>
How Women Speak their Gods
in Watchi-Ewe Spirit Possession
Nadia
Lovell
University of Kent at Canterbury
I will start with a brief exploration of current anthropological approaches to the phenomenon of spirit possession, and examine the applicability of such models to various ethnographic contexts. I shall then explore the particular case of spirit possession among the Watchi-Ewe of southern Togo (for lack of time and space, I will not dwell upon biomedical explanations attributing possession, for instance, to a lack of calcium, nor will I examine theories proposed by ethnomusicologists as to the significance of music in influencing physical movement and metabolic functions).
The terminology of possession is nebulous, and needs to be briefly examined: most writers seem to agree that possession corresponds to an altered state of consciousness (Rouget 1985, Crapanzano 1977, Bourguignon 1973). But the consensus seems to stop there, as possession diversifies into a multiplicity of distinct practices. Possession can be violent or lethargic, it can be controlled or chaotic, involve speech or be experienced in silence, be combined with mediumship or prophetic announcements.
Possession by spirits is but one manifestation of the phenomenon as a whole (according to Bourguignon, 52% of societies in which trance occurs associate this with the presence of spirits (1973). It has attracted the attention not only of anthropologists, but also of psychiatrists. There are obvious areas of influence between these disciplines. Defined in broad terms, anthropological approaches to spirit possession fall into at least three identifiable categories: 1. the structural functionalist model of explanation; 2. the phenomenological approach and 3. the semantic model. Like most analytical models, these are not mutually exclusive, and most studies concerned with possession, while retaining a particular theoretical profile, tend to combine various aspects of each approach.
The structural functionalist model, spearheaded by Lewis, has had a constant flow of followers (and critics) since it was originally propounded in 1971. The argument is based on two fundamental principles: the marginality of women in society, and the peripherality of spirit possession as a religious phenomenon, in stark contrast to the central position of the dominant religious form (Islam in this case). Affliction caused by spirits, frequently leading to possession, is identified in the east African context as a primarily female phenomenon, although a few men are occasionally similarly afflicted. Lewis's contention is that spirit possession, often preceded by severe and extended illness, is a socially anchored psychological response to a repressive and authoritarian system where the majority of women, and a minority of men, are deprived of power. Women, and the few men who do not fit into the dominant male ideology, seek to overturn this imbalance by the only available means at their disposal: spiritual acts. Hence the notion of peripheral cults. Those who cannot enter the mosque as freely as men can, those who do not have access to the Quran to the same extent as men do, due partly to the exclusion of women from the literate tradition, develop alternative forms of religious expression and activity which are necessarily inferior in status to the main, male, predominant model. If access to the dominant god is restricted, women resort to peripheral spirits to express their religiosity, while simultaneously expressing their wishes and desires under the guise of their spirits. Since men also believe in spirits, they cater to the requirements of their womenfolk without jeopardising their own supremacy. Small concessions are thus made, for the sake of the spirits rather than for women themselves. This face saving exercise, Lewis explains, thus allows for the alleviation of tensions while preserving the fundamentally male dominated fabric of society. The power of the weak may occasionally be allowed a voice, but the war of the sexes rages on without respite for either party.
Despite fierce criticism, primarily from female anthropologists working on the topic, this theory of marginality and peripherality survives almost untouched, attracting new adherents among every generation of anthropologists. Lewis' latest analytical venture into the world of spirit possession mirrors his original thesis all too clearly (Lewis 1993, see also Gellner 1994, Constantinides 1985, Gussler 1973).
There is, in Lewis's work, an acknowledgement of the psychological deprivation experienced by the `weak', but the religious experience of spirit possession is placed unequivocally within a social context. Others, such as Crapanzano or Zempléni (1977), concentrate far more on the psychoanalytic aspects of spirit possession, viewing possession as a liberation from the constraints of the conscious. Possession provides an escape from guilt. Although more sophisticated than its predecessor, the psychoanalytic explanation still relies on the notion of possession as a form of resistance, as the expression of protest from the margins, and a mechanism for the weak to compensate for inequalities.
Alternative explanations have targeted primarily the notion of marginality, arguing that such theories exclude what may not, in effect, be excluded (Giles 1987, Boddy 1989). Seen from a male perspective, including that of a male anthropologist and male Muslim actors, Zar and Pepo cults may indeed be considered marginal to the tradition of orthodox Islam. Seen from the perspective of women, Islam comes to include the religious forms which they share with men as well as their own adherence to possession cults. A similar argument is made by Ardener when he describes Bakweri concepts of nature and the wild, which are strongly gendered, and encompass different signifiers for men and women. Moreover, spirits are not as external to Islam, or other 'dominant' religious forms, as observers have stipulated, but are inherently part of the religious traditions. Furthermore, in the cases of Islam or early Christianity, a preference of emphasis on text rather than practice as valid religious form will also influence theoretical insights. The writings of Giles and Boddy also reflect on the difficulties involved in defining the criteria by which certain cults come to be deemed peripheral, while others are considered central to the religious order of society. If possession cults are peripheral, to whom are they considered to be so? Surely, the notion of peripherality depends on the stance of the observer, rather than that of the practitioner. The question is one of perspective, of observer and observed, of who defines whom.
However, some of the criticism is itself structuralistic in its approach. The structural organisation of possession cults is thus re-examined: women may indeed be the primary targets of spirits, and the membership of cults may be predominantly female, but their position in society at large must be scrutinised in order to actually determine how marginal they are. What transpires in Giles study, for instance, is a context where many women are pivotal to the well being of large sections of their communities, male and female, and where they hold prestigious positions on a par with those of male religious leaders. The redefinition of previous theoretical paradigms is therefore crucial, but its constitution remains more or less intact.
The essentialistic, structuralistic, emphasis on the organisation and function of possession cults as cathartic mechanisms denies the dynamic processes involved in the experience of spirit possession, overlooks the fundamental aspects of possession allowing for the expression of identities which are not one's own, and which can be appropriated, albeit temporarily. The multiplicity of identities so central to spirit possession thus comes to be reduced to a simple problem of contrasts between fixed constellations of human beings defined unequivocally as men and women.
Theoretical debates about possession have recently shifted to a more dynamic approach which focuses on the question of embodiment, putting the main emphasis on its experiential aspects. The emphasis here is on performance, on possession and trance as processes which emphasize identities in the making. Thus if possession cannot be remembered, let alone described, by those who experience it, the performative aspects of possession, and the significance of the experience itself, have come to the forefront of the analytical debate. As Deren (1953) reminded us in one of the earliest exegetic accounts on possession, she could not recall what had happened to her during a violent possession trance, but she certainly knew that it had taken place. The big gap in her recollection of events, the knowledge that her memory, as it usually recorded `reality', had omitted to register the particulars of her episode of possession and trance, were in themselves reminders of the event, signifiers of an out of body experience. The gap between memory, as recorder of reality, and knowldege of reality, through bodily experience, is here made all the more evident. Phenomenology addresses precisely the problem of negotiations of meanings, and is concerned primarily with the experience of actions, at physical and intellectual levels. The experience of ritualistic possession is thus explained in transcendental terms: spirit possession overrides social categories, bridges the worlds of living and dead, of gods and humans. The body transforms itself into an intermediary for and repository of knowledge, it becomes `good to think with', precisely because it is at the cross-road of experience and thought: humans can thus express ideas, think with, and through, their bodies. If a spirit becomes a person, as Lambek (1981) proposes, it is because the body is the meeting point between thought and action.
Boddy's work falls more or less within such a framework. The experiential, physical aspects of affliction, possession and infibulation among Hofriyati women are examined against a cognitive set of explanations: women describe their experiences of these phenomena in various terms, providing a complex, composite image of their relationships to their spirits, to men, and to each other. Thus although they acknowledge that infibulation does, at one level, conform to male ideals of femininity, it is simultaneously central to the constitution of their own identities as women. Women do not suffer an imposition of expectations, they are central agents in creating these identities in the first place. Thus if there is, in some contexts, a rapport de force between men and women, it is equally maintained on both sides. Spirits are only one manifestation of these relationships.
