because, as J. Comaroff explains, “the total penetration of the world capitalist system…denies determination to forces outside itself” (Comaroff 1985: 154, as cited in Morgan 1987: 141). Specifically, what is at stake here is local determination. As the present study will show, the inhabitants of Ambanja are not passive victims of capitalistic forces; rather, ritual forms enable them to transcend and transform labor relations, the meaning of work, and future economic development of the region.

Moving Beyond Class-Based Analysis: Understanding the Migration Experience in Northwest Madagascar

A class-based analysis is an essential element in the study of inequality from a Marxist perspective, but this concept, as Strathern (1984: 3ff) has illustrated, does not necessarily translate well cross-culturally. As Comaroff and Comaroff (1991: xi) ask, “how is [consciousness] mediated by such distinctions as class, gender, and ethnicity?” Throughout Madagascar, ethnicity is generally a far more important unit of analysis. At times it may correspond with Western conceptions of class; elsewhere it operates independently. For example, on a national scale, the Merina and Betsileo of the central high plateaux have the greatest access to Madagascar’s resources, and it is they who form a significant proportion of this nation’s elite. This is an issue that is of prime importance to the Sakalava of Ambanja. As case studies throughout this work will illustrate, ethnicity is a much stronger determinant than class for local power relations in this town.

Gender is also a significant category, one that extends beyond the limits of social class or ethnicity. It is especially important to the understanding of the migration process in northwest Madagascar. An assumption running throughout studies of voluntary migration in Africa is that it is predominantly an adult male experience. This process is rooted in colonial policies and subsequent demands of capitalist labor, where indigenous peoples have experienced shifts in the division of labor by gender. Men have become the primary wage earners, while women remain in rural areas, caring for the land, animals, homestead, and offspring. Laborers send remittances home to support kin, and the presence of women in the villages ensures the reproduction of future generations of labor (see, for example, Cohen 1969; Epstein 1958; Mayer 1971; Meillassoux 1982; Murray 1981; Powerdermaker 1962; Richards 1951).

Madagascar provides important contrasts. As Little (1973) and Schuster (1979) have shown, migration can be a liberating experience for women who relocate on their own to towns. In Madagascar, women migrate, often alone. The majority of my female informants in Ambanja were the sole heads of their households, and they were the primary wage earners, often supporting three generations: themselves, their children, and their aging mothers. Women experience the freedom that accompanies migrating alone, but they must also endure the obstacles of migrant status. If migrants are to succeed economically and socially, they must establish networks that will enable them to settle and become integrated into their new community. An important focus of this study is to unravel how women accomplish this. Age is yet another dimension: children may also figure prominently in the migration process, although they are relatively invisible in studies on this topic. This problem will be explored in chapter 9.

More generally, capitalism shapes the nature of social relations, and anthropological approaches offer important tools for analyzing the manner in which indigenous forms of social structure respond or are transformed, as well as how capitalist relations are perceived and experienced. Throughout Madagascar a premium is placed on kinship, since it is kin-based networks that one may exploit in times of need. In Ambanja, strangers are very vulnerable. They can rely on no one, and so the newly arrived migrant engages in a passionate search for kin or others from the same ancestral land who can help them find work as well as housing, loans, food, child care, and, perhaps, over time, land. Tera-tany status has these advantages, but integration for migrants is very difficult, because Sakalava dislike outsiders. Although marriage is one way in which to become integrated locally, marriage unions in Ambanja are generally short-term and tenuous, making this an unsatisfactory solution.

Kinship and wage labor define opposing categories of social relations in Ambanja. Wage labor is perceived simultaneously as an economic and social relationship, one based on inequality and involving payment in cash for services by an employer to an employee. Wage labor may generate income, which for migrants is essential to survival, but it does not carry with it the permanent and obligatory social ties that characterize kinship. Kinship, on the other hand, is defined by reciprocal relationships that involve the dual exchange of goods and services.

Capitalism and, more generally, colonialism, as social and political forces, have also affected social structure by undermining indigenous power. In the precolonial era, Sakalava defined themselves in relation to their royalty. Throughout the colonial period the French sought to undermine royal authority. As a result, structural rules that defined the relationship between commoner and royalty broke down. This process in turn affected Sakalava principles of commoner kinship. More recently, the opposed categories of insider and outsider have become the most important social categories for daily interactions in Ambanja; these are shaped by national and local ethnic factionalism. Tromba possession is among the few institutions that mediate between the competing statuses of insider and outsider. It also offers its participants a means through which to opt out of wage labor and capitalist relations by embedding them in an alternative social world defined through fictive kinship.

