The healer and the witch

Two words for the same woman

Somebody in the village knows what to say over a sick child.

It is generally known who that is. She does not advertise and she does not need to. A neighbour names her, and by evening a father is at her door with a fever in his arms. She says the words over the child. Sometimes the child gets better, and the family remembers who was standing over the cot when it happened.

That is a competence, and in a village a competence is never private.

If a form of speech can reach into a body and cool it, the same form of speech can reach into a body and do the opposite. There is no second and gentler power held in reserve for the good uses, no separate faculty issued to people who mean well. There is one capacity and it points both ways.

Which is why she is never more than one bad season away from being the answer to a different question.

A child does not get better. A family that has waited years for a birth is still waiting, in front of the same neighbours who have watched them wait. A quarrel over a boundary hardens into something nobody can climb down from. Misfortune of that kind does not stay abstract for long in a place where everyone’s business is everyone’s, and it will not stay authorless. Somebody did this. The candidates are not drawn at random. They are drawn from the people already known to have the capacity, which is to say from the people who get sent for.

So there are two words available for her, and she does not choose which one she gets.

Called one thing, she is the woman a father runs to at nine in the evening. Called the other, she is the reason his child died. The practice is identical. The syllables are identical. What has changed is that other people, on an occasion of their choosing, have given the same capacity a different name.

A name of that kind is not a description anybody can argue with. It is an event. What follows one, in the record, belongs to the next chapter and is deliberately not written here.

This chapter is about the arrangement itself: how it works, and who documented it.

Whose sentence this is

The arrangement is not this treatise’s discovery, and the credit has to be given precisely, because a borrowed proposition given carelessly turns into an announcement.

In 2004 a doctoral dissertation was accepted at the School of Oriental and African Studies in London. Helen Macdonald had written it on witch accusation in Chhattisgarh, from fieldwork begun in September 2000 in the village she calls Ballabgarra, with a base in Raipur, and she called it Resolution and Rupture: The Paradox of Violence in Witch Accusations in Chhattisgarh, India. One sentence in it is the sentence this chapter was built around.

It has to be given with its frame, because the frame changes what the sentence is.

“Certainly there are ‘family resemblances’ we can identify in the Rajgond case: those who protect society against illness are often assumed to possess the same power as witches (Nadel 1952; Mair 1969; Nitibaskara 1993); children are vulnerable to witch attack; witchcraft is a salient feature of ādvāsī culture; witches only attack their affines.”

The words are Macdonald’s. The proposition is not presented as hers. It is one clause in a list of four, it is scoped to “the Rajgond case,” and it carries three citations to earlier anthropology: S. F. Nadel in 1952, Lucy Mair in 1969, Ronny Nitibaskara in 1993. It is a comparative family resemblance she registers, rather than a finding her fieldwork produced.

Read as a standalone doctoral finding it becomes something it is not. What the sentence is: a well-attested proposition in the anthropology of witchcraft, registered by a fieldworker in Chhattisgarh as applying to the case she studied.

The proposition itself still does the work, and it is worth reading slowly, because it is easy to mistake for a platitude about superstition and it is nothing of the kind. It is a claim about identity. The power that cures and the power that kills are not two powers that resemble each other. They are held, by the people who live with both, to be one power. The healer and the witch are doing the same thing with the same capacity. What separates them is a name, and the name is assigned by other people.

That is this chapter’s spine. It is a borrowed spine, and this chapter says whose.

One power, two names

Take the proposition at its full strength, as Nadel, Mair and Nitibaskara framed it and as Macdonald found it holding in Chhattisgarh.

If protective power and destructive power are one substance, then the question of which name a practitioner gets is never settled by what the practitioner does. It is settled by who is deciding, and when, and what they need. A woman who is consulted about a child’s fever on Tuesday is, in her own person, exactly as capable of causing a fever as of ending one. That is not an accusation. It is the structure of the belief, and everyone inside it understands it perfectly well, including her.

Which means the practitioner lives permanently on the wrong side of a hinge that other people control.