Possession provides the multi-dimensional language par excellence: it plays with notions of the self, transforms to transcend the worlds of the visible and concealed, and enables the possessed to bridge and merge the domains of humans and spirits. Superhuman agents, such as gods and spirits, and non-human agents, such as ancestors and ghosts, are brought together with, and within, their human hosts. Yet if spirit possession is talked about as the amalgamation of various identities, it is also linked to a process of interpretation of events. Phenomenology may be concerned with the embodied experiences of the possessed, and with the contextual stratification's in human identities, but these events and images have to rely on certain conventions of interpretation in order to be defined as possession in the first place. If identities, therefore, are constantly negotiated, and are defined depending on a variety of contexts, the same also applies to the interpretation and meanings given to such events. However, if phenomenology provides a useful toll for understanding the subjective experience of possession, it fails to address the question of how the subjective experience of the possessed, mediated through the body, is transformed into an intersubjective knoelwdge and interpretation. If physical bodies constitute the fundamental repositories for the occurrence of possession, how does this event provide meaning to more than one person at any one time?
Thus it is that possession has also been approached as a semantic puzzle, as a conundrum of knowledge, a manifestation of particular forms of experience, where knowledge and meaning are simultaneously enacted, displayed and created. Semantic interpretations overlap with phenomenology, insofar as bodily experience is central to both, and the dynamics of discourse are, in both cases, incorporated into analysis. But whereas phenomenology concentrates primarily on the notion of embodiment, on how the body mediates between various levels of consciousness and perceptions of human and spiritual worlds, semantic approaches view experience as a field for the creation and communication of specific forms of knowledge. The focus here is on the explication of meanings, the negotiations of paths from the wilderness to the ordered, from the unknown to the graspable. The dynamic processes leading to a state of possession are acknowledged, and the experience is certainly considered to involve several levels of consciousness or discourse, but the fundamental point, in my view, remains this: that spirit possession is a state negotiated, semantically, through speech and observation, between the possessed and the observer (Irvine 1982). Knowledge of possession is generated through the experience itself, and is constantly on the move. The presence of witnesses is therefore crucial to the mechanism of what we call possession, irrespective of cultural context, as the attribution of supernatural agency to human actions is negotiated between actors and spectators. As Irvine expresses it: "...interpretation is a creative process ... involving active collusion among participants" (ibid. 1982:257). In this context, the definition of spirit possession is not fixed. Rather, it can be seen as one mechanism among others implicit in the formation of knowledge. Possession may take place within one body, but bodies represent only one way ofcommunicating meaning
***
Watchi gods are called Vodhun . The polytheistic nature of this religion has already attracted much attention (Maupoil 1943, Verger 1957, Augé 1988). Gods are characterised by their distinct identities, partly defined by the identity of their guardians, and by their propensity to inflict illness upon humans when dissatisfied with the level of attention which they receive. Specific Vodhun are linked with particular illnesses: Sakpata, for instance, is the patron of smallpox or other similar diseases, Toxosu controls encephalitis, Down's Syndrome, Albino births and other `anomalous' categories, Hevieso is linked with severe swelling of the limbs and certain forms of paralysis, Mami Wata generates wealth, and can thus also be the cause of ruinous financial ventures if dissatisfied. Afflictions suffered by members of the community can therefore be attributed to cosmological agency. However, Vodhun are engaged in a dialectical, highly symmetrical, relationship with their human counterparts: each party has to comply with certain rules and fulfil specific obligations towards the other in order for each to be satisfied. Whereas humans have the power to forget their gods, these in return have the power to punish through illness, possession and, occasionally, death (on the negotiations of identities between humans and gods, see Barber 1981 and 1990).
The following scenario is described as a common manifestation of the presence of a Vodhun : a woman or young girl will suddenly enter into a violent trance, her eyes bulging, her limbs flailing around her body, her mouth contorted in a rictus of pain and surprise. Onlookers can readily identify this as the onslaught of an attack by spirits. However, during this initial trance, no-one will usually interfere, as the identity of the particular Vodhun involved is yet unknown. The woman will be left to experience her possession to the full, there will be no music, no chanting, nor any accompaniment by other women prone to being possessed. Some older women in the community may attend her, protecting her from being physically hurt during the trance. The Vodhun causing the episode of possession is believed to lead its future initiate to a shrine bearing its name. The woman or girl simply follows the path indicated to her by the spirit. Once she has arrived to the designated shrine, she collapses in exhaustion. Formal initiation into the secret society of the Vodhun which has just made itself known to her is de rigueur, and will usually begin a few days later. The uncontrolled, violent trance originally experienced will in future be replaced by a more purposeful and controlled trance, occurring primarily during ceremonial rituals in honour of the Vodhun . Any further uncontrolled trance experienced by the same devotee after initiation will be seen as an indication of a new demand being made by the god.