Understanding Indigenous Notions of Power

If we are to comprehend the nature of inequality in Ambanja, it is necessary to construct a clear definition of power; specifically, we must explore how it is indigenously conceived. Arens and Karp note that anthropological discussions of power rely heavily on Weber’s definition (Macht) (Weber 1947, as quoted in Arens and Karp 1989: xv). Weber focuses on social actors engaged in rational and “mutually acknowledged” exchanges that are based within established and accepted arenas of authority (ibid. 1989: xiii ff). Arens and Karp argue for the need to break from universalist (and thus ethnocentric) definitions of power to more culturally specific ones, especially in reference to cosmological systems. For the purposes of this study, the greatest limitations of Weber’s definition lie in his focus on individual action over collective experience and his preoccupation with formal authority. This study will also explore how power is symbolically expressed.

Arens and Karp suggest that anthropologists consider an alternative definition provided by D. Parkin:

Power rests [not simply] on the acquisition of land, myth, and material objects but rather that which comes from unequal access to semantic creativity, including the capacity to nominate others as equal or unequal, animate or inanimate, memorable or abject, discussor or discussed. (Parkin, ed. 1982: xlvi)

Even though Parkin’s definition focuses on power through discourse, it still offers an appropriate alternative, since it marks a break from a materialist orientation characteristic of Weberians (and neo-Marxists). For the sake of this analysis, I would replace Parkin’s term semantic with symbolic. Furthermore, although nominate implies formal political action, the process of determining status is essential in Madagascar, specifically in reference to ethnic factions that characterize life at national and local levels.

Arens and Karp also prefer Parkin’s definition because “this perspective moves us away from an exclusive emphasis on the exercise of power and provides room for examining the relationship between power and consciousness” (1989: xiv). These authors are especially interested in Weber’s distinction between authority and power and stress the importance of exploring the latter: “Power is multicentered and, further,…the idea of a center may itself be produced through the ideology of power.” As the examples in their volume illustrate, “the source of power resides in the interaction between natural, social, and supernatural realms” (1989: xvi, xvii).

In Ambanja, it is clear that power may be exercised in different arenas: through formal structures (for example, by the French colonial government or the present state of Madagascar), ritual authority (tromba mediumship), and individual action (by Sakalava tera-tany over non-Sakalava vahiny). It also may be manifested through legal action or ritual. Tromba provides an important example for the study of power. It is formal and institutionalized, yet the authority previously investedin living royalty has been transferred almost completely to the spiritual realm. As will become clear below, a consciousness of power relations may be articulated in ritual forms that can have temporary or long-lasting effects.

Historical Considerations

As Bloch (1986: 157ff) has illustrated in his study of Merina circumcision, rituals may change form over time in response to transformations in power relations on a national scale. In turn, historical and other forms of knowledge, as well as power, may be embedded in ritual, so that ritual may operate as a force of resistance and change against the state or other forces (Apter 1992; see also Fry 1976 and Lan 1985; Sahlins 1985). This pairing of history and power is an important theme underlybing tromba possession, since it is not a static religious institution, but one that has responded in unique and creative ways to ever-changing political, economic, and social forces. These transformations are rooted in the manner in which the Sakalava of Ambanja conceive of their recent history, which they divide into three major periods: the precolonial (prior to 1896), the colonial (1896–1960), and postcolonial (1960 to the present). Transformations to the precolonial order is a theme of chapter 3; here I would like to provide a brief summary of recent national developments since they inform contemporary Sakalava perceptions of urban life.

The evolution of the Sambirano into a lucrative plantation economy is rooted in the colonial period. Three policies under the French administration are significant here. First, following conquest, the French removed the Bemazava-Sakalava from the most fertile regions of the valley in order to make room for large-scale plantations. Second, when the Bemazava refused to work as wage laborers, local planters recruited migrant labor from other areas of Madagascar. Third, as part of French pacification policies, the colonial government sought to undermine the authority of Bemazava royalty (see chapter 2).

These policies lie behind more recent changes that have occurred within the last three decades. The postcolonial era in turn may be broken down into two major periods. From 1960 to 1972 marks the period of the Malagasy Republic (République Malgache). President Philibert Tsiranana, who served throughout this period, retained many of the colonial structures and continued to employ French advisers and administrators. Also, major industries remained in the hands of private citizens, many of whom were French expatriates. In the final years of his presidency, Tsiranana witnessed the rise of malagasization,^[5] a nationalist movement that was socialist and anti-French. Originally its advocates demanded that all school curricula be taught in the Malagasy language so as to define and reflect a national—rather than foreign—culture; later they demanded the expulsion of French technical advisers working in the upper echelons of the government. The collapse of the Tsiranana administration was precipitated by a strike of university students in March 1972, followed by a peasant uprising and military coup (the May Revolution) (Covell 1987: 45–46). Tsiranana was expelled, and the country was renamed the Democratic Republic of Madagascar (Repoblika Democratika Malagasy, République Démocratique de Madagascar).^[6]

