Skip to navigation | Skip to content

Fatal Words: Restudying Jeanne Favret-Saada

I. Introduction

If I count correctly, I have transferred the draft of this article onto seven consecutive computers without ever trying to get it published. I always liked it, though. It is about Jeanne Favret-Saada’s Deadly Words, a book about witchcraft in Mayenne, Western France, that was first published in 1977. The book fascinated me when I was a student, and indeed it fascinated many other anthropologists of my generation. It stood for an involved and engaged type of ethnography, one that could affect the ethnographer to a remarkable degree: shattering her scientific certainties and changing her way of being in the world.

I was less sure about Favret-Saada’s theory of witchcraft, although it is intimately linked to the way she let herself become affected by her research. Witchcraft, she argues, is based on words. Since using words can be deadly where witchcraft is concerned, nobody would think of speaking about witchcraft merely to inform an ethnographer. To find out what witchcraft is about, you have to let yourself become entangled in its net. Only by accepting the diagnosis of being bewitched, seeking healing from bewitchment and later becoming an apprentice healer herself was Favret-Saada able to access information about this topic.

The book gives a radical answer to questions most anthropology students, I certainly among them, grapple with. What happens if we try to objectify experience, be it our own or that of the people whose lives we describe? Favret-Saada’s plea for subjectivity and involvement during fieldwork was taken up by the Writing Culture generation, who often cite her as a witness against the dangers of objectifying methods. For me, then, Favret-Saada certainly stood on the right side. I found her book somewhat hard to read, but I still recommended it to others and placed it in the – rather small – virtual bookshelf of anthropological classics I carried around in my head.

Until I went to Mayenne, Western France, in 1998 and 2003. My experiences there convinced me that her theory, which anchors witchcraft in words, was remote from the experiences of the people I lived with. I slowly realised how great a part of her theory emerged from her methodological choices. In this article, I try to substantiate that claim, and to analyse some of its consequences. I wrote the first draft of this eleven years ago. Since then, I have sporadically thought about publishing it, but I never did. Somehow my mere four months of research experience in Mayenne, time I mostly spent feeding cattle, repairing fences and putting fertiliser out on the fields, did not seem to have enough weight against an anthropological classic. It seemed indecent to criticize a deeply honest scholar when I was not even able to present a better theory than hers, but simply thought hers was wrong. Furthermore, I had come to Mayenne as an economic anthropologist interested in consumption patterns; what I learnt about witchcraft seemed too anecdotal and too far removed from my own area of expertise for me to write about it.

Click to enlargedobler-3

Photograph by the author

I later found that a number of colleagues had had similar experiences. Some had even written papers refuting earlier anthropological sojourners in the region where they did fieldwork – but very few of these papers have actually been published. Given the number of topics and places in the world, few anthropologists deliberately set out to study what a predecessor has authoritatively dealt with. By implication, we are often not experts on topics others have studied before us in the same region, and can never be as sure of our findings as we might be if they came from our ‘real’ field.

Since I wrote the first draft of this article, Jeanne Favret-Saada has, in 2009, published a third book on witchcraft, Désorceler, whose final chapter was translated and has appeared in HAU (Favret-Saada 2012). An English version of the book, The Anti-Witch has been announced by HAU Press and might have come out by the time this article goes online (Favret-Saada 2015). The work builds on her earlier accounts, brings some topics out much more clearly and widens the scope, but it does not contain any fundamental revision of her theory. I will quote from it whenever I find her thoughts formulated in a clearer or more differentiated manner, but just as her later book builds on her first, I developed my own thoughts mostly in relation to Deadly Words and Corps pour Corps. The following should thus not be read as a review of The Anti-Witch, although it might accompany a critical reading of it.

II. Deadly Words

Let me begin by outlining Favret-Saada’s theory of witchcraft as she presents it in Deadly Words. She starts from the observation that in modern France, witchcraft is not a belief system. It cannot be described in Evans-Pritchard’s classic terms: “Azande believe that…”. Instead, she describes witchcraft implicitly as one of three concurrent “discourses” on misfortune and healing, together with official medicine and Catholic faith. One cannot ascribe these three discourses to pre-existing social groups. The average peasant is no more an expert in witchcraft than in biomedicine. Average Mayennais are, to use Alfred Schütz’s terminology taken up by Michael Lambek in his work on Mayotte (Lambek 1993), ‘men (and women) on the street’ in all three knowledge systems. They only become ‘well-informed citizens’ or even ‘experts’ in any one discourse if concrete problems force them to do so. As long as witchcraft does not concern them personally, people tend not to look for deeper knowledge about it. Peasants in Mayenne are sceptics towards all three discourses, and are not readily convinced of witchcraft’s reality. Jeanne Favret-Saada’s first methodological axiom was not to become trapped in assumptions about peasant belief: “Toujours me rappeler que les paysans ne sont pas crédules”, as she writes in her field diary (1981: 27).

Witchcraft beliefs thus are always subject to doubts, and people fear to be seen as backwards and irrational if they take witchcraft for real. Towards outsiders who are not supposed to believe in witches, these fears translate into a certain double talk. Witchcraft is always somewhere else. In the neighbouring department, or perhaps in some village behind Meslay, people still believe in witches – but certainly not here.

So peasants are no easy prey for the ethnographer. Favret-Saada pointedly describes how her attempts to discover what witchcraft was about remained fruitless as long as she asked questions meant to generate theoretical knowledge. People told her folkloristic recipes and anecdotes, but they were never personally involved in any of that. This only changed when the role she was assigned ceased to be neutral. The first person who openly spoke to Favret-Saada about his own bewitchment was hospitalised in a psychiatric clinic and, as she only realised later, took her for a healer who might be able to help him (1981: 100; 111ff.; 140ff.). Then, somebody interpreted a series of misfortunes that befell Favret-Saada during her fieldwork as a sign of bewitchment. She took this diagnosis to a healer who started a therapy. In the process, she became an assistant and apprentice to this healer (1980: 175f. for a short summary). At this point, her first book more or less stops. She does not give us any details about her personal experiences as a healer. The reason is clear even from the first book, and she formulates it much later: her experience was “all but untellable. It was so complex that it defied memorization and, in any event, it affected me too much.“ (2012: 7) Even though she does not write about the healing sessions in the first book, its perspective is clearly shaped by the untold experience.

There are many masterful and fascinating ethnographic descriptions in Deadly Words. Favret-Saada is at her best when she narrates conversations with people who believe themselves to be bewitched. The different episodes of her relationship with the Babin couple, for example, give close insights into the radical fears and hopes of people bewitched. The main part of Deadly Words, however, is not concerned with ethnographic description, but with theory and the attempt to make theoretical sense of information gathered during her earlier fieldwork. The basic theoretical idea is expressed in the title: witchcraft consists in words. Words, sometimes very harmless words, are remembered and taken up when social tensions emerge. When mishap combines with distrust, every word can be taken as a sign of ill will and a proof of witchcraft at work. So witchcraft is not about ritual, recipes or magic books. It is about the link between power and the spoken word. In every word, power can lurk and wait to destroy you; each conversation is part of life-threatening action. To casually speak about witchcraft would be as dangerous and as careless as bringing a candle into a powder magazine.

Click to enlargedobler-2

Photograph by the author

The only way an ethnographer can become a meaningful interlocutor is by becoming involved. She must have her own place in the workings of witchcraft. The real starting point of Favret-Saada’s research thus marks the end of her ethnographic descriptions: the moment she lets herself become seized by witchcraft. Deadly Words describes her transformative journey from intellectual outsider to radically involved insider. Radical involvement, however, was necessary not because of general methodological concerns, but because the nature of witchcraft in Mayenne made it so. Favret-Saada’s theory on witchcraft and her methodological argument are inseparable.

III. Doubts

My first discomfort with Jeanne Favret-Saada’s interpretation of witchcraft came on an early spring evening in 1998, sitting at André and Yvette’s kitchen table on their farm in Mayenne. We had returned from a day in the woods, cutting young chestnut trees for fences, and were chatting while sipping cidre bouché and eating bread with Yvette’s homemade paté. I had come to Mayenne in search of a field site for my PhD, but I soon decided to stick to my initial choice, the Ile d’Ouessant. Still, I stayed in Mayenne for two months, working and living with my peasant hosts. In the evenings, André, a 61-year old, drawing on a familiarity that had grown out of some weeks of common labour, often asked what I thought about things – about the Euro, about people who crossed you in the streets without greeting, about Pre-Concile Catholicism, to name but a few.

On that particular evening, my father called me on the phone, telling me he was about to be hospitalized for kidney stones. This turned our conversation towards healing. André said he no longer trusted hospitals. His magnétiseur, however, was really strong. He had not only made his own calculi disappear, but he was also strong enough to do away with witchcraft. When I seemed interested, André, sometimes seconded by Yvette, told me many stories about witchcraft, often concerning people I knew and sometimes themselves. One of these stories was still on-going and made them rather uneasy.

This did not really square with what I remembered of Favret-Saada’s theories. André and Yvette were certainly concerned about witchcraft; they were also insecure as to what the right interpretation of events should be; but they were not at all reluctant to speak about it. While both of my hosts readily acknowledged the dangers of witchcraft, they did not seem to perceive any danger attached to words about it. Had words about witchcraft become so much less dangerous in the thirty years that had passed since Favret-Saada’s fieldwork?

Back home, I read her book anew, and I became aware of other discrepancies between my own experiences and her account. My doubts grew when I read and learned to admire her second book on her fieldwork, Corps pour Corps (Favret-Saada 1981). Deadly Words gives us a theorized, ‘scientific’ interpretation of witchcraft; Corps pour Corps is an edited version of her early field diaries from 1969 and 1970. It is beautifully written, much more witty and inspiring, and sometimes even more thorough than the more scholarly Deadly Words. I sometimes find her portraits of peasants (let alone priests or medical doctors) condescending; she has little patience for what she sees as sanctimoniousness or insincerity. Yet in reading her diary we come very close not only to her own experience, but also to that of the peasants and rural petty bourgeois she is speaking too. Corps pour Corps has not been translated and is much less well known, but it is the better book.

I found the discrepancies between Deadly Words and my experience mirrored by discrepancies between monograph and field notes. Reading Corps pour Corps convinced me that my experiences might be valid, and that the discrepancies emerged partly from Favret-Saada’s methodological choices, partly from the way she theorised upon her material in a specific scholarly tradition.

So in 2003, when I had a summer off between my PhD and a new project, I decided to go back to Mayenne for another ten weeks and see what I could make of these discrepancies. I partly worked with André and Yvette on their farm again, partly drove around on a bicycle to visit their friends and neighbours, and generally spoke to as many people as possible about health and healing. I encountered many people who believed in witchcraft and some who did not. People who did not know who I was rarely shared personal experiences; those who had met me with Yvette and André were less reluctant. Never once did I encounter fear to speak about witchcraft. Partly, this might have been caused by my status as an outsider unlikely to bewitch them or their farms; but even a group of neighbours involved in a dispute about a common fence once openly discussed local witchcraft cases. The less I asked, the more was usually said.

Click to enlargedobler-8

Photograph by the author

Strangely enough, Jeanne Favret-Saada had the same experience. Generally, she made sure she never did interviews about witchcraft less than twenty kilometres from her own village, wanting to keep a secure space for her family. This methodical choice, she tells us, was justified because villagers, as well, only seek healing from specialists who live far from the village (Favret-Saada 1980: 20f.) – which, although true, seems slightly beside the point. In Corps pour Corps, however, she gives many examples of casual conversations about witchcraft in her own village. In one telling scene, she waits in the queue for the local telephone box when an acquaintance openly speaks to her about witchcraft. She concludes the scene by telling herself: “I definitely should spend a day dilly-dallying in the village every week”.

This short remark sums up what I find problematic in Favret-Saada’s methodology. Her research at first consisted in questions and answers, in prompts and stories exchanged between researcher and informant. She did interviews about witchcraft and realised they were lacking something. The only possibility to overcome the reluctance to talk was to immerse herself into the context in which words about witchcraft were used. In her own words, she had to become affected by witchcraft. There is no doubt that she became radically involved in the healing process, a transformative experience in which she could finally understand words in their context. The context she was affected by, however, was a very special one, and one far removed from the experiences of most people. She never seems to have tried the easy way: to learn about witchcraft by participating in everyday life.

My own short period of research on witchcraft certainly did not reach the depth of affect, understanding and knowledge she acquired by assisting a healer; yet my own very different approach also enabled me to become involved. I worked and lived on a farm and shared the everyday life of normal people. I cut down oaks and mended fences, drew cider into bottles, kept a safe distance from the bull and was part of the interminable feasts of Sunday lunches with the family. André and Yvette, and some of their friends, were quite literally at home with me. Witchcraft was a part of that home, and when the right context emerged, they shared with me their thoughts about it. I did not need to be involved in witchcraft, as long as I was involved in their lives. Even for friends who did not know me, knowledge about my place in society sometimes established familiarity by proxy. So I became affected by something different from what had affected Jeanne Favret-Saada: not by the deadly fight of dewitching, but by an everyday in which the search for healing found its place.

What affected me is absent in Favret-Saada’s work, just as my experiences fail to encompass what affected her. As a consequence of that absence, Deadly Words separates witchcraft from everyday life – first in Favret-Saada’s method and then in theory. I will come back to this essential point towards the end of the article. But in order to substantiate the claim that her method – which is what the book’s fame rests on – was responsible for an incomplete empirical account of witchcraft and for a one-sided theory, I have to go back to the phenomenon in question and to Favret-Saada’s representation of it. Since it is a more authoritative account than my own fieldwork, I use Corps pour Corps to argue my points – even those I learnt about during my own fieldwork.

IV. Witchcraft and its context

Dewitching is only one form of healing unacknowledged by biomedicine. Favret-Saada mentions five principal groups of practitioners: dowsers, who find hidden watercourses using rods or other techniques; toucheurs or cerneurs who use inherited secrets to make warts disappear, heal snake-bites and bee stings or get rid of psoriasis or scab; diviners who are able to diagnose the reason of an illness or of ill luck without necessarily having the powers to fight it; magnétiseurs who heal people or livestock by virtue of their inner force; and finally the désenvoûteurs, anti-witches, who fight witches and the manifestations of their power.

I cannot go into any ethnographic detail regarding these groups. There clearly are vast differences between them, and local people see each of them as a distinct group. Dowsers and diviners only diagnose, toucheurs, magnétiseurs and désenvoûteurs also heal. Toucheurs mechanically apply a technique somehow divulged to them; magnétiseurs and désenvoûteurs use an inner force and thoroughly invest themselves in the healing process. All, however, share a heterodox standpoint. Common sense and the official institutions of French society view them with scepticism. Désenvoûteurs are in this regard the most extreme case: in the eyes of the sceptical majority, they pretend to use unknown powers to heal the imagined consequences of impossible acts.

Among all these practices, Favret-Saada is mostly interested in witchcraft and dewitching, a class of phenomena she sees as clearly distinct from the others. “To touch”, she writes in Deadly Words, “theoretically bears no relation to witchcraft” (1980: 45). By extension, this goes for the other practices unacknowledged by biomedicine, as well. But what exactly does the sentence say? The phrasing is as ambiguous in the original French as in the English translation (“…n’a, théoriquement, aucun rapport avec la sorcellerie” 1977: 84). Does it imply that witchcraft is a separate class of phenomena which our theories have to describe as radically different – or that the difference is discernible only in theory, but gets blurred in practice? I would subscribe to both interpretations, and it is quite possible that Favret used ambiguity in order to express them both.

Yet throughout her books, she is most interested in witchcraft and its theoretical distinctiveness from the phenomena surrounding it. Anxious to distance herself from folklorists who collect superstitions and strange rituals, she stresses that we can only understand witchcraft if we concentrate on its inner logic, and describe it without reference to the surrounding field. Witchcraft is a fight over life and death; its weapon is the inner power of the witch and the désenvoûteur, and its casus belli are words.

I will come back to the war metaphor. For now, I want to emphasize some of the practical relations between witchcraft and the wider field of healing – relations that complicate the theoretical distinction, as well.

First of all, the hierarchy of healers outlined above is very important to understand why peasants set aside their scepticism and decide to seek out a healer who can, perhaps, do away with witchcraft. Dowsers or toucheurs are much less difficult to believe in. The water they find and the warts they remove are concrete, visible things. These specialists do not claim any hidden knowledge or power – they just do the trick. If it works even though there is no explanation, their actions widen the realm of the possible. If people want to convince you (or each other) that witchcraft may exist, they usually cite facts that do not have much to do with witchcraft proper. A dowser who has found a hidden watercourse; a relative who heals burns or makes warts vanish; a magnétiseur who has healed cattle by looking at the floor plan of a stable or who heals ulcers without touching the patients – such stories circulate among neighbours, relatives and friends on many occasions. Healing unaccounted for makes witchcraft more plausible, and diagnoses and treatments at hospitals seem much more arcane to most peasants than the actions of toucheurs or magnétiseurs.

Favret analyses this practical relation between the belief in different healers and the belief in witchcraft. She also shows how different ways of healing interlink in practice. She describes how désenvoûteurs often start out as toucheurs and gradually develop additional powers. In one episode (1980: 41ff.), a patient sceptical about anti-witches agrees to have a toucheur called, who then repeatedly heals him in the manner of a magnétiseur, and finally announces that “there is somebody behind this”. He first takes on the important role of ‘announcer’, turning the witchcraft diagnosis into a socially acceptable reality; after this, he transforms into a désenvoûteur and has family members guess about the witch.

Most dewitchers are in fact mainly magnétiseurs or diviners. There may be some who have stopped their usual practice to specialise in dewitching only, but I have not heard of a single one. The typical désenvoûteur has a general practice as magnétiseur or as diviner, and most patients come because they or their livestock are sick. While acknowledging the importance of other forms of healing in Deadly Words, Favret-Saada stresses their theoretical difference.

A second important context of witchcraft weaves through the pages of Corps pour Corps, but is almost absent from Deadly Words: Religion. Favret-Saada identifies the Catholic Church as a crucial negative force, as one of “the national bodies of ideological control” … “ready to pillory [the peasants] just as soon as an example of witchcraft went wrong and led to some tragedy” (2012: 440; see also 1980: 6ff). The great tradition of Catholicism is an ignorant opponent of the little tradition of witchcraft. While this is certainly true, Deadly Words downplays the links between witchcraft and local Catholicism. She presents the Church as one of the forces people turn to if affected by ill fortune, but one that often remains powerless and has to leave the field to the specialist: “The priest and the doctor have faded out long ago when the unwitcher is called” (1980: 8). Favret-Saada’s own laicism may have played as important a role here as her desire to bring out the special and unique character of witchcraft.

Corps pour Corps, however, makes clear that the Church and official religion are more than a first port of call. Even people who do not want to mingle with the Church think that the local Saints known for their healing power may help in times of illness. (This has not changed today, even though Catholicism in France has transformed greatly over the last forty years. The prayer books in the chapels of Saints, for example, still clearly show the anxieties and hopes for healing invested in religion.)

Religion helps against witchcraft, too. Mme Lebrec, in Corps pour Corps (1981: 201), “confesses that she has never prayed so much as during these times. This non believer became”, in Favret-Saada’s slightly mocking words, “a wholesale user of candles, holy water and masses for personal concerns.” “Before, we did not often go to Church. The children had their First Communion, we went to Church on Easter and communicated, that was it. We had to start it again! I have made a vow that, if I could simply escape it, I would start going to Church again. I’ve not missed a single Mass. Christ, this was really necessary!” (1981: 300) To the wife of a garage owner, returning to the Church seemed a better alternative than seeing a healer: “If we had problems, we would surely go to church again – that is better than believing in sorcerers” (1981: 100).

Rites and blessings of the Church help, even when its officials see witchcraft as mere superstition. In many therapies against witchcraft, prayers, holy water or blessed salt play an important role. Even Mme Flora, Favret-Saada’s own désenvoûteuse, encourages her always to use a lot of holy water (1981: 319). If a priest questioned her about it, she should just tell him her children were ill. Another priest tells Favret-Saada how peasants “have had him”. They thought their farm bewitched and invited the priest to bless it; when he entered, the peasant suddenly realised the priest had not yet seen the farm, and showed him around (a quite common practice for invitees.) The priest only afterwards realised he had been tricked into a magical action: “il m’a fait cerner la ferme” (1981: 13). By circumambulating its boundaries, he had bestowed the Church’s blessing on the farm and prevented evil from entering.

These are just two examples of the context surrounding the practices Favret-Saada is most interested in. Given the clear connections between dewitching and other forms of healing and blessing, why should Favret-Saada stress the theoretical differences between them? She does so to avoid a trap many of her predecessors have fallen into: by grouping witchcraft with other folk beliefs and miscellaneous rituals, they have turned it into harmless material for gossip about popular superstitions. Witchcraft, she insists instead, is far from harmless. Only if we take its radical nature seriously can we understand it at all.

Her identification of power as the core element in the effectiveness of ritual forms is indeed a very important contribution to our understanding of rituals. Evading one trap, however, Deadly Words falls into another: the book separates witchcraft from its context. To use an analogy: we cannot understand brain surgeons by studying family doctors. If we do not carefully study the relationship between family practice and specialised surgery, however, we will no longer be able to understand surgery – nor how people come to believe in it. Jeanne Favret-Saada gives us many examples of the links between different practices, but she stresses difference rather than connectedness. This goes hand in hand with her main interest: she wants to describe the core of witchcraft, the single most important element that sets it apart from all other practices, and that remains if we remove all unnecessary context. Witchcraft is a fight to the death between the dewitcher and the witch (e.g. 1980: 11f.; 194ff). In this fight, one sets his or her inner power against the others. Only one of them can prevail; if the witch is too powerful, the healer will have to stop fighting or she herself will die. Next to words, death for her is the second core element in witchcraft.

V. Death, War and the Meaning of Witchcraft

Here, again, what she sees as the core may look to others as the tip of the iceberg. It might be the purest form of witchcraft, but it is a relatively uncommon one which very few people in Mayenne actually experience.

A monk who had a great reputation as a spiritual healer, even though he was not officially an exorcist, tells her of a different way of dealing with witchcraft. She writes his words down in her field diary, but they do not find any echo in the monograph. When people who think themselves bewitched seek his aid, he asks them about their troubles and makes them join in an “Our Father”. Many who see themselves as bewitched, he tells Favret-Saada, stumble over the words “as we forgive those who trespass against us”. He then advises them to try and forgive. Unless they forgive their aggressors, he cannot do anything for them. Once they are able to forgive, the monk tells the ethnographer, the effects of witchcraft disappear (1981: 185ff).

Click to enlargedobler-1

Photograph by the author

This has very little to do with Corps pour Corps, an eye for an eye. Reconciliation takes the place of a fight for life or death. If we assume (as Favret-Saada does) that nobody is in fact a witch, but that claims of witchcraft are based on an anti-social interpretation of certain words or gestures, reconciliation is as effective a means of neutralising witchcraft’s effects as the annihilation of the adversary. Witchcraft’s social drama may result in reconciliation and reintegration or it may lead to the utmost fissure embodied in the death of the opponent. Both ends make a new beginning possible.

The monk may be an exception, but many dewitchers work in similar ways. One of today’s most successful healers “takes witchcraft on himself”. He never divulges the name of a witch and cannot be induced (at least to my knowledge) to “throw the evil back” on him or her. He states that if he did that, witches would have been successful in their attempts to promote fissure and increase mishap in the world. Although Favret-Saada records similar statements, she does not try to integrate them into her theory of witchcraft.

Two books she read during her fieldwork stand in for, and help illuminate, her fascination with this particular variant of witchcraft: Carl von Clausewitz’s On War and Jules Michelet’s La Sorcière. Clausewitz becomes the metaphor for understanding the fight between good and evil and a model for describing the strategy and tactics involved. Michelet, on whose book she writes a paper to escape the implications of her first meeting with the healer whose assistant she would later become (reprinted in Favret-Saada 1981), seems to have marked her profoundly. Michelet is often read as the classic interpreter of witchcraft as an expression of emphatically female psychological forces and longings. Favret-Saada instead sees in him a savant who recognizes his own image in the witches’ desire for knowledge and the interdiction to know, and who refuses to relegate them to a different order of beings than himself. Michelet “rejects the position of extraterritoriality” (1981: 362). In this, he becomes a role model to her as an ethnographer. Even the title of her own work, “Les mots, les sorts, la mort”, echoes Michelet’s prose: “Elle est sorcière, et fait le sort, du moins endort, trompe les maux” (Michelet 1966: 31); “La mort, la mort, la mort, c’est ce qu’on sent dans chaque ligne” (Michelet 1966: 36).

If confronted with witchcraft, both the ethnographer and the peasant have to decide whether or not to believe in it. The ethnographer, however, has bound her whole career and a good part of her person to a different discourse and a different practice. We cannot believe in witchcraft without betraying this endeavour. We can of course always mask our convictions by explaining that beliefs can create realities; but I do not personally know any ethnographer who would be ready to accept it as a fact that, say, a witch can make mushrooms grow on her neighbour’s field which channel milk from the udders of her neighbour’s cows into those of her own. Yet as ethnographers, we have to take the peasant’s beliefs seriously. The rules of our specific science oblige us to look for sense where, according to the rules of science in general, only error can exist. We have to believe the peasants without believing in what they tell us. (And yes, we have to let our own categories be changed and sometimes shattered by those of the people with whom we live. My own capacity to do so, however, has limits; beyond them, search for truth turns into the telling of interesting stories which I cannot reconcile with my own reality, and which thus have lost any critical potential; pace ontologists.)

In this regard, and only in this, the ethnographer’s task is similar to that of a therapist. She has to take her patients seriously and seek the truth in what they say, but she must not let herself be convinced that the world is really like her patient sees it. Just like many a therapist, the ethnographer will sometimes find herself in situations in which she can no longer decide who is right.

Jeanne Favret-Saada became a therapist for real. This offered her a way to escape the dilemma of whether or not to believe in her interlocutors’ interpretations of the world. When a couple afflicted by witchcraft asked her for help, she offers them a “healing through words”, a psychotherapy (1981: 140). Later she chooses to work as the assistant to a local healer instead. I do admire her for this engagement. Even the most hesitant attempt to find a middle ground between the misery of her subjects of enquiry and her own world view is more humane and more honest than anthropology’s traditional attitude: “I listen to what you say, but don’t expect me to help you or to give any advice where witchcraft is concerned”.

Yet her therapeutic attitude, strengthened by a semiotic orientation in her theoretical background, has consequences for Favret-Saada’s theory of witchcraft. Her guiding questions go directly back to Michelet: “What are [peasants] trying to express by means of a witchcraft crisis?” and “In what way are the bewitched right when they say they are suffering?” (1980: 4; 13). Just like a Freudian psychotherapist unveils fundamental experiences behind a patient’s sentences, the ethnographer has to unveil the something else behind witchcraft beliefs. This, however, cannot be done by merely decoding it and translating it into the language of science. Favret-Saada emphatically argues against a view that sees witchcraft as mere representation, and she questions functionalist interpretations of witchcraft as motivated by envy or possible economic advantage. For her, witchcraft is a tangible expression of what is at play in social situations, and of existential psychological threats – in struggle, death and death wishes.

VI. Methods, Theory and Fieldwork

Favret-Saada’s methodological journey, as she tells it in Deadly Words, begins when she realizes how little she learns from interviews about witchcraft and gradually understands that “in witchcraft, words wage war. Everybody talking about it is a belligerent, the ethnographer like everyone else. There is no room for uninvolved observers” (1980: 10). People will only speak to you if you allow yourself to become involved in witchcraft. Favret-Saada’s own arrival story and the learning process she went through simultaneously prove her theory about the role of words in witchcraft.

Yet the theory takes shape at a very early stage in her fieldwork, in November 1969, when she reads psychiatrist Morel’s thesis on witchcraft in Orme. “One is silent about witchcraft, says Morel, because speaking increases the sorcerer’s power. Magnificent idea. See if it is the spoken word [la parole] which casts the spell, and what writing (the ‘books’) is doing in this.” “One does not accept to speak about witchcraft unless to the person who can undo it. I take this on his word.” (1981: 27).

These words were written long before her conversion to a radically involved ethnographer. Her theory about the role of words in witchcraft remained unaffected by her methodological shift, and indeed by her fieldwork data. This does not invalidate it; we have all experienced early intuitions that later became the core of elaborate explanations. Based on my own experiences in Mayenne, however, and on my reading of Corps pour Corps, I think some parts of her theory actually fly in the face of reality.

To be sure, words are important. The social fact of witchcraft beliefs is to a good extent shaped and changed by words. After a diagnosis, when the victim looks for the culprit, every utterance by a potential witch is meticulously searched for expressions of ill will. Witchcraft produces a state of heightened social alert. The wrong words might lead you to disaster; every slight might be the expression of hidden mortal enmity. Yet this is not constrained to words alone. Gestures, attitudes, facial expressions can be just as dangerous and alarming. Words influence the triadic relation between witch, bewitched and witchdoctor, but they do so along with other, non-verbal means of communication.

So why does Jeanne Favret-Saada insist on the role of words? In the early phase of her fieldwork, she clearly feels the inadequateness of the words spoken in an interview with an outsider intellectual from Paris. In the methodological turn that lies at the centre of her books, she finally escapes that role ascription. She changes sides and becomes entangled in witchcraft. Now, as a victim and later as therapist, her questions about witchcraft make sense to her interlocutors, and they engage her in serious and meaningful dialogues. At the same time, her new role allows her to distance herself from French Anthropology of the 1960s and 1970s, a tradition she wants to escape yet remains influenced by.

Her books present this new role as a radical shift, and indeed it is. Yet two elements of her methods remain unchanged. First, the new role, too, mostly relies on words. She “discreetly recorded” around thirty sessions on audiotape, “in order to constitute a body of material on which I could later work” (2012: 441). Her background and her own preferences induce her to make a methodological choice that in the end leaves her little else to work with but words. Secondly, like the interviews of the first phase, the séances of the second were about one thing alone: about witchcraft.

This is what I realised when sitting at André and Yvette’s kitchen table. After some weeks working on the farm, I was no longer just a foreign intellectual who asked too many questions. I had become part of their everyday life. Talking with me about witchcraft came as naturally to André and Yvette as talking about hospitals or the price of fertilizer. Becoming involved is necessary, but entanglement in witchcraft is not the only way of becoming involved.

So just like Jeanne Favret-Saada, I found a place in the local society that made speaking to me about witchcraft seem a sensible thing to do. This place, however, differed greatly from Favret-Saada’s. The place she found gave her unique access to a certain kind of information on witchcraft. What she saw was shaped by radical involvement, by constant war and by the anxieties of therapy. Every word she heard about witchcraft was uttered as a weapon in this war. From her vantage point, witchcraft indeed had little to do with peasants who removed warts or with priests for whom the very idea of witchcraft embodied superstition and backwardness. Words and death became the centre of witchcraft.

From my vantage point as a farm worker, things looked rather different. Witchcraft was embedded in everyday life. If somebody was ill, he might go to the hospital, to a healer or on a pilgrimage, or try any combination of these. Every conviction about specialist knowledge was temporary. Over the seventeen years I knew André until his death in 2014, he sometimes put his trust in healers, sometimes in hospitals, but he always remained doubtful. He – like most people I met in Mayenne – was never reluctant to speak about witchcraft, if he thought you were interested. « J’en parle à ceux qui veulent le savoir », he told me, « qui veulent entendre – aux autres, c’est pas la peine ». Nobody I spoke to seemed overly concerned with words, whereas rituals were vividly discussed. These were not the rituals we find in early manuals of French folklore, with their vast repertoire of crossroads, new moons, graveyards and such; but rituals paired with the power of those who used or prescribed them.

So between the two vantage points, which one allows us to see witchcraft more accurately? The mere question is treacherous. Only a combination of both, the specialist perspective and the one anchored in everyday life, can really grasp what witchcraft is about. For this very reason, I am not able to offer a new theory of witchcraft. Yet this still leaves Favret-Saada’s theory skewed. Her radical involvement went much further than anything I would be capable of, but in the end it remained incomplete: it only extended to witchcraft and left aside everyday life.

VII. Conclusion

While I was rewriting this article for publication, Favret-Saada’s third book on witchcraft was announced in HAU’s book series in an English translation. Its last chapter has already appeared in English translation in HAU as “Being affected” (Favret-Saada 2012). The chapter is best seen, Favret-Saada tells us, as “the culminating point of a three-volume anthropology of witchcraft, two volumes of which describe in detail the situations in which I, the ethnographer, was ‘caught’, as well as the ways in which I came to understand and accept (or not) the demands made of me” (2012: 437). She takes a very clear position towards anthropological method and links her own experiences to more general questions of ethnographic epistemology. Favret-Saada’s aim is so close to my own wishes for ethnography that I am reluctant to criticize her, lest I be misunderstood as opposing it. The solution she presents, however, seems to me to miss an essential ingredient of an anthropology that lets itself be seized: the involvement in mundane everyday life. Her chapter thus allows me to draw my argument together, trying to do justice to her work and to my own experiences.

Favret-Saada starts out by criticising anthropology. Her representation of anthropology seems slightly unfair. “In the field”, she writes, “my colleagues seemed to combine two types of behavior: an active stance, involving regular work with paid informants whom they would interrogate and observe; and a passive one, in which they attended events linked to witchcraft (disputes, visits to mediums, etc.). The first type of behavior can scarcely be described as ‘participation’; and in the second case, participation seems to mean trying to be present, which is the minimum requirement for observation. In other words, what mattered, for these anthropologists, was not participation, but observation” (2012: 438).

If this describes fieldwork, she has every right to be upset. I do not think it does, though. The meaning of participant observation is to let oneself become entangled in social relations; to learn about work by working, to understand what trust means by trusting and being betrayed, to experience what grief is by sharing the distress of others. Even though I have many misgivings about my discipline, I cannot recognize it in the following description: “[…] ordinary ethnographic communication – i.e., verbal, voluntary and intentional communication that seeks to discover the informant‘s system of representations – is among the most impoverished forms of human communication. It is especially inept at providing information about non-verbal and involuntary aspects of experience” (2012: 443). Indeed. There is no disagreeing with the criticism; I just think it hits a straw man. The strength of ordinary ethnographic communication has always been that, like human life in general, it is so much more than verbal, voluntary and intentional. From Malinowski to Judith Okely and from Marcel Mauss to Maurice Bloch, most anthropologists have seen anthropology’s capacity to engage everyday knowledge and practice as the heart of the discipline.

Favret-Saada then builds on her view of ethnography to pit participation against observation. “If I ‘participated’, fieldwork would become a personal activity, that is to say the opposite of work, but if I attempted to ‘observe’, meaning keeping myself at a distance, I’d have nothing to ‘observe’” (2012: 440). Her choice is evident: “I am struck by the clarity of my methodological choices: everything happened as though I had undertaken to make ‘participation’ an instrument for knowledge” (2012:440).

The idea of an opposition between participation and observation has been a part of anthropology’s methodology for a long time. As ideal types, participation and observation certainly stand for quite different relations to the world we set out to study (see e.g. Förster 2001). And yet, we can only completely separate them in practice if we regard observation as the activity of a complete outsider. Observation without participation can then indeed become meaningless if we want to understand what human beings are doing.
If we are part of a social situation, however, participation cannot exist without observation. Human beings always do things and, simultaneously or consecutively, reflect on them. Not everything we do becomes an object of such reflection; much enters our stock of sedimented knowledge without any conscious observation on our part. Ethnography certainly has to take that side of human knowledge very seriously indeed. We might also temporarily immerse ourselves in a social situation to such a degree that we no longer distance ourselves from the experience so as to observe what we participate in. Yet such oneness never lasts. Sooner or later we are forced, or feel a wish, to observe what is happening. We distance ourselves from the situation by analysing it, evaluating it and transforming, in phenomenology’s terms, pre-predicative experience into predicative. Participant observation is the modus of sociality, not an invention by anthropologists.

Jeanne Favret-Saada sees ‘being affected’ as a precondition for any meaningful ethnography. Ethnography, she writes, should not start from representations, but from the affect itself. She is equally opposed to participant observation and to empathy. Both, for her, are ways of staying outside and guessing what people on the inside feel. Instead, if we allow ourselves to be affected (be it by witchcraft or by any other meaningful human experience), we share an experience of which we can then ask in what way it is meaningful to us and to others. “When two people are affected”, she writes (2012: 442), “things pass between them that are inaccessible to the ethnographer” [i.e. the ethnographer unaffected by witchcraft]. Allowing oneself to be affected, however, has consequences: “one risks seeing one’s intellectual project disintegrate. For if this intellectual project is omnipresent, nothing happens. But if something happens and the intellectual project is somehow still afloat at the end of the journey, then ethnography is possible” (2012: 443).

I agree with Jeanne Favret-Saada that ethnography needs these disturbing and unsettling experiences; only from them can valid descriptions arise of experiences others have had. She does not tell us, however, how to bridge the gap between our own experiences and that of the other – how ‘things pass’ between them and the ethnographer once she is similarly affected. This is of course a general problem of anthropology based on experience. When, say, Loic Wacquant learns boxing in order to understand what boxing means – how can he be sure that his ethnography is about more than just his own experience of boxing? Favret-Saada makes a very convincing case for radical ethnography, but she does not tell us how to find the way between our own experience and the experience of others. Empathy is certainly not enough – but what is?

Rita Kesselring (2015) has recently argued that bodily and other experiences that affect us during fieldwork can form the basis for recognizing what is important in other people’s lives. Shared experiences – allowing oneself to become affected – can become the basis for shared (and not always verbal) references to an overlapping perception of the world. We do not understand the other because we have had similar experiences, but because we realize jointly that we have had similar experiences.

Then, in Favret-Saada’s words, things indeed pass between the ethnographer and the people with whom she lives. Any such ‘passing’ needs a social context that both share. The more we learn of the other by sharing context, the broader the field in which things can pass between us becomes. Favret-Saada has, in her fieldwork, privileged a very narrow context: that of witchcraft alone. She does not seem to consider that fieldwork could actually mean sharing the everyday life of peasants, not out of an interest in witchcraft, but out of interest in them. In her entire account, mundane everyday life is nothing but the context a reader needs to understand witchcraft. It is not the field within which witchcraft exists. Her ethnography, in consequence, is about witchcraft and healing, but it is not about the ordinary lives of concrete people who become implicated in witchcraft.

One final example from “Being affected” can illustrate this. Favret-Saada writes (2012: 440): “I, myself, wasn’t quite sure whether or not I was bewitched. Of course, I never believed in a propositional sense that a witch might harm me with charms or spells, but then I doubt that the peasants did either. Rather, they asked of me that I personally (rather than scientifically) experience the real effects of the particular network of human communication that is witchcraft. In other words: they wanted me to enter into it as a partner, to stake the contours of my then existence in the process.”

I am quite sure that the peasants Favret-Saada met wanted her to be present in a meaningful manner. But I am equally sure that many people I met in Mayenne were actually sure in a propositional sense that a witch might harm them with charms or spells. How can Favret-Saada, in spite of the extent to which she was affected by her fieldwork experience, not care whether witchcraft is indeed real for real people? This is only possible because she was primarily affected by witchcraft, with Mme Flora’s and her own experience of it – not by people’s everyday life in general. Once we step outside the narrowly defined field, we can again inscribe witchcraft into its social context, and words might lose their menace.



Favret-Saada, Jeanne 1977, Les mots, la mort, les sorts. La sorcellerie dans le Bocage. Paris, Gallimard.

Favret-Saada, Jeanne 1980, Deadly Words. Witchcraft in the Bocage. Cambridge, University Press.

Favret-Saada, Jeanne 1981, Corps pour Corps. Enquête sur la sorcellerie dans le Bocage. Paris, Gallimard.

Favret-Saada, Jeanne 2009, Désorceler. Paris, Éditions de l’Olivier.

Favret-Saada, Jeanne 2012, Being affected. HAU: Journal of Ethnographic Theory 2(1): 435-445.

Favret-Saada, Jeanne 2015, The Anti-Witch. Chicago, University Press.

Förster, Till 2001, Sehen und Beobachten. Ethnographie nach der Postmoderne. Sozialer Sinn 3(1): 459-484.

Kesselring, Rita 2015, Why the Body Matters: Moments of Dislocation in Ethnographic Research. Basel Papers on Political Transformations. Online at (accessed February 2015)

Lambek, Michael 1993, Knowledge and Practice in Mayotte: Local Discourses of Islam, Sorcery, and Spirit Possession. Toronto, University of Toronto Press.

Michelet, Jules 1966 [1862], La Sorcière. Paris, Garnier-Flammarion.

Please join our mailing list to receive notification of new issues