The original trance experienced by the possessed is seen as an indicator of the presence of a Vodhun , and failure to acknowledge this presence will lead to ever more powerful attacks of possession, ultimately resulting in death if no positive action is undertaken to formalise the relationship between Vodhun as spirit and devotee as recipient.. Initiation is a formal acknowledgement of the presence of the god, making initiates adhere to rules set down by the Vodhun . Thus possession irrevocably leads to initiation, but other forms of devotion are also possible. Devotees can worship their gods through other means (which I will not dwell upon here), but possession, once experienced, has to lead to institutionalised initiation.
However, if violent possession trance is considered by the Watchi as an indicator of a Vodhun's presence, it is only one among several other triggers for initiation. Prolonged illness, repeated experiences of misfortune, or the inheritance of a spirit, will also lead to initiation. The procedure for identifying the Vodhun is quite different in this context, as it will involve the use of divination. However, once a devotee identified in this way has become initiated into her god's secret society, she, too, will be prone to possession by spirits. Hence although initiation may not have been triggered by possession, it is in itself a port of entry to this experience. The cosmological link between devotee and deity is emphasised through the initiatory process. The following equation can sum up these relationships:
Violent trance/possession > initiation
Initiation > possession
Vodhun are seen as gendered entities: they are referred to as being either male, female, or both. However, the latter definition is highly situational: Some are acknowledged to have dual gender identities, these properties can come to operate at different times. They are often described as being 'sometimes male, sometimes female' or, alternately, as being `male and female at the same time' (definitions of androgenous: physical attributes/appearnce being neither male nor female; hermaphrodite: referring to dual procreative properties). In prayers and sacrifices, the male and female sides can be addressed jointly or independently (Vodhun Sotowo being one example). Most Vodhun , however, are readily identified as being single sexed, and are always accompanied, cosmologically and in their shrines, by a spouse of the opposite sex. Thus Hevieso, male god of thunder and lightning, and Agbi, female deity of blood, form a pair. Vodhun Sotowo is considered to have a male and a female side, while Sakpata, god of smallpox, is sometimes described as having dual sexual attribute, and sometimes considered to be only male.
Initiates are described as the spouses of Vodhun . The terms used, Vodhun si or Vodhun sro, are themselves both ungendered, and apply equally to male and female initiates, although it is clear that more than 95% of these are women. Sro is also used by men and women in non-religious contexts to refer to one's spouse. However, while Vodhun count both men and women among their devotees, only women have the propensity to become possessed.
Structurally speaking, all the elements of Lewis's thesis seem to be in place: women become possessed, men do not; the former are superior in number in the secret societies associated with Vodhun deities; and women are often referred to in `secular' contexts as being inferior to men in status and prestige. Superficially, we could be dealing with a peripheral cult, catering to the needs of the weak and the oppressed, allowing a voice to marginal beings through the interference of their spirits. The presence of a few men would serve to strengthen this argument: men who do not comply with the male ideals of the community use initiation (but obviously not possession in this case) to compensate for their weaknesses. Having thus joined a predominantly female world, they can escape the pressures demanded of their male counterparts.
There are obvious difficulties in applying such theories to Vodhun practices. Lewis's original intention to produce a model universalistic enough in its approach to be applicable to most occurrences of spirit possession runs into serious difficulties in this case. First of all, Vodhun as a religious complex constitutes the primary system of belief for all Watchi, and is therefore central to men and women alike. We also have to examine more closely the relationship between possession, initiation, and other processes leading to initiation. If men can be initiated into Vodhun secret societies, why are they not prone to becoming possessed? And if, as stated earlier, possession leads to initiation, and vice versa, why does this rule not apply to devotees of the male sex? A better understanding of the mechanisms leading to initiation is essential. We need to look more closely at the processes involved in the adherence to Vodhun cults, and the way in which these connections between gods and humans are validated.
