Descola: Beyond nature and culture: The traffic of souls

Publisher’s note: This article is chapter 14 of the forthcoming translation by Janet Lloyd © by The University of Chicago. All rights reserved. Published 2013. The University of Chicago Press, Chicago 60637. The University of Chicago Press, Ltd., London. Originally published as Philippe Descola, Par-delà nature et culture (Paris: Editions Gallimard, 2005) © Editions Gallimard, Paris, 2005. We are very grateful to The University of Chicago Press for permission to publish this chapter.

Beyond nature and culture

The traffic of souls

Philippe Descola, Collège de France

Translated by Janet Lloyd


Between identification, a means of specifying the properties of existing beings, and relations, a means of specifying the general form of the links between those beings, two kinds of connection are possible. Either the plasticity of a relational schema makes it possible for it to structure interactions in a variety of ontologies, which will then present a family likeness despite the heterogeneity of their essential principles; or, alternatively, one of the modes of identification is able to accommodate several distinct relational schemas and this introduces into an ontological configuration widely distributed in space (a cultural region, for example) the kind of concrete diversity of customs and norms from which ethnologists and historians love to draw their material. The second case is what we shall now be considering. However, the combinations made possible by the conjunction of a mode of identification and a relational mode are too numerous for us to consider them all in a systematic and detailed fashion, especially since some of them turn out not to be possible for reasons of logical incompatibility, as we shall soon see. So let us limit ourselves to considering the variations of ethos that various relational schemas imprint upon one particular mode of identification: this will be animism. The demonstration will certainly not be complete, but it will at least provide the beginnings of a proof that anthropology can always hope to find when it enters into some detail in a comparative study of a number of cases. As Mauss, mobilizing John Stuart Mill in his support, declared, “a well made experiment is enough to demonstrate a law” (Mauss 1950: 391).

If I have chosen animism for this experiment, that is because, in one of its geographical variants, it raises an exemplary problem in the interpretation of the question before us. Whatever theoretical line they take, all the specialists on the Indians of Amazonia sketch in an ethnographic picture of the societies that they study in which the features of the animist ontology are easily recognizable. However, that is no longer the case when it comes to describing a specific style of social philosophy that is valid for the whole of Amazonia. Here, total disagreement reigns, with each anthropologist tending to project on to other peoples of the region the values and practices that he or she has observed in one particular ethnographic context. Ever since Lévi-Strauss, in one of his earliest articles, drew a parallel between inter-tribal trading and warfare in the lowlands of South America, it has become common to say that the paradigmatic relationship in this region is one of exchange: men exchange marriageable women (the model being a swap of sisters), goods (often identical ones) and the dead (in vendettas and warfare); women exchange among themselves plant cuttings, foodstuffs and tamed animals; chieftains exchange the right to polygamy in return for a duty to be generous; hunters exchange offerings to the animals that they hunt in exchange for their meat. In short, everything seems to circulate in an unending round of reciprocity (Lévi-Strauss 1943). More recently, as we have seen, some anthropologists have laid the emphasis on an altruistic variant of exchange, defining Amazonian sociability as a mutual production of persons amid generous conviviality, while others have, on the contrary, insisted on the cannibalistic dimension of the incorporation of others, as a typical mode of interaction. Should we accept, along with Viveiros de Castro, that “generalized predation” is “the prototypical modality of Relationship in Amerindian cosmologies” (Viveiros de Castro 2002: 164), or should we believe Overing and her disciples, who regard an intimacy based on sharing as the dominant feature of the Amazonian socius? Is the ethical horizon of these populations a fair exchange between partners of equal status, an ideal “togetherness” irrigated by mutual help and gifts? Or is it a bellicose seizure of others? The self-evident answer is that all these relational modes are certainly present, but distributed in different collectives.1

Predators and prey

I shall seek an illustration of what predation may amount to once it becomes a dominant relational schema by turning to a people that has gradually become familiar to my readers. Until they were “pacified” by missionaries between 1950 and 1970, the various Jivaro tribes were reputed to be of a bellicose disposition and seemingly anarchic in their collective life. Their ceaseless wars were a source of perplexity to observers and a motive for anathema. Yet they did not indicate any disintegration of the social fabric or an irrepressible propensity for violence. On the contrary, they constituted the principal mechanism for structuring individual destinies and links of solidarity and also the most visible expression of one key value: namely, the obligation to acquire from others the individuals, substances and principles of identity that were reputed to be necessary for the perpetuation of the self. Headhunting among the Jivaro tribes—the Shuar, the Achuar, the Huambisa and the Aguaruna—and likewise the unending vendettas between members of the same tribe were, in effect, expressions of one and the same need to compensate for every death within a kindred group by capturing real or virtual persons from close or more distant neighbours. Shrinking the heads of enemies made it possible, by means of a long and complex ritual, to strip the dead person of his original identity in order to transfer that identity to the murderer’s local group, where it would become the principle for the production of a child yet to be born. By dint of shrinking the head, which preserved the dead man’s physiognomy and, along with it, his individuality, the victorious warrior captured a virgin identity that would allow his kin to multiply without incurring the obligations inherent to a marriage alliance. Consequently, the enemies who were beheaded had to be neither too close nor too distant since they provided an identity that was culturally usable yet at the same time perceived as different: they were invariably Jivaros, but were selected from a neighbouring tribe that spoke a different dialect and with which no relationship of kinship had been established in the recent past.

Vendetta warfare did not involve capturing heads, but the principle that governed it was nevertheless identical. Whatever the vengeful motives involved in order to spark off an armed confrontation between two kindred groups, the assassination of an enemy belonging to the same tribe in effect often led to the seizure of his wives and young children. The wives took their place alongside the victor’s earlier wives, while the young children were adopted and treated by him as his own offspring. Thus, even if the capture of the women and children of neighbouring local groups was never an explicit and sufficient reason for undertaking a vendetta, it was in many cases an expected or even hoped-for outcome. For the victorious warrior, the advantages gained were twofold: the death of his enemy was regarded as payment for a real or imagined slight and, at the same time, he enlarged his domestic group without incurring the obligations of reciprocity upon which a marriage alliance was founded. To be sure, both headhunting and vendettas were likely to lead to reprisals, but these were obviously not sought for as such and efforts would be made to avoid their consequences. The violent and reciprocal appropriation of others within the Jivaro group was thus the product of a rejection of pacific exchange, not a deliberately engineered result of an exchange of human lives in the course of a bellicose interaction.2

Furthermore, both vendettas and head-hunting were carried out against persons that the Jivaro classified as affines even if, in actual fact, the enemies killed in the fighting might be consanguineous or, on the contrary, have no genealogical link with the murderer. A few words about social organization will help to explain this identification of enemies with relatives by marriage. The traditional Jivaro habitat is widely dispersed, with each house belonging to a single, very often polygamous family, and constituting an autonomous, political and economic unit separated from neighbours by distances that it takes between a few hours and one or two days to cover, either on foot or by canoe. Here and there, though, one comes across larger local groups comprising ten to fifteen houses strung out along a river, the members of which are more closely linked by consanguineous kinship and marriage alliances. The latter follow the rule of union between bilateral cross cousins. Now, like other Amazonian societies of the same type, these small endogamous networks tend to regard themselves as ideal consanguineous communities, for the links of affinity within them are in practice obliterated as a result of manipulations of the kinship terminology. These tend to divide affinity and consanguinity between the sexes in such a way that an exclusively masculine affinity is matched by a marriage alliance based on paradoxically consanguineous unions.3 By dissociating affinity from actual marriage, the Jivaros give themselves the means to convert it into a logical operator for thinking through relations with the outside world, as can be seen, for example, in the practice of transforming into affines consanguineous relatives if these reside outside the endogamous network. The local group’s Utopian closure on itself in effect presupposes a symmetrical opposite: namely an affinity that is clearly objectivised, given that it is free from any consanguineous contamination. Although relations outside the endogamous network are usually hostile, they are graduated according to the scale of social distance or relative otherness. This finds expression in the form of a schematization that is increasingly marked by the affinity relationship the further one moves away from the focal point where it effectively orientates the marriage alliances.

Internal wars usually break out following conflicts between local neighbouring groups over some real or supposed infringement of the rules of marriage alliance. When a quarrel breaks out within the endogamous network over matters linked with rights over women, payment of compensation and the mediation of a great warrior generally suffice to prevent the outbreak of a vendetta between close kin. If an amicable arrangement proves impossible, it is usually because the guilty party or the victim of the infraction comes from another endogamous network. For endogamous closure is an ideal. In fact, though, thanks to a strict application of the principle of uxorilocality a variable percentage of exogamous unions always make it possible to introduce into a local group men who are natives of a neighbouring network. These foreign sons-in-law find themselves in a difficult situation to the extent that affinity instituted by alliance with distant kin is far looser than the more fundamental affinity instituted by a prescriptive exchange. So when a serious incident occurs, the transplanted in-law naturally enough tends to flee and seek help and protection among his direct consanguines. Through marriage alliances, each local group thus maintains a tenuous network of links of affinity with adjacent groups that may serve as the basis of a temporary coalition or, on the contrary, provide the pretext for a factional confrontation. In short, in these conflicts between neighbours, the enemies are unequivocally identified as real affines, who are sometimes described collectively as “givers of women.”

In contrast to a vendetta, inter-tribal warfare has as its sole objective the capture of heads from neighbouring Jivaro tribes in order to celebrate the tsantsa ritual. The difference between “ordinary” Amazonian trophy-heads and the shrunken Jivaro heads is that the former rapidly lose any traces of a specific physiognomy, whereas the latter—for a while, at least—perpetuate the unique representation of a face. That is the sole objective of extracting the skull, desiccating the tissues and modelling the features so as to obtain a resemblance of the victim. When the tsantsa is produced, its role is that of an easily transportable condensed identity. However, the tsantsa is not a miniature effigy of a particular person, but a formal expression of a purely existential individuality indicated by no matter what facial distinctive trait, provided the head is that of a nonrelated Jivaro. For the Jivaros, an individual identity is contained not so much by the physical features of the head, but by certain social attributes of the persona: a name, speech, the memory of shared experiences and face-paintings. In order to be used in the ritual, the tsantsa must therefore be relieved of any referential residues that might remain to prevent it from embodying a generic Jivaro identity: it is never called by the patronym, if indeed that is even known, of the one whose head has been taken; its face is carefully blackened to obliterate the memory of the patterns painted on it; and finally all its orifices are sewn up, thereby consigning the sense organs to an eternal phenomenal amnesia.

The depersonalization of the tsantsa renders it suitable for a rite the discontinuous phases of which extend over rather more than a year. In the rite, the tsantsa functions as a logical operator—both as a term and as a relationship—in a series of permutations between terms and relations that are themselves affected by variable values. First called “profile,” then “soft thing,” the head either simultaneously or consecutively occupies different positions, from the point of view of gender and kinship, in a series of univocal or reciprocal relationships, which may be either antagonistic or complementary, with the killer, his kin and affines of both sexes and a number of other ceremonial groups. By the time this topological ballet is completed, the tsantsa has played every role in a symbolic procreation: non-parent, giver of women, taker of women, wife and finally embryo. The very real fruit of this simulated alliance—a child to be born within the murderer’s kindred group—is thus perfectly consanguineous without being incestuous. As a virtual existence obtained from strangers, the child owes his procreation to the staging of an ideal affinity, the only kind truly satisfactory for the Jivaros, because free of any obligation of reciprocity: this is, in short an affinity without affines. Seen from this point of view, this inter-tribal war is really indistinguishable from the intra-tribal warfare of which it constitutes a logical, or even historical extension. For repeated confrontations between coalitions of different blocks of local networks can only consolidate antagonistic regional identities, thereby contributing to the continuous process of tribal differentiation that is necessary for the perpetuation of head-hunting. Between stealing women and children from potential affines who have been excluded from the kinship community and stealing identities that will produce children from non-kin with whom one simulates an ideal affinity, the difference is one of degree, not of nature.

Whether waged against close enemies or distant ones, Jivaro warfare is the motor for the fabrication of collective identities. In a society without chiefs, without villages and without lineages, it renders possible a temporary coagulation of factions, a renewal of solidarities that have slackened as the result of such a dispersed habitat, and a stimulation of the social link brought about by the federating sensation of sharing a common enemy. It is through that warfare that groups of relatives acquire their substance and the principles for their renewal, by means of poaching persons and identities, all of them rare and precious, from affines either real or symbolic, who are treated as prey. To be sure, armed clashes are not permanent, but in everyone’s mind warfare is always present. At any moment a smouldering conflict is ready to burst into flame, providing the main topic of conversation and orienting the political dynamic of alliances and the interplay between factions. In a society in which the word for “peace” is unknown and the only collective rituals are those that announce or conclude the exercise of collective violence, warfare is by no means an unfortunate accident: it is the very stuff of social life.

Likewise, it is through warfare that individual masculine identities are forged. As soon as boys reach adolescence, they are pressed to enter into contact with an arutam spirit, in the course of a visionary trance induced by severe fasting and continuous absorption of green tobacco juice and other hallucinogenic liquids. This terrifying experience enables the adolescent to establish a personal and secret relationship with the ghost of a deceased Jivaro warrior who will pass on to him his strength and protection. Arutam first appears in a frightening guise—a glowing head jerking from side to side, a couple of intertwined giant anacondas, or a gigantic harpy-eagle—which noisily disintegrates as soon as it is touched, then returns in human form in order to deliver a message of assistance. The young man will from then on identify with his arutam, in particular by painting his face with red dye in a design that recalls the monstrous figure in which the spirit first revealed itself to him. The immediate effect of this identification is an irrepressible desire to manifest the bravery unleashed upon him by his encounter with his protector-spirit, by plunging wildly into warfare. However, the quest for arutam needs to be regularly renewed, for the power that a man obtains from it disappears every time he takes part in a victorious expedition or kills an enemy. It then leaves him defenceless. Since the physical survival of a warrior is subjectively dependent on his ability to restore his skill at killing, the mechanism of the acquisition and subsequent loss of arutam thus contributes to a kind of uncontrollable increase in his individual propensity to accomplish his destiny in the exercise of violence.

The predatory attitude that the Jivaros manifest in their relations with others, the need that they feel constantly to incorporate the bodies and identities of their neighbours in order to persist in being themselves, even while being partly determined by that which they capture and assimilate, and their stubborn rejection of any freely accepted reciprocity: all these are traits that reappear in their relations with non-humans. In this domain, the Jivaros set a higher value on their violent appropriation of substances and fluids than on the free play of their circulation. Yet, as we saw at the beginning of this book, many plants and animals are regarded as persons who share some of the ontological attributes of the humans with whom they are linked by relations of consanguinity and alliance. However, non-humans are not integrated into a network of exchange with humans and they are allowed nothing in exchange when their lives are taken. To be sure, the Jivaro hunters do address anent incantations to the game that they hunt, to the spirit-masters of the animals and to the prototypes of each species, so as to establish with them a relationship of connivance: hunting is regarded as an expression of the complicity between relatives through marriage alliances, in which the ultimate end, the killing, is masked by ludic formulae. Hunting anent are absolutely explicit in this respect: the animals are always described as brother-in-law with whom one communicates in the slightly jokey tone of forced affability that is usual in such a relationship; and sometimes the sisters of the hunted animal are even referred to as potential wives for the hunter. But treating one’s prey as an affine is really nothing but a deceit designed to disguise the basically inegalitarian nature of the relationship between the men and their animal victims. The point is to allay the mistrust of the animals so that they will not elude the hunter’s darts or make him pay for his cannibalistic intentions. As in many societies in which hunting plays a predominant role, it is not unknown for excesses to be punished. If one kills more game than is needed, one risks a snakebite or a fatal accident in the forest. But in such a case, this is purely revenge on the part of the animals—or rather their master-spirit and protector—and is designed to punish a hunter’s hubris; there is no question of it being a process of voluntary exchange founded on parity between the two parties.

Even relations with plants are not free from this predatory ideology, so one should not see it as a simple rationalization of the productive destruction that characterizes any form of hunting. Manioc, the main foodstuff for the Jivaros and the most common plant in their immediate environment, is reputed to suck in through its leaves the blood of those who brush by them, but it mainly attacks the women who cultivate it and also their young children. It is a threat that is not taken lightly and the death of a baby is often attributed to anaemia provoked by manioc vampirism. Consequently, the women have to sing special anent incantations to this plant, in an attempt to switch its thirst for blood toward other, undesirable, visitors to the garden. The women treat the manioc as a child, but one who will eventually be eaten by those who have raised it. Meanwhile, the manioc is itself a child that seeks to bring about the death of human children whose sole nourishment for several years is, precisely, constituted by a kind of manioc porridge. Beneath its benign appearance, gardening in truth implies a mortal competition between the human and the non-human young. For the women, it is a matter of reproducing and raising young plants, whose flesh the humans will consume, meanwhile taking care to prevent the manioc plant from retaliating by consuming the blood of the human young who come into contact with it.

The capture of real or virtual persons from close or distant enemies, the furtive seizure of game and the cunning warfare against the cannibalistic manioc thus all, in different domains, express an identical rejection of exchange in relations with others. This predatory tension is what structures the relations that the Jivaros maintain with a whole mass of subjects of many different kinds, in that it integrates their experience of the world in many domains ordinarily distinguished by the misleading analyses of dualism and it is applied, without distinction, to both humans and non-humans, to both kinship relationships and to techniques of subsistence, and to both territorial organization and ritual. One property of relational schemas is to embrace vast areas of practice without discriminating between terms according to their ontological status or the situations in which they relate to one another. These schemas are thus at the source of the stylistic effect perceived by an observer of a “culture” that is different from his own. It is an ineffable and perhaps illusory feeling, but it can be traced back to the thematic patterns of behaviour that feeds the stereotypes that every group of humans adopts toward its neighbours.

This example presents an opportunity to return to consider the way in which an ethos comes to be incorporated as a way of acting according to behavioural principles that are, however, never made explicit. For the schema of predation upon affines is not regarded by the Jivaros as an explicitly transmitted norm. Given that the concepts of predation and affines are not expressed by any words in their language, their tendency to behave toward others in this way is something very internalized that has become implanted, as time has passed, ever since their earliest days and has been constructed not so much through the assimilation of a system of “collective representations” as by successive inductions based on constant observation of the conduct of adults. There are plenty of opportunities for children to be alerted to the behaviour patterns that they sense: the differences in the way that various persons are treated, the interminable discussions about the ongoing vendettas in which the shifting cartography of intimacies and alliances can be sensed, the commentaries that punctuate hunting stories or that accompany the cutting-up of the game, participation in ceremonies that are still mysterious but in which contrastive blocks of oppositions emerge, some heavy-handed joke or even an anodyne remark that remains imprinted on the mind: all these play their part in supplying reference points, prompting automatic responses, infiltrating attitudes, in short, instilling the confidence necessary to enter as an actor into the world into which one has been born.

Among the Jivaros, as elsewhere, this process is fuelled by affective responses, through apprenticeship and the reinforcement of models of interrelations and interaction that occur in the first instance on the occasion of events that are remarkable because of the emotions that they arouse. This applies to warfare, of course, with all its attendant mourning and victories. It also applies to the relations of lethal complicity with hunted animals that are forged by the handling of corpses that are still warm and the excitement of the first experiences of tracking and killing one’s prey. And, for people who know nothing of “natural deaths,” it applies to the obsession with shamanistic aggression to which, from time to time, physical accidents or misfortunes testify. Here, predation upon others is not just a synthetic norm of behaviour or some anthropological idea: at an early age, Jivaro children are bound to come into contact with it both physically and mentally. By experiencing the pain of a loss and a desire for revenge, the excitement of triumph and the pleasures of resentment, every Jivaro learns to cultivate all these identifications and antagonisms that an ethnographer then dutifully logs.

There are now so many rich and detailed ethnographical works that interpret the logic behind the actions of this or that ethnic group in the lowlands of South America according to the schema of generalized predation, that the case of the Jivaros no longer seems exceptional. Among the most striking examples are the Juruna and the Arawete of the Xingu valley, the Parakana of the Tocantins valley, the Mundurucu of the Tapajos valley, the Piraha of the Madeira valley, the Wari’ of Rondonia, the Yanomami of Brazil and Venezuela and, further south, the Nivacle of the Gran Chaco.4 All these peoples confer the position of an intentionalsubject upon a large number of members of the cosmos. These thus find themselves in a situation of formal equality at the ontological level, while the relations between them are, on the contrary, defined in effect by a circumstantial asymmetry, with each of these humans and non-human subjects striving to incorporate the substance and identity of others, in permanent denial of any reciprocity. A similar situation is not unknown in North America, as is testified by, among others, the Sioux of the Plains and the Chippewa of the South-Western edge of the Great Lakes.5 Other cases are also to be found, for instance among the Kasua of the Mount Bosavi region of New Guinea and the Iban of Sarawak.6 However, these seem more rare, although it is hard to say whether the apparently greater concentration of predatory animism in the Americas results from particular features of the continent’s development in isolation from the rest of the world or simply from the greater attention that ethnographers studying autochthonous peoples there pay to certain details of their relations with plants and animals.

The symmetry of obligations

We need not look far afield to find a perfect counter-example to the Jivaros. Whereas the latter do all that they can to escape the obligations of exchange, the Tukanos of Colombian Amazonia on the contrary strive to respect such obligations meticulously in all their interactions with other inhabitants of the cosmos.7 Yet these two ethnic groups, each of which is composed of several tribes, do share many characteristics in common. In the first place, they are relatively close spatially, separated by no more than five hundred kilometres which, given the scale of Amazonia, is a mere nothing. The environments in which they live are also similar: they are dominated by the equatorial rainforest. Here, there are, to be sure, certain local differences in the availability of certain resources, but this imposes the same kinds of ecological constraints upon both groups. The Tukanos and the Jivaros have responded in similar fashion to these constraints. In both cases, they are dispersed in residential units of relatively small numbers of people; their itinerant slash-and-burn horticulture consists mainly of manioc (sweet in the one case, bitter in the other); they acquire their proteins by means of a combination of hunting and fishing, hunting being more important for the Jivaros, fishing for the Tukanos. And, finally, the way they see their environment is altogether similar: both categorize humans, plants and animals as “people” (masa, in the Tukano languages) or as “persons” (aents in the Jivaro languages) all of whom possess an analogous interiority. This makes it possible for most of the species to lead the same kind of social and ceremonial life, despite the differences in their physicalities. It is on this basis that humans can maintain with plants, animals and the spirits that protect them individual relations governed by a code of behaviour similar to that which prevails among the Indians themselves.

Both the Jivaros and the Tukanos unquestionably belong to the ontological regime of animism. But the principles and values that guide their relations with others could not be more different. The Desana, one of the sixteen tribes that make up the Tukano group, offer a good starting point for an examination of those differences, for an ethnographic study of them has provided Reichel-Dolmatoff with the material for the “thermodynamic” model of the cosmos mentioned at the start of this book, with which many societies in the Amazonian north-west are now credited (see, for example, Reichel-Dolmatoff 1971, 1996). According to this model, the universe was created by Father Sun, an omnipotent and infinitely distant being for whom the actual daily sun is, as it were, a delegate to this world. The fertilizing energy that emanates from Father Sun animates the entire cosmos and, through this cycle of fertilization, gestation and growth, of humans, animals and plants, ensures their vital continuity. It is likewise the source of other cyclical phenomena such as the revolutions of the heavenly bodies, the alternating seasons, the variations in nutritional resources and periodical recurrences in human physiology. However, the quantity of energy produced by Sun is finite and is deployed in an immense closed circuit that encompasses the entire biosphere. In order to avoid entropic losses, exchanges of energy between the various occupants and regions of the world therefore have to be organized in such a way that the quantities of the energy that humans extract can subsequently be re-injected into the circuit. For example, when a Desana hunts and kills an animal, a portion of the potential of the local fauna is cut off and is transferred into the human domain when that game becomes food. It is therefore necessary to ensure that the needs for human subsistence do not endanger the good circulation of the flows of energy between the different sectors of the world. And it is the responsibility of the Desana to keep a watch on the situation and compensate for the losses that are caused by what they take from non-humans.

The most common means to achieve this result is sexual abstinence. By checking his carnal desires, a hunter effects a retention and accumulation of sexual energy that can then rejoin the general stock of fertilizing power that is in circulation in the universe and thereby benefit the reproduction of hunted animals. This balance can also take more direct forms. For the Desana, the relation between the hunter and his prey is above all of an erotic nature: in the Desana language, “to hunt” is rendered as “to make love to the animals” (Reichel-Dolmatoff 1971: 220). So men try to win the favours of their prey by means of love philtres, aphrodisiac perfumes, and seductive invocations. Charmed by these ploys, the animals fearlessly allow themselves to be approached and even visit men in their dreams or day-dreams, in order to copulate with them; and this reproductive operation helps to multiply the members of the species to which the animals belong. Although the Jivaros and the Tukanos conceive of their relations with animals as being governed by relations of affinity, the content that they ascribe to those relations could not be more different. Whereas the Jivaro hunter treats his prey as a brother-in-law who is potentially hostile and to whom nothing is owed, the Desana hunter treats it as a spouse whose line of descent he is fertilizing.

Even more direct is the principal process of energy feedback. Human souls are traded against animals that can be hunted. After his death a Desana generally enters the “Milk House,” a region of the cosmos conceived as a kind of uterine paradise. In contrast, the souls of those who have not respected the exogamous prescriptions go off to great underground or underwater houses where the Vaí-mahsë live. These are the spirits that govern the destinies of hunted animals and fish. There the human souls become animals, as a kind of enforced compensation from those who have not respected the rules of exchange between humans. But this is not the most common mechanism for the renewal of the fauna. The most common operation is the responsibility of shamans. These periodically pay visits to the Vaí-mahsë in the course of trances brought on by narcotic drugs, in order to negotiate a provision of forest animals for the members of their communities to hunt. Every animal thus made available for hunting must be compensated for by the soul of a dead human that will change into an animal of the same species, destined to be included in the stock of animals amassed by the animals’ master. The humans destined to become animals after their deaths usually come from neighbouring groups but are selected by a consensus. It is said that shamans from the various Tukano tribes meet in the house of the Vaí-mahsë to decide together who, among the members of their respective collectivities, will have to die to ensure that hunting continues to be good. Negotiations for a future exchange of souls in return for game, arranged in an amicable fashion by the shamans of several tribes, thus precede the exchange of souls that each shaman negotiates with the animals’ spirit-master.

The negotiation that the shaman conducts with the Vaí-mahsë aims to bring about a scrupulous equivalence between the objects involved in the transaction. Once the two parties have reached agreement, the shaman enters the house where the animals are kept, suspended from the rafters like quarters of meat in cold storage. He then shakes one of the posts to dislodge the hunk of game that he fancies for his group. But if he shakes the beams too roughly and detaches more animals than the agreed quantity, a new bout of bargaining is required in order to achieve parity. In this way, humans and animals enjoy an equal status in the living community’s access to energy; both groups help to maintain a balanced flow of it, and their functions are reversible in this quest for a perfect balance, based on strictly equal transfers. The freely accepted obligation of mutual dependence is equally central to the non-human communities. So the spirit-master of the terrestrial animals and that of the fish regularly visit each other for festivals and dances accompanied by all their families. These are opportunities to exchange women and to render one another’s respective communities fecund. As can be seen, egalitarian exchange is at the heart of the relations that the Desana weave with non-humans; its demands colour all their actions affecting the environment.

It is true that certain aspects of Reichel-DolmatofFs proposed model of the Tukano cosmology have prompted disagreement, in particular the correctness of translating the Desana notion of boga (“current”) by the thermodynamic concept of a closed circuit of a finite quantity of energy. Another Tukano specialist has recently proposed an alternative model, based on his study of the Makuna, in which recyclable energy is replaced by an open-ended flow of “spiritual” forces, which sometimes increase and sometimes decrease (Cayón 2002). According to Luis Cayón, every tribal territory of the Tukano group is animated by a particular essence regarded as one of the manifestations of the mythical hero Yurupari, whom all of them recognize. This essence resides concretely in the musical instruments that are used in the periodic Yurupari rite but are ordinarily deposited in some stream or river. The essence thus travels through the rivers of all the territories and thereby, through the interconnections of the hydrographic network, mingles with the essences from other tribal groups. On the occasion of the Yurupari ceremony, which all the tribes celebrate at the same time, the forces of fertility circulating in the rivers reach a high level of concentration and bring fecundity to the forest, to the rivers and streams and to the non-human inhabitants of the cosmos. So, for the Makuna, vital power comes not from Father Sun but from the submerged instruments of Yurupari: since the quantity of energy carried in the rivers fluctuates depending on the rainfall, it falls to humans to divide it up in a balanced fashion, thereby allowing non-humans to benefit from it. In this undertaking, a crucial role is played by the specialists of ritual, for it is they who are responsible for fertilizing the non-human occupants of the territory in the course of a ceremony known as “the healing of the world” (ümüãri wãnōre). It is also up to them to go and negotiate with the Master of the animals for the more than usual quantity of game required for organizing a great collective festival. This they obtain in exchange for offerings of coca and tobacco that the Master immediately converts into fertile power for the animals.

Ordinary men, too, take an active part in encouraging animal life. As may be remembered, it is a Makuna hunter’s duty to send the spirit of a slaughtered animal back to the house of its species so that it can be reborn there. They manage this by dint of an incantation that they chant silently before eating any game. In this incantation they retrace the mythical origin of the particular species that they are about to eat. It is a symbolical way of reconstituting its collective genesis in such a way as, practically, to reconstitute the essence of an individual that has temporarily been taken away from its fellows (Århem 1996; Cayón 2002: 206). It is even said that, thanks to this process, two new subjects of an animal species are born for each animal killed, an increment for which Reichel-Dolmatoff’s homeostatic model makes no allowance. The exchange made with non-humans thus takes the form of an obligation on the part of the Makuna to regenerate those that they destroy. It is a way for the animals to perpetuate themselves and for the humans to continue to feed on them. In short, even though Cayón and Reichel-Dolmatoff diverge as to ethnographic details, they are certainly in agreement on the fact that parity in the exchanges made between the Tukanos and their non-human neighbours is indispensable for the survival of the world. As Cayón remarks, “the fact that reciprocity is the axis of the system is beyond question” (Cayón 2002).

The social organization of the eastern Tukanos is governed by the same principle of reciprocal dependence as that which rules their relations with animals. The traditional form taken by the habitat is that of a large house for several families that make up an agnatic descent group, known as a maloca in the Spanish of this region. The physical and symbolic reproduction of the local communities results from matrimonial exchanges and the distribution of ritual functions within a group composed of at least sixteen exogamous units (which I have been calling “tribes,” for the sake of convenience) (Jackson 1983, especially Ch. V). But that term is not really appropriate. It is true that each of those exogamous units is characterized by a distinct language and a specific name (Desana, Makuna, Tatuyo, Barasana, and so on); each claims descent from its own founding hero; and each holds the privilege of making and using certain types of ritual objects. However, each unit also observes a strict rule of exogamy that stipulates that it should obtain its wives from groups that speak a different Tukano language, or even from groups that speak Arawak or Carib. Furthermore, at present, none of these “tribes” occupy a continuous territory. A maloca is composed of men who communicate together in the language of their own linguistic group, and of women who come from several adjacent linguistic groups, who continue to speak their own languages, for multilingualism is general throughout this region.

Clearly, each of the exogamous linguistic groups does not form matrimonial alliances with all the other sixteen, in the first place because there are certain pairs of linguistic groups between which unions are prohibited (phratries); and secondly because marriages are usually arranged with neighbouring groups: the Desana with the Pira-Tapuya, the Bara and the Tuyuka, for example, and the Barasana with the Tatuyo and the Tuyuka. Matrimonial exchange thus makes it possible to structure the whole inter-tribal system, since the women identify both with their husbands’ groups and with the group in which they themselves were born. In this way, they serve as intermediaries between clearly differentiated local units, ensuring that they all become integrated. This close complementarity of different linguistic groups is reinforced by the idea that each possesses its own economic specialization (hunting, fishing, or horticulture), which complements those of the others, even if all of them are polyvalent in the techniques of subsistence. Thus the Desana regard themselves as “hunters” and, for preference, they marry women from the Pira-Tapuya unit, which is classified as a tribe of fishermen. Furthermore, each of the units engaged in such exchanges is associated with one sex in particular, depending on the nature of its specialization. Thus the Desana “hunters” consider the Pira-Tapuya “fishermen” as a whole as a feminine element, while regarding themselves, collectively, as a masculine unit (Reichel-Dolmatoff 1971: 17-18).

As well as linguistic exogamy, there are other factors that combine to create solidarity between the peoples of North-West Amazonia, welding them into an inclusive regional organization. One factor is mythology, which unites all these linguistic groups in a common origin and assigns to each of them a territory and a place set in a hierarchical order according to the site and order of their appearance in the cosmogony. This is related in stories the structure of which is common to them all. It describes a series of episodes in a mythical journey to the sources of the rivers made by a group of primordial anacondas that halted at various sites characterized by a chaos of rocks and rapids. In each of these sites, one of the anacondas emerged from the waters and a portion of its body was transformed into a group of human ancestors each of whom then gave birth to one of the numerous patrilineages that compose each linguistic group. This process of progressive and itinerant segmentation is often represented as a canoe journey that produced all the successive ancestors of the descent groups of the various Tukano “tribes.” The most prestigious in the symbolic hierarchy are those who were the first to land in the lower reaches of the hydrographic network.

Moreover, all the Tukano linguistic groups (and a few non-Tukano ones too) celebrate the cult of Yurupari in the course of a series of ceremonies during which masculine initiations take place; but the principal objective of these is to renew contact with the founding heroes and the ideal norms of existence that they established long ago (S. Hugh-Jones 1979). Every time a maloca organizes one of these ceremonies, members of the various neighbouring “tribes” are invited, along with their musical instruments that contain the Yurupari essence of their own particular descent groups. This complementarity of linguistic groups in a rite commemorating the aetiology of the totality that encompasses them all reaffirms the vigour of the intrinsic links that unite them. The regional division of crafts likewise confers upon each “tribe” a reputation of excellence and hence exclusivity thanks to their production of a kind of object necessary in the daily lives of all of them. Canoes come from the Bara, cassava presses from the Tuyuka, basketry sieves from the Desana, drug-pipettes from the Tatuyo, stools from the Tukanos, and so on. This specialization engenders a system of artificial rarity that is very common in Amazonia and that encourages a generalized circulation of artefacts that accentuates the sense of a voluntary mutual dependency. Finally, these links of mutual dependency are strengthened by the systematic practice of paying long visits to one another, visits that sometimes last for several weeks, and by regular drinking festivals during which the invited affines offer their hosts vast quantities of smoked meat and fish, just as the latter will to the former on a subsequent similar occasion. These systematic exchanges of food and hospitality between residential units that are totally autonomous where their subsistence is concerned help to strengthen sociability and a sense of belonging to a single group. Despite the diversity of their languages, each maloca, each descent group, each Tukano linguistic group thus feels it is an element within a meta-system and that it owes its material and symbolic survival to regular exchanges with the others who are part of the whole system. As in their relations with animals, it is the logic of parity in compensation that governs relations between humans here.

No doubt the eastern Tukanos and their neighbours in the Amazonian NorthWest have carried to a degree seldom attained elsewhere their obsession with maintaining a close network of relations of equitable exchanges with the many kinds of persons that compose their world. Although it may elsewhere assume a slightly less systematic form, the constant attention paid to maintaining a balanced reciprocity of transfers as a cardinal schema of action, is by no means rare in the animism archipelago. Good examples of similar behaviour are found in the Guianas, in particular among the Wayapi and the Akuriyo and, on a wider scale, by the kind of confederation that is formed by the indigenous peoples of the upper Xingu, which is similar in many respects to the regional system of the Amazonian North-West.8 The hunting peoples of the Siberian forest provide another illustration. As Hamayon observes, here “the very act of hunting, of killing game . . . is governed by a logic of marriage alliances . . . modelled on behaviour toward a human partner.” In fact the two relationships are two of a kind, to the extent that “the hunting system [is] analogous to the matrimonial system” (Hamayon 1990: 374). The same principles of equivalence seem to be at work among the Moi peoples of the high forests of central Vietnam: here, the Reungao establish extremely formal alliances with the spirits of animals, plants and meteorological phenomena, some of which are characterized by obligations analogous to those that stem from kinship links and association pacts between humans (Kemlin 1999: 165- 283). In all these cases human and non-human “others” are treated as “alteregos” with whom it is only possible to live amicably if an agreement of egalitarian exchange is scrupulously observed.

The togetherness of sharing

Our second counter-example is likewise situated no more than a few hundred kilometres away from the Jivaros, but this time to the south: the Campas form a pluri-ethnic community in which generosity, solidarity and the predominance of common welfare over the interests of individual parties have been elevated to the rank of a supreme canon of behaviour that is far superior to the rules of equal and complementary exchange that the Tukanos like to respect. “Campa” is the generic name given to a cluster of tribes that speak Arawak languages in the upper central Amazonia of Peru—the Ashaninka, the Matsiguenga and the Nomatsiguenga—who, together with the Piro and the Amuesha (or Yanesha) make up the sub-Andean Arawak group. All of them live in a foothill equatorial forest similar to that of the Jivaros, in the valleys of the Urubamba and the Perene. Moreover, they are all diversified producers living in dispersed small autonomous local communities that combine swidden horticulture with fishing, hunting, and gathering. Finally, the Campas all agree on the fact that animals, plants, and the spirits that protect the former or embody them are social beings, endowed with an interiority and faculties of understanding similar to those of humans. All these persons with different appearances are primarily distinguished by their detachable bodies, which are assimilated to cushma, the long cotton tunics traditionally worn by the Indians of this region. But despite all these resemblances, a greater distance than that which separates the Campa ethos from the Jivaro ethos could not be imagined.

The cosmologies of the Campas tribes are all organized according to the same dualist principle that divides human societies, animals and spirits into two distinct and mutually antagonistic ontological domains.9 One domain possesses a positive value and includes all the entities that share a common essence: namely, the Campas tribes and some of the forest tribes that surround them (in particular the Cashibo and the Shipibo-Conibo, who speak Panoan languages), the deities of the heavens (Sun and his father, Moon), the master-spirits of hunted animals and the animals themselves. The other domain is totally negative and is defined by its radical difference from the first one. It encompasses all humans who come from the Andes, whether Indians or Whites, sorcerer-animals and their masters, who are bad spirits. Most of the hunted species and their masters stem from a race of good spirits whom the Campas call “our people” or “our fellows” (ashaninka) and who are reputed to be well disposed toward the Indians. These live on the periphery of the known world, immediately above or below the terrestrial strata, along the margins of the territory and on the mountain peaks. They have a human appearance that is invisible to the Campas and so, when visiting, they adopt the form of lightning, thunder or various animal species. Some of them control important resources. Otters, grey herons, and egrets are the masters of fish and ensure that these swim back up the rivers every year in the spawning season so that the Campas can fish for them in the shallow waters of the dry season. The swallow-tailed kite is the father of edible insects: the shaman pays regular visits to its wife to ask her to allow her children—who are regarded as the shaman’s brothers—to accompany him so that humans can feed on them. Most of the birds that the Campas hunt are themselves embodiments of good spirits. Their slaughter is only an illusion; after the hunter has asked the bird for its clothing, out of compassion for him it deliberately presents its carnal envelope to his arrows, at the same time preserving its immaterial interiority, which is immediately reincarnated in an identical body or else resumes its invisible human appearance. The bird thus suffers no damage and its act of benevolence requires no reciprocation except, perhaps, a feeling of gratitude. Certain very common species of game birds, in particular toucans, penelopes and hoccos, are not reincarnations of spirits but instead are protected by them. And those good spirits offer them freely to the humans, for them to hunt. The reason for this generosity is the fact that the good spirits, their animal transformations and the species that they control are all identical to humans at the ontological level. The Campas regard them as close kin, and the gift of their bodies is seen simply as evidence of the dutiful generosity that people of the same kin owe one another. The solidarity that such a link presupposes is expressed in exemplary fashion when the good spirits associated with hunting descend, in their invisible form, among the humans so as to dance and sing with them. In so doing, they are not seeking compensation for any services rendered, but simply wish to show their affectionate closeness and their desire to share in a conviviality that is free from any obligations.

The status of the mistress of peccaries makes it possible to contrast this dutiful generosity with the imperative of exchange that characterizes hunting in the Tukano groups; among the Campas, this is a feminine entity, described as a generic sister, who keeps the peccaries in an enclosure at the top of a mountain.10 From time to time a shaman comes to intercede with her, asking her to part with one member of her herd. She then tugs out a tuft of bristles from the back of one of the animals and blows it away so that it will eventually produce many more peccaries, which she will then send down to the humans, for them to hunt. This is an action of pure benevolence. It certainly creates certain moral obligations for the hunters. In particular, they must make sure that they kill the peccaries with a single arrow shot, so as not to cause them to suffer. However, unlike among the Tukanos, no compensation is demanded. The same goes for fishing: the fish, filled with pity, allow themselves to be caught on the fisherman’s hook and line, after he has repeatedly and sadly mumbled, “My bag is empty, my bag is empty” (Rojas Zolezzi 1994: 205).

The good spirits have no sexual activity. This is a feature that sets them firmly apart from the usual figures of the masters of game in Amazonia and the animist world in general. Among the Tukanos, the spirit-masters of animals are characterized by their superabundance of sexual energy and, as we have seen, they send their protégés to copulate, in dreams, with the hunters, a ploy that is perfectly understandable, given that the spirits are responsible for the reproduction of the animal species. To that end, they need the assistance from the reproductive powers of humans who, for their part, are happy to oblige in exchange for the vital force that they absorb when they consume the animals. The good animal spirits of the Campas are quite different. Although they exist as two sexes, they reproduce without coitus. In their human reincarnation they are said to possess atrophied genital organs and their women give birth by parthenogenesis, simply by shaking out their tunics. Furthermore, also in contrast to a Tukano hunter who seeks to win the favours of the animals by making himself attractive to them with charms and perfumes that enhance his erotic attraction, the Campas men endeavour to purify themselves as completely as possible before setting out on a hunting expedition. They expunge all residual signs of their sexual relations with women, in particular any defilement left by contact, even of an indirect nature, with menstrual blood. The horror that the good spirits feel for anything that draws attention to the physiology of reproduction and its cycles, their disgust at the uncontrollable desires and the flow of the substances necessary for existence indicate clearly that the relations between humans and these entities that supply them with game have nothing to do with the exchange and recycling of fertilizing energy and principles of individuation that characterize such relations in the Amazonian North-West. The bodies that the good Campas spirits deliver up to the hunters are nothing but carnal envelopes stripped of any subjectivity or principles of animation, and this manifestation of generosity in no way affects the perennial integrity of these beings that are forever unaffected by the contingences of organic life.

Nevertheless, this Campa world is not without negative aspects. It teems with evil spirits that live in close proximity to the humans and are a constant danger to them. These are known as kamari and they assume as many different forms as the good spirits do. Most of them have monstrously large sexual attributes. Some have a gigantic penis that causes the deaths of the women and men whom they violate, while others take the forms of attractive incubuses and succubae that beat their partners to death after coitus. Moreover, many evil spirits adopt animal forms that may be permanent, as in the cases of insects, bats, or felines, which the Campas are careful not to approach or kill. Others, though, are transient: these are species that are normally edible—toucans, monkeys, birds—whose outward appearances the kamari adopt and then, if the humans laugh at them, transform themselves into incubuses or succubae. Evil spirits of the class known as peári sometimes even take on the disguise of some ordinary hunted animal which, if it is killed and eaten, causes those who consume it to die. In all such cases, the human victim then becomes an evil spirit of the same kind as the one that attacked him or, worse still, changes into a White. Finally, kamari may be masters of sorcery, which they use to harm the Indians. Shamans then do their best to cure the latter with potions and by rubbing them with medicinal herbs.

The Campas’ relations with non-humans are not confined to accepting the benefits of food that the spirit-masters of animals lavish upon them, for at the same time a cohort of evil spirits preys upon the Campas and may slip into the skins of even the animals with the most inoffensive appearances. On the one hand, hunters receive the gift of meat that they ask for, without offering anything in exchange; on the other, they themselves are hunted, powerless to avert their own fate as game. However, it would be mistaken to interpret this reversal as a sign that predation or exchange might be recovering their rights. For that to be the case, the Campas would either have to be the active instigators of this violent alienation, which they are not, (for they are its victims and try by every means to protect themselves from it), or else the persecutions that they suffer would have to be regarded as a compensation to which they consent in exchange for the game that they are given (which is clearly not the case). The good spirits and the evil spirits, the Campas, and the people of the Andes, the generous provisions of meat and the animals that have become sorcerers are all divided into two hermetically sealed ontological domains that are in perpetual conflict. One domain is ruled by the constantly reaffirmed values of sharing and solidarity; the other, which is the agent for the evil that every lucid mind can detect in the world, embodies a cruel and senseless otherness that nothing can moderate.

No system of relations between humans can be ruled exclusively by a logic of gift; and the Campas are no exception to that rule. The altruism and prodigality that the good animal spirits manifest when they offer their bodily remains are less manifest in the rules that govern symmetrical exchange in the system of Dravidian kinship or intertribal bartering than they are in the ethos that is characteristic of daily life in which trust, generosity and a horror of constraints predominate. The Campas have carried to extreme lengths their desire to eliminate dissent and otherness in their community, by reducing to a minimum the differences between the individuals that are indispensable if a relationship, be it reciprocal or predatory, is to be established. This point has been emphasized in particular by ethnographers of the Matsiguenga. Writing about them, France-Marie Renard-Casevitz notes that they manifest “a constant concern to reduce oppositions between the self and others that might affect the entire social field.” Meanwhile, Dan Rosengren observes that, among them, “sharing is highly valued . . . and almost imperative” and that “emphasis is put upon harmony and social balance, as positive values to strive for” (Renard-Casevitz 1985: 88; Rosengren 1987: 63-64, 161).

The Campas are famous for their heavy reproof of internal violence, for it is a source of lasting animosities and a factor that undermines social cohesion. This is illustrated by the oral jousts between Matsinguenga men forced apart by some disagreement, in which verbal provocations and offers of peace alternate. They are brought to an end when one of the protagonists, deciding to turn his aggression upon himself, starts to beat himself repeatedly and is immediately imitated by his opponent. Violent or mean individuals and those who indulge in scandalous behaviour become the subjects of public disapproval. This is first expressed by a woman, who mentions the facts but without naming the culprit; then, if the reprehensible behaviour continues, other women gradually join in the denunciation. If the situation drags on, a quarantine is imposed and the individual who has deliberately cut himself off from the network of solidarities is ignored, as if he were invisible, by the entire community. If all these measures fail, the woman who initiated the complaint has no option but to commit suicide so that her death will wipe out the separation and the disorder that her accusations have created (Renard-Casevitz 1985: 88). The principle of generosity reputed to govern the behaviour of game animals is expressed as it were in reverse, in that all positive attempts that are not followed by the desired results are interpreted as an indication of a personal failure caused by an untimely initiative that has placed someone else in a situation in which he is forced to stand apart from me in response to my intention.

The Amuesha have given a particularly clear form to this philosophy of sharing and harmonious conviviality, for they, like Aristotle, consider that love is the source and principle of the existence of all things. They distinguish between two forms of love: muereñets means the giving of oneself in the creation of life and is characteristic of the attitude of the deities and religious leaders in an asymmetrical relationship; meanwhile, morrenteñets denotes the mutual love that is indispensable for all sociability and is expressed by a constant uncalculating generosity that is exempt from any expectation of reciprocation (Santos-Granero 1991: 201-205, 295-6). This is a far cry from a constructed and negotiated distinction that makes it possible to regard “others” as a term in a reciprocal relationship, as the Tukanos do, or as prey that is necessary for one’s own reproduction, as the Jivaros do. The model of the behaviour most favoured by the Amuesha and likewise the Campas seems, rather, to be the relationship between parents and their children, in which you unstintingly give affection, care, and protection to those who depend upon you.

Obviously, it is within local communities, in kindred groups welded together by mutual aid and daily interactions that the schema of generosity and sharing is most clearly manifest, both in the precepts taught to children and also in the customary practice of one and all. However, a disturbing parallel is detectable within the vaster group of sub-Andean Arawak tribes. These maintain two different kinds of relationships with two kinds of non-human groups: on the one hand the gift-giving animals that donate food to humans; and, on the other, the evil spirits that practice predation. In parallel, their relationships with two antagonistic networks of humans also stand in marked contrast to each other. The fact is that these people of the foothills have never ceased to engage in warfare along their Andean frontier, even as they reject it within their own midst, where they favour a system of regional interactions and alliances, mostly founded upon the trading that takes place between linguistically linked ethnic communities that share the same concept of civic virtues and social concord. The interethnic complementarity of the products exchanged is reminiscent of the craft-specializations of the Amazonian North-West: the Shipibo are renowned for their painted fabrics, the Matsiguenga for their bows and arrows, the Piro for their canoes, the Nomatsiguenga for their fine cottons, while the Amuesha and the Ashaninka produce not only much sought-after ornaments but also salt. The links developed through the circulation of material goods cement this mosaic and reinforce the sense of belonging to a community federated by common values. Nothing could provide a better illustration of this than what the explorer Olivier Ordinaire has called “the moral Decalogue,” a ritual litany that was recited whenever two members of different Campas tribes met and that enumerated the reciprocal duties that they owed each other on account of their belonging to the same community (Ordinaire 1892: 144-145).

Fernando Santos suggests that condemnation of endo-warfare is characteristic of a pan-Arawak ethos, and that may be so (Santos-Granero 2002: 44-47, except for the Guajiros, however, who engage in permanent vendettas). But in the case of the Arawak of Peru, internal peace was matched by a remarkable ability to see off external enemies, by mobilizing the Campas tribes in large military coalitions along with some of their Pano allies. This exo-warfare was purely defensive, its purpose being to defend their territorial integrity against the attempts to annex land on the part of all kinds of invaders from the Andes. These range from the Inca armies of the early sixteenth century to the columns of Maoist guerrillas of the present day and include the forces that the viceroy of Peru and subsequently the young Peruvian Republic dispatched, without success, into the foothills forests, to subdue these intractable Indians to the sovereignty of the central authorities. So it is hardly surprising that the puna runa, the “highland peoples,” just like the evil spirits and their animal incarnations, should have been seen as perfect embodiments of an otherness that was as radical as it was harmful, for ontologically they were all identical since they all proceeded from the same mythological origin. Incas, Spaniards and hostile animals all had to be opposed and confined to the margins of the Campa territory: their negativity had to be expelled from a Campa land ofhomogeneous togetherness. Here, the perpetuation of an ideal of closeness without indebtedness or calculated expectations comes at a price: namely, respect for rules of exchange and complementarity between honourable neighbours whose help may be needed to prevent the Campas from being wiped out by other neighbours who treat them as prey.

The Campas are by no means the only representatives of the archipelago of animism to have sought to put this ideal into practice, and some have done so more successfully than they have. Thousands of kilometres away from the Peruvian rainforest, the northern Algonquins present an example of a people that engages in similar relations with both humans and non-humans but does so free from the threat of predation and likewise of the constraints of exchange that make it possible to face up to such predation.11 In the early pages of this book, we saw that the Cree and the Ojibwa groups regard the subarctic region, despite the seemingly strict limitations that it imposes on human life, as a benevolent environment that is inhabited by entities that are attentive to the needs of humans. It is always out of a feeling of generosity that a hunted animal delivers itself up to the hunter. Moved by compassion for humans in the grip of hunger, it presents him with its carnal envelope, as a gift, without expecting any compensation. That manifestation of generosity is of no consequence since, as among the Campas, the animal’s soul is soon reincorporated in an individual of the same species, always providing that its corpse receives the appropriate ritual treatment. Relations between humans obey an identical schema. Warfare was banned between the bands of Montagnais, Naskapi, Cree, and Ojibwa, and the sharing of all possessions and resources was an absolutely imperative rule, especially among the co-residents of small winter hunting camps.12 As Emmanuel Désveaux writes in his study of the Ojibwa of northern Ontario, “the sociological horizon of the Indians knows nothing of otherness” (Désveaux 1988: 264). A similar attitude prevails further north, among the Inuit, as it also does far away, among the Chewong of Malaysia and the Buid of the Philippines (Rasmussen 1929; Howell 1989 [1984]; Gibson 1986). As for the disinterested trust, the spirit of liberality and the commitment to sharing that Bird-David attributes to the Nayaka and the Pygmies and that she considers to be typical features of the relationship that hunter-gatherers weave between themselves and their environment both human and nonhuman, we should recognize that these amount to far more than a possible correlation with a particular mode of subsistence. For they denote a general schema for the treatment of others to which animist ontologies offer a special point of anchorage, whatever other techniques they employ to make the most of their environment.

The ethos of collectives

The prevalence of a relational schema in a collective leads its members to adopt typical behaviour patterns, the repetition and frequency of which are such that ethnographers who observe and interpret them feel justified in describing them overall as normative “values” that orientate social life. The need for sharing among the Matsiguenga and the Ojibwa, the bellicose spirit of the Jivaros and the obligation of exchange among the Tukanos all provide examples. But no relationship is absolutely predominant for, all together, they constitute the panoply of methods at the disposal of humans for organizing their interactions with other occupants of the world.

To return to the example of the Jivaros, it would be absurd to claim that everything in their daily existence stems from violent incorporation. The schema of predatory assimilation constitutes, rather, a moral horizon that orientates many fields of practice, each of which reflects it in its own way. It tolerates and encompasses other relational schemas that are elsewhere preponderant but here are relegated to particular niches which, however, are always under threat from insidious contamination by the dominant schema and the influence that this exerts. Thus, the Jivaro kinship system, which is of the Dravidian type, is founded on the ideal model of an exchange of sisters between cross-cousins. This form of union, which is, in practice, very common, establishes and perpetuates within localized kindred groups an island of reciprocity and solidarity between real affines; and this is probably indispensable for the development of a predatory attitude toward more distant affines, whether these be real, potential or ideal. It indeed seems likely that the generalized hostility toward all that lies more than one day’s march away necessarily engenders, in reaction, a central kernel in which symmetrical exchange makes it possible to count on a relative security. However, the fall-out rate from the system is considerable: brothers may become deadly enemies if they become rivals for the same potential spouses or if they feel slighted when, in accordance with the levirate rule, the widows of one of them are distributed among the deceased’s brothers. Similarly, a son-in-law may attack his father-in-law if the latter refuses to let him marry the sister of his first wife. In such cases, murders and seizures of women are not uncommon. Despite all the measures taken to minimize the fracturing effect of affinity at the heart of a local network, the possibility of this is always present, as a fermenting agent of dissension capable of blowing sky-high the fragile balance of reciprocity between the closest members of a kindred group. Fair exchange is thus formally present in the logic of the Jivaro alliance system, but it remains peripheral to the Jivaro ethos.

Conversely, predation is not absent from the Tukano groups, even if warfare between them has long since disappeared, possibly as the result of a deliberate choice to favour pacific exchanges instead. We know, at any rate, that the Tukanos used to draw a clear distinction between on the one hand raids to procure wives from linguistic groups with which wives were not normally exchanged and, on the other, murderous more long-distance expeditions. The first type of raid seems to have been quite common. Generally, no bloodshed occurred and the raid was assimilated to a hunting expedition and considered as a possible alternative to ordinary exogamous exchanges. In most cases these abductions were subsequently regularized through negotiation between the two parties, and this could then lead to the establishment of a cycle of matrimonial alliances of the classic type. Exchange would thus recover its primacy following an occasional act of predation (C. Hugh-Jones 1979: 223; Århem 1981: 160). Although very rare, the murder of a man in a distant Tukano tribe constituted a far more drastic form of violence in that it affected the procreative power of another group and thus caused a loss harmful to the whole system. However, unlike in Jivaro head-hunting, this gratuitous destruction cannot be assimilated to an act of predation since it implied no gain of energy or genetic power for the murderer’s group. For this reason, a warrior who was “a killer of a man” (masa sĩari masa was regarded as the very most negative figure in any possible interaction between Tukanos (C. Hugh-Jones 1979: 64).

As can be imagined, the Tukanos’ relations to non-humans are likewise not exempt from a predatory dimension. Emphasizing this aspect, Århem even chose to describe what he called the eco-cosmology of the Makuna as a world envisaged from the point of view of a hunter, that is to say as a network of eaters and eaten (Århem 1996).13 He defines the limits of the system by two poles: at one extreme, the supreme predators (jaguars, anacondas, certain other rapacious species, and Yurupari spirits), which feed on all living beings and are not prey for any of them; and, at the other extreme, edible plants, the very lowest level in the food chain. Between these two poles lie most of the organisms whose fate is to be at once predators and prey. That is, in particular, the case of humans, whose souls, when they die, are captured (literally “consumed”) by the spirit-founders of their clan, so that they can be reborn in another form. Such formulations are hardly unexpected since all animist cosmologies seem to derive their functional principles from the model of the food chain, regardless of the nature of their most favoured relational schema. Even Århem admits that these relations between the eaters and the eaten are regarded by the Makuna as exchanges, not as acts of predation: “In this cosmic society, where all mortal beings are ontological ‘equals,’ humans and animals are bound by a pact of reciprocity. . . . The relationship between the human hunter and his prey is thus construed as an exchange, modelled on the relationship among affines” (Århem 1996: 191-192).14 The subordination of predation to exchange could find no better expression. Finally, regarding the Campas, one just needs to recall that the gift schema only occupies a dominant position at the heart of human and non-human kindred groups because it is set against a background of predation from which they can protect themselves only by maintaining a system of exchanges with neighbours identical to themselves.


The three cases studied in this chapter prompt a more general interpretation of the nature of what I have called a “collective.” Even if such an entity acquires part of its apparent homogeneity from the mode of ontological identification that characterizes it, that is not enough to differentiate it from other entities that are similar to it in this respect. So the limits of a collective are above all defined by the prevalence within it of a specific relational schema. But the resulting unit does not necessarily tally with the customary divisions into ethnic groups, tribes, linguistic groups, etc.

The example of the Jivaros will serve to illustrate this point. The way in which I have been describing them up till now might suggest that, despite internal dialectal and cultural differences, they constitute an altogether separate group. However, some of their southern neighbours, such as the Shapra and the Candoshi, share with them not only the schema of predatory appropriation but also the institutions associated with it, and do so despite differences in language and in many features of their social organization and their material culture (Surrallés 2003). On their eastern frontiers, in contrast, the Jivaros maintain enduring relations of commercial exchange and sometimes intermarriage with communities speaking the Quichua language, the sacha runa, even though the Quichua do not share the Jivaro predatory ethos (see Whitten 1976). At first sight, the scale of contrasts between the forest Quichuas and the Jivaros seems neither greater nor less great than that which differentiates between the Jivaros and the Candoshi or the Shapra. Nevertheless, it is reasonable to treat the latter two peoples as if they were part of a “Jivaroid” continuum, whereas the Quichuas, despite many resemblances, have attained a higher level of differentiation. This is borne out by the customary behaviour of the interested parties. Although the Jivaros may “Quichuarize” themselves in a peaceful fashion through marriage, and vice-versa, such an incorporation is always prompted by an individual initiative. In contrast, the Candoshi and the Shapra maintain with the Jivaros a collective relationship of essential otherness that is sufficiently close for them to be included in the code of head-hunting and abduction of women, whether as victims or as aggressors. The Shapra and the Candoshi are thus essential players in the constitution of the Jivaro “self,” whereas the Quichuas, for their part, offer the alternative of “becoming different” to all those tempted by a change of identity.

The unification of a mosaic of peoples through the sharing of a dominant relational schema is even clearer in the interethnic cluster of the Amazonian North-West. We should bear in mind that the flows of reciprocity peculiar to this region include not only the eastern Tukanos but also Arawak groups (Baniwa, Wakuenai, Tariana, Bare, Kabiyeri, and Yukuna), a Carib group (the Carijona), and the Maku, hunter-gatherers speaking an independent language who trade game in exchange for the products cultivated by the riverside communities of sedentary horticulturists. It is true that linguistic exogamy is limited to the Tukano tribes, with the exception of the Cubeo, who dispense with it. But all the components of the meta-system subscribe to the same conviction: namely, that the harmony of the cosmos can only be maintained by dint of a constant and balanced exchange of goods, principles of individuation and reproductive elements between the various communities of humans and non-humans that inhabit it. As for the various Arawak peoples of the Peruvian foothills, I hardly need to repeat that they know they belong to the same network of solidarities, structured by their shared values of generosity, egalitarianism and openness toward others—values that are all the more cherished and respected because, in every way, they stand in opposition to the negative attributes ascribed to the Andean invaders.

In short, it is not so much linguistic limits, the perimeter of a commercial network, or even the homogeneity of modes of life that mark out the contours of a collective. Rather, it is a way of schematizing the experience shared by a more or less vast collection of individuals, a group that may well present internal variations—of languages, institutions, and practices—that are sufficiently marked for one to consider it, on a different scale, as a transformational group composed of separate units. Even if it cannot be a complete substitute for the habitual categories—culture, civilization, ethnic or linguistic group, social milieu, and so on—which may well remain useful in other analytical contexts, such a definition at least makes it possible to avoid the snags of essentialism and to sidestep the almost automatic tendency to apprehend the particularities of human groups on the basis of the characteristics to which they themselves draw attention in order to distinguish themselves from their close neighbours. This way of proceeding is the reverse of that which Benedict adopts in order to reveal her “patterns” of culture; instead of casting one’s eye over a group with pre-assigned limits, to which one ascribes an abstract and transcendent unity that is a mysterious source of regularity in behaviour-patterns and representations, it is better to seek out a field covered by certain schemas that bring together the practices of collectives of very variable sizes and natures, the frontiers between which are not fixed by custom or by law but simply reflect the breaks that separate them from other ways of being present in the world.

Stripped of any functional or purposive dimension (such as a desire for togetherness) that notion of a collective is also somewhat different from Latour’s definition of one: namely, a specific association of humans and non-humans as put together or “collected” within a network at a particular given moment and in a particular given place. Likewise, for me, a collective is a group combining entities of many kinds. But it is not, strictly speaking, one organized as a network whose frontiers—inexistent in effect if one decides to include all their ramifications—can only be drawn by the analyst’s arbitrary decision to limit his field of study to data that he is in a position to take into account. If, instead, one recognizes that the limits of any collective are co-extensive with the area of influence of this or that schema of practices, then its definition will depend above all on the manner in which the humans in it organize their experience, in particular in their relations with non-humans.15 The task traditionally assigned to anthropology, namely to set in order and compare the discouraging multiplicity of circumstances in this world, will in this way perhaps be rendered less difficult, providing grounds for hope for those who persist in believing in the worth of such a mission and a sign of encouragement for those who wish to devote themselves to the task.


Albert, Bruce. 1985. “Temps du sang, temps des cendres: Représentation de la maladie, système rituel et espace politique chez les Yanomami du Sud-Est (Amazonie brésilienne).” PhD dissertation, Paris-X Nanterre.

Århem, Kaj. 1981. Makuna social organization: A study in descent, alliance, and the formation of corporate groups in the north-western Amazon. Uppsala: Acta Universitatis Upsaliensis.

———. 1996. “The cosmic food web: Human-nature relatedness in the north-west Amazon.” In Nature and society: Anthropological perspectives, edited by Philippe Descola and Gísli Pálsson, 185–204. London: Routledge.

Baer, Gerhard. 1994. Cosmología y shamanismo de los Matsiguenga (Perú Oriental). Quito: Ediciones Abya-Yala.

Boltanski, Luc and Laurent Thévenot (1991) 2006. On justification: economies of worth. Translated by Catherine Porter. Princeton: Princeton University Press.

Brown, Joseph E.c 1997. Animals of the soul: sacred animals of the Oglala Sioux. Rockport, Massachusetts: Element Books.

Brunois, Florence. 2001. “Le Jardin du Casoar, la forêt des Kasua. Influences des relations au milieu forestier tropical sur l’identité et les savoir et savoir-faire écologiques de la société Kasua. Grand Plateau de Papouasie Nouvelle-Guinée.” PhD dissertation, École des Hautes études en Sciences Sociales, Paris.

Cayón, Luis. 2002. En las aguas de yuruparí: Cosmología y chamanismo makuna. Bogotá: Ediciones Uniandes.

Descola, Phillipe. 1992. “Societies of nature and the nature of society.” In Conceptualizing society, edited by Adam Kuper, 107–126. London and New York: Routledge.

———. 1993. “Les affinités sélectives: alliance, guerre et prédation dans l’ensemble jivaro.” L’Homme 126–128: 171–190.

Désveaux, Emmanuel. 1988. Sous le signe de l’ours: Mythes et temporalité chez les Ojibwa septentrionaux. Paris: Editions de la Maison des sciences de l’homme.

———. 1997. “Parenté, rituel, organisation sociale: Le cas des Sioux.” Journal de la Société des Américanistes 83: 111–140.

Fausto, Carlos. 2001. Inimigos fiéis: Historia, guerra, xamanismo na Amazônia. Säo Paolo: Editora da Universidade de Sào Paolo.

Franchetto, Bruna and Michael Heckenberger, eds. 2001. Os povos do Alto Xingu: História e cultura. Rio de Janeiro: Editora UFRJ.

Freeman, Derek. 1979. “Severed heads that germinate.” In Fantasy and symbol: Studies in anthropological interpretation, edited by R. H. Hook, 233–246. London and New York: Academic Press.

Gibson, Thomas. 1986. Sacrifice and sharing in the Philippine highlands: Religion and society among the Buid of Mindoro. London: The Athlone Press.

Gonçalves, Marco Antonio. 2001. O mundo inacabado: Açao e criaçao em uma cosmologia amazônica. Etnografia pirahä. Rio de Janeiro: Editora UFRJ.

Grenand, Pierre. 1980. Introduction à l’étude de l’univers wayäpi: Ethno-écologie des Indiens du Haut-Oyapock (Guyane française). Paris: S.E.L.A.F./C.N.R.S.

Hallowell, Alfred I. (1960) 1976. “Ojibwa ontology, behavior and worldview.” In Contributions to anthropology: Selected papers of A. Irving Hallowell, edited by Raymond D. Fogelson, 357–390. Chicago: University of Chicago Press.

Hamayon, Roberte. 1990. La chasse à l’âme. Esquisse d’une théorie du chamanisme sibérien. Nanterre: Société d’ethnologie.

Howell, Signe (1984) 1989. Society and cosmos: Chewong of peninsular Malaysia. Chicago and London: The University of Chicago Press.

Hugh-Jones, Christine. 1979. From the Milk River: Spatial and temporal processes in northwest Amazonia. Cambridge: Cambridge University Press.

Hugh-Jones, Stephen. 1979. The palm and the pleiades: Initiation and cosmology in northwest Amazonia. Cambridge: Cambridge University Press.

Jackson, Jean. 1983. The fish people: Linguistic exogamy and Tukanoan identity in the northwest Amazon. Cambridge: Cambridge University Press.

Jara, Fabiola. 1991. El camino del kumu: Ecologia y ritual entre los Akuriyó de Surinam. Utrecht: ISOR.

Kemlin, Emile. 1999. Les Reungao: rites agraires, songes et alliances. Une société proto-indochinoise du Viêt Nam au début du XXe siècle. Textes réunis et présentés par Pierre Le Roux. Paris: Ecole française d’Extrême-Orient (coll. Réimpressions, no. 11).

Lévi-Strauss, Claude. 1943. “Guerre et commerce chez les Indiens d’Amérique du Sud.” Renaissance 1: 122–139.

Lima, Tânia Stolze. 1996. “O dois e seu multiplo: Reflexöes sobre o perspectivismo em uma cosmologie tupi.” Mana 2 (2): 21–47.

Menget, Patrick. 1993. “Notas sobre as cabeças mundurucu.” In Amazônia: Etnologia e história indigena, edited by Eduardo Viveiros de Castro and Manuela Carneiro da Cunha, 311–321. Sào Paolo: NHII-USP/FAPESP.

Mauss, Marcel. 1950. Sociologie et anthropologie. Précédé d’une introduction à l’oeuvre de Marcel Mauss par Claude Lévi-Strauss. Paris: P.U.F.

Ordinaire, Olivier. 1892. Du Pacifique à l’Atlantique par les Andes péruviennes et l’Amazone. Paris: Plon, Nourrit et Cie.

Rasmussen, Karl. 1929. Intellectual culture of the Iglulik Eskimos. Copenhagen: Glyldendaske Boghandel, Nordisk Forlag.

Reichel-Dolmatof, Gerardo. 1971. Amazonian cosmos: The sexual and religious symbolism of the Tukano Indians. Chicago: The University of Chicago Press.

———. 1996. The forest within: The worldview of the Tukano Amazonian Indians. Dartington: Themis Books.

Renard-Casevitz, France-Marie. 1985. “Guerre, violence et identité à partir des sociétés du piémont amazonien des Andes centrales.” Cahiers Orstom (série Sciences humaines) 21 (1): 81–98.

———. 1991. Le Banquet masqué: Une mythologie de l’étranger chez les Indiens Matziguenga. Paris: Lierre and Coudrier.

Ritzenthaler, Robert E. 1978. “Southwestern Chippewa.” In Handbook of North American Indians. Vol. 15, edited by B.G. Trigger, 743–759. Washington: Smithsonian Institute.

Rivière, Peter. 2001. “A predaçào, a reciprocidade e o caso das Guianas.” Mana 7 (1): 31–54.

Rojas Zolezzi, Enrique. 1994. Los Ashaninka, un pueblo tras el bosque: Contribucíon a la etnología de los Campa de la Selva Central Peruana. Lima: Pontificia Universidad Católica del Perú.

Rosengren, Dan. 1987. In the eyes of the beholder: Leadership and the social construction of power and dominance among the Matsigenka of the Peruvian Amazon. Göteborg: Göteborgs Etnografiska Museum.

Santos-Granero, Fernando. 1991. The power of love: The moral use of knowledge among the Amuesha of central Peru. London: Athlone Press.

———. 2000. “The sisyphus syndrome or the struggle for conviviality in Native Amazonia.” In The anthropology of love and anger: The aesthetics of conviviality in native Amazonia, edited by Joanna Overing and Alan Passes, 268–287. London and New York: Routledge.

———. 2002. “The Arawakan matrix: Ethos, language and history in native South America.” In Comparative Arawakan histories: Rethinking language, family and culture area in Amazonia, edited by Jonathan D. Hill and Fernando Santos-Granero, 25–50. Urbana and Chicago: University of Illinois Press.

Sterpin, Adriana. 1993. “La chasse aux scalps chez les Nivacle du Gran Chaco.” Journal de la Société des Américanistes 79: 33–66.

Surrallés, Alexandre. 2003. Au coeur du sens : Perception, affectivité, action chez les Candoshi. Paris: Editions du C.N.R.S. Fondation de la M.S.H.

Taylor, Anne Christine. 1983. “The marriage alliance and its structural variations in Jivaroan societies.” Social Science Information 22 (3): 331–353.

———. 1985. “L’art de la réduction.” Journal de la Société des Américanistes 71: 159–173.

———. 1993. “Les bons ennemis et les mauvais parents: Le traitement de l’alliance dans les rituels de chasse aux têtes des Shuar (Jivaro) de l’Equateur.” In Les Complexités de l’alliance, edited by Elisabeth Copet-Rougier and Françoise Héritier-Augé, 73–105. Paris: Editions des Archives contemporaines.

Varese, Stefano. 2002 [1973]. Salt of the mountain. Translated by Susan Giersbach Rascon. Oklahoma: University of Oklahoma Press.

Vilaça, Aparecida. 1992. Comendo como gente: formas do cannibalismo wan. Rio de Janeiro: Editora UFRJ.

Viveiros de Castro, Eduardo. 1986. Araweté: Os deuses canibais. Rio de Janeiro: Jorge Zahar/ANPOCS.

———. 2002. A inconstância da alma selvagem, e outros ensaios de antropologa. Säo Paolo: Cosac and Naify.

Weiss, Gerald. 1975. Campa cosmology: The world of a forest tribe in South America. New York: American Museum of Natural History.

Whitten, Norman E. 1976. Sacha Runa: Ethnicity and adaptation of Ecuadorian jungle Quichua. Urbana: University of Illinois Press.


1. That is why it is pointless to oppose, as some do, two approaches to Amazonian sociability that are irreconcilable: on the one hand, that of the “hawk” camp, the partisans of ontological predation, led by Viveiros de Castro and myself, on the other, the “dove” camp, which defends the aesthetic of conviviality, led by Overing (Santos-Granero 2000). For I, for my part, have never suggested that predation is the only way of treating “others” in Amazonia.

2. This presentation of Jivaro head-hunting is inspired by the works of Taylor (1985, 1993). For a more detailed analysis of Jivaro forms of warfare, see Descola (1993).

3. The mechanisms most commonly employed to this end are the assimilation of cross-sibling relationships and relationships of conjugality, the affinization of masculine consanguines by men, the consanguinization of affines of both sexes by women, the obliteration of affinity between co-residents of opposite sexes in the same generation and its accentuation in alternate generations. On this subject, see Taylor (1983).

4. On the Jurana, see Lima (1996); on the Arawete, see Viveiros de Castro (1986); on the Parakana, see Fausto (2001); on the Mundurucu, see Menget (1993); on the Pirahä, see Gonçalves (2001); on the Wari’, see Vilaça (1992); on the Yanomami, see Albert (1985); on the Nivacle, see Sterpin (1993).

5. On the Sioux, see Brown (1997) and Désveaux (1997); on the Chippewa, see Ritzenthaler (1978).

6. On the Kasua, see Brunois (2001); on the Iban, see Freeman (1979).

7. These are the eastern Tukanos (Desana, Makuna, Tatuyo, Barasana, etc.) who, together with the western Tukanos (Coreguaje, Siona, Secoya, Mai Huna, etc.) make up the Tukano linguistic family. When used from here on, the term “Tukano” will apply exclusively to the entire group of eastern Tukanos.

8. On the Wayapi, see Grenand (1980); on the Akuriyo, see Jara (1991); on the Upper Xingu, see Franchetto and Heckenberger (2001).

9. My sources on the Ashaninka are Weiss (1975), Rojas Zolezzi (1994), and Varese (1973); on the Matsiguenga, Renard-Casevitz (1985, 1991), Baer (1994), and Rosengren (1987).

10. According to Weiss (1975: 264), this figure is feminine, but Rojas (and Elick) present it as masculine, as is the Master of the Cervidae (Rojas Zolezzi 1994: 180, n24).

11. Admittedly, the obsession with predation is not totally absent among the northern Algonquin groups. Here it is represented by Windigo (or Wiitiko), a cannibalistic monster in human form that terrorizes the Indians. However, unlike the evil Ashaninka spirits, who are said to be responsible for very concrete evils, in the anecdotes that tell of encounters with the Windigo, the latter is always overcome by the humans (Désveaux 1988: 261-265).

12. According to Hallowell, the sharing of one’s possessions is one of the “supreme values” of the Ojibwa culture (1976 [1960]: 385).

13. When he emphasizes the predatory aspect of the Makuna cosmology, Århem implicitly distances himself from the “exchange” interpretation that I had produced of the Tukano model in an earlier publication that he cites but does not openly criticize (Descola 1992). The point did not pass unnoticed by Peter Rivière, who declared himself in agreement with Århem on the fact that, contrary to what I had suggested when I opposed the Jivaros to the Tukanos from the point of view of their relational schemas with others, Amazonian cosmologies are transformations of one fundamental model in which predation and exchange are closely combined (Rivière 2001). Neither Århem nor Rivière seems to have noticed that in my view the predominance of predation or of exchange in a collective by no means excludes expression of the other schema which, however, is subordinate to the dominant one.

14. To dissipate any ambiguity, he adds, “Men supply the Spirit-owners of the animals with ‘spirit-foods’ (coca, snuff, and burning bees wax). In return, the Spirits allocate game animals and fish to human beings” (Århem 1996: 191-192).

15. This notion of a collective is closer, in its extension if not in its meaning, to what Boltanski and Thevenot have called “cities,” that is to say social models founded on conventions that are shared by sub-groups of individuals within industrial societies and that allow these to set up differentiated common worlds (Boltanski and Thevenot 2006 [1991]). “Cities” resemble collectives that are identifiable from their combination of dominant schemas of identification and relations; in the very midst of the categorical entities of classic sociology (classes, sexes, income levels, professions, political opinions), “cities” carve out contrasting forms of coexistence and social links (the “ideal city,” “the domestic city,” “the city of opinions,” etc.), which blur the conventional frontiers between groups and redistribute the criteria for drawing distinctions.