There is a second consequence, and it is the one that makes the gender argument work. The apparatus of accusation does not choose between a woman’s healing and a woman’s harm. It polices her on both sides at once. Her competence is the evidence. The same neighbours who came to her door for help are the pool from which her accusers will be drawn, and they will not need to discover anything new about her in order to accuse. They already know what she can do. They came to her because they knew.

Macdonald’s Resolution and Rupture is a doctoral dissertation submitted at SOAS, University of London, in 2004, and it is cited here as what it is: examined scholarship built on fieldwork, which has not been through a university press. Both places this chapter draws on are secured, the family-resemblance passage and the table of accusers, and both are anchored in the sources note below. What is not secured is Macdonald’s own assessment of the Nadel-Mair-Nitibaskara proposition beyond registering it, which would need the surrounding chapter read rather than the page.

One limit has to be declared before the chapter goes further, because the reader is entitled to know the terms.

None of the scholarship in this chapter is about the Śābara corpus by name. Not one of these researchers writes about śābara mantras, the Śābaratantra, or the printed chapbooks chapter 17 traced through Bombay and Banaras. They write about healing, possession, amulets and accusation in the ritual world of village and small-town India. What licenses their presence is that the corpus is used by somebody, somewhere, and this is the world in which that use happens. The connection is contextual. It is not textual, and nothing below should be read as though it were.

The field locations are carried and named where each researcher appears: Chhattisgarh for Macdonald, eastern India for McDaniel, Garhwal in Uttarakhand for Bindi, Hyderabad for Flueckiger, Tamil Nadu for Kalpana Ram, the Saora highlands of southern Odisha for the Adivasi and SCSTRTI material.

Where the record is print, and where it is ethnography

There is a generalisation available here, and the evidence will not carry it. It goes in this shape: every role in this world that carries a title, a text, a signature or a price is male, and women’s practice is oral, domestic, improvised and charismatic. It is quotable and it is unevidenced, and this chapter’s own sources refute it, the Odisha state institute’s journal most flatly of all.

Here is what the evidence does support.

The roles that leave a documentary trace of the kind this treatise has been reading are held by men with great consistency. The lineage in chapter 6, the twelve Kāpālikas headed by Ādinātha and the twelve pupils headed by Nāgārjuna, is a roster of men. The teacher-to-pupil transmission of chapter 8, the dīkṣā, the sealed passage of a syllable from one qualified mouth to one qualified ear, runs between men. The bazaar press of chapter 17 is a male commercial world: Gaṅgāviṣṇu and Khemraj Śrīkṛṣṇadās on the imprints, Jvālāprasād Miśra and Baladevaprasād Miśra on the commentaries. Those are printed books, signed colophons, numbered catalogue entries, and in all of them the names are men’s names.

That is a real and consistent finding, and it is a finding about print, lineage lists and the trade. It is not a finding about ritual life.

Extend it beyond that and it breaks immediately, in this chapter’s own sources.

The ojhā is not male by definition of the term. The only statutory definition of ojhā in Indian law, which chapter 20 prints in full, is gender-neutral: a person who claims to possess power. The ActionAid and Odisha State Commission for Women study, 295 pages of Odisha fieldwork, never genders the term either. And Madhu Mehra and Anuja Agrawal, in their three-state study, record evidence of women as instigators of witch-hunting, while noting that men outnumber them in that respect. Women operate this system as well as suffering it, and the next section but one gives the count.

So the honest version is this. Where a role is recorded in a printed book with a signature on it, the record this treatise reached is overwhelmingly male. Where the record is institutional ethnography rather than print, titled women’s offices appear at once, with names, competences and routes of entry. The generalisation refused above is a fact about a bibliography written in the grammar of a fact about women.

Kudanboi

Start with the community whose name this book carries, because the answer is there.

The Saora ritual specialist has a name in the ethnographic literature, and the name is gendered in pairs. In Adivasi, the journal published twice a year in Bhubaneswar by the state’s tribal research institute, Harihar Das and Devi Prasad Das set out, in 2007, the Saora magico-religious specialist as the Kudan, who is male, or the Kudanboi, who is female. The specialist goes into trance, communicates with the aggrieved god or spirit, and conducts the sacrificial ritual the case requires. The village priest, a separate office, is the Buyya.

Read what is in that sentence. A named office. A named female holder of it. A defined competence. A published Indian institutional source printing the term in a journal that has been running since 1955.

The same institute’s ethnographic volume goes further. Particularly Vulnerable Tribal Groups of Odisha, edited by A. B. Ota and S. C. Mohanty, SCSTRTI, Bhubaneswar, 2015, carries an article by P. K. Das Patnaik titled “Bejunis: their Initiation into Shamanhood.” That is an Indian account, in an Indian state institute’s own publication, of women entering a shamanic office by a described route.

Two limits. The Indian institutional sources do not agree on the terminology: Adivasi in 1997 gives the Saora magician as Disari, Adivasi in 2007 and the THRTI volume The Saora, edited by N. Patnaik, Bhubaneswar 1989, give Kudan and Kudanboi, and Elwin’s 1955 ethnography uses a different set of terms again. Three Indian sources and one English ethnography do not converge on a single word. That divergence is reported here rather than resolved. And the Bejunis article was located at citation level, not read through; nothing is attributed to Das Patnaik beyond his title and subject.

What the divergence does not touch is the point. Whatever the word, the office exists, women hold it, and Indian scholarship has been printing that fact in Bhubaneswar for decades.

Ask the Indian sources first and the generalisation refused above never gets written.

Called by dream

June McDaniel’s 2023 article in Religions is the most useful single map of women’s religious roles in eastern India, because she does the elementary thing nobody else had done in one place: she counts the kinds.

Santal priestesses. Sādhikās, women following a formal spiritual discipline. Trance mediums, whom she records under the local designation “bhar ladies.” Āyes. These are distinct offices with distinct competences and distinct routes into them, and treating them as one undifferentiated category of folk healer destroys the information.

They are also, every one of them, titled roles held by women. That is worth saying flatly, because a typology of titled women’s offices is exactly the evidence that any claim about their absence has to answer, and it has been in print since 2023.

One route matters more than the rest for this treatise. McDaniel records Santal priestesses who are called by dream rather than by initiation.

Hold that against chapter 8.

Chapter 8 built the orthodox mantra apparatus carefully, because everything after it in this treatise is a departure from it: the ten saṃskāras by which a syllable is made fit for use, the qualified teacher, the sealed transmission, dīkṣā as the gate through which a person passes in order to become someone whose speech works. That is a credential system, and a credential system has a door, and a door has a keeper. Somebody decides who is admitted. Somebody can refuse.

A dream has no keeper.

There is no committee that can decline a dream, no lineage that can withhold it, no register in which it must be entered, no fee. A woman who is called in a dream holds an authorisation that the textual apparatus is structurally unable to gatekeep, because the apparatus was never given jurisdiction over her sleep. It is a different mechanism altogether, and it is the only one in this treatise that no institution can control.

That paragraph is this treatise’s inference and not McDaniel’s argument. What McDaniel records is narrower and points a different way: having a deity choose to visit a woman in a dream or a trance and give her knowledge is the socially permitted route, and it is permitted partly because it protects her from being called a witch. The absence of a gatekeeper is this treatise’s reading of what that permission implies. It is flagged as a reading.

McDaniel also records the cost. What the article says, in its own abstract, is that among the Santals a religious woman must avoid being called a witch, because women’s power is seen as dangerous and religious social roles are traditionally forbidden to them. Dangerous and forbidden, in her terms and in no other. The two hold each other up. It is forbidden because it is feared, and feared because those who hold it acquired it without permission.

The typology and the abstract’s wording are taken at article level. No page or section anchor within the published article was secured, and nothing is put inside quotation marks that has not been read on a printed page.

The mediator, and the reading the specialists declined

Serena Bindi’s study of Garhwal, published in Social Compass in 2022, adds a correction that has to be taken seriously because it cuts against the reading most readers will arrive with.

Bindi’s married women are indispensable mediators. They are not marginal figures tolerated at the edge of a male ritual system. The system does not function without them, and everybody involved knows it.

Then comes the part that matters. It would be easy, and it would feel satisfying, to read women’s exorcistic and possession practice as covert resistance: a hidden weapon of the weak, a way of speaking that a subordinated group uses because the open channels are closed to it. That reading has a long and distinguished history in the anthropology of possession and it is available here for the taking.

Bindi declines it. On her evidence the practice establishes these women’s position rather than smuggling a protest past it. It is constitutive. It is how a married woman becomes a person of standing in her community, openly, in a role her community recognises and requires.

So this chapter will not reach for the resistance reading anywhere, in any register. The specialist working in the field has already looked at it and put it down, and a treatise that has spent eighteen chapters insisting that the primary literature outranks the popular assumption cannot start overruling the primary literature the moment the popular assumption is more attractive.

Bindi is not alone in declining it, and the second refusal comes from the scholar who should have been in this chapter from the start. Kalpana Ram’s Fertile Disorder: Spirit Possession and Its Provocation of the Modern, Honolulu: University of Hawai’i Press, 2013, works from fieldwork among women in Tamil Nadu. Its argument, on the publisher’s own statement of it and on Soumhya Venkatesan’s review in the Journal of the Royal Anthropological Institute in 2015, is that possession is neither a covert politics nor an exotic survival. What it does instead is expose instabilities in the modern conception of agency itself: the assumption that action requires a conscious, deliberating, self-present chooser. A woman who is possessed is not smuggling a protest past her family. She is acting inside a model of the person that the vocabulary of resistance cannot describe.

That is a stronger version of the same refusal, and it comes from further inside. Two researchers, two states, north and south, both declining the reading a Western-educated reader arrives with.

Amma’s room

Joyce Burkhalter Flueckiger’s In Amma’s Healing Room: Gender and Vernacular Islam in South India, Bloomington: Indiana University Press, 2006, follows one woman, a Muslim healer working in a poor district of Hyderabad, and it is the closest thing in this literature to a portrait.

Amma dispenses written-word amulets. She does so in three distinct ways. What those three ways are will not be described here, and no amulet in this treatise is described operationally, in keeping with the line this book has held since the prologue: no working mantras, no vidhi, no materials, no timings. That there is a threefold practice is a fact about how her healing room is organised. How to do any of it is not on offer.

Two findings from Flueckiger carry the chapter.

The first is about what it costs her to hold the role at all. In taking up a public healing role that is normally held by men, Amma crosses a line her community does not ordinarily let a woman cross, and the consequence is continuous. She does not cross it once and then occupy the position. Her status as a healer, as a woman healer, and as a pīrānimā, the wife of a pīr, depends on her husband’s standing, and it has to be renewed again and again by narrative and by discipline.

The wording above is this treatise’s own and not Flueckiger’s. What her chapters four and five support is the substance, and no phrase of hers is reproduced here inside quotation marks.

Note also what Amma is. She holds a title. She has a room, a clientele and a price. She is the counter-example that the generalisation refused above walks straight past.

The second finding is a proposition about paper: Muslim amulets work for everyone.

A piece of paper crosses a line a woman cannot

Put those two findings side by side.

Chapter 12 established, by frequency count rather than by assertion, that the invoked authorities of a popular Śābara compilation include पैगंबरों, the Prophets, sitting in the same list as Mahādev and Gorakhnāth. The religious boundary in that corpus is crossed, and it is not perceptibly defended. An appeal, a duhāī, an appeal made in someone’s name, will be addressed to whichever authority is held to have power, and the question of which tradition that authority belongs to does not arise.

Flueckiger’s line says the same thing from the other end. A Muslim healer’s written amulet works for a Hindu client. Nobody at the transaction is troubled by this. The object moves across the line as though the line were not there.

Now measure the other boundary.

Amma is a woman doing a job men do. She has to justify her occupation of it continuously, in front of the same people, for as long as she holds it, and she does so partly by borrowing the standing of her husband. The religious line is porous to the point of invisibility. The gender line does not yield, not once, not for a woman of demonstrated competence with a full room and a working practice.

A piece of paper moves between religions more easily than a woman moves into a role.

That is the sentence to carry out of this chapter, and it is not this chapter’s discovery. It is Flueckiger’s own. She proposes that gender boundaries are impermeable where religious boundaries are the opposite: her clients accept a common cosmological and ritual universe, in which a Muslim amulet works for anyone, while the line Amma herself has to keep re-crossing never softens. Crediting her costs one sentence, and it converts what would look like a synthesis into a sourced finding.

What chapter 12 adds is a second, independent instance of her first half. The Prophets sitting beside Mahādev in a Hindī compilation is the same permeability, in print, in another century, in a corpus she never worked on.

The register and the practitioner

There is one more thing to do with Amma, and it closes a circuit this treatise opened eight chapters ago.

Chapter 11 read the corpus’s characteristic vocabulary as a sociological signal. Vācā, the utterance named as an utterance. Phuro, let it be fulfilled. And the third clause, the naming of an authority under which the demand is made. The argument there was that a speaker who must assert that his word is true, and who must name somebody else’s power before his word can act, is a speaker without institutional standing. A person with credentials does not certify his own speech. His position does it for him.

Amma is that structure walking around in a person.

She continually asserts her legitimacy, because it is not conferred and stays unconferred. She leans on her husband’s standing, because the standing is not hers to hold. Truth-assertion and borrowed authority: the two moves that chapter 11 found running through the whole register are the two moves by which a woman without a sanctioned role keeps a healing practice open.

The register and the practitioner match. A grammar built out of asserting and borrowing is the grammar of people who have to assert and borrow.

That is inference, and it is flagged as inference. Nothing in Flueckiger connects Amma to the Śābara corpus, and the compilation counted in chapters 10 to 12 is anonymous: it carries no speaker, no gender, no attribution of any kind. We do not know who says these words. We know what kind of speaker the words are shaped for.

Sixteen female accusers

One count in Macdonald’s own study cuts across the whole picture, and it is not a small one.

Macdonald counted them. Her table is headed “Breakdown of Sixteen Female Accusers.”

Sixteen women who accused. Counted, set out in a table, in the same study that supplies this chapter’s central sentence. Women are not only the accused in this system. They operate it. They bring cases, they name names, they testify, and they act on grievances of their own that are as real and as ordinary as anyone else’s. Mehra and Agrawal find the same thing across their three states in a different form: women instigate, and men outnumber them in doing it.

A chapter in which women appeared only as victims would be false to the evidence. The women in this chapter are practitioners with expertise, working under constraints they did not set and understand better than anyone, and they are also participants in a community’s quarrels with their own interests and their own capacity to do harm. Both facts are in the record. Neither cancels the other.

What this chapter contributes

The question is fair and it should be answered in the chapter rather than in a research file.

The structural finding here is Flueckiger’s. The proposition at the spine is Nadel’s, Mair’s and Nitibaskara’s, registered by Macdonald and scoped by her to the Rajgond case. The typology of women’s offices is McDaniel’s. The refusal of the resistance reading is Bindi’s and Kalpana Ram’s. The female shamanic office among the Saora is printed by the Odisha state tribal research institute and was found by opening its journal. Take those away and there is no chapter.

So this is a synthesis chapter, and that is what it should be called.

What it adds, and the list is short and honest, is three things. It sets Flueckiger’s permeability finding beside a frequency count from a Hindī compilation she never worked on, which is an independent instance of the same pattern in another century and another corpus. It reads McDaniel’s dream-calling against chapter 8’s dīkṣā apparatus and observes that one authorisation has a gatekeeper and the other does not, which is flagged above as this treatise’s inference and not hers. And it matches chapter 11’s grammatical register against Flueckiger’s portrait, which is flagged as inference too.

Three connections between existing findings. No new fieldwork, no new archive, no new text. A synthesis that credits its sources is worth more than a discovery that does not.

There is a further thing it does. It is the chapter in which this treatise stops writing about a world of men, and the material that makes that possible comes from Indian institutional sources.

Named and unopened

The list is longer than it should be.

Sudhir Kakar, Shamans, Mystics and Doctors: A Psychological Inquiry into India and its Healing Traditions, New York: Alfred A. Knopf, 1982, reissued in paperback by the University of Chicago Press in 1991. This is the standard Indian account of vernacular healing and of who consults a healer for what, and it belongs at the centre of this chapter. Its contents are now fixed, and three chapters are the load-bearing ones for this chapter and for chapter 20’s ojhā: “Lord of the Spirit World” and “Other Shamans,” in the part on local and folk traditions, and “Tantra and Tantric Healing.” Two archive.org copies exist and both are lending-restricted; neither was circumvented. Chapter-level anchors are available. Page level requires a physical copy, and nothing here is cited to a page.

Veena Das, Life and Words: Violence and the Descent into the Ordinary, University of California Press, 2007, and Affliction: Health, Disease, Poverty, Fordham University Press, 2015. Affliction is the more on-point of the two: illness, healing practitioners and the everyday in poor Delhi neighbourhoods, which is the same social world as the ojhā. Life and Words is the right source for how violence enters ordinary speech, which is the accusation material in this chapter and the next. Both are lending-restricted and neither was opened. Both are named here and used nowhere.

Kalpana Ram is used above at book level only, where she should be used at page level.

Three Indian scholars whose work bears directly on this chapter’s subject, all named, none read at page level. Naming three holes is worth more than filling one of them with a citation this research cannot stand behind, and nobody working from this chapter should cite any of the three at page level until the pages are opened.

Three more belong to chapter 20’s subject and are fixed in the sources note below so that they can be found. Shashank Shekhar Sinha, “1857: Witch-hunts, Adivasis, and the Uprising in Chhotanagpur,” in the Economic and Political Weekly of 12 May 2007. Soma Chaudhuri, Witches, Tea Plantations, and Lives of Migrant Laborers in India: Tempest in a Teapot, Lexington Books, 2013, and her article “Women as Easy Scapegoats: Witchcraft Accusations and Women as Targets in Tea Plantations of India,” in Violence Against Women, 2012. Govind Kelkar and Dev Nathan, Witch Hunts: Culture, Patriarchy and Structural Transformation, Cambridge University Press, 2020. None of the three was read at page level. Chaudhuri’s monograph is lending-restricted and was not circumvented. Kelkar’s book is cited above in the one correct form, with the year the Cambridge imprint carries.

Erndl, Gold and Narayan were located during research for this chapter and are not on point for charm and healing transmission. Nothing is cited to them, and nothing should be added later on the strength of their reputations.

Why the four stay

One question follows from all of that, and it deserves a direct answer rather than a silence: if Indian scholarship is preferred, why is this chapter still carrying four researchers who are not Indian?

Because on the four specific propositions they carry, no Indian equivalent was reached, and each of them is cited at page or chapter level for something narrow.

Macdonald carries the family-resemblance proposition and the table of accusers. The table, sixteen female accusers counted and broken down, has no Indian counterpart located in this research, and it is the single piece of evidence that keeps this chapter from describing women only as the accused. McDaniel carries the typology of women’s offices in eastern India, which is the count nobody else has made in one place. Bindi carries the Garhwali finding that mediation constitutes standing rather than concealing protest. Flueckiger carries the Hyderabad portrait and the permeability proposition, which is this chapter’s spine and which she reached first.

Kalpana Ram, Kakar and Veena Das belong beside them. The Saora Kudanboi material belongs under them, as the treatise’s own community’s case, and it is the one place where an Indian institutional source corrects the foreign scholarship rather than sitting alongside it, because it names an office that the generalisation refused above would rule out.

Preference for Indian scholarship is a rule about where to look first. It is not a rule about whom to cite once you have looked, and a chapter that dropped Flueckiger to make a quota would be a worse chapter and a less honest one.

What is carried forward, and by whom

The rest of this belongs to the next chapter, and it is deliberately not written here.

Chapter 20 carries the harm ledger: the scale with its caveats attached, the statutes, the killing in Pune in August 2013 and the ordinance that followed six days later. What this chapter hands over is a role.

The ojhā is the formal one. He is the remunerated one, the one a village sends for, the one who appears in the case files under that name. The statute that defines the office does not specify a sex, and the field studies record women acting in this system too. What the documentation this treatise reached shows, with consistency, is men holding the paid and named position and men being sent for. What a person with that standing does with it, and what the record shows happening to the people they name, is chapter 20’s subject and chapter 20’s alone.

The healer and the witch hold the same power.

Only one of them was ever asked which she was.