The violent trances that originally indicate possession by a Vodhun , although seemingly random in the choice of a recipient for the spirit, appear to follow paths which informants describe as predictable and, with hindsight, easily identifiable. The Watchi will agree that violent possession should be avoided, as it is very dangerous for the possessed to find themselves in such a state. However, the presence of the god, and its impending demands during the original, unpredictable, and uncontrolled possession trances, could easily have been detected through the use of divination, for instance, long before the occurrence afflicting the future initiate. The reason stated is simple: initiation is seen strictly in terms of inheritance, and female initiates are said to inherit membership through a deceased maternal grand-mother, herself a former initiate in one of the secret societies. Upon the death of such a person, divination will be used to decide who will, ipso facto, succeed her as a new member. The new adept will often be in her infancy and, as it is regarded as relatively expensive, initiation will often be deferred. Although contrary to the ideal held by most people, this situation is nevertheless relatively common. Ultimately, if the demand of the Vodhun for a new disciple to replace the deceased has not been fulfilled, the situation will result in violent trance, leading the subject to the shrine of her deceased maternal relative.
If women are readily identified with this bond of inheritance, male initiates follow a different route. Not for them this inheritance through a maternal grand-mother! The link of matrifiliation is, in their case, much more direct: men are born into their mothers' shrine. Women who become pregnant while themselves undergoing initiation will immediately see their offspring, male and female, become initiates. In these instances, the process cannot be avoided or delayed, and the new devotee will undergo the same procedures as its mother. However, while girls born into the shrine will in future, experience possession, male offspring will not. Exegesis attributes this to the `anomalous' position of boys in the overall context of initiation. Women who become pregnant have broken a taboo prohibiting sexual relationships during initiation, but girls born in these circumstances would, in any case, have inherited their position through their grand-mother, and therefore been subject to possession, whereas boys cannot inherit this link. Although members of the secret societies, the position of men, and their ritual performances, are circumscribed by rules different from those applicable to women adherents.
I shall leave aside this aspect of the argument for the time being, in spite of the obvious interest it presents in understanding the complex structures underlying Vodhun secret societies as institutions, and concentrate attention on how the Watchi themselves consider and refer to spirit possession.
When a devotee is possessed, she is referred to as having been `taken' or `abducted' (tso) by the Vodhun . The same terminology is used when a future bride is abducted by the relatives of her future husband. There are obvious sexual connotations associated with both, although a wife will also refer to herself as having `taken' her husband, albeit without the involvement of any forced manoeuvres. An initiate does not refer to he Vodhun in these terms. She does not `take' or `abduct' her god. During the period of initiation, several rituals will be held in order for the Vodhun to `enter the head' of the devotees (ade Vodhun ta me). When possessed, the idiom is lightly different: `Vodhun lado lame nu', literally translated as `the Vodhun is inside the flesh' or `the Vodhun enters the body', and has overt sexual meanings. The same idiom is used to describe sexual intercourse between men and women. After initiation, women may be possessed, or entered, by the Vodhun without visibly going into trance. There is nevertheless an acknowledged loss of self to the deity.
Since initiation is inherited, possession by spirits is inherently more complex than originally described. Although it is indeed acknowledged by possessed and observers that it is the Vodhun that santches and possesses, the act further indicates and revives the matrifiliation which underlies the recruitment to membership of the shrines. During possession, women are thereby also instrumental in remembering links with the past, and are pivotal in invoking ancestors important to the community at large. Initiation, and by extension, spirit possession, thus reveal layers of sedimentation relating women to the past, but also extending into the future as they themselves, in the present, guarantee to provide future devotees through their offspring. Residues from the past are rendered evident through possession, and women's bodies become inscribed with, and in, history. Although their primary association is with their own female ancestresses, possessed women are said to become the mouthpieces of all ancestors attending rituals of possession held by humans. Indeed, when referring to their gods, the Watchi will often substitute the term Vodhun for the shorter denomination hun, which means blood. The relationship between a devotee and her deceased mother's mother is also described in terms of blood. A new disciple of the secret society will refer to her "having taken the cord of blood". The significance is polysemic: she indicates her adherence to her god's secret society, while simultaneously making explicit the maternal kinship which underpins this institution. "Taking the cord of blood" refers as much to her grandmother's blood as to that of her god.