The Socialist Revolution, which spanned the period from 1971 to 1975, was marked by violence and drastic changes in political orientation under three separate presidents. Following the assassination of Colonel Ratsimandrava, who held office for less than a week in 1975, Didier Ratsiraka was named president (and continues to hold this position today; he was reelected in 1989).^[*] Ratsiraka’s early policies included the rejection of institutions based on French models and, subsequently, the withdrawal of much foreign capital. Until 1987, Ratsiraka’s administration had been fiercely isolationist and advocated self-sufficiency. The nation’s strongest allies have been located in (what was then) Eastern Europe, the People’s Republic of China, and North Korea. It was only recently (in mid-1986) that Madagascar chose to strengthen its ties with Western powers. (For more detailed discussions of the recent political history of Madagascar see Althabe 1980; Covell 1987: 29–75; Rabenoro 1986; and Rajoelina 1988.)

As this book was going to press, national elections were held. On March 27, 1993, Albert Zafy was sworn in as Madagascar's new president.

Although Madagascar by name is a socialist state, its economy may be defined more clearly as a form of state capitalism, where the ownership of all major industries and land holdings rests with the national government.^[7] In the Sambirano, capitalism, and not socialism, shapes the bulk of economic relations. Informants from the Sambirano report that only a few changes in the economic order occurred following the Revolution, the greatest involving land reform policies. In this regard, the holdings of foreign planters were seized by the state, and portions were then resold to local residents, with preference shown for those who could prove that they were tera-tany—that is, whose ancestral land was the Sambirano. Since the Socialist Revolution, state-owned plantations have been referred to as enterprises (FR: entreprises). The enterprises are among the few profitable export businesses of Madagascar, and several operate fairly independently from the government. Also, state- and privately owned businesses exist side by side in the Sambirano. In the transition to state ownership, workers experienced no major structural changes in their places of employment. Labor unions have been legalized, but they are associated with the national party, AREMA, and so loyalty to the state is mandated by the unions. Some workers report that the most significant change was that their hours were shortened, which in turn shrunk their monthly earnings.

Investigating Possession: Social Change, Marginality, and Religious Experience

An assumption running throughout anthropological studies of religion is that through the study of belief and ritual we may grasp indigenous conceptions of the world; ritual especially provides fertile ground for the expression of social tensions or for coping with sudden change (see, for example, La Barre 1970; Linton 1943; Wallace 1956; and Worsley 1968). Studies set in colonial Africa explore the effects of forced resettlement, migration, and urbanization on the web of social life, and the manner in which religion functions as a means for expressing conflict (Colson 1971; Epstein 1958; Gluckman 1954; Mayer 1971; Mitchell 1956; Powdermaker 1962; Richards 1951; Scudder 1966). Furthermore, it is well documented that the incidence of spirit possession often rises dramatically in times of social disruption and crisis. Colson, for example, found that possession occurred with greater frequency among the Gwembe Tonga (in what is now Zambia) when they were relocated by force after the British colonial administration decided to dam and then flood their valley (1969, 1977). The works of Colson and others remain influential in studies of possession, since this phenomenon continues to be investigated in light of the tensions and uncertainties of everday life. A problem underlying these studies, however, is that indigenous peoples are generally portrayed as powerless victims of change who do not fully comprehend the forces that are responsible for new predicaments in which they find themselves.

Ritual, Symbolic Action, and Power

More recent studies of religion have sought to go beyond a functionalist analysis, relying instead on a neo-Marxist framework.^[8] These studies are historically situated and set within the complex of unequal relations that characterize colonized societies in the Third World, and they are concerned with economic oppression and exploitation. A primary focus is the dispossessed, the marginal and powerless members of societies who are victims of a world capitalist system. Examples include investigations by such authors as J. Comaroff (1985), Nash (1979), Ong (1987), and Taussig (1980a, 1987), in southern Africa, Bolivia, Malaysia, and Colombia, respectively.

These authors do not reject religious experience as evidence of false consciousness (see Marx 1964: 43–44); instead, religion reveals an indigenous awareness of the inequalities that characterize capitalist economic relations. Although this level of consciousness is not fully articulated in everyday action and discourse, it is, nevertheless, richly interwoven in the symbolic imagery of religious expression. Nash and J. Comaroff both illustrate how religion lends order to the world, symbolically (and temporarily) empowering its participants. Nash describes how Bolivian tin miners attempt to control the hazards of their work through magic and by appeasing Tio, a hungry and greedy devil who lives underground. Comaroff argues that among the Tshidi of southern Africa, the rituals associated with the Church of Zion are evidence of attempts to reorder the world, where military and other symbols of power are incorporated into the regalia and ritual language of the church. Ong shows how spirits that possess young female Malay workers on the shop floor are a manifestation of the exploitative labor relations that characterize the electronics industry. While possessed, these women attack microscopes and other objects that are the instruments of their oppression (see also Lim 1983). Taussig, in turn, describes the disorder and madness that capitalist exploitation imposes on landless laborers in Colombia and the manner in which an indigenous awareness of such chaos is inherent to healing sessions and other ritual spheres.

Thus, as each of these studies illustrates, ritual action may be rich in symbols that reveal a keen awareness of the manner in which capitalist relations define or undermine the local social order. In Ambanja, perceptions of social and economic change are reflected within the context of tromba possession in several ways. First is the theme of historical knowledge, where Sakalava may reflect on and reinterpret who they are in relation to other peoples. In the context of tromba, power and historical knowledge are intrinsically linked, enabling Sakalava tera-tany to manipulate social and economic forces that shape their world. Second, healing rituals, involving tromba spirits, are an important arena for the expression of the meaning of affliction for tera-tany and vahiny.

Third, land, work, and identity are important local themes that are shaped or redefined through the logic of tromba. In the Sambirano, the meanings associated with land are multifaceted, since it is both an economically and symbolically defined category. Madagascar is a nation of peasants: it is estimated that approximately 80 percent of the population acquires some proportion of its subsistence needs from farming (World Bank 1980). Natural disasters—such as droughts and cyclones, coupled with population pressures and deforestation—threaten the viability of a large percentage of the population maintaining rights to arable land (cf. Downs and Reyna, eds. 1988). Severe economic problems also make it difficult for many to find wage labor. Migrants come to the Sambirano hoping to acquire work or land. In this region land is valued as a means of production, since ownership ensures that one may remain economically independent and thus free from capitalist relations that characterize wage labor. But ethnic identity and land are also intertwined, since access to land is defined by the opposing categories of tera-tany and vahiny. Vahiny seek to be rooted symbolically to the local ancestral land because it provides a means to achieve insider status. This is something that may occur through participation in tromba possession. In addition, the importance of this opposition of insiders and outsiders is reflected in the multiple meanings assigned to the word “work” (asa) by Sakalava, migrants, and employees of the French colonial and present governments. The meaning of work may also be redefined through possession.

The Question of Possession and Marginality

The relationship between powerlessness and symbolic expression is central to many studies of spirit possession. This argument is most widely known through the work of I. M. Lewis (1966, 1971, 1986, 1991). Lewis argues that possession gives marginalized people a voice and simultaneously enables them to displace the blame for their actions onto their spirits. This is possible because of the manner in which possession operates: it involves a foreign spirit or entity taking control of a living person, whose spirit, soul, or essence is temporarily displaced. Whatever is said or done during possession is the fault of the spirit, not the person. Lewis notes that it is women who predominate in possession cross-culturally, and so his theory is a corollary to a more general theory in gender studies which assumes the nearly universal subordination of women. A key assumption of this argument is that in any given culture, it is men who dominate formal power structures (Rosaldo and Lamphere, eds. 1974; Ortner 1974; and Reiter 1975).^[9]

The question of why possession is so common cross-culturally (Bourguinon 1973) has been a perplexing problem for anthropologists. Numerous authors have sought to break from I. M. Lewis’s argument, proposing other lines of inquiry. Dominant themes include exploring possession from psychological or psychiatric angles, either by analyzing the cultural construction of self and personhood (Crapanzano 1973, 1983; Crapanzano and Garrison, eds. 1977; Obeyesekere 1977, 1981), possession as pathology (Krippner 1987), or by challenging the pathological model and, more generally, assumptions about culture-bound syndromes (Bartholomew 1990; Hahn 1985; Karp 1985). Others have proposed a biological model; among the better known is the calcium deficiency hypothesis (Kehoe and Giletti 1981; Raybeck et al. 1989). Social roles and function provide another focus, including possession’s therapeutic role (I. and P. Karp 1979; Prince 1964) and its communicative and performative aspects (Ackerman and Lee 1981; Lambek 1980; Lebra 1982; Leiris 1958; Rouget 1980; V. Turner 1987: 33–71). Others draw on feminist theory (Boddy 1988; 1989) or political-economic approaches (Lim 1983; Morsy 1978, 1991; Ong 1987).^[10]

Functionalist arguments about marginality, nevertheless, remain implicit in the majority of writings on possession, since many authors share I. M. Lewis’s assumption that possession provides the powerless with a means to symbolically express social, economic, or political oppression. A second assumption that dominates the literature is that possession is a temporary experience confined to ritual contexts. In the literature framed by a political-economic approach, for example, possession is viewed as a means to express one’s powerlessness relative to established structures, but it has no long-term effects that alter relationships with spouses, employers, and others (see, again, Ong 1987, for example). Psychological studies focus on possession as a means to redefine identity, yet the subjects under study are still assumed to be somehow marginal to the greater society. For example, Crapanzano, in his study of the Hamadsha, presents an interesting case, since it involves men rather than women. He argues that adherents experience shifts in their identity and social roles, yet this institution remains a “peripheral” or “fringe phenomenon” relative to the greater Moroccan society (1973: 7) because they are members of the urban poor who are also psychologically disturbed.

There is, however, a growing body of literature that has begun to challenge the presumed universal association between marginality and the possession experience. Giles argues that women engaged in possession activities in Mombasa are not marginal members of their society, nor is possession, as Lewis advocates, peripheral to orthodox Islam. She advocates that possession by pepo (or sheitani) spirits is a central aspect of Swahili society, and historical analysis reveals that in the past it was more widespread than it is today. Pepo cults provide their members with close-knit groups that offer social support that extends beyond possession ceremonies into daily life. She also stresses the importance of gender differences for understanding the significance of possession in the community, and that possession cults are among the few places where women and men may interact on equal footing. This is an argument that Boddy (1989) has also taken up in her study of zar possession in the Sudan. She stresses that there is insufficient evidence to support the “fantasy” of a dominant male culture that subordinates the culture of women—it is a matter of perspective which is mediated by the gender of the informant (1989: 6) (and, perhaps, the anthropologist).

In Madagascar, several authors have analyzed possession in reference to social and economic change. An assumption running throughout these studies is the assumed centrality of possession in the local culture. Althabe (1969: 95–118), for example, illustrates the manner in which tromba possession among Betsimisaraka laborers of eastern Madagascar operates to integrate outsiders into communities located near a large sugarcane plantation. Marriage patterns are exogamous and virilocal; in this context a woman who is a “stranger” may become a spirit medium for the ancestors of her husband’s kin. Thus, possession operates as an institution for incorporation, whereby the new wife becomes more closely linked to her affines. In turn she is empowered, since it is through her that the ancestors speak. Fieloux and Lombard (1989), writing from southwest Madagascar, illustrate the usefulness of an analysis that focuses on the local political economy for understanding possession (again, see Morsy 1978, 1991). They describe changes in bilo possession which reveal the unease felt by local Masikoro, among whom the introduction of a new cash economy is undermining animal husbandry. Individuals who have grown “rich” and “fat” from their involvement in the cotton industry have lost sight of kin-based social relations, and bilo, as a ritual form, reflects these concerns and transformations.

This study of tromba possession in northwest Madagascar seeks to offer other important counterexamples, where the dominant assumptions about religious experience—and, more specifically, about possession and its participants—are inappropriate. As I will show, tromba possession is not a peripheral domain in the community of Ambanja; rather, it is a central and defining aspect of Sakalava culture. Furthermore, tromba possession may be permanently empowering, its significance extending beyond the ritual context into everyday life. This case from Madagascar also offers an example of the structural legitimacy of possession, in which kinship is a central operating principle. As a result, tromba carries with it important ramifications for those who occupy competing statuses of insider and outsider. What I will show is that to be active in tromba is a sign that one is not marginal.

The Body as a Natural Symbol

As Mauss reminds us, “The body is the first and most natural tool of man” (1979; quoted in Scheper-Hughes and Lock 1987: 1). Medical anthropology and sociology are rich in accounts that detail how the body serves as a field on which affliction is experienced and expressed (see, for example, Blacking 1977; J. Comaroff 1985; Crapanzano 1973; Martin 1987; Ong 1987; Taussig 1980b; B. Turner 1984, 1987). It is assumed here that spirit possession has a performative aspect and thus, in the context of tromba, the human body is a vehicle for the expression of conflicts through the manipulation of a symbolic order that is shared by members of this community. The aim of this study is to analyze the logic (or, as Taussig [1987] would say, the chaos) of the symbolic meanings that are played out on the body of the medium. Also, tromba possession may be a transformative process that affects different levels of identity.

