Exorcising angry deities and spirits of the dead: Spiritual and earthly battles of married women in Uttarakhand (India)

Abstract

Ethnographic analysis of the two main exorcistic practices in which married women are involved in Garhwal (India) shows that the theme of domestic conflict is central to both rituals. Addressing classical debates in anthropology about possession, this text raises two main questions: are these practices forms of feminine resistance to patriarchal social rules? And what is the notion of the person and her or his action in the world underlying these practices? Although these rituals may sometimes bring benefits to the women participating in them, women do not seem to perceive themselves nor to act as individuals who are resisting social structures, but more as part of collective networks of human and spiritual persons. As for the effects of these rituals, they are geared towards the preservation of family unity. This is achieved by the fact that, while evoking human conflict, these ritual devices subordinate it to the problem of divine conflict. Yet these practices do not only have an integrative function vis-à-vis conflicts that potentially endanger the family unit, but they also firmly establish the position of married women and their irreplaceability in the fabric of social life. Being close to the deities of their natal villages and easily affected by spirits of the dead, married women have an essential role in mediating with them and therefore in building and preserving the network of human and non-human beings on which the well-being of the family depends.

Introduction

The category of exorcism can be a valuable tool for the social sciences of religion for analysing a mode of relationship that humans develop with the spiritual world: an antagonistic relationship with a spiritual being, from whose presence a space (body, mind, place, collectivity of people) must be liberated. ¹

Defined in this way, this category allows us not only to compare the different forms taken by this relationship in different regions and communities worldwide but also to see how, even within the same social group and locality, this kind of relationship can take multiple forms. The analysis of the multiple exorcistic practices which exist within the same society – with regard to the social actors, the cultural themes, the occasions on which each of them must be performed – enriches our understanding, on the one hand, of such a society itself and, on the other, of the practice of exorcism and its sociological forms in the contemporary world. This is the approach adopted in this article which, among the multiple forms of exorcism existing in the north Indian state of Uttarakhand, will focus on the two main types of exorcism involving married women.

If one adopts such a definition of exorcism, it is indeed possible to say that the landscape of exorcism in this region of South Asia is extremely varied, being composed of innumerable practices. In nearly two decades of research in the state of Uttarakhand, I have not encountered a single household in which one or even several such practices had not been performed. We could observe that their relative complexity is an important way of differentiating between them. From that point of view, they form a continuum. At one extreme of such a continuum, there are ordinary and extremely simple gestures that do not require great preparations or ritual specialists. One example is a very frequent act which can also be performed by a relative of the affected person: grains or seeds are rotated on the head of a person in order to extract, often on the basis of a simple suspicion, a possible nefarious influence before throwing it elsewhere, along with the grain or seeds. At the other end of the continuum, we encounter rituals of dialogue, negotiation, and sometimes of actual fighting with a spiritual being, rituals that can last several days and nights, require the presence of many specialists, and must be performed numerous times before being declared effective.

In the more complex practices, female subjects far outnumber male ones as victims to be liberated. We could even say that a woman’s life is filled with exorcisms.

As I have shown elsewhere, adolescent girls, at an age between puberty and marriage, are now in some mountainous areas of the state the most frequent victims of complex forms of exorcism, called rakhvālī (literally ‘caretaking’), aimed at ridding them of the presence of spirits. Such practices are based on, and at the same time reproduce, specific physiological concepts concerning the bodies of women in general and adolescent girls in particular (Bindi, 2015).

At the time of marriage (or sometimes a few years later, when the couple try to have a child), in many places in the state, women are subjected to a ritual of liberation from a spiritual entity residing in their natal village (which is different from the village to which they move at marriage) or in the space separating the two villages. This extraction, which takes dissimilar names and forms in the different areas of the state where it is practised (chal pūjā; fareda), is accompanied by an animal sacrifice for the entity to be expelled and is considered indispensable in order for a woman to later conceive a child (Bindi, 2009; Polit, 2012: chapter 4; Sax, 1990).

However, marriage does not mark the end of a woman’s involvement in exorcisms. Rather, it is probably at this stage of her life, even more than throughout the rest of her existence, that she is involved in exorcistic practices of various kinds. The following pages will focus on two such practices, which are at present widespread in the region. One of them, involving deities, has been the subject of excellent ethnographic analyses in the Chamoli district, neighbouring the one where I conducted my research, but these have so far focused mainly on the Dalit caste populations (Polit, 2012, chapter 4; Sax, 2009), while my ethnography is mainly based on accounts and observations collected from the two highest caste groups, Rajput and Brahman; the other, involving spirits of the dead, has not so far, to my knowledge, been the subject of any in-depth anthropological investigation. ²

A joint analysis of these two practices has never been undertaken, despite the fact that they are the two main exorcistic practices in which married women are involved. Such an analysis is fruitful first and foremost because it shows how the practices share an important feature. What is central to these two forms of exorcism is not only the conflict between humans and spiritual beings, as is the case, by definition, in all exorcism rites. It is also the conflict between humans and, more specifically, the conflict of the married woman with other members of the marital family, and the suffering that results from it. Human conflict is not only present in indirect ways or in the interpretation given by the anthropologist (cf. later in the text). The discourse on domestic conflict is an integral part of the ritual device itself: it is one of the central themes that the ritual actors (oracles, possessed mediums, and other ritual specialists and musicians) talk about and discuss before ritual performances and during the rituals themselves.

Focusing on how these exorcisms, each in their own way, place at their centre the theme of domestic conflict in which married women are involved, this analysis contributes to one of the classic debates in anthropology of possession and related rites of exorcism: that which concerns the interpretation to be given to the heavy involvement of women in possession practices, which is a widespread phenomenon worldwide. In this regard, some of the questions that this text raises are: can we see these practices as forms of resistance for married women, as many anthropological analyses do? What is the notion of the self and its action in the world that underlies these practices and is reproduced by them? How do these exorcistic rites and representations shape domestic conflict and the way conflict is experienced by people in this region?

The data that provide the basis for this analysis were collected during several ethnographic surveys, carried out between 2004 and 2020 in the Uttarkashi district of the Garhwal region. Garhwal is part of the north Indian state of Uttarakhand, bordering China and geographically located in what is referred to as the Central Himalayan region. Investigations were first conducted as part of a doctoral thesis (Bindi, 2009) on oracular possession practices and, subsequently, as part of three research programmes devoted to the types of presence of invisible entities (ANR Presencedesprits, 2013–2018) and the coexistence of secular and religious practices of coping with the experience of mourning in the Himalayan region (IdEx Ghosting, 2019–2022 and ANR Phantasies, 2020–2024). Such research, carried out both individually and in the company of local assistants, necessitated long stays in the villages, which allowed me to build an intimate relationship with one or two families per village, and then, through these families, to be accepted by the entire village community and thus be able to observe the daily life of these collectivities. Caste is here, as elsewhere in India, a very present social unit. While there exists a rather high percentage of Dalits or scheduled castes, most of the population belong to the two highest castes, Rajput and Brahman. It is with members of these two groups that I conducted the most extensive research.