This link with the past is further indicated in the role of initiates associated with a specific Vodhun called Tro. These disciples can, after further specialisation, become spirit mediums and diviners. It is then their exclusive prerogative to enter into direct contact with their Vodhun , but also with the founding ancestors of the community. These women speak in tongues, and use their skills primarily as healers and counsellors to members of the public. While the Vodhun is still referred to here as "entering the head" of its devotee, she in turn will refer to herself as "speaking the god" or "speaking Tro" (fwo Vodhun ; fwo Tro)
Let us now re-examine some of the theories presented earlier in this paper. The one concerned with the marginality of women and the peripherality of their religious practices will be omitted from my overview at this stage. Rather, I wish to concentrate on theories of embodiment, and on the semantic aspects of possession involved into he Watchi context.
The experience of possession by spirits has been referred to as involving a merging of identities (Boddy 1989, Deren 1953), or a displacement of identity (Lambek 1981). The distinction is important as, in the first case, the identities of person and god become one., the host acting as the spirit while, in the second case, spirit possession leads to the displacement of a person's identity, to be replaced during possession by the identity of the spirit. This would account, accordingly, for the amnesia described by most hosts prone to this kind of experience. While the spirit enters the body of the possessed, the identity of the host is displaced, while she herself "is absent from her body" (Lambek 1981:41). This strengthens Lambek's argument that host and spirit are indeed seen as separate entities, which have to negotiate a space (the host's body) as an arena for dialogue and communication. Spirits remind humans of their presence through possession, requiring sacrifices, libations, prayers, and devotees (possession then takes on the guise of a human sacrifice), while humans attempt to avoid these obligations in the first place, trying to maintain their distinctiveness from spirits. In possession, the interface between humans and gods is at its most obvious: both are forced to enter into a satisfactory alliance with one another. Spirits will be ensured continuity, human beings peace of mind.
The question of merging or displacement seems difficult to answer in light of my Watchi material. The two do not appear to be mutually exclusive. Women hosts of Vodhun during possession will admit to not remembering any specific details about their episodes of possession, which might indicate a displacement of the self in favour of that of the possessing spirit. One of the idioms used, that of "entering the head", is similar to the description Lambek provides. The spirit rises to the head of its host, `taking temporary control of all bodily and mental functions" (ibid:40). However, the Watchi also use the idioms of abduction and intercourse, which negate the notion of displacement, and indicate instead a cohabitation of spirit and human, a joint collaboration in intercourse, and the performance of a procreative act. The host becomes a vessel for the spirit, but the identities of both coexist inside her at this particular time. Although Vodhun are explicitly said to be unable to procreate, all their initiates are referred to as their children as well as their spouses. The turn of phrase is contextual. The terminology of affinity will be used by devotees and observers alike when referring to the relationship between deity and possessed (hence making possible and legitimate the idiom of intercourse), while most people will refer to themselves as the children of Vodhun when describing the protection provided them by their gods. In addition, devotees refer to the gods and their spirits as being "in the hand" of the possessed (Vodhun le asi nye), a term which designates ownership (a form of possession) in secular contexts. Hence initiates (and some cult leaders) 'have' deities. Another twist is added: female devotees are also addressed directly as 'Vodhun ' by members of the community at large, which does indicate a merging of identities, although no possession is involved in these cases.