Scheper-Hughes’s and Lock’s concept of “the three bodies” is helpful for exploring the symbolic nature of tromba possession, since tromba is significant to individual, social (as well as cultural), and political experiences. As a medium enters and exits trance, she experiences shifts in different levels of identity, shifts that may be temporary or permanent. These include selfhood,personhood, and ethnic identity. On the first or “individual” level, the first two forms of identity come into play. As these two authors stress, the manner in which the notion of self is defined (and therefore experienced) is not universal but culturally specific. In the context of tromba, selfhood is subjectively perceived as shifting in nature as the medium moves in and out of trance. In this sense it operates on a very personal (or psychological) level. Possession also affects notions of personhood: in this sense, the medium is perceived as a social being who experiences structural shifts in relation to her kin and friends, and these shifts affect her role and status in the community.^[11]

On the second level, or in reference to the “social body,” indigenous concepts of ethnic identity are significant. As a medium for royal spirits, a woman embodies what it means to be Sakalava, and this is evident in her words, attire, and actions. Thus, when possessed, her individual body provides a map of the social body. In addition, an important shift occurs if she is a migrant, because, through tromba possession, her ethnic identity is altered: when she becomes a medium for tromba spirits she is redefined ethnically as Sakalava. Since the concept of ethnic identity is central to this study, unless specified otherwise the term identity refers to ethnicity.

The third level is the “political body”: in this sense possession may be seen as politically charged (see, for example, Lan 1985). The political nature of tromba is defined through ancestral authority, a form of power that is respected locally and legitimated through recent national policies, including malagasization. As a vessel for Sakalava spirits, a medium embodies the Sakalava past: she helps record, shape, and redefine what it means to be Sakalava in the context of social change. She carries the authority of Sakalava ancestral power, and thus it is she who may manipulate relationships between individuals and groups in the Sambirano. This occurs most often during healing ceremonies, where she helps individual clients gain control over events that disrupt their lives. In other instances mediums may dictate the thrust of future economic development in the valley. The economic importance of the Sambirano, at a national level becomes more apparent when it is evaluated against the backdrop of the country as a whole.

The Logic and Methods of Inquiry

The Setting

Madagascar is the fourth largest island in the world (approximately 1,500 kilometers long, with a surface area of 587,000 square miles), yet until recently it has received relatively little attention from scholars outside Madgascar and France. Geographic and political isolation and low productivity contribute to the severity of economic difficulties inherent in life in Madagascar (World Bank 1980; Pryor 1988). In many ways this nation is considered to be a backwater of the Third World. Reasons for this are reflected in recent socioeconomic statistical data for this country. The national population for 1984 was estimated at 9.9 million, with a doubling time of twenty-six years; and the average life expectancies for men and women are 54.2 and 57.8 years, respectively. In addition, the national income has shown a sharp decline over the past twenty years, especially since 1970 (Covell 1987: xiii ff). It is only within the last two years that Madagascar has received much coverage in the popular press (see, for example, The Economist 1989; Jolly 1987, 1988; Payer 1989; and Shoumatoff 1988a, 1988b). Although these recent articles focus primarily on environmental depletion through the destruction of precious rain forests, relatively little has been written on the increasing severity of social and economic problems faced by Madagascar’s citizens.^[12]

Most English-speaking anthropologists are familiar with Madagascar primarily through the works of Bloch (1989) and Kottak (1980), who have written, respectively, on the Merina and Betsileo of the high plateaux. The geographical interests of these two anthropologists reflect a more general trend in the social sciences in Madagascar, where, until recently, the highlands have been a focus for the most intensive research.^[13] Another main focus has been village studies of discrete groups of Malagasy speakers. As a result, two main goals of my research have been to provide a much needed study of coastal peoples of Madagascar and to explore the significance of Malagasy identity in a polycultural, urban setting. By focusing on migration, this study seeks to reflect the reality of life as it is experienced by many Malagasy today. Today Malagasy of diverse origins must cope with living together, especially in urban environments, as opposed to the ideal, where peoples of common origins inhabit the same territory.

Migration is very much a part of Madagascar’s history and it has shaped ideas about Malagasy identity. Although Madagascar, by virtue of its geographical location, is considered to be an African nation, the Malagasy language is linguistically categorized as Austronesian. Malagasy trace their origins to locations as diverse as Polynesia, the islands of Southeast Asia, East Africa, and the Persian Gulf. The influence of these diverse cultures on what it means to be Malagasy is revealed through social, structural and cultural similarities, including rules of descent, burial practices, styles of dress, cuisine, and physical features. Linguistic analysis also reveals loan words from kiSwahili, Arabic, Malay, French, and English (Southall 1986). Cultural origins are a topic of heated debate among Malagasy intellectuals, who generally prefer to look eastward to Indonesia and Malaysia for their origins. These categories are important for current conceptions of ethnicity.

Another purpose of this study is to analyze the relationship between tromba possession and social change. Spirit possession is among the most common themes found in the literature on Madagascar (see, for example, Althabe 1969: 115ff; Estrade 1977; Fieloux and Lombard 1989; Ottino 1965; Raison-Jourde 1983: 48ff; S. Ramamonjisoa 1984; Rason 1968; Russillon 1908). As my own work progressed in Ambanja, these studies seemed, for the most part, outdated, since they stress the nature of tromba as it should be, set within the royal context. They do not address its dynamic and changing nature over time (for notable exceptions see Althabe [1969], and Fieloux and Lombard [1989], as noted above ). Today, tromba possession is common throughout Sakalava territory, but its incidence is unusually high in Ambanja. Tromba is the point at which social, religious, political, and economic forces converge, and so this institution provides a means to explore the dynamic nature of daily life at all levels in this community.

Research Methods

The data for this study were collected during a twelve-month period (January 1987 to January 1988), originally for the purposes of dissertation research. Thus, the ethnographic present is 1987. The first two months of research were dedicated to conducting archival work in the Bibliothèque Nationale, the Archives Nationales, and university and seminary libraries located in the national capital of Antananarivo and in the provincial capitals of Toamasina (Tamatave) and Antsiranana (Diégo-Suarez or Diégo).^[14]

Anthropological fieldwork was based primarily in Ambanja and the surrounding countryside of the Sambirano. Comparative data were also collected during short-term visits to Diégo in the north and Mahajanga on the west coast, the east coast town of Antalaha, and other locations listed below. The primary methods used for data collection were participant-observation of many facets of daily and ritual life, and personal, often multiple, interviews with a wide variety of informants. These interviews were conducted in Malagasy, primarily in the Sakalava dialect, although some were carried out in other dialects or in French. Two assistants aided me throughout the course of this work. My primary assistant was a young Sakalava woman in her late twenties who was born into a tera-tany family. The second was a high school student who was the child of migrants.^[15]

Throughout the course of the year I attended a total of nine large-scale tromba ceremonies, involving as many as twenty mediums at one time. Since many mediums work as healers, I also attended more than twenty private consultations with clients, and I was a regular observer and client in the homes of four mediums. Throughout the course of the year I encountered a total of ninety-eight mediums (of whom four were male). I conducted in-depth interviews with twenty of these (including two men), followed by additional visits and follow-up interviews.

A special quality of tromba possession is that each spirit has a distinct personality, a characteristic it shares, for example, with possession found in Korean (Kendall 1985, 1988), Brazilian (Bastide 1978; S. and R. Leacock 1972; Krippner 1987), Haitian (K. Brown 1991), and west and north African traditions (I. M. Lewis et al., eds. 1991). As a spirit enters a medium’s body, it displaces her own spirit, so that she does not remember what came to pass when she was possessed. This gap in knowledge is most clearly illustrated by the following episode. In October I was traveling with a group of five women and a five-year-old girl. One day the child fell sick with malaria, trembling and delirious from the fever. My adult companions felt powerless to help her because, as one woman explained, “I know nothing [tsy heko é!] about how to cure this type of sickness; only my tromba [spirit] knows how [mahay ny tromba].” She and the other women assembled and called up her spirit to heal the child. After the session had ended and the medium had left trance, others told her what her spirit had done. (The child eventually recovered.)

Since the knowledge of the spirits is sacred, mediums are reluctant to discuss tromba in detail, fearing that their spirits will become angry and will make them sick. Others who have heard a spirit’s words or who are knowledgeable about tromba possession also prefer not to discuss them in any detail. For this reason, I interviewed mediums in and out of trance, collecting life histories from both the living and the dead (cf. Wafer 1991). Mediums’ kin, neighbors, friends, and clients were also valuable informants who provided complementary data.

Life histories provided a means through which to explore different areas of mediums’ experiences, including possession, migration and work, social ties, education, and health. Several attempts were made to conduct formal household surveys. These were designed to enable me to collect similar data from a broader spectrum of the population. I was forced to abandon this technique since it met with great resistance (and, at times, anger) from my informants and research assistants. Malagasy believe that the government already asks too many questions of private citizens, and I did not wish to alienate potential informants. Instead, much of the information originally requested on the survey was collected during informal discussions on related topics.

Sakalava in Ambanja draw sharp distinctions between what I will refer to as popular tromba, or the form that possession takes in town, and royal tromba. The latter involves a handful of mediums who are possessed by the greatest royal ancestors, and they live on Nosy Faly, a small island located just off the northwest coast of Madagascar. It is here where the Bemazava-Sakalava tombs are located. These special mediums work solely as advisers to living royalty. Acquiring information on royal tromba was by far the most difficult part of the research, since it was only after eight months in the field that I was able to gain the trust of local royalty. Data on royal tromba were collected through interviews with the Bemazava king, his wife, and advisers and during a visit to Nosy Faly where, under close supervision, I was able to witness a series of royal rituals and briefly interview two royal mediums. I was also invited by the Antakarana king to attend a series of ceremonies in his ancestral village, Ambatoharan̂ana. (The Antakarana are the northern neighbors of the Sakalava; they share many cultural traditions.) Finally, I made several visits to the nearby island of Nosy Be to consult with royalty of the northern Bemihisatra-Sakalava dynasty.

In addition to tromba, there are other forms of possession that occur in Ambanja. Although tromba is primarily an experience of adult women, njarinintsy spirits, for example, frequently possess adolescent schoolgirls. These spirits are dangerous and volatile, and they are viewed as a serious illness that must be cured. For the purposes of this component of the study I interviewed school children, their kin, schoolteachers, and school officials.

Clinical medicine, as well as Protestantism, Catholicism, and Islam provide additional insights into local ideas about possession. According to Sakalava logic, possession and madness are believed at times to overlap. The most severe cases may be brought by kin to psychiatrists for treatment (Sharp, in press). More often, however, kin and patients seek assistance from special Protestant groups that have full-time specialists who exorcise spirits through power derived from the Holy Spirit. I made several visits to a bush hospital run by a group of exorcists, and I attended a number of their curing ceremonies in Ambanja. My own residence proved to be a focus for such activities, since I rented a house that belonged to the Lutheran church. The Malagasy pastor who lived there during ten months of my stay was trained as an exorcist. He allowed me, as a member of his household and congregation, to observe him at work and we had many lively discussions on tromba possession. Religions of diverse faiths were also important to this study, since they provide a means for male and female migrants to become socially integrated. I therefore sought out parishioners, pastors, priests, and congregation elders from a variety of faiths, interviewing them on possession and, more generally, on life in Ambanja. Clinical medicine was another important area of exploration, since it provided comparative data that assisted in the evaluation of attitudes toward healing. For this reason I observed the actions of local physicians, nurses, and pharmacists and interviewed them about possession as well as the nature of affliction in the Sambirano.

Since the enterprises are central to the lives of the majority of Ambanja’s inhabitants, I interviewed local government officials and enterprise staff and laborers. I also interviewed a wide spectrum of male and female laborers of diverse backgrounds. In so doing, I sought to comprehend the nature of the migrant experience in general terms and to understand variations that might occur as a result of one’s ethnic or geographical origin. I made periodic visits to the enterprises, joining the weekly tour with the occupational health doctor in order to watch her and members of her nursing staff work with patients and so that I could interview managers and workers. To grasp the rhythm of the workday and the labor demands associated with agricultural work, I accompanied a woman to work one day and assisted her as she sorted cocoa and coffee.

The Organization of the Study

This study is divided into three parts, reflecting three different levels of analysis. Part 1, “Historic, Political-Economic, and Social Levels of Experience,” provides background information and a framework for exploring the forces that have shaped indigenous notions of identity. Specifically, it focuses on the economic development of the Sambirano and the rise of local polyculturalism over the past century. Chapter 2 is primarily a historical overview of the development of the local political economy of the valley. Chapter 3 describes the nature of factionalism on national and local levels and the relevance of the conflicting statuses of insider and outsider, or tera-tany and vahiny. Chapter 4 provides case studies that illustrate the experiences of migrants in Ambanja and the factors that facilitate or impede their integration into the local tera-tany community.

Tromba is the focus of Part 2, “Spirit Possession in the Sambirano.” Within this section, chapter 5 is an overview of tromba and other forms of possession. Chapter 6 focuses on tromba as a locus of power in the Sambirano. Chapter 7 explores the structural principles that are associated with tromba and which enable outsiders to become insiders in Ambanja.

Part 3, “The Conflicts of Town Life,” seeks to integrate these two foci—the shifting nature of identity and tromba possession—through discussions on affliction in this polycultural community. Chapter 8 examines the problems that adults face and the role that mediums play as healers of personal and social ills. Chapter 9 explores problems that are specific to children, who form an unusual and especially vulnerable group of migrants. Chapter 10 concludes this study with a discussion of an alternative therapy offered by Protestant exorcists who treat patients troubled by problematic cases of possession.

Notes