Marriage, residence, and conflict

The majority of the population of the Garhwal region live in the mountainous areas, whose economy consists mainly of agricultural production, cash remittances from relatives who have migrated to the city to work in various sectors, and, to a lesser extent, livestock farming. Although employment in other jobs, either in the public or private sector, remains limited, activities related to religious tourism and trekking are steadily increasing in some areas.

Although an increasing number of families today tend to abandon it, the residential model that still dominates in rural areas prescribes that on marriage a woman moves into the house where not only her husband but a whole group of blood relatives live with their respective spouses. This housing pattern is an important factor which, in addition to consolidating the bond between family members, contributes to producing conflicting relationships and suffering.

Women who, by marriage, become part of their husband’s lineage and co-resident group are among those most exposed to such suffering in Garhwal. It is impossible to count the number of times I was present during conversations concerning a woman’s difficult relationship with members of the marital family. The most recent of them dates back to the days in which I wrote this text. A friend, a migrant from rural areas and now living in the state capital Dehradun, explained to me how some of his friends, now in their 30s, like many couples in urban areas today, had chosen to cohabit without getting married, mostly in order ‘to spare their girlfriend the suffering implied by the relationship with the marital family and especially with the mother-in-law’.

Far from being limited exclusively to rural and relatively isolated regions such as the one in this study, the difficulties associated with the condition of the married woman seem to be a topic of concern throughout the subcontinent. This is testified also by the many articles in the media reporting cases of young women whose suffering in the marital family has led to suicide, for example, ‘What’s Behind Suicides by Thousands of Indian Housewives’ (Pandey, 2021) published on the BBC News website, Delhi, in December 2021. An awareness of this problem can also be seen in the Indian Penal Code, which contains several laws punishing domestic violence against women, including forms of psychological violence (Vashi, 2020). One of them, called ‘Of cruelty by husband or relatives of husband’ (498A., Indian Penal Code), is specifically devoted to punishing the violence suffered by a woman at the hands of her husband or her husband’s relatives.

Multiple anthropological analyses have shown how such violence takes shape within a set of highly unequal power relations that characterise the Indian family model in much of rural India (Polit, 2012) and Nepal (see, among others, Bennett, 1983). As Polit (2012) shows in her monograph devoted to the lives of Dalit women in Garhwal at various stages of their life, married women are heavily subject to the authority not only of men but also of other women who have been integrated into the family before them: the wives of their husband’s older siblings (called jeṭhānī) and, above all, the mother-in-law (sās). The latter exerts significant control over her sons’ brides, particularly those recently added to the family, whose position is particularly fragile since, before the birth of a healthy male child, they are not considered complete social persons (Polit, 2016: 59). She decides how much they should sleep, when and what they should eat, for how many hours they should work, and the relationship they should have with their husband.

What makes the condition of married women in Garhwal even more difficult is village exogamy, which applies to this area as to most of rural India and prescribes marrying outside one’s village. This causes the woman to leave her native village and family (mait) after marriage to move to her husband’s village and home (sauryas). Married women, living therefore outside their natal village, are called dhyani by the members of their natal family and village, a term often translated as ‘outmarried daughter’ (Sax, 1990, 1991). However, in Garhwal, a dhyani’s emotional and material relationship with her mait remains particularly strong. As we shall now see, it is widely believed that it is this strong link that married women maintain with their natal lands and its human and non-human inhabitants that enables them to ‘spiritually attack’ their enemies in their husband’s house.

The attack of the angry deities

In the first complex of exorcistic acts and representations that we will examine here, married women are not only considered victims to be liberated through exorcistic processes but also aggressors, or as bearing at least some responsibility for the spiritual aggression that is at the root of the problem. This complex is articulated around the practice called ghāt. The word ghāt, which literally means ‘attack’ and is translated by English-speaking informants as ‘curse’, refers to a kind of witchcraft attack that consists of enlisting the complicity of certain local deities to unleash them against someone in order to cause them harm and misfortune.

Only a few deities are believed to be willing to lend themselves to this kind of activity, which, as we shall see, is far from unambiguous. In the area where I carried out my research, the deities considered to be ‘prone to the attack’ (ghāt vālā devtā) are all male (and therefore I will refer to them in this text with the pronouns he or his). The most famous ones are Agornath, Bhairav, Eda, Pokku, Goril, Narsingh, and Caciuru. Their worship is extremely widespread, as the practice of attack leads them, as we shall see, to continually implant themselves in new families and villages. In many families they are part of the domestic pantheon, and in some villages, they have a temple built for them within the village territory and the entire community celebrates their cult.

As readily emerges in informal discussions with the inhabitants of the region, regardless of castes and genders, a particularly strong relationship of mutual fondness exists between these deities and married women coming from the families or villages where the deities reside. In the area where I conducted my research, women often speak of such deities as members of their own family or village. ³ Sax (2009) points out that, in the course of oracular consultations, oracles sometimes refer to such deities with the expression ‘the beloved of married women’. The normative ideology, conveyed by Sanskrit ritual treatises (Inden and Nicholas, 1977; Sax, 1991), rules out the possibility of married women continuing to worship the deities of their natal lineage. Yet, in Garhwal, people assume that they bring the deities of their natal village and family with them into the marital family, particularly when these deities are among those prone to the attack. ⁴ Or rather, it is the deities who follow married women into the husband’s home to protect them from the marital family. According to some people, they move materially with them, residing in their hair or their favourite piece of jewellery, ready to intervene in case of need.

It is widely believed that such deities are particularly sensitive to the tears of their ‘outmarried daughters’ and cannot resist their weeping. It is enough for a married woman to cry with despair for the deity to act against her enemies in the marital family. Although tears are considered sufficient, there is also an expression that, uttered aloud or in one’s heart, unfailingly solicits such a deity’s help: ‘I have no one!’ (mera koi nahim hai) (see also Sax, 2009). Some friends and informants have told me that, after having uttered it mentally in situations of great fear, they literally heard the phrase ‘Here I am’.

Although the married woman is not the only one who can perform it, ⁵ this way of cursing that invokes the love of her deity in a moment of strong emotion is considered a mode of occult aggression closely associated with the position of the married woman residing in her husband’s household.

While this type of attack is extremely easy for women to perform, it is believed to produce widespread effects: a spiral of suffering and pain that is very difficult to stop. First of all, a deity set against someone will attack not only them but their entire family, including lands and animals and, often, the generations to come. Moreover, such a deity, once unleashed, will act independently of the will of the person who invoked them. To stop this, deity’s action is considered to be a difficult task. Many respondents have affirmed that this is due to the particularly passionate nature of such deities who act in a state of strong emotion: they feel love for their daughters and thus wish to protect them; consequently, they feel anger in the face of any injustice their daughters may suffer. According to some informants, these gods are also particularly gluttonous, and the animal sacrifices obtained for these practices keep fuelling day by day their craving for meat and thus their aggressive actions.

What emerges from these different interpretations is that these deities are seen as autonomous and extremely dangerous social agents. In my years of travelling through Garhwal, I collected numerous stories, told both by the victims themselves and by others, of how such a deity continued to harm the entire family for generation after generation (even if the perpetrator of the attack had repented, forgotten, or died) until the deity managed to kill the entire extended family. This is the case with Prahlad, whom I met in 2015. Left alone after all his relatives had died of different kinds of illnesses and accidents, he had recently lost his young wife as well. In our numerous interviews, Prahlad attributed his plight to an attack launched several generations earlier against the family by one of his own aunts by marriage, who was now long since dead.