In the Watchi case, exegetic idioms referring to the experience and description of spirit possession are varied, contextual, and also overlapping. Thus while a displacement of the self may be discerned at some junctions of possession episodes, the discourses which describe the event amalgamate several modes of explanation. Perhaps it would be appropriate to refer to levels of explanation, in the same way that we refer to levels of consciousness while experiencing the event. A further point needs consideration: a devotee of the Vodhun may become host to more than one spirit, and one god may decide to possess more than one of its disciples. However, these explanations appear singularly one dimensional when considered against the identities of Vodhun deities. By their own admission, the Watchi acknowledge the dual, and at times multiple, sexual identities of their gods. It could be argued that initiates, as recipients for the gods during possession, see their gendered identity displaced by the deity, which itself is no longer clearly gendered. The non-fixity of the affinal terminology, which makes no distinction between male and female as spouse to the god (or to a person of the opposite sex in a purely human context), points us in this direction. Thus if humans and gods become merged, gender also shifts, and male and female become composite parts of one another, an amalgamation of several bodies. This is reflected in another idiom used as a term of address by men and women during possession rituals: ame ame, 'human being', is the only form which prevails during this time. No references indicating the sex, or gender, of a person are made. And if there is a duality of gendered identities, there is also a symmetrical duality which applies to humans and gods themselves. Identities as male/female, human/spirit are re-articulated in the process of possession, thus negating any distinctions between them while simultaneously expressing their separateness as entities. This is clearly expressed in the reference the Watchi make to themselves and their gods in the context of possession: humans are sometimes themselves, sometimes their spirits, they are objects of desire (Vodhun take them, as in the case of sexual intercourse), and subjects of their Vodhun at the same time (they hold the gods in their hands). One becomes the other, and vice versa, but each party is also transformed in the process to become something quite different. Humans also speak their gods, providing more than a mouthpiece for them as the gods would have no means of expression otherwise. Thus if there is, at one level, a play on the dualities of humans and gods (as expressed through spirit possession), and on the dualities of gender, there is also a constant reassertion of the absence of those dualities. Identities are not fixed, they are generated and negotiated in the public fields of language and experience.
BIBLIOGRAPHY
Augé M. 1988, Le Dieu Objet. Editions Flammarion, Paris
Barber K. 1981, How Man Makes God in Africa: Yoruba Attitudes towards the Orisa. In Africa 51(3):724-745.
Barber K. 1990, Oriki, Women and the Proliferation and Merging of Orisa. In Africa 60(3):313-338.
Boddy J. 1989, Wombs and Alien Spirits: Women, Men and the Zar Cult in Northern Sudan. University of Wisconsin Press, Madison, USA.
Bourgignon E. (ed.), 1973 Religion, Altered States of Consciousness and Social Change. Ohio State University, Columbus, Ohio.
Constantinides 1985, Women Heal Women: Spirit Possession and Segregation in a Muslim Society. In Social Science and Medicine 21.
Crapanzano V. 1977, Introduction. In Crapanzano V. and Garrison V. (eds.), Case Studies in Spirit Possession. Wiley Inter-Science Publications, London.
Deren M. 1953, Divine Horsemen: The Living Gods of Haiti. Thames and Hudson, London
Gellner D. 1994
Giles L. L. 1987, Possession Cults on the Swahili Coast: a Re-examination of Theories of Marginality. In Africa 57(2):235-257.
Gussler J. 1973, Social Change, Ecology and Spirit Possession among South African Nguni. In Bourguignon E. (ed.), Religion, Altered States of Consciousness and Social Change. Ohio State University, Columbus, Ohio.
Irvine J. 1982, The Creation of Identity in Spirit Mediumship and Possession. In Parkin D. (ed.), Semantic Anthropology, ASA Monograph 22, Academic Press, London, Orlando and New York.
Lambek M. 1981, Human Spirits: A Cultural Account of Trance in Mayotte. Cambridge University Press.
Lewis I.M. 1971, Ecstatic Religion. Harmondsworth, Penguin Books.
Lewis I.M., El Safi A. and Hurreiz S., 1991 Women's Medicine: the Zar-Bori in Africa and Beyond. Edinburgh University Press for the International African Institute.
Maupoil B. 1943, La Géomancie à l'Ancienne Côte des Esclaves. Institut d'Ethnologie, Paris.
Rouget G.1985, Music and Trance: a Theory of the Relation between Music and Possession. University of Chicago Press.
Verger P. 1957, Notes sur le Culte des Orisa e Vodun à Bahia, la Baie de Tous les Saints au Brézil et à l'Ancienne Côte des Esclaves en Afrique. Institut Français d'Afrique Noire, Dakar, Sénégal.
Zempléni A 1977, From Symptom to Sacrifice: the Story of Khady Fall. In Crapanzano V. and Garrison V. (eds.), Case Studies in Spirit Possession. Wiley Inter-Science Publications, London.
A version of this paper was given at the AFRICAN STUDIES ASSOCIATION: Orlando 1995.