Favret-Saada: Death at your heels

This work is licensed under the Creative Commons | © Jeanne Favret-Saada. Attribution-NonCommercial-NoDerivs 3.0 Unported. ISSN 2049-1115 (Online)

Death at your heels:

When ethnographic writing propagates the force of witchcraft

Jeanne Favret-Saada, Ecole pratique des hautes études

Translated from the French by Mylene Hengen and Matthew Carey

Since the publication (in 1977) of her book on witchcraft in the Bocage of western France, the author has received numerous requests to help dewitch people. These requests came from readers convinced, despite the fact that they were raised in a cultural context devoid of witchcraft, they had been bewitched in the same way as Bocage peasants. How could they rid themselves of the series of incomprehensible misfortunes they faced? In this article, the author analyzes three dimensions of “belief”: 1) the prepositional content of a subject’s belief, 2) the range of different possible attitudes towards this content, and 3) a given subject’s shifting attachment to a particular attitude. Beyond the potential cultural differences between bewitched Bocage peasants and readers of an ethnography (who may, of course, overlap), a common, transcultural ground exists, without which the transfer of witchcraft from one world to another would seem incomprehensible. Witchcraft thought’s common core allows us to infer the figure of the dewitcher, the structures of dewitching, and understand why urban readers, convinced they have death on their heels, turn to an ethnographer of witchcraft to save them.

Keywords: witchcraft, dewitching, belief, ethnography

When I left for the Bocage of Western France to study witchcraft, in 1969, I situated my project within the framework of an anthropology of belief. After a few months of fieldwork, I began to question the well-foundedness of this term, insofar as it was used as a concept, supposedly devoid of ambiguity. Throughout the 1970s and 1980s, and despite the best efforts of a few colleagues (Needham 1972, Pouillon 1979, Sperber, 1982) and myself (Favret-Saada 1977), the supposed concept of ‘belief continued to resist our critiques: we were asking the right questions, but nobody knew how else to speak of the term. From the 1990s onward, cognitive anthropology, social psychology and a few branches of the philosophy of mind buried our efforts under a heavy blanket of facts. Meanwhile, the history and the sociology of religion replaced the contested notion of “belief” with a verbally substantiated form, “believing” (le croira), as if referring to the believer’s engagement were sufficient to clarify the notion.

This article was written in 2011, at which point my book Désorceler already existed as a common reference for anthropologists and psychotherapists alike. Among these scholars figured, for the first time, psychoanalysts who had abandoned the idea that the only valid form of therapy was their own. Désorceler built on my earlier work to develop an analysis of the dewitching process.1 The witch-medium was seen to liberate her bewitched patients from a cycle of repeated and incomprehensible misfortunes by progressively modifying their vision of reality.

The fact that this process is accomplished unbeknownst to the bewitched is a central aspect of the dewitching treatment: chapter 4 details the processes of this “unwitting therapy.” The chapter examines how it allows a bewitched family to trade their position as passive victims condemned to ruin and death, with that of social beings capable, when threatened, of enacting a measured form of violence.

Désorceler helps us understand the gradual means by which an accusation of witchcraft comes to be directed against a “neighbour”: the bewitched person is initially convinced that the witch is a close kinsperson. After a few séances, however, the dewitcher shifts them away from this position, and, together (yet for the bewitched, unwittingly), they decide on a new guilty party—this time a non-relative, someone compatible with witchcraft “tradition.” This allows us to renew our understanding of the use of “belief,” in anthropology.

This word, as found in ordinary speech, combines three distinct facets: 1) the propositional content of a subject’s belief, 2) the range of possible attitudes towards this content, and 3) a given subject’s shifting attachment to a particular attitude. (Consistency is strikingly uncommon.)

The plurivocality of the term argues against its use as an analytical concept within anthropology. Instead, an ethnographic account should aim to separately examine the three distinct facets outlined above.

“La mort aux trousses” is based on accounts from non-Bocage readers asking me to dewitch them. Even people who were not raised within a witchcraft paradigm are still inclined to resort to the pragmatic system of dewitching, if they find themselves caught up in repeated and incomprehensible misfortunes.

When an ethnologist finds herself working on rural witchcraft in Europe, she is initially inclined to suppose that people are firm believers, that their conviction is stable, and that there exists a clear distinction between believers and non-believers—to which latter category, it goes without saying, belongs the ethnographer. Unfortunately, this dividing line does not exist, or if it does, it is buried deep within oneself. In any case, it never ceases to re-define itself—including deep within. The only actual differences between the researcher and those she studies are dependent on the strength of the local witchcraft tradition—which, if we take the time to study its history and current modalities, also reveals itself also to be malleable and subject to reinvention.

This situation means that the anthropology of “belief” is essentially an exercise in science-fiction, based on an (apparently ineradicable) conviction that the noun “belief,” the verb “to believe,” or its substantiated version “believing” (le croire) can all be used as analytical concepts. This, however, is not possible, and for a simple reason: “to believe” is an attitude verb, which can express certainty as much as it can express supposition, this to infinite degrees—from quasi-certainty to quasi-skepticism. Hence, to say that someone “believes” X is to have in mind only one of the numerous attitudes that he is capable of having towards X, without specifying which one. And it leads us to assume that this mental state (“to believe” X, with that specific degree of certainty), is stable. We may as well admit it: so long as anthropologists persist in speaking of “belief,” they are nothing more than ordinary speakers exchanging words for the sake of pleasure, rather than social scientists in search of accuracy.

The situation would be less problematic if anthropologists ceased to cling to the idea of “belief” as an analytical concept and took the time to list and describe the thousands of fluctuating native attitudes concerning witchcraft, and then drew their conclusions from there. But that is not what occurs: in practice, and without explanation, scholars insist on pruning the term “belief” until it approximates the univocality that academia demands of its concepts. They assign their informants one attitude, and one attitude only—complete conviction—and then struggle to understand how the natives manage to avoid the torments of cognitive dissonance.

The case of the Bocage peasants demonstrates why this is such a gross empirical error. Now that I have finalized my ethnography of Bocage witchcraft, I can affirm this much: except in rare instances, no one in the Bocage believes in spells with complete certainty (Favret-Saada, 2009).2 Access to a dewitching séance demands only that those requesting it be caught in a spiral of incomprehensible misfortune, and that they not dismiss the spells ’efficacy as inconceivable. Then, the work of the dewitcher proceeds via supposition, save for those exceptional moments I just mentioned. Whenever I drove the bewitched back to their houses, following a séance with Madame Flora during which they had experienced two or three moments of certainty, I was struck by the speed with which they regained their footing—returning to a suppositional stance—with, of course, some variation. Working from these experiences, I reconsidered the content of the dewitching séances and concluded this: the work of the dewitcher consists in making the consultants shift from various degrees of supposition, to a minute or two of certitude, without demanding their conversion to a stable state. This ephemeral certitude, so laboriously obtained and which must each time be regained, is in fact absolutely necessary for the consultants to be able to undergo the existential shift that will free them from the cycle of misfortune.

Let us consider a similar situation: the fascination that fictional or reported cases of witchcraft everywhere provoke. This fascination seems normal enough in those “caught in the spells,” as they say in the Bocage: natives who, we assume are already convinced of the spells ’reality. But it is almost as widespread among non-natives, including among those whose cultural traditions do not include witchcraft. Should we then say that, because they are fascinated, they too “believe”? No, or in any case, not in the same way as bewitched Bocage peasants. How, then? We cannot say. For the implicit pact of fiction (“it’s just a story”) or of ethnographic writing (“they are only natives”), exempts the listener or reader of such a tale from all intellectual responsibility. This is the reason why readers allow themselves to sympathize with these unfortunate heroes struggling within an infernal cycle of misfortune: they are not required to object to the idea that envious sorcerers3 use their supernatural powers to siphon off the strength, wealth and lives of the bewitched, and so they can enjoy the story and its delicious emotional impact. The conclusion is this: allowing oneself to be fascinated by tales of witchcraft does not equate to believing it, in the sense of having the certitude that the tale describes a real state in the world. Instead, it boils down to the allotting of a certain epistemological weight to the supposition that: “maybe, after all . . . ”—in short, believing in it, without knowing for sure, or in a minor key (“what if . . . ?”).

However, in my experience, nothing is more fragile than these pacts between author and reader, for at least two reasons. On one hand, the ideas and conceptions upon which witchcraft rests are not primarily statements about the world or on the nature of the beings that populate it: their essential purpose is based on a strategy for escaping the cycle of incomprehensible misfortune (this is why I spoke of “witchcraft therapy”). This is also why some readers feel, for whatever reason, the need for this kind of therapy. Some write to me asking me for my dewitcher’s contact information, or if I know any other reliable specialists; others (including doctors, academics and even some well-known researchers at the CNRS) asked me, quite matter-of-factly, to dewitch them. Conclusion: confronted with an incomprehensible cycle of unfortunate events, anybody—whatever his culture—may adhere to the ideas of Bocage witchcraft.

What is more, the Bocage ideas of witchcraft address—in their own particular manner—the universal demands of life in society. Francois Flahault’s (2001) work on fairytales has demonstrated that these ideas are present in all human societies, including pre-literate ones. They are present in the form of tales whose plot, characters and motifs remain surprisingly stable across space and time, though of course the ways in which they are told vary from one region to another. Examined as a whole, the tales are evidence of the social labour involved in reminding successive generations of the exact clauses of the “social contract.” I use the term “social contract” as a sort of commodity, in order to highlight the fact that the notion of reciprocity (Lévi-Straussian “exchange”) is the cornerstone of social life—a claim which, as we shall see, is also at the heart of witchcraft thought.4

Reduced to its core—to the small number of ideas indispensable to a discussion of witchcraft—this style of thought establishes certain fundamental guidelines of social interaction that the witch is supposed to transgress.5 From this schema, it is then easy to infer the characteristics of the dewitcher and the nature of his action.6Granted, this ideational kernel is specific to one particular region—the Bocage of Western France-—and to one type of economic activity—family businesses dealing with living beings. But many of those who wrote to me following the publication of my books had identified with the books ’content, seeing themselves as victims of repeated and incomprehensible misfortune. They had understood that, despite local curiosities, Bocage witchcraft thought formulated certain universal facts of life in society, and they were asking me to lay out, from my experience as an ethnologist, the plan of action they so needed in order to recover. Their identification was in fact facilitated by the fact that the discourse of Bocage witchcraft focuses exclusively on the victims ’point of view, and on that of their champions, the dewitchers: even in narratives internal to the witchcraft tradition, they are the only possible narrators—witches wall themselves up in a silence broken only by the occasional snicker.

The notion of “force,”7 as it is used in the following table, calls for two comments. First, I settled for opposing a lower case “force” to a capitalized “Force,” as natives do not have words to distinguish between the two, merely phonic or gestural markers (a certain click of the tongue, a worried look over the shoulder.) Second, Bocage natives never specify what exactly this “Force” consists of, where it comes from, why it is present or absent, and how those who have it, acquire it.8 For them, it is enough to affirm that it is present in our corporeal world, and that two types of people—witches and dewitchers, manage to channel enough of it to allow them to produce empirically noticeable effects.

Five pairs of antagonistic traits define the bewitched in relation to the witches, and situate them within fields that tend to be distinguished by academic thought: morality, ontology, and psychology in the broad sense of a hexis, a stable way of being. Each of these actors is entirely identified with the characteristic in question: the bewitched is only goodness, whereas the witch is only malice, etc. The moral of the story could be summarized thus: any human who aims for the Good, and whose desires are limited to the shared human world, runs the risk of being reduced to passivity and weakness, because there exist, in his surroundings, people who seem just like him, but who are endowed with almost infinite avarice and incomparably superior powers. Stated in more general terms: one cannot be simultaneously good and strong, good and active, good and prosperous, good and, in the long term, alive. So long as there exists a Force to which only a few have access, human action is reduced to the alternative: “kill or die.”

Table 1. The Bocage version of a witchcraft attack
Pairs of oppositions marking the axes The bewitched
A being respectful of the rules of reciprocity.
The witch
A systematic transgressor of the rules of reciprocity.
Axis 1:
Good vs Evil
The bewitched declares himself entirely Good, going so far as to turn the other cheek to his adversaries. The witch prefers the ease of evil to the strictures of the Good.
Axis 2:
force vs Force
The bewitched has only normal “human” force, which he invests in his family and his farm. The witch has access to a Force that we will classify as “magic,” for lack of a better word, as it operates without mediation or delay.
Axis 3:
Limitedness vs Illimitableness
Finiteness vs Infinity
The bewitched is content with his lot.
His desires are inscribed within the boundaries of a shared world, where ontologically similar beings coexist, each one gifted with a limited force, whose origins are identifiable.
The witch experiences an infinite amount of greed for the goods of others, even if he does not need them; he lives in a world where there is no room for another, in a world of illimitable wholeness.
Axis 4:
Visibility vs Invisibility
The possessions, force and actions of the bewitched are visible to all: he has nothing to hide.
N.B.: up until now, this section has described the condition of an ordinary farmer; one who has not yet had any contact with a witch.
The witch’s evil spells are invisible, as is his Force: it is only perceptible via its effects. However, it circulates via ordinary means of communication (the gaze, speech and touch).
Axis 5:
tendency to Passivity vs tendency to Action
The bewitched can only passively suffer. He is prey and can only deplore the damage produced in his domain by the “force” of the witch. The witch behaves as an active predator, always greedy for the possessions, health and lives of others. Were he not subjected to his own Force, which compels him ceaselessly to commit acts of harm, he would be the Master of Evil, of Life, and of Action.
The sum of an actor’s positions on these axes defines the issue at stake. Good, force (lower case), Limitedness, Visibility, Passivity = vulnerability, weakness, ruin, death. Evil, Force (upper case), Illimitableness, Invisibility, Activity = invulnerability, Force, monopoly, murder.
Consequences of the contact between these two sets of structural oppositions The bewitched is steadily drained of his force. Ultimately, this spiral of loss leads to his ruin and/or death. The witch benefits from the force he extracts from his victims. Once he is done with one victim, the witch turns to the next. The waxing of his powers is, therefore, unending and may even lead to eternal life. As the saying goes, “The Good Lord doesn’t want him, and nor does the Devil.”

One need only recall this schema to understand the situational shift that takes place when a bewitched person appeals to a dewitcher. For the dewitcher channels together Force and the Good, bringing about an intersection between two forms of two forms of absence or deficiency: the bewitched’s lack of Force and the witch’s lack of Good9. So, unlike the bewitched: the dewitcher is good but not too good, he is not reduced to a position of passivity, and his other capacities (Infiniteness and Invisibility), allow him to combat the witch on his own turf. Finally, neither witch nor dewitcher is, in any sense, absolute Master of Force. Both have access to a certain amount of Force, and this is what will allows them to pit their respective strengths against one another, with unforeseeable results.

Whilst an anthropologist might, as a matter of habit, look to identify the Bocage dewitcher’s symbolic legitimacy, such an approach would inevitably draw a blank. Unlike, say, the rituals of affliction described by Africanists in the 1960s, the dewitcher’s activity not does require a process of initiation, which publicly confirms that a transmission of knowledge and symbolic powers has indeed taken place. A dewitcher’s legitimacy depends on the access to the Force he displays in front of his clients—i.e., on their perception of it during the séances. They note the dewitcher’s exceptional character, his ingeniousness, and the intensity with which he tackles the cases brought to him. And behind this individual stands a tradition that describes the limits and contours of his action.

It is worth noting that the rituals of affliction described over the past twenty years also, in fact, lack symbolic legitimacy—much like dewitching. In contemporary Africa, both in major cities and in a countryside unsettled by wave after wave of displaced populations, there are nowadays too many competing sources of authority for any particular form of anti-witchcraft activity easily to claim legitimacy: self-proclaimed dewitchers calling themselves “traditional healers,” old-fashioned countryside healers (now often disdained), prophets, Pentecostal or Catholic priests, etc. Indeed, one ends up wondering whether the idea of symbolic legitimacy had in fact any traction “back in the day,” during the golden age of hermeneutic anthropology, or whether it was merely a mirage of twentieth century social science.

We now understand why victims of repeated and incomprehensible misfortune, whatever their cultural tradition, can find a hold in Bocage witchcraft: those who have death at their heels cannot afford to fuss over cultural differences. They stop thinking about bewitched peasants as savages and easily identify with their representations of witchcraft. Because these representations do not, at heart, separate morality and ontology from psychology, people latch onto them in the hope that they will allow them to rebuild both themselves and a new world rid of the cycle of misfortune. And this leads them straight onto the thing that matters most: the end result of dewitching, an end to their misfortune. They are, for the most part blind to the work of the dewitcher: that of introducing, unbeknownst to the patient, a chiasmus between Force and the Good and then of teaching the patients to reproduce this chiasmus in their relationship with the witch, as well as in their daily lives. And they are just as blind to the fact that in order to escape, they will have to dip into the well of illimitableness (i.e. into the pit of hatred and aggression) that everyone carries within.

In a more general manner, Bocage witchcraft seems both to provoke a fundamental suspicion about the frailty of the social contract, and simultaneously to provide a means of consolidating this latter through a perpetual activity of group and self-reconstruction—activities whose outcomes are in fact equally precarious. In sum, perhaps the Bocage has given rise to a cultural complex that cultivates modest hope as to the solidity of the social, and which draws on a minimal ontology, rather than one of these enchanted native worlds, with their ultra-sophisticated thought and their luxuriant “forests of symbols” that anthropologists so unfailingly discover and then exalt.


Boyer, Pascal. 2001. Et l’homme créa les dieux. Paris: Gallimard.

Favret-Saada, Jeanne. (1977) 1997. Les Mots, la mort, les sorts. La sorcellerie dans le Bocage. Paris: Gallimard.

———. 1980. Deadly words: Witchcraft in the Bocage. Translated by Catherine Cullen. Cambridge: Cambridge University Press.

———. 2009. Désorceler. Paris: Editions de l’Olivier.

Flahault, Francois. 2001. La Pensée des contes. Paris: Anthropos-Economica.

———. 2010. “Le sorcier, à la manière d’un vampire...” L’Homme 193: 185-194.

Needham, Rodney. 1972. Belief, language and experience. Oxford: Basil Blackwell.

Pouillon, Jean. 1979. ”Remarques sur le verbe ‘croire’.” In La fonction symbolique, edited by M. Izard and P. Smith, 43–51. Paris: Gallimard.

Sperber, Dan. 1982. Le savoir des anthropologues. Paris: Hermann.

La Mort aux trousses : Quand l’écrit ethnographique propage la force sorcellaire

Résumé : Depuis la parution de son livre sur la sorcellerie dans le Bocage (1977), J. Favret-Saada a reçu de nombreuses demandes de désorcèlement. Elles venaient de lecteurs convaincus d’être ensorcelés selon le même schéma que les Bocains, bien que ne disposant pas de la sorcellerie dans leur héritage culturel. Comment rompre la série des malheurs incompréhensibles auxquels ils faisaient face ? Dans cet article, l’auteure analyse trois dimensions de la « croyance » : 1) les contenus de pensée auxquels le sujet est dit « croire », 2) la variété des attitudes envers ces contenus, et 3) les fluctuations d’un même sujet dans son investissement d’une attitude donnée. En-deçà des différences culturelles évidentes entre les ensorcelés bocains et les lecteurs d’une ethnographie, il existe une base commune transculturelle sans laquelle ce transfert de la sorcellerie d’un monde dans l’autre serait incompréhensible. Ce noyau de la pensée sorcellaire permet d’inférer la figure du désorceleur et les linéaments du désorcèlement, et de comprendre le fait que des lecteurs urbains convaincus d’avoir la mort aux trousses fassent appel à l’ethnographe de la sorcellerie pour les tirer d’affaire.

Jeanne Favret-Saada is an anthropologist and author of Les Mots, la mort, les sorts. La sorcellerie dans le Bocage (1977, Gallimard; English translation, Deadly words: Witchcraft in the Bocage, translated by Catherine Cullen, Cambridge University Press, 1980) and Corps pour corps. Enquête sur la sorcellerie dans le Bocage (1981, Gallimard, co-authored with Josée Contreras). She has since published numerous other works, including Le Christianisme et ses juifs: 1800-2000 (Le Seuil, 2004), once more in collaboration with Josée Contreras, Algérie 1962-1964: essais d’anthropologie politique (Bouchene, 2005), Désorceler (Editions de L’Olivier, 2009) and Jeux d’ombres sur la scène de l’ONU. Droits humains et laïcité (Editions de L’Olivier, 2010).

Publisher’s note: this article is a revised and refereed version of an article that was originally published in French under the title: “La mort aux trousses,” in the journal of psychoanalysis Penser/rêver (Editions de L’Olivier), n° 20, “Le temps du trouble,” autumn 2011. Hau is extremely grateful to the journal for allowing us to translate and publish it in English.


1. The term dewitch (désorceler), and its derivatives, has also been translated as unwitch, notably in the author’s book Deadly words: Witchcraft in the Bocage, translated by Catherine Cullen (Cambridge University Press, 1980).

2. Bocage peasants use the words “désorceler,” “désorcèlement,” “désorceleur,” rather than the standard French: “désensorceler,” ”désensorcèlement” “désensorceleur.” For this reason, I will use the terms “to dewitch,” “dewitching,” “dewitcher.”

3. In the Bocage, no one casts spells, but some peasants say they have been struck by them.

4. “Witchcraft thought” is a handy substitute for my own interpretation of conversations I noted during my stay in the Bocage.

5. Bocage people, having not read Evans-Pritchard, do not make a distinction between “witch” and “sorcerer”: they avoid any precision in this respect, speaking of the witch or sorcerer as: “The one who has done this to me.” (Cast a spell.)

6. François Flahault (2010) recently published an article on Dewitching (Désorceler): “ The witch, just like a vampire...” This text helped me to formulate the argument I present here, though it differs in a number of ways. Specifically, Flahault does not seem to have grasped how important it is that the witch be made aware of the bewitched’s accusations and that he react to them. Even if the accusations reach him indirectly, the witch never ignores the fact that he has been accused. It follows, contra Flahault (p.188) that the structure of the communication is quite different from that of gossip, and is instead one of true confrontation.

7. Bocage natives speak uniformly of “force,” using it to describe physical, moral, or magical strength, the energy necessary to keep oneself alive, the ability to produce children, the ability to keep one’s family alive, the ability to manage the productive-reproductive potential of one’s farm.

8. As Pascal Boyer (2001) argues, natives are not theologians responsible for defending their doctrines.

9. The establishment of a chiasmus between Force and the Good is witchcraft’s fundamental driving force: cf. “l’embrayeur de violence” in chapter 4 of Désorceler.