Oracular verdicts and ritual remedies

The set of these representations is largely transmitted and reproduced through the consultations of local oracles. By this term I refer to specialists of what can be referred to, employing an anthropological category, as oracular mediumship or oracular possession. They can be both men and women, although women can only access some specific forms of this kind of activity. They are seen as having been selected by some deity, through various processes, for becoming her or his ‘medium’ or ‘possessed’. On specific occasions, it is considered to be such a deity who, having taken control over them, expresses herself or himself though them, either through their words or their body movements. This extremely widespread form of divination provides answers and solutions for individual and collective problems (for a description of the many forms that oracular possession can take in the area, see Bindi, 2009, 2016, 2022). The analysis of numerous oracular sessions I was able to record between 2005 and 2020 shows that the consultants’ problems can be attributed to different types of causes and that the same problem can often be attributed to several causes simultaneously. It also shows that, to employ a now-classic distinction (Forster and Anderson, 1978), one or more causes of a ‘personal’ type – that is, attributing the responsibility for the problem to a human or divine person – can coexist with impersonal causal concepts – those that attribute the problem to a mechanical cause (e.g. a physical disease). ⁶ In such a context, the idea that the problem was caused by the occult violence exerted by another person can never be ruled out. If one looks at oracular interactions and divinatory verdicts, it constitutes a cause (or one among the causes) of most of the undesirable events that touch the lives of the inhabitants of the region.

The oracles always begin their consultations by asking whether there is any conflict within the family. And, in a great many cases, the unhappiness of a married woman is the first cause hypothesised by the oracles for the suffering of their client and her or his family. Often, the oracle alludes to this indirectly, speaking of a ‘deity who came with a dhyani’. In very many cases, the family consulting the oracle is thus asked during an oracular session to reflect on any reasons that might make the women of the family unhappy and have given them cause to call upon their deity to attack other members of the marital family.

When family members come to accept that the cause (or one of the causes) of their sorrows is a ‘deity who came with a dhyani’, they have several options to solve the problem. The first solution proposed by the oracle is one based on what is considered the most important moral principle of the local society: family unity (cf. on this point Sax’s detailed analysis among the Dalit castes devoted to the worship of one such deity, called Bhairav, Sax, 2009). Sometimes it is sufficient that those who, according to the oracles, are involved in the conflict go together to the nearest temple to offer a joint celebration to the deity concerned. However, in many cases, the oracle indicates that the deity demands to be included in the family pantheon and to receive regular worship. To use anthropological categories, exorcism, or deliverance from the nefarious action of the deity, here requires an ‘adorcism’. This now-classical anthropological category was initially proposed by De Heusch (1971) to refer to practices aimed at appeasing or welcoming spiritual entities into a person or place (as opposed to exorcism, which aims to expel the spirit) and building a potentially positive relationship with them.

The inclusion of ‘a deity who came with a dhyani’ consists of a ritual that both requires and produces family unity. Its execution requires that the family members make peace and that they cooperate materially in the performance of the ceremony, which is quite an elaborate one. After having installed the deity in his new temple, the family will worship him recurrently and organise regular sessions (called jāgar) in which the deity will dance and speak in the body of one or several people they have selected as their mediums among family members (this can be the woman who has called the deity or another person). Animal sacrifices are considered essential for deities during such ceremonies. As a result of such worship, the deity will give harmony, peace, and prosperity to the family. ⁷

After consulting an oracle, some families agree to begin worshipping a deity who came with a dhyani. Therefore, the worship of such deities keeps expanding within the villages of the Uttarkashi district, where I conducted my research (as also happens in other districts, see Sax, 2009). A paradigmatic case that ends with such an inclusion, albeit two generations later, is described by Sax (2003: 179):

A low-caste woman named Shanti married and had three children. Some years later her husband died, and after that his first wife (Shanti’s elder co-wife) died, leaving Shanti to care for her two step-children as well as her own three children. One day her deceased husband’s younger brother Manori accused her of neglecting her step-children. She argued with him, then returned to her natal village, went to Bhairava’s thān, and cursed him for making a false accusation. Fourteen years later Manori’s granddaughter Seema, then about nine years old, began to suffer stomach pains. Medical doctors were unable to provide a diagnosis, and the pains continued. Seema’s father Chandravir consulted several oracles and determined that the curse from fourteen years ago had finally taken effect. Manori Lal and his son Chandravir established a shrine for Bhairava in April 1997. I was present for the entire ritual, during which the god was summoned, the ghost of Shanti (who had long since died) was exorcised, and Seema’s pains ceased.

However, other families refuse the inclusion of such deities. This refusal may be due to the fact that the woman (or more rarely, the man) who, according to the oracle, is involved in the conflict does not admit to having carried out the attack and does not agree to reconcile with her or his enemies. According to my interlocutors, this refusal may also happen out of fear, since people feel that, once they have started to worship one of these deities, any forgetfulness towards them may be lethal, as the latter tend to be extremely touchy and passionate. Others still might fear that their arrival in the family may offend any deities already residing in the house and attract their punishment.

When no agreement is found and the deity is neither worshipped nor included in the family pantheon, another possible solution for the person or people who consider themselves victims is to perform a celebration that is meant to appease (shānt karnā) the deity and induce him to leave the family alone. Although this is rarely admitted, most of the ritual practitioners whom I interviewed revealed to me how such a ritual of appeasement and liberation is almost always accompanied by a counterattack. To put it another way, the aim of such a practice is, simultaneously, to calm a deity’s anger towards one’s own family and to turn his attack back on the person who performed the curse, through a specific rite called pet denā ⁸ in the Uttarkashi region.

However, the basic idea behind all of these different options is that, in all cases where the two parties fail to come to an agreement, the wrath of the deity will sooner or later fall on the woman who called upon him (while also continuing to afflict her marital family). If, after invoking the help of a deity, a woman should fail to import his worship into the marital family, she will have the impression that she suffers his wrath, sometimes for her entire life. Her own illnesses and the illnesses of her loved ones, economic difficulties, and quarrels within the family will be interpreted by her and others as symptoms of this deity’s vengeance.

Still others, such as the Rajput family of my friend Vinoj where I stayed for several months (May–June 2018; January–February 2020), agreed to include one of these gods in the family but, at the time the deity possessed their medium, ⁹ made him swear that ‘he will not go with the dhyanis’. This means that the god will not indulge in the practice of attack in the future, even in exchange for animal sacrifices. It is believed that, by not taking sacrifices and not attacking, such deities will gradually and eventually become less greedy and aggressive.

Ultimately, a deep ambiguity surrounds the performance of a ghāt. Considered to be a natural and almost inevitable action for a woman, it is simultaneously seen as a fundamentally amoral practice because it threatens the central moral value of family unity, and is a dangerous act for its perpetrator.

Stories of intra-family curses are common and often end tragically, with the person who first cursed (or one of their loved ones) being punished by the deity. ¹⁰ This is the case with Saraswati, who has now returned to live with her natal family. She confessed to me in January 2020 that, a few years after her marriage, she had asked her own deity Goril to harm her mother-in-law. As a result, the latter had become seriously ill. However, the marital family never accepted this version of the story. They instead attributed the illness of Saraswati’s mother-in-law to a deity – a goddess – already residing in the family. According to this interpretation, it is because this goddess felt that the family had neglected her that she was punishing the mother-in-law. As a result, they never agreed to worship Saraswati’s Goril. It was because of this, she explained, that Goril gave her infertility. Her failure to conceive a child in the 6 years after her marriage caused the family to repudiate her and her husband to marry another woman.

Affliction by the spirits of the dead

A second set of exorcistic/adorcistic practices in which married women are the main protagonists are the rituals called ghaḍiyālā. This ritual practice must be understood within the framework of concepts concerning victims of sudden, untimely, and violent deaths. The spirits of these people (called ćhāyā, pāpī, ūprī ātma, alpāyu, or ādhogati) cannot make a normal transition to ancestral status (pitr devtā). They roam the earth homeless, try to attract the attention of the living, and express their ‘hunger’ for food (ritual offerings) and affection by causing some people unpleasant physical sensations and perceptual disturbances (pain, discomfort, perceptions of voices or visions).

It is through oracular sessions, such as those mentioned above, that these sensations and perceptions are interpreted as being due to the action of these souls, and that people learn what ritual action must be undertaken to calm the spirit. Such ritual actions are partially different for each category of untimely death. The spirits of women who died after marriage, particularly those with children, constitute a specific category. To appease them, it is necessary to organise a ghaḍiyālā ritual, which will enable the deceased to converse with her relatives and to negotiate her entry into the domestic pantheon. In order to do this, she will be incarnated in a person whom she has selected to become her medium during the ceremony (then called pashuā ou otariya). In very many cases, the person selected will be the new wife of the dead woman’s widower.

The typical situation that gives rise to a ghaḍiyālā ritual is well illustrated by the case of Pradeep, whom I met in May 2018. After the untimely death of his first wife Sangeena, Pradeep entered into a second marriage with Manisha. During the first 2 or 3 years of their union, Manisha often suffered from different kinds of pain in several parts of her body: stomach, head, joints. After some time, the family members took her to an oracle who resided in a nearby village. The oracle ascribed Manisha’s suffering to the spirit of her husband’s first wife, Sangeena. Through such symptoms, this spirit was asking for a ghaḍiyālā in order to incarnate herself in Manisha’s body and start a conversation with the family about her own possible inclusion in the family pantheon.

The organisation of the rite requires the presence of a specialist, the ‘guru of the ghaḍiyālā’, who is usually a man belonging to one of the specific Dalit groups who perform such practices and, within this group, to a family where this is a hereditary activity. It is therefore usually from his father, his grandfather, or from some other elder of his family that he has learnt how to play the drum called hudaka and the specific rhythms and chants which are needed for the ceremony. The ghaḍiyālā takes place only in specific months which are considered particularly suitable for contact with the dead. During the course of the ritual, which lasts several days and nights, it is very often a man’s current wife who is possessed by his first dead wife.

A salient feature of this rite is the fact that it establishes a strong parallel between two conditions and two figures: the married woman and the dead woman. From the beginning, the offerings that are made to the spirit of the deceased woman are symbols of the married state, such as glass bracelets and vermilion powder. The parallelism between the two conditions is also at the heart of the song that the ritual specialist, after the ritual offerings to the local and lineage spiritual entities and deities, begins to chant while playing his ritual drum (sometimes accompanied by the sound of a metal cymbal played by an assistant). This song is considered to be an invitation for the spirit of a deceased person to be incarnated in a medium’s body. In very many cases, it is performed to invite the spirit of the deceased wife to be incarnated in the second wife’s body. There are many different songs, and the same song may vary significantly in different valleys and localities, both because the type of Garhwali language spoken in each valley is different and because ritual specialists make personal adaptations and innovations. However, most of the chants I have collected from a wide area within the district focus on the feelings of longing experienced by both married women and dead women. For their part, dead women long for life and for the love of their loved ones; for theirs, married women living with their husbands’ families long to be with their natal families. As an example, here are some of the lyrics of one of these songs.

Basant’s season has arrived, you can tell by the choi whistle. ¹¹

Basant season is the most beautiful of all. It is the season of burans ¹² blooming on the mountains. O sister, even the pigudi ¹³ now this is how they bloom!

In seeing the pigudi bloom, the heart is sad, and tears come to the eyes.

Panchamī day is approaching and with it the Magh festival, when the girls return to visit their birth homes and celebrate their joy.

It is the time when girls who have no one left at the birthplace (mait) cry, hidden in a quiet corner of the forest.

In the month of Magh the invitation to return to their native homes is delivered to the girls, and they will then come to their parents’ homes.

Then, you too will have itchy feet and wait for the invitation to come from your birthplace. If you could come, then we could give you your due share. But you will no longer be able to come.

Why was your life destroyed? Who is the impostor who killed you in your full youth?

And now that your life is gone, your body and blood have also dissolved. Who is the impostor who destroyed your life?

You were not able to live out your youth while all your friends stayed here alive.

O sister, the gaze of these Matri is evil, and the groups of the Rath Devtā, all are evil. Who has taken away your life and youth? (Sadhu Ram, Purola, recording on 3 March 2007)

The text of the song refers to the frequent visits of married women residing in the marital home (sauryas) to the natal home and village (mait), a practice that, as I have shown elsewhere (Bindi, 2012), constitutes a central cultural motif in Garhwal. ¹⁴ The most emblematic of all the occasions on which married women visit their natal village is ‘the festival of Magh’, which is also referred to in the song. On this occasion, which is celebrated during the month of Magh (January–February), a male member of their natal family goes to invite married women and accompanies them to their natal village, where they spend a few days.

Establishing a parallelism between the married woman and the dead woman, the song describes the longing that a woman feels for her natal family, the desire to visit her loved ones, and the joy of meeting people from her birthplace. The expression employed in the song, that is, to have ‘itchy feet’, is used in the area to indicate someone’s strong desire to urgently go elsewhere. The song then describes the unhappiness of those women who cannot go to their birthplace: those who have no one left in their home village, and those who have died. Neither can experience the joy of return. The song then denounces the injustice of these two situations. In the conclusion, it particularly denounces the violence with which death takes away the one who died prematurely. With the expression ‘the gaze of these Matri is evil, and the groups of Rath Devtā all are evil’, the accusation is particularly directed towards two categories of spiritual entities, one female (Matri) and the other male (Rath Devtā). They are often considered responsible for premature deaths because, moved by their attraction to a certain person (or sometimes by other emotions), they instigate her or his death in order to keep this person close to them.

However, it is in the next stage of the ritual that the parallelism between the two conditions of married woman and deceased woman, and the reflection on the unjust suffering they both have to endure, reaches a climax. It is indeed at this stage that, with the beginning of possession, a married woman identifies with a dead woman, who is, in most cases, the deceased first wife of her own husband.

Accusations and apologies

To illustrate the next part of the ritual, which involves possession and dialogue with the spirit of the dead woman, we will take the example of a ritual I observed and filmed in December 2015. The rite was organised for Puspati, the deceased bride of Adit, who was then remarried to Sumita.

Towards the end of the chanting, Sumita began to move in the manner which, as I discovered after attending multiple rituals of this kind, is characteristic of the spirits of dead women: she fell to her knees, turned her head upside down, and let her hair, now loose, swing before the ritual specialist. The woman began to cry and scream, manifesting great distress. Several people in the audience – consisting of Puspati’s siblings (Puspati’s parents were not present as they are dead), all members of the marital family, and many inhabitants of the marital village, including old men and children – began to weep. Amid the sobs, Sumita began to scream, addressing herself at first to the ritual specialist, about her own loneliness. She recounted the circumstances of her own death, which occurred as a result of an accident in the forest during one of the many activities that make up the daily lives of women in the region. She then approached her own brother, to hug him. She tenderly called Puspati’s son, born before her death, who was about 4 years old at the time of the ritual. The father accompanied the child, trembling but not crying, into his mother’s arms, where he remained for a few minutes. Finally, still in tears, Sumita turned to her husband, her husband’s brother, and her mother-in-law, crying out to them her sadness, loneliness, and hunger. A few minutes later, she began to accuse them of the suffering she had endured in their home, when she was alive. Addressing her by Puspati’s name, they scrambled to make apologies and reassurances of the love they felt for her. In the meanwhile, Sumita (again speaking for Puspati) continued sobbing and screaming, accusing her mother-in-law of having made her life unbearable. She mentioned the inhumane work she was subjected to and the food deprivation. Turning then to her husband, she accused him of being absent and abusive. The members of the marital family continued to make apologies to Puspati, while those who attended the rituals continued watching and, in some cases, crying.

The ritual continues with numerous other stages that include, among other things, the interventions of other deities, who possess their respective mediums, and start talking about important matters concerning the family; the dialogue and often the dance and hugs of the dead woman’s spirit with other spirits of dead who possess their respective mediums; the dialogue of the ritual specialist and family members with the dead woman’s spirit to agree (when possible) regarding her requests in terms of offerings and ceremonies to come; the eventual acceptance of offerings by the dead woman’s spirit. The ritual is in fact conceived as a kind of negotiation between the family organising the ritual and the spirit. The dialogue with the living is supposed to ‘reason with’ the spirit, telling her that her suffering is understood but helping her detach herself from such emotion; reassuring her of the love that the members of the marital and natal family feel for her, persuading her to accept the offerings prepared for her and thus to become integrated in the domestic pantheon, within the category of deified female spirits, relatively benevolent towards the family, called Hantya Devi. Depending on what has been agreed upon with the spirit, who makes specific requests during the ritual, the new Hantya Devi will be installed in a small silver or wooden icon and will usually be kept in one of these places: in the family’s altar at home; in another part of the house; in a small altar located outside the village; or, more rarely, in a necklace which will be worn by her chosen medium. After this first ritual, a new ceremony will be organised every 3–5 years. These periodic ceremonies continue to take place, in some cases for many years, whenever the spirit demands it, through new suffering caused to the person that the spirit has selected as her medium or through other misfortunes affecting the family, which will be interpreted as the spirit’s request for a new ceremony. During each ceremony, the chosen medium (as we have seen, most often a married woman) will be possessed again by the spirit (who is often the dead first wife of her husband) and will continue to converse with those present, especially members of the marital family.

Although in ways that are different from those seen in the case of the attack by the protective deities of the ‘outmarried daughters’, the married woman’s conflict with her marital family and the suffering that results from it constitute, if not the central theme, one of the dominant motifs of the ritual we have just described. Introduced by a chant focusing on the feelings of longing shared by married women and dead ones, the expression of the suffering that characterises the woman’s married life then continues into the ritual phase of possession. This is the only occasion in years of research in the region where I have observed such an open dialogue on conflictual domestic relations between a woman and her marital family. I therefore found it extremely surprising the first time I was able to observe it. Nevertheless, subsequent participation in 16 such ceremonies allowed me to see that this is an unfailing part of this ritual.

Instrumental readings of female possession

This strong involvement of women in possession practices is not specific to the Garhwal region. The socio-anthropological literature has shown how, in many regions of the world, women are protagonists of forms of exorcism and adorcism far more often than men (cf. the reviews of this literature proposed, respectively, by Boddy, 1994; Bourguignon, 2004; Kirmayer and Young, 1998; Plancke, 2012; Ram, 2013). A considerable part of this analysis deals explicitly with the issue which we have observed in Garhwal: the frequency of possession of married women and how problems and tensions concerning marital relations lie directly or indirectly at the heart of these possessions. This has been shown to be a common phenomenon both in South Asian contexts (Freed and Freed, 1964; Harper, 1963; Kapferer, 1991; Nabokov, 1997, 2000; Ram, 2013: 73–105) and in other contexts (Bilu, 2003, regarding possession in Judaism; Crapanzano and Garrison, 1977, considering a variety of African, Asian, South and Central American contexts; Lambek, 1980, in Mayotte; Lewis, 1969, in Somalia; Harris, 1957, in Kenya).

One of the anthropological interpretations of female possession in general that has found the greatest acceptance is the one that sees it as a woman’s reaction to the subordinate position she occupies within most societies, one that after marriage will become particularly harsh and unbearable.

It would be impossible to enumerate all the different forms which this interpretation has taken, divergences which might reflect local variations as well as differences in theoretical positions. Nevertheless, we can see, at the risk of over-simplification, that they have tended to embrace two main perspectives. One perspective, which has some continuities with the clinical views of possession (cf. on this point Kirmayer and Young, 1998), emphasises the highly unconscious nature of possession as a form of somatisation, an idiom of distress, or an unconscious body language which expresses women’s deep suffering or their resistance (or often suffering and resistance at the same time) to the numerous forms of oppression imposed on them by societal norms, which are controlled by men. One of the main supporters of this position is Bourguignon (2004) who, basing her interpretation both on her own fieldwork in Haiti and on examples from anthropological literature all over the world, has argued that, ‘for women, possession trance constitutes a psychodynamic response to powerlessness by providing a means for the gratification of wishes ordinarily denied to them’ (p. 557). Socialised to accept their subordinate status, women can express their wishes only indirectly by reproducing ‘their subordination ritually, dramatically, and symbolically in possession trance, which they experience as the takeover by an external, ego-alien, powerful entity’ (Bourguignon, 2004: 570–571). Possession is, in this view, an unconscious process, in which women can express their wishes because they attribute them to an external entity with whom they identify, so that they can ‘deny to themselves as well as to others that they are powerful’ (Bourguignon, 2004: 570). This allows them to gain power, either temporarily or sometimes afterwards as well, if, as mediums of a spiritual entity, they manage to become diviners, for instance, or religious leaders of a cult.

A different perspective, destined to become a classical one, was initially proposed by Lewis (1969, 1971). On the basis of some detailed fieldwork carried out in Somalia (Lewis, 1969) as well as compiling examples from numerous other societies, Lewis (1971) proposed a cross-cultural account of possession cults which, inspired by a structural perspective, analysed various types of possession existing in different societies or within the same one, as well as their relations to the respective power-positions of those who are possessed. This perspective leads him to distinguish central cults and peripheral cults. The first, which are linked with a positive experience of spirit possession and to spiritual entities which uphold the society’s moral order, are typically controlled by men. The second, which are associated with the involvement of amoral spirits, lead to illness and require healing, are, according to Lewis, linked to women or other people of marginal or subordinate status. This approach, which has had great success and has been applied by researchers to multiple contexts, has tended to see in the possession of female (or other supposed marginal) subjects a strategy, often conscious, through which they express themselves, rebel, and sometimes obtain benefits (for a recent discussion of this approach and its applicability in Asian contexts, cf. Bindi, 2015; Riboli, 2021; Seale-Feldman, 2022; Torri, 2021). For example, in the Somali ethnographic context where Lewis carried out his fieldwork, married women used to be possessed by zar spirits especially in moments of tension, typically when their husband was about to take a second wife, and this strategy allowed them to stop him or obtain some other kind of compensation.

Far from being present only in the academic discourse, these are the two main explanations for women’s possession which nowadays prevail in South Asian societies. As Ram (2013: 78) has pointed out, for educated elites in Tamil Nadu (social workers, nongovernmental organisations, doctors, teachers), it has become part of their everyday common sense to tend to explain the prominence of women among the possessed by seeing possession as a woman’s attempt to extract concessions from those around her and an outlet for social tensions (a situation which seems to have many similarities with the Nepali context described by Seale-Feldman, 2022).

However, within the academic world, the urge has emerged in the past few decades to go ‘beyond instrumentality’ (Boddy, 1994: 407) in the study of possession. These interpretative models of possession as an instrument employed by people – more or less consciously or unconsciously, successfully, or unsuccessfully – to fight oppression and express suffering have been the subject of a lot of criticism. It has been emphasised how such reductive, naturalising, or rationalising approaches – which endeavour to make sense of possession by reducing it to explanations in universal terms – do not pay sufficient attention to the specific context where possession is found, to local concepts and experiences (see on this point the excellent cross-cultural review of studies of possession by Boddy,1994).

What has also been stressed is the need to question the implicit presuppositions that influence the accounts of possession we produce and from which we draw conclusions (Crapanzano, 1977: 7). In several texts, Olivier de Sardan (1993, 1994, 1996) has deconstructed interpretations of possession, especially in African contexts, which tend to give a unilateral meaning, be it political, religious, or therapeutic. He has denounced the ‘epistemic violence’ (Olivier de Sardan, 1996) that some such interpretations exert on field data, because they are often the result of a mere projection of the researcher’s ideas rather than faithful reconstructions of the meaning that local actors give to possession practices. ¹⁵

In the past few decades, methodological caution and concern for context and experience have led scholars to pay attention to ideas of selfhood and action and to explore local gender imagery and dynamics. Some authors have denounced the implicit views of the subject and its actions in the world presupposed by the interpretations we have considered, which do not correspond to the way local subjects perceive themselves and experience the world. For example, the concept of somatisation or idiom of distress presupposes a Western view of the subject in which mind and body are separated and bodily symptoms display an unconscious deeper reality. In relation to the notion of resistance, a resistant subject is one who sees himself as autonomous and endowed with individual agency (for a recent discussion of this second kind of criticism, see Seale-Feldman, 2022). Moreover, the notion of resistance implies a perception of women as marginal, a concept which needs to be placed under scrutiny as we will see later in this text.

Several authors have also underlined how paying attention to the contexts and local-actor understandings also means taking seriously the givenness of spirits in the daily lives of their hosts. This is, among others, the case of Keller (2002) who, in a book nourished by feminist and postcolonial theory, has re-examined works about female possession; Ram (2013) with her phenomenological approach to possession in South India; and Boddy (1989) in the monograph about the zar possession in Sudan. It has been argued that spirits, gods, ancestors, and other external forces that constitute indigenous explanations of possession should be considered and not just regarded as mystifications of the ‘real’ forces, such as psychological pressures, social tensions, or chemical imbalances. As Ram (2013: 86) writes, another vision of possession is possible if we try to take spirits seriously and if we try ‘to comprehend how the world might be lived in by a being who is other than a self-enclosed, relentlessly conscious, and knowing subject confronting a world that is entirely external’.

These calls for interpretive caution and for context-sensitive and experience-near accounts of possession are well-founded and destined to renew scholarly analysis of possession phenomena in the years to come. However, in the case of the two practices analysed in this text, it seems difficult to totally reject one main point which the two previously mentioned perspectives on female possession have in common: the idea that tensions and conflicts are central to married women’s possession and the rituals associated with it. Indeed, it is not (only) the anthropological interpretation that sees the centrality of conflicts in these rites. It is the practices themselves that give a central and explicit space to the conflict between a woman and members of her marital family and the suffering that results from it. If, therefore, these two practices oblige us to put conflict – and domestic conflict in particular – at the centre of analysis, how can we take this centrality into account without doing ‘epistemic violence’ to the actors’ conceptions and experience?

On married women’s ‘supposed’ resistance

In the first kind of exorcism (and adorcism) analysed, which is related to the practice of cursing, an interpretation in terms of married women’s resistance to the patriarchal order would, at first glance, seem particularly appropriate. The practice of ghat and the deities associated with it inspire genuine terror in the people of the region. Some of my interlocutors refuse to talk about the practice, for fear that such deities, known for their susceptibility, might take revenge on them. While agreeing to discuss it during interviews, many refuses to mention the names of Goril and Bhairav, who are two of the most feared among these deities. It could therefore be assumed that the possibility of cursing constitutes an instrument of resistance for married women in Garhwal, prompting members of the marital family to guard their behaviour towards them more carefully, and that through this practice women can resolve inter-personal conflicts in their favour.

However, the life stories analysed do not allow us to support this interpretation, at least if we define resistance, as Lewis has done, in individualistic and conscious terms. The case of Saraswati, whose attempts to get her deity Goril accepted by her marital family ended up backfiring on her, shows very well that the practice of cursing is not always a tool of resistance in this sense. Although these deities are quintessentially the protectors of the dhyani and their worship expands essentially through their movements towards the marital household, their behaviour eludes a linear logic of individual empowerment of the woman invoking them. Viewed, as we have seen, as autonomous agents, they have the tendency to protect the dhyani, but only if the unity of the family is also preserved, that is, if a rite of adorcism requiring the participation of the whole family is performed. Such a rite is the only way to calm the dangerous wrath of the deity and to transform his power from a threatening into a benevolent one. Once installed in his new temple, the deity no longer belongs exclusively to the woman or her natal family but will protect the entire marital family against external enemies. Therefore, on closer inspection, such practices do not give empowerment merely to supposedly marginal or subordinate women but to the family unit as a whole (cf. on this point also Polit, 2012 and Sax, 2009).

This does not mean that these practices are, as I and others have argued with regard to other rituals, a way of forcing women who escape or resist the norm back into the dominant moral order (Bindi, 2015; Nabokov, 1997, 2000). The consequences of an accusation of witchcraft directed towards a woman vary from case to case. Sometimes, indeed, women succeed in having their deity accepted in the marital family. It would be erroneous to generalise the effect of such representations and practices on the lives of married women. Such effects depend, as I have shown in other contexts commenting on the outcome of a therapeutic itinerary (Bindi, 2012), both on the actions of the various actors involved and on conjunctural aspects (e.g. specific temporalities). However, what is certain is that, while allowing the expression of conflict and some benefits when the woman obtains the cooperation of the family group, the rituals are neither described nor perceived by women as forms of individual resistance.

Given the ambiguity of the results obtained and the widespread awareness of the dangerousness of the practice, it is therefore appropriate to ask why women do utter curses. One answer is provided by Boddy (1989) who, in her study of possession by the zar spirits in Sudan, argues that possession is a kind of habitus, embedded and naturalised, that constitutes an almost automatic reaction of women in order to articulate some difficult or complex experiences. For a Garhwali woman who is confronted with the difficult experience of domestic conflict, cursing would therefore be a normal reaction (on this point see also Polit, 2012).

Nevertheless, understanding such a practice requires above all abandoning a Western notion of the subject. Marriott and Inden (1977) have been emphasising for a long time the ‘dividual’ and relational nature of the person in India, who is perceived as relational and part of a collective, as opposed to the Western conception of the person as an independent individual. In Garhwal too, people see themselves as part of a group, constantly connected to and influenced by others (for a similar relational conception of the self in Tamil Nadu, cf. Ram, 2013; in Nepal cf. Seale-Feldman, 2019).

This conception of the self has a direct impact on the way people conceive their capacity for action in the world. Women (even recently married ones) do not see themselves, nor are they seen, as passive or even marginal. Therefore, their action in the world, even actions that can negotiate or fight dominant norms (as understood in the common anthropological usage of the category of ‘agency’), is not perceived as a quality primarily linked to individuals but rather as collective, always exercised by networks of people who are connected to each other. Moreover, in Garhwal, these ‘others’ with whom people feel connected include not only humans (e.g. parents, husbands, siblings, friends) but also supernatural beings (e.g. the deity of one’s natal village, or the dead members of one’s family). To put it another way, both human and divine persons are members of the collective networks which, for people in Garhwal, constitute their way to act in their world.

Even more than in the previous case, we would be tempted to see in the ghaḍiyālā rite an illustration of the idea that possession allows for the indirect expression of frustration and tension by subjects who are in a position of structural weakness: in this case, married women, and in particular those who have married a widower. However, on closer inspection, even if the ritual allows the expression of sorrows and complaints, this does not seem to be particularly related to individual suffering or to a strategy that would bring individual benefits.

I was able to build a relationship of intimacy and confidence with both Sumita and Manisha, the two women who are at the centre of the two ghaḍiyālā rituals described here. Neither lived in a particularly complicated family situation. Manisha had a happy marriage to Pradeep and did not seem to have any particular difficulties with either her mother-in-law or her two sisters-in-law. Sumita also had a relatively peaceful married life. Indeed, she often expressed relief at having finally been able to marry: a marriage conducted at a late age by local standards (25 years old) because of what she called ‘a problem with her horoscope’, that is, a configuration of the planets at her birth that prevented her from easily finding a compatible partner (as the two partners’ horoscopes have to be studied by astrologers to see if they match, a very common practice in India).

Moreover, even more importantly, for both of these women, possession by their husband’s first wife was not a surprise: it was actually expected. The fact that a woman who has married a widower will experience, during the first years of marriage, the ‘presence’ of the first bride in her own body is known and accepted. The only thing that is not certain is the exact moment when the spirit of the dead will manifest itself, and with what symptoms. An example to the contrary will illustrate the expected nature of this practice. In July 2015, I met Prahlad, a young man of Rajput caste who was widower at the age of 27. Between the years 2015 and 2018, Prahlad organised multiple ghaḍiyālā rituals (some of which I had the opportunity to film) to allow his dead first wife to embody herself in him. Although most of the oracles and other ritual specialists consulted disagreed with him, and despite his diagnosed typhoid fever, Prahlad explained to me on several occasions that he was sure that his numerous physical symptoms, accompanied by altered perceptions (voices, visions) and appearances of his wife in dreams, were signs of his wife’s willingness to incarnate the ritual in him. He was well aware that the procedure was abnormal, but they had contracted a love marriage (still rare in an area where arranged marriages dominate), and the great love that bound them together was the reason why his wife would incarnate herself in him, which was quite unusual. During the various rituals he organised, Prahlad exhibited signs which locally indicate a state of possession but, according to those who attended the ritual and whom I interviewed, it was not his wife but other entities that possessed him. Two years ago, he told me by telephone that he had found a new bride. And he added, ‘Now it is to her that my wife will come’ (an expression locally used to indicate possession). Vishal, a local ghaḍiyālā’s ritual specialist (guru), confirmed this idea in an interview done at the same time (even if he expressed some scepticism as to future possession due to the fact that Prahlad’s new wife is unable to walk due to a severe physical disability, which made the possibility of possession uncertain).

In light of this evidence, it is difficult to read these possession practices as forms of expression of individual suffering or as a strategy of resistance. Rather, ghaḍiyālā is, somewhat similarly to an initiatory ritual, an expected and prescribed ritual which is structurally part of the local social organisation. Its main preoccupation is the unity the family, including both the living members and the spirits of the dead, while its ritual actions are all directed to the aim of appeasing deceased members, who will then protect the family. The expression of family conflict and married women’s suffering is only important as part of this larger collective strategy.

This reading in collective terms is confirmed by the very structure of the ritual action. The chant does not singularise the situation of the invoked spirit, nor that of the possessed woman; rather it calls forth generalisable experiences and emotions common to entire categories of beings: the accidentally dead and the married women. As for the conversation with the dead woman’s spirit that follows the chant, as much as the spirit directly addresses her husband’s family and sometimes mentions concrete events that have occurred, her complaints refer to relatively common experiences: a general lack of attention, affection and consideration, feelings that most married women can easily identify with. While during this rite-specific family problems are often raised by the many intervening deities who speak through their mediums, it is rare for such problems to directly concern the new bride, in whom the first bride is embodied.

Are married women and their possession ‘marginal’?

The data presented so far has enabled us to show that the two main rites of exorcism (or, more precisely, rites articulating exorcism/adorcism) in which married women are involved in Garhwal have an integrative function in relation to the conflicts that punctuate domestic life. They are oriented towards overcoming dissent and preserving collective unity and the values that underpin it. This theoretical proposition is, in this sense, in line with the countless works that, while adopting dissimilar perspectives, have highlighted the ways in which rituals can contribute to restoring and preserving social order. On the one hand, in the field of religious studies, numerous scholars have emphasised the integrative roles that rituals play in society. To mention but a few classical theories, Durkheim emphasised the capacity of ritual to unite people, to give them a sense of belonging to society and to remind them of their commonality and their past. In the work of Gluckman, and in some analyses by Turner, social conflict is recognised within the strategic limits of ritual, where it can be systematically subsumed within a reaffirmation of unity (for a detailed analysis of how the field of ritual studies has addressed the relation of ritual and conflict, cf. Bell, 1992: chapter 7). On the other hand, in the rapidly growing interdisciplinary field of conflict resolution and transformation studies (for a review of this field, cf. Kriesberg, 2011), several scholars have begun to take a specific interest in the way rituals play a crucial role in social conflicts, facilitating a fundamental transformation of conflict, for instance, in the post-war period, by opposing violence with an alternative framework (Zemskov-Züge, 2015).

However, these exorcistic practices do more than merely helping reproduce the socio-moral order. Both seem to me to play an important function in defining the positionality of married women in the social fabric. We should start here by placing under scrutiny the idea that women are subordinate or marginal subjects. Responsive, like other scholars, to feminist anthropology, Boddy (1994: 415) invites us to question this idea, in her review of scholarship about possession: ‘By whose evaluation are women marginal? Does economic or political subordination fully determine women’s positions in other spheres of life’. We have seen how, in Garhwal, women have limited access to political and economic power in the context of the unequal domestic relations that characterise a married woman’s life. This does not mean, however, that married women, including those newly married, are either considered, or perceive themselves to be powerless (cf. on this point also the already mentioned ethnography of Bennett, 1983). And this is not only because of their power of giving birth.

It is useful here to reflect on which conception of power is adopted in anthropological analyses of possession (see on this point Karp, 1989, and the discussion about agency and female possession in African contexts by Plancke, 2012). Anthropological interpretations of possession as a tool used by marginalised female subjects to resist or to express their suffering tend to adopt a definition of power – one which is widespread in social studies – as control over people and resources. Possession is therefore interpreted as a means of exerting such control, under conditions of inequality. However, it has been shown that power can also be conceived as ‘energy and potentiality’ (Karp, 1989), as the potential for ‘receptivity’ and ‘permeability’ (Keller, 2002), or as the ‘capacity to be affected’ (Ram, 2013). In many contexts worldwide, women possess more than men the quality of permeability, both to spiritual energies and to emotions. In South Asia, this is a typically feminine quality (which goes along with a set of physiological ideas about the female body; Bindi, 2012, 2015). It can on some occasions be considered neutral, or even dangerous (Bindi, 2012, 2015; Seale-Feldman, 2019, 2022). In the two ritual practices we have analysed, this quality is valorised as it gives married women the potential to become the mediators with spiritual beings, which can be both a threat for the social order or, if pacified, a source of protection, well-being, and prosperity for the family. It is married women who, because of their proximity to their natal land and its deities, have the power to call on them, thus igniting a process which can eventually lead to their transformation into very powerful and feared deities who protect the household. It is mostly married women who, because of their ‘capacity to be affected’, are afflicted by and become the medium of the soul of their husband’s first wife, allowing the ritual transformation of such potentially malevolent souls into benevolent members of the family pantheon. In this regard, possession and dispossession practices involving married women do not seem to fit the idea of Lewis for whom feminine possession was often a peripheral one, involving spirits which are not central to the society’s moral order. Quite the opposite. These two categories of spiritual entities – deities who came with dhyanis and pacified spirits of the dead – are, according to many informants, the deities they feel are most close to them and the most helpful in everyday matters.

If married women have the crucial power of giving birth to new family members, through these exorcistic practices, they also have the power to give ‘ritual birth’ to spiritual entities who will contribute vitally to the family’s prosperity and well-being.

Yet the fact that domestic conflict and suffering is a central theme in these exorcistic practices leads to a further final reflection. Garhwali married women’s possessions seem also to open up a space where some of the ambiguities lying at the heart of the moral and social order are displayed. In her ethnography of zar possession in Sudan, Boddy highlights how, through the mouths of the women they possess, spirits use a kind of antilanguage that burlesques society’s assertions about gender and morality (Boddy, 1989). By decentring or reshaping accepted meanings, zar possessions open ‘a space for reflexion and ambiguity’ (Boddy, 1994: 417) and constitute a sort of ‘meta-commentary’ of society on itself and its values. They can be regarded as an important moral activity, which is at once challenging and reinvigorating for the moral order of the society. And this is ‘gendered moral activity’, as its responsibility and power belong exclusively to women (who are the subjects mostly concerned by zar possession). Similarly, Ram suggests how, in the south Indian state of Tamil Nadu, episodes of female possession by the souls of dead women might at times force the community to reflect not only on a normative understanding of what it is to live and die well, but also on some of the contradictions and ambiguities which are inherent in social life and human existence itself. Such possessions

keep alive all sorts of unsettling possibilities: that bad deaths exist; that the hierarchically dominant group can and does overstep the limits of justice in its treatment of the subordinate group; that there is an authority higher than that of the dominant social group or gender; and that death does not extinguish the force of injustice. (Ram, 2013: 101)

Possession practices of married women in Garhwal could also be regarded as a meta-commentary on the society, and particularly on family normative values and their contradictions. During such possessions some of the complications which this socio-moral order entails are brought into the foreground, namely, domestic conflicts and profound suffering for married women. It is not only through linguistic exchanges but also through their performative material and sensorial aspects (such as gestures and music) that possession-based rituals engage participants and allow for a heightened collective attunement on these difficulties. Such processes can be seen as a gendered activity bearing, in Boddy’s words, both a challenging and a reinvigorating effect on the morality and norms governing social life.

Conclusion

An extremely widespread practice in the north Indian state of Uttarakhand, exorcism can here take various forms, all of which are endowed with specific characteristics. These pages have allowed us to show how, in the forms of exorcism that involve married women, the theme of domestic conflict and suffering is central to the ritual practice itself. However, such practices cannot be considered to be individual forms of resistance. Despite the fact that they may sometimes benefit the women participating in them, these women do not seem to perceive themselves, nor do they act, as individuals who resist social structures or norms, but more as part of a collective network of people, made up of both human and non-human beings.

As for the effects these forms of exorcism have on domestic conflicts, we have shown how they are geared towards the preservation of family unity. Both these ritual devices evoke human conflict, specifically in the oracular sessions preceding the first form of exorcism, and generally during the second one. Nevertheless, pacifying spiritual entities in both practices of exorcism constitutes the main problem to be solved. Domestic unity and harmony are both a means to such an end and one of its effects. Deities and spirits of the dead – spiritual beings in the grip of their emotions – must be calmed and soothed. And the only way to do this is to perform collective ritual actions: celebrations and offerings involving the whole family. Once pacified through such ritual actions, these spiritual entities will ultimately bring unity and well-being to the family members. To put it differently, it is by giving expression to human conflicts and by subordinating them to divine ones that these exorcisms actively contribute to the preservation of family unity.

Indeed, while geared towards the preservation of the socio-moral order, these exorcisms, where domestic conflicts and married women’s suffering take centre stage, can also be seen as a society’s meta-commentary on its own internal ambiguities, with a challenging and reinvigorating effect on moral values and social norms.

Married women are assigned a crucial role in this process of the preservation of the social group through rituals of exorcism. Far from being marginal to the fabric of social life, they have, in addition to the power of giving birth to human children, the power of giving ‘ritual birth’ to these spiritual entities who, once pacified, will protect family unity. Being close to the angry deities of their natal villages and easily affected by spirits of the dead, married women have an essential role in mediating with them, in transforming them into protective entities, and thus in the creation and preservation of the network of human and non-human beings on which the well-being of the family depends. Therefore, exorcistic practices not only have an integrative function vis-à-vis conflicts potentially endangering the unit of the family, but they also firmly establish the position of married women and their irreplaceability in the fabric of social life.