Donos demais: maestria e domínio na Amazônia

Carlos Fausto

Translated by David Rodgers
Translation from Mana, Rio de Janeiro, v.14, n.2, p. 329-366, Oct. 2008.

This text is concerned with a specific Indigenous category – usually translated as "owner" or "master" – which, in Amazonia, transcends the simple expression of a relationship of property or domain. The category and its reciprocal terms designate a generalized mode of relating that characterizes interactions between humans, between non-humans, between humans and non-humans and between persons and things. It is a key category for understanding Indigenous sociologies and cosmologies, which, notwithstanding its importance, has received relatively little attention. Through the analysis of ethnographical evidence, this text seeks to explore the consequences of imagining the Amerindian universe as a world of owners and the owner as a model for the magnified person, in order to discuss notions of ownership, domain and power in lowland South America.

Key words: Amazonia, Property, Power, Person, Individualism

RESUMO

O texto versa sobre certa categoria indígena – usualmente traduzida por "dono" ou "mestre" – que, na Amazônia, transcende em muito a simples expressão de uma relação de propriedade ou domínio. A categoria e seus recíprocos designam um modo generalizado de relação, que caracteriza interações entre humanos, entre não-humanos, entre humanos e não-humanos e entre pessoas e coisas. Trata-se de uma categoria-chave para a compreensão da sociologia e da cosmologia indígenas que, não obstante, recebeu relativamente pouca atenção. Esse texto visa explorar, a partir de um conjunto de evidências etnográficas, as conseqüências de imaginar-se o universo ameríndio como um mundo de donos e o dono como o modelo da pessoa magnificada, com o objetivo de discutir noções de posse, domínio e poder nas terras baixas da América do Sul.

Palavras-chave: Amazônia, Propriedade, Poder, Pessoa, Individualismo

This text discusses an indigenous Amazonian category – usually translated as 'owner' or 'master' – which far transcends a simple expression of a relation of ownership, authority or domination. The category and its reciprocal terms designate a mode of relationship that applies to humans, non-humans, and things. I argue that it comprises a key category in terms of our comprehension of indigenous sociology and cosmology, despite receiving relatively little attention thus far. In fact, Seeger first called our attention to this kind of relationship three decades ago:

]]> The concept of the owner-controller permeates Suyá society, even though there is relatively little property in the material sense of the word [...] But it is a fallacy of ethnocentrism to maintain that ownership and property are unimportant (1981:181-2).

The reasons for this omission are linked to a widespread view of the South American lowlands as a realm of equality and symmetry, in contrast to the hierarchy and asymmetry predominant in the Old World and the Andean region. This horizontal conception of social relations – conceived either under the rubric of sociopolitical equality or of symmetrical reciprocity – has marked the literature from the early chroniclers to modern ethnology. The notion of owner fits uncomfortably into this sociopolitical imagery, not only because of the asymmetry of the relation that defines it, but also because of its potential evocation of private property. Consequently, the mastery-ownership relation has ended up consigned to ethnographic footnotes, or reduced to a simple ontological category, that of the masters or owners of nature.

This paper looks to show that, on the contrary, mastery is as central to understanding indigenous sociocosmologies as affinity. Here I return to a problem that I first attempted to examine some ten years ago via the notion of 'familiarizing predation' – a schema through which predatory relations are converted into asymmetric relations of control and protection, conceptualized as a form of adoption (Fausto 1997). Now I propose to imagine the Amerindian world as a world of owners and the owner as a model of the magnified person (Strathern 1991).

The owner-master category

As far as I know, all Amazonian languages possess a term – historically, fairly stable – that designates a position involving control and/or protection, engendering and/or possession, and that applies to relations between persons (human or non-human) and between persons and things (tangible or intangible). This term is highly productive and applies over a wide spectrum.

Seeger writes that for the Suyá "most things have owner-controllers: villages, ceremonies, songs, houses, gardens, belongings, pets and so forth. The importance of kande is pervasive" (1981:182). The term kande, 'owner-controller,' not only refers to the possession of tangible and intangible wealth (such as ritual knowledge) but also to the potential ability to produce these goods. It also forms expressions designating social functions endowed with prestige and political power: thus war leaders were called weropakande, 'owners of our village,' while the ritual specialist is known as mërokïnkande. Indeed, Seeger states that kande is "the most important concept in Suyá thinking about power" (1981:181).

The category's importance is also apparent in the multilingual system of the Upper Xingu. Viveiros de Castro (2002a:82-3) claims that it constitutes "a fundamental notion of Xinguano culture," applicable to a wide range of contexts, its concrete model being paternity. Among the Yawalapiti, wököti designates the ritual sponsor, the specialist song master, the master of animals and plants, the chief, as well as designating an owner in the sense we would ordinarily attribute.² In all these cases, the category defines a relation between a subject and a resource: the owner is the mediator between this resource and the collective to which he or she belongs.

Among the Kuikuro, a Carib people of the Upper Xingu, the category oto applies to the same semantic field and likewise takes filiation as its basic schema. Parents are 'our owners' (kukotomo) because they take care of us and feed us. To be the owner of a collective structure – owner of the ritual path (ama oto), the men's house (kuakutu oto), the central plaza (hugogó oto) and the village itself (eté oto) – implies care since the owner must maintain these structures and feed the people who help in this task. Being an owner implies prestige and responsibility: sponsors of rituals become public people who feed the collectivity, including the spirit-owners of the festival, considered the sponsor's 'children.' As in the Suyá and Yawalapiti cases, the owner-master category also applies to the depositaries of intangible knowledge: the ritual specialists are called 'song masters' (eginhoto), the healers are 'incantation masters' (kehege oto), the sorcerers are 'spell masters' (kugihe oto) and so on.

Among Tupi-Guarani peoples, the vernacular terms for the category 'owner' are cognates of *jar and have been well-known since the 16th century. Viveiros de Castro tells us that the Araweté cognate ñã connotes ideas such as "leadership, control, representation, responsibility, and ownership of some resource or domain" (1992:345). Among the Parakanã, the most common reciprocal term of -jara is 'pet' (while in the Upper Xingu, as we have seen, the reciprocal is 'child'): the concrete schema for the ownership relation is familiarization of the young of animal prey (Fausto 2001:347-8). This is also true for other Tupi-Guarani peoples, such as the Wayãpi, for whom "all jar have 'their young,' which they treat like eima, or wild pets" (Gallois 1988:98). The same category appears in Hans Staden's account of his capture by the Tupinambá, who said to him: xé remimbaba in dé, "you are my pet " (2008 [1557]:52).

This category of owner-master is extremely widespread in the region and corresponds to what Hultkrantz (1961) termed "the supernatural owners of nature." Until recently, ethnology limited itself to these figures when speaking of owners or masters, depicting them as hyperboles of the species they represent or the anthropomorphic form through which they appear to shamans. Despite their importance, these figures need to be reinserted in the overall set of ownership relations, since, as Cesarino notes in relation to another Pano people, the animal owners "replicate the same configuration that characterizes the Marubo maloca owners (shovõ ivo): both are chiefs of their houses, in which they live with their families and have their own ways of being" (2008:25). The masters of the animals are therefore owners in their own environment, containing a collectivity within themselves: they represent and contain a species.

Among the Kanamari, a Katukina people of Western Amazonia, recursivity is the main trait of the category warah-, meaning owner, chief, body, trunk, or main river.³ Luiz Costa emphasizes the intrinsic relationality of the category: "a person is always a 'chief/body/owner' in relation to something, someone or some people" (2007:63). Warah expresses a relation of container-contained, singularity-plurality, such that "the name of a person followed by –warah designates not only that person's body, but also, in the case of chiefs, all those people who call that person 'my body-owner' ('my chief'), along with all the belongings of the person whose name forms the noun phrase 'X-warah' […]." (Costa 2008:4). This structure is replicated at different scales: between the soul and its body, between a people and their chief, between the village chief and the chief of a hydrographic network, and so forth. The cosmic limit of this relation is the primordial Jaguar, which, at the 'moment zero' of the cosmos, contained within its body all the different singularities in virtual existence.⁴

So what general features of the category 'owner-master' can we extract from these examples? First of all, we need to shift our emphasis from the ontological category to the relation it implies (Déléage 2005:191). Beyond exploring the concept of owner-master in Amazonia, we need to analyze a relational schema that applies to innumerable contexts. If the category presumes a relationship, it demands a reciprocal category: this seems to oscillate between 'child' and 'pet animal,' both implying an underlying idea of adoption. The prototypical relation of mastery-ownership is, then, adoptive filiation, a relation that is not given but constituted, frequently through the dynamic I have called familiarizing predation (Fausto 1999). I have illustrated the pregnancy of this relational schema in the domains of shamanism, warfare and ritual in previous works, making it unnecessary to repeat these examples here. It is merely worth recalling that the same schema accounts for relations as diverse as those between the shaman and auxiliary spirits, the warrior and the captive child, the killer and the victim's spirit, or the ritual officiant and the ceremonial objects.⁵ Combining the findings of these earlier works with the examples given above, we can infer that the mastery-ownership relation:

a) frequently applies to the possession of certain material items (principally ceremonial objects) and immaterial items (especially ritual knowledge);

b) does not always designate the parent-child relation, although it almost always applies to the relation between parents and adoptive children, in particular war captives;

c) never applies to autonomous living enemies, but it may designate the relation between the killer and his victim after the killing;

d) never applies, either, to game animals, although it designates the relation to pets and, very frequently, the shaman's relation to auxiliary spirits;

e) often applies to the relation between chiefs and their followers and, as we shall see later, was used to designate new relations in the context of conquest and colonization;

]]> f) does not apply solely to relations between humans (or humans and non-humans), but also designates relations internal to the non-human world.

One of the important features of this relation is its asymmetry: the owners control and protect their creatures, being responsible for their well-being, reproduction and mobility. This asymmetry implies not only control but care. Hence, the master of the animals among the Chimane of Bolivia is defined as chojca-csi-ty, "the one who watches over them, who looks after them, who cares for them" (Daillant 2003:317). From the perspective of whoever is adopted-captured, being or placing oneself in the position of an orphan or a wild pet is more than just a negative and inescapable injunction: it may also be – as we shall see later – a positive way of eliciting attention and generosity.⁶

The asymmetry of the ownership relation is very often conceived as a form of encompassment, sometimes expressed as a relation between container and contained. For example, the masters of animals usually keep their animals in an enclosure or container, releasing them slowly to be hunted by humans. For the Chimane, the master keeps "his animals in corrals [...] releasing them, from time to time, via a door" (Daillant 2003:303). An Arara shaman once explained to Teixeira-Pinto that the owners (oto) of animal species keep their 'creatures' (iamït) in a box, like the closets made by whites (1997:97).⁷

Shamans also store their auxiliary spirits in containers. Some keep them inside baskets and feed them with tobacco; others insert them into their own body in the form of resins or stones, literally containing them. For instance, the Wayãpi õpi-wan caterpillars, anthropomorphic auxiliary spirits of the shaman, are contained within his body, wrapped in tiny slings, just as the shamans are wrapped in the webs linking them to the masters of animals (Gallois 1996:46-47). This is likewise the case of the Kanamari dyohko, solidified bits of plant resin that are kept by shamans within their own bodies, but may also be placed in baskets to be thrown as magic darts or to wander in the forest in the form of jaguars (Costa 2007:381-383). The same applies to the Zápara magic stones fed with tobacco by the shaman, who keeps them in a bag, though they may also be incorporated into the owner (Bilhaut 2007:57-61). The topology is always complex since the shaman's auxiliaries appear as internal and external parts of the owner-master simultaneously.⁸

This topology also involves an interplay between singularity and multiplicity: the owner is a plural singularity, containing other singularities within himself as a body (Costa 2007) or a maloca (Cesarino 2008). The owner-master is, therefore, the form through which a plurality appears as a singularity to others. It is in this sense that the chief is an owner. When speaking in the central plaza, the Kuikuro chief refers to all the inhabitants of his village, irrespective of sex or age, as his 'children' (kangamuke). All other distinctions are obviated for him to appear as an inclusive singularity, a magnified person (Heckenberger 2005:259-263). The chief-form – the body, the bow-in-the-hand, the speech commemorating the unique history of the Kuikuro people (Franchetto 1993) – appears to the eyes of the messengers from other villages as a people, an otomo (the collectivized form of the term 'owner'). In this sense, rather than being a representative (i.e. someone occupying the place of another), the master-chief is the form through which a collective is constituted as an image: it comprises the form in which a singularity is presented to others.⁹

As a singular image of a collectivity, the master-chief form also applies to the owners of animals. The prototypical example is the figure of the master of peccaries. Here the master is a chief who contains a collectivity of peccaries, conceived as his children or wild pets. For the master to appear as a magnified singularity, the band must appear as an anonymous collectivity without its own agency. This is why I have argued elsewhere that the master represents the jaguar-part, while the band represents the game-part, the passive aspect of the peccaries (Fausto 2007:509).¹⁰ In Amazonia, every magnified singularity appears to the eyes of others as a predator, usually as a jaguar, anaconda or harpy eagle.

The owner is, then, a double-sided figure: in the eyes of his children-pets, a protective father; in the eyes of other species (especially humans), a predatory affine. Jaguarness is one of the traits associated with the figure of the master in Amazonia. Even the mild-tempered and non-aggressive Upper Xingu chief covers himself in parts of a jaguar body when he ritually greets dignitaries from other villages: a belt and hat made from the animal's pelt, a necklace made from its claws. In an sense, every master is a jaguar. And it is easy to understand why: the main device for producing encompassment and hence for magnifying the person is cannibal incorporation. Predation is an asymmetric vector of identification-alteration: by eating, one contains the other and its alterity within oneself.

Possessive (in)dividualism

Thus far I have shown how the notions of owner and ownership are indispensable to our comprehension of indigenous Amazonia. The absence of private ownership of important material resources has blocked our conceptual imagination of ownership relations, as though their model par excellence were exclusive private ownership of goods, corresponding to a consumerist and expansive conatus. In the Amerindian case, though, the possession of objects must be seen as a particular case of the ownership relation between subjects, and the thing-artefact as a particular case of the person-artefact. As Sztutman writes, mastery is "a cosmological notion that is reflected on the sociopolitical plane, referring in very general terms to this capacity to 'contain' – to appropriate or dispose of – persons, things and properties, and to constitute domains, niches and groups" (2005:261).

I begin with the double problem confronted by Locke in his refutation of absolutism and patriarchalism in Two Treatises of Government: on one hand, he looked to lay the foundations for individual freedom and, therefore, the limits of Government; on the other, he sought to base private property on natural law, despite positing an originary state in which the world was given in common to all. Locke located the solution to both problems in the concept of self-ownership, the originary and exclusive relation of a person to him or herself, which simultaneously founds both freedom and property: "Though the Earth, and all inferior Creatures be common to all Men, yet every Man has a Property in his own Person. This no Body has any Right to but himself" (Locke 1988:287 – Book II, Chapter V, § 27).

If self-ownership makes despotism and slavery contrary to natural law, how then do we pass from this self-relation to the relation between persons and things? How are legitimate ties established between a subject and an object to the exclusion of other subjects? For Locke, the extension of self-ownership to things is achieved through labour. Objects are contaminated, so to speak, by the action of the body, an action that belongs exclusively to the agent and that removes things from their natural state and annexes them to the self as its exclusive property. This reasoning, known as the labour-mixing argument, implies that labour is mixed with things, adding to them something that belongs to the subject of the action (Locke 1988:288; Bk.II, Ch.V, § 27).¹²

The Lockian theory of ownership presupposes a theory of personal identity, since it is necessary to found the subject to which an originary right over the self is attributed. What ensures its continuity in time and space? How do we know if it is always the same and not other? According to Locke, personal identity is founded on the continuity of consciousness, on the subject's reflexive relation with him or herself. A person is a "thinking intelligent being, that has reason and reflection, and can consider itself as itself, the same thinking thing, in different times and places" (Locke 1987:xxx; Bk.II, Ch.XXVII, § 9). Here sameness and selfhood merge, since both depend on a relation to self, a self-identity. Mêmeté etipseité, to use Ricouer's vocabulary (1990), become indissociable in the construction of the person. The self must be identical to itself ("the same thinking thing in different times and places") in order for it to become the object of a judgment: without identity the pairing of moral responsibility and legal accountability cannot be constructed; without reducing difference to zero, sociability cannot be founded on appropriative individuals who are free because they own themselves.

Locke's theory of property activates a series of cosmological and anthropological premises. We have a divinity who fabricates a world peopled by subjects (human beings) and useful things (animals, plants, land...) given in common to humanity. These subjects have two main attributes: firstly, a self-identity that is maintained over time and makes their acts accountable (to God and to Men); secondly, they are owners (the cause) of their acts by being owners of their own body, which is the means by which these acts have efficacy on the world. Actions on the world – conceptualized under the category of labour – lead to the appropriation of useful things, meaning that what was given in common becomes individuated and owned by some to the exclusion of others. In social life, this process leads to a distinction between owners and non-owners where the former, through their ownership of things that are added to their own body, acquire a surplus of agency. The owner thus becomes the model of the agent and appropriated goods are transformed into indices of the person's capacity for agency.

How would an indigenous narrative compare with this Lockian account of the constitution of the person and society, freedom and obligation? Were we to narrate it in an Amerindian key, what kind of world would emerge? Undoubtedly indigenous cosmologies activate very different premises, not because these peoples lack a conception of ownership or a mechanism of appropriation but because their cosmologies are based on very different principles. As an experiment in conceptual imagination, let us now try to tell the same story from another viewpoint.

A world of owners

In the beginning, the world was not given by a divinity to all humans in common for them to appropriate it. The ontology of mythic time does not establish two major classes of beings: on one hand, autonomous subjects (self-owners); on the other, appropriable things (potential properties). No definitive separation exists between subjects and objects. The mythic world is pervaded by a background of continuous subjectivity, a communicational flux involving all existents. In contrast to the original identity with God-Substance, this state is, as Viveiros de Castro argues (1998:41; 2007:51), a state of infinite differences, internal to the person, characterized by an ontological regime of metamorphosis.

In this primordial state, difference is presupposed, though not yet posed, since what myth narrates is precisely the positing of difference – that is, the production of discontinuities between species, between human collectivities, between sky and earth, between day and night, between dry land and water, differences that, combined, will constitute the world as we know it. And it is precisely the potential owners, beings with creative and transformative capacities, that will engender-fabricate the post-mythic world through their actions and their lapses.

The First People lived just as shamans do today, in a polymorphous state in which no boundaries yet existed. It was the time of origins (illud tempus) when Heaven and Earth were still connected and the distinctions between species not yet recognized. Only when these divisions solidified did the First People finally remove themselves from Earth, leaving their forms behind as reminders of what this Dream Time had been like. After their withdrawal from the Earth, each of the First People became the 'Master' or arache of the species they engendered (Guss 1989:52).

Not only 'natural' attributes are determined in this process of speciation; the 'cultural' attributes of each species are also defined. This definition often derives from the transfer of ownership from one being to another. Many indigenous etiological myths narrate not so much an origin-genesis as the way in which attributes that typify human sociality were appropriated from animals. Culinary fire is the most famous example: in the Tupi-Guarani myths, the theft of the fire owned by vultures makes humans become eaters of cooked meat in opposition to necrophagy; in the Gê myths, the theft of fire from jaguars leads to the distinction between a raw (cannibal) diet and cooked food, capable of producing identity between kin (Fausto 2002a, 2007).

The Gê narratives resonate particularly well with the argument of this text since here transspecific dynamism is founded on adoptive filiation. We pass from an initial relation between brothers-in-law to a relation of familiarization between a boy and a jaguar.¹³ In the Kayapó-Gorotire version, "the jaguar's wife (who was an Indian) didn't like the boy, who she called me-on-kra-tun ('foreign' 'or abandoned child'); nonetheless, jaguar, who had no children, decides to adopt him" (Lévi-Strauss 1964:75). The adoptive father goes hunting everyday to feed the child, leaving him alone with an anti-mother. So he can defend himself, the jaguar gives the boy a bow and arrow with which he ends up killing his adoptive father's wife. He then takes the jaguar's belongings: the bow, cotton thread and cooking fire (Kayapó-Kubenkranken version). By refusing adoption, the boy reinstates the enmity that thereafter marks the relation between jaguars and humans, while reconstituting the kinship ties with his own kind.

The post-mythic world that emerges from this initial dynamic is a world of multiple domains. These domains structure the cosmos, meaning that one of the premises informing human action over what we call the natural world is that everything has or can have an owner. As Descola has shown (1986), nature is domestic because it is always the domus of someone. For the Achuar, the forest is the plantation of Shakaim, wild animals are the young kin of the 'mothers of the game,' and cultivated plants are cared for by Nunkui, the spirit-woman who gave rise to cultigens. The non-human world does not belong to everyone, nor is it the land of no one. As the Kuikuro would say, the world is not tatutolo engü, 'everyone's thing,' which would be the same as saying 'there's nobody to care for it.'¹⁴ It is precisely because there are owners who zealously protect their possessions that the Kuikuro, before embarking on a collective fishing trip, proffer a lengthy incantation in which they name all the masters of the water, asking them not to prey on those who are fishing or to hide their fish (Fausto, Franchetto & Montagnani 2007). This incantation is intended to induce a generous disposition in the owners, persuading them to relinquish their precious things.

If the indigenous world is a world of owners, what kind of domains do they own? Referring to the Guarani-Kaiowá, Mura suggests that:

from the viewpoint of the tradition of indigenous knowledge, it is impossible to imagine the existence of places, paths, living beings and inanimate beings as neutral, autonomous and owner-less. All the elements composing the current Cosmos possess owners, constituting domains and reflecting an extremely significant logic in the Universe's hierarchization [...] (2006:234-235).

The world is therefore divided into different domains, or spaces of domesticity, pertaining to humans and non-humans, each with its own owners-masters. Gallois (1984) even suggests that, for the Wayãpi, there are a finite number of domains, each of which could in theory be exhaustively described. Referring to the Araweté, Viveiros de Castro argues somewhat differently that:

]]> The juridical notion of ownership is the least important aspect and not even always present. The Araweté do not have a general conception of the cosmos as a set of domains possessed by different ñã [owners] with whom man must come to terms or fight (1992:345-6).

Although the Araweté cosmos is peopled by dozens of ferocious cannibal owners (Viveiros de Castro 1992:77-82), humans do not have to negotiate with them continually in their everyday activities. Indeed, the relation that really preoccupies the Araweté is between the living and the dead or, more precisely, between the living and divinities via the dead. The Parakanã are also unconcerned with negotiating any act of predation or appropriation since they do not postulate the existence of animal or plant masters, meaning no risk is involved in preying on other species. However, various Amazonian peoples not only conceive the world to be made of multiple domains, they are also conscious of the fact that, to live, humans are compelled to ignore their limits: planting, hunting and fishing require them to invade these alien spaces, almost always with predatory intentions. The Miraña of the Caquetá river, for example, conceive the forest to be the domain of the masters of animals, who "reign there in the same way that the human maloca masters reign over their people" (Karadimas 2005:342). Human intrusions into this space "are perceived as bellicose acts against animals, undertaken in an identical fashion to the war expeditions of the past" (2005:344).

Even the production of certain artefacts can be dangerous, since, as Guss indicates, it implies a "transfer of ownership" (1989:61), a transfer conceptualized by the author as a "conversion of wild objects into domestic ones" (1989:95). For example, before Yekuana men can take the canes used to make their famous bicolour flat baskets, they must ask a shaman to negotiate with Yododai – the master who plants the canes and protects them. Permission obtained, a series of rules must then be observed while cutting the cane and making the baskets, a process during which a design emerges that is associated with Odosha, the prototypical figure of predation (1989:106-7;130-2). The process of conversion-domestication described by Guss is also, therefore, the fabrication of a jaguar-artefact.

In sum, everything in principle has or may have an owner: the forest, animals, rivers and lakes, but also an animal species, another plant species, or a particular stand of bamboo, a curve of the river, a certain tree, a particular mountain. Claiming that the current cosmos is structured by ownership relations does not mean, though, conceiving it to be organized exhaustively into discrete spaces (territories and jurisdictions), as though resulting from a series of enclosure acts decreed at the end of mythic time. The passage from the continuous to the discrete describes this process Amerindian mythology implies the constitution of a world traversed by ownership relations, but not a cosmic cartography of discrete and exclusive properties. These ownership relations are multiple and potentially infinite. Neither are they given once and for all: they have a post-mythic dynamic in which beings can appropriate or can become appropriated, inserting themselves within a new ownership relation. Objects are fabricated, children are engendered, capacities are acquired, animals are captured, enemies are killed, spirits are familiarized, human collectivities are conquered.

This dynamic operates in the macro-relations between collectivities and in the microproduction of the person, which is constituted and deconstituted continually by appropriating others and being appropriated in turn. This person is not a unitary 'self-identical self,' therefore. Merely announcing, though, that it is a 'distributed' or 'relational' person is insufficient. Locke's theory of personal identity is not based exactly on a self-enclosed individual in contrast to the relational person of Amazonia, Melanesia or wherever. As Balibar points out, Locke takes identity to be a relation, which implies "that it presupposes difference, or that it is [...] a certain way of dealing with difference [...] by reducing it to zero" (1998:247).¹⁵ Locke's model also includes a distributed person (Gell 1998) insofar as property-objects are indices of agency. The Lockian proprietor is a magnified person to the extent that, thanks to a relation to self, the world can be appropriated. The own (guaranteed by self-consciousness) and ownership (based on the private property of the body) leads to appropriation, which magnifies persons by 'extending' them to things and 'annexing' things to persons.

The crucial distinction in terms of Amazonian indigenous peoples is the fact that the founding relation here is not self-identity: the Self and the Same do not merge in the construction of the Amerindian person.¹⁶ Hence the term ownership is not the most appropriate, since the very essence of the owner is to be altered. The multiplicity and fractality of ownership relations imply internally composite subjects, 'self-different'persons (Viveiros de Castro 2002b:377). The model of the agent is not, therefore, that of the owner who annexes things to an immutable Self, but the master who contains multiple singularities. Consequently, while both the Lockean and Amerindian models are appropriative, the risk of the former is, as Kant would say, the 'a-social sociability' of possessive individualism, while the risk of the latter is the cannibal sociality of possessive singularity. The mechanisms for limiting appropriation also differ: on one hand, the moral responsibility of the forensic person; on the other, the sociality of kinship and the body of kin.

Magnification and power

Just as the spectre of private property has blocked our conceptual imagination of ownership relations, so our capacity to think of power in the South American lowlands has been obscured by the State model and the focus on coercion. It is essential to lessen the gulf created by centuries of polarized images, conceptual black holes that suck in our imagination whenever we think about power in the indigenous world. It is as though we were continually forced to choose between an anti-state model (negatively obsessed with the State) and a model of teleological centralization (positively obsessed with the State). We need to construct an ethnographically informed language to conceptualize asymmetric relations in the region without dissolving them into a swamp of symmetrization, or transforming them into seeds containing the tiny protoplasm of a state apparatus (a statelet from which a process of state-genesis is waiting to burst).

As an alternative, then, I suggest mastery as a relational schema for producing magnified persons that contains the mechanisms both for generating potency and for undermining power. In the microanalysis of this relation, it is crucial to identify the mechanisms for constituting and deconstituting relations that imply control. This seems to me a fundamental step if we are to escape the essentialist language of Clastres, full of beings-for and beings-against, in particular his metaphysics of primitive society qua the absolute.¹⁷

As far as I know, Strathern was the first anthropologist to associate a critique of the notion of control with the Anglo-American concept of property: "This notion of control implies something like an exercise of proprietorship, either over attributes 'belonging' to one self or else over attributes 'belonging' to others and yielded by them. The concept already prejudges the manner in which persons impinge upon one another" (1988:141). Our task, then, is not to prejudge but to investigate, in each ethnographic context, how persons impinge upon one another. The verb 'impinge' means to 'to go against,' 'to impose,' whose participle is 'impact.' We can ask then: what impact do these masters have?

If the classical Weberian definition of power as "the possibility of imposing one's own will within a social relation despite any resistance" (Weber 1984:43) fails to apply adequately to the Amerindian context, this stems more from the notion of 'own will' than 'imposing.' Magnified persons are constituted precisely by incorporating relations with alien-subjects endowed with other-wills, imposing their perspective but under the constant risk of losing it. The master's potency is the capacity to extract an action from his wild pet. This is coercion, as Strathern would say (1988:272). But here we find an ambivalence, since it is impossible to know who caused the action and who is acting. Who is the agent of the Araweté warrior's song, the killer or his victim (Viveiros de Castro 1992:241-245)? Who is the Parakanã curer, the dreamer or the dreamt enemy (Fausto 2001:357-369)?

This paradoxical image, in which antagonic elements are condensed and appear to be simultaneously singular and multiple, is the very source of the ritual efficacy of these figures, as Severi argues (2007). In Amazonia, this efficacy suffers from a constitutive instability, since we can never know who adopted who and who controls who: to be powerful, shamans and warriors must ensure that the subjectivity of their wild pets is preserved, which means that they can never become entirely tamed and domesticated (Fausto 1999a:949). This explains the ambivalence of shamans and warriors in Amazonia, forever on the verge of adopting the perspective of the others contained within themselves.

The alteration induced by mastery (the fact that the master is inevitably 'affected' by his wild pet) combines with the multiple relations contained within a magnified person, which produces the latter's relational dispersion. As Rodgers states, "the shaman is a multiple being, a micropopulation of shamanic agencies sheltering within a body: hence neither are his 'intentions' exclusively 'his,' nor can he ever be certain of his own intentions." (2002:121). This plurality also characterizes the killer, who contains relations with different kinds of humans (his victims), but also with non-humans, since his predatory potency must be fabricated before the homicidal act through his 'jaguarization'. Among the Jívaro, for instance, warfare success depends on the prior encounter with the arutam, the image of an ancestor with a jaguar affection, which "will lodge in the recipient like an internal double" (Taylor 2003:237).¹⁸

The fact that the plural and altering nature of Amazonian mastery produces an instability in the ownership relation helps explain why it has rarely crystallized into an institutional locus of power. The very constitution of these functions seems to contain the means for undermining them, since potency depends on an uncertain relation with other-subjects who are never entirely loyal. It would be a mistake, though, to ignore the fact that there were (and still are) institutionalized forms of chiefdom in the region. The question, therefore, resides in knowing how the centrifugal tendencies of the mastery-ownership relation can be blocked, turning them into a mechanism for concentrating and localizing power.

My intuition is that this happened where a limit was posed to the multiplicatory and alterative logics of warfare and shamanism. As I have argued elsewhere, indigenous warfare involves an almost unlimited expansion of the number of killers and vital attributes that can be obtained and transmitted by the warriors (Fausto 2001a:305-306, 330-331; Fausto 1999b:272-275). This amplification is linked to the low degree of hierarchization of men in terms of warfare exploits, since warfare involves multiplying the regenerative capacities to be captured rather than ranking men according to their predatory power. Significantly the highest crystallization of power is found where this logic is curtailed. This is the case, for example, of some Chacoan peoples, where membership of the warrior rank was limited to those who had actually scalped an enemy and brought back the trophy. The victim could be handed over to a companion so he could acquire this status instead, but each trophy corresponded to just one warrior (Clastres 1982:222; Sterpin 1993). In the Aztec case, by contrast, with a much more rigid class system, passing on a sacrificial captive to another person was a crime punished by death (Clendinnen 1991:116).

These are examples of how a mechanism of dispersion can be converted into a mechanism of concentration. Similar processes may well have occurred in the transition from shamanic systems to temple-priest systems, a transition in which the emergence of vertical shamanism, associated with ancestrality, was perhaps an intermediary phase (Hugh-Jones 1994; Viveiros de Castro 2002c:471-2). If so, spatial territorialization (the temple) and temporal territorialization (ancestrality) would have corresponded to the conversion of multiple ownership relations into a pyramidal system of domination. This hypothesis perhaps helps us to conceptualize the prior existence of predatory mega-machines in the Americas – state theocracies that conserved the cannibal principle as a constitutive element of power, subsuming ancestrality and predation within a single hierarchical structure.

Returning from the terrain of hypotheses to the firm soil of ethnology, I turn once more to the Kanamari category warah-, which Costa (2007) translates as owner-body-chief. As we have seen, it serves to express the same structure at different scales: souls contained in bodies, bodies contained in chiefs and chiefs contained in other chiefs. Is there a limit to this magnification? Kanamari mythology flirts with the image of a universal jaguar, a global body containing all the differences found in the post-mythic world. In concrete terms, the limit was the region's main river, the Juruá, itself conceived as a –warah not matched by any form of Kanamari chiefhood.¹⁹ The structural locus, however, was there, waiting to be occupied by another –warah. As announced in Lévi-Strauss's celebrated diagram in The Story of Lynx, the Kanamari structure anticipates a place for other owners – or indeed, for owners of another kind, since this position came to be occupied by whites, as occurred elsewhere in Amazonia.

The mastery relation served to conceptualize the asymmetries that have branded colonial and post-colonial history. This is a recurrent phenomenon that reappeared in native interactions with missionaries, slave raiders, rubber bosses and, more recently, government agents. The relational schema served as a pivot connecting the system of captives derived from indigenous warfare and the colonial slavery system (Karadimas 2001; Santos-Granero 2005), just as it would later serve in the debt-peonage system of the rubber economy. The mastery relation worked to connect a system focused on extracting the regenerative capacities of persons with another system focused on extracting surplus labour and the production of goods. The hierarchical structure of mastery combined with its double face (predation and protection) also served to connect with the structures of colonial power, especially in the context of mission settlements and, later, the tutelary system (Machado 1994).

There are various ethnographic examples of whites being compared to owners-masters. I explored this point elsewhere in describing how the Western Parakanã equate the whites with powerful dream enemies, who are conceived in turn as the dreamer's wild pets since they cure and give songs without requesting anything in return. I also showed that there was a curious inversion of this relation, indicated by the vocative expression used: the dreamer addresses the bestial human enemies as wetom, 'my father,' or more frequently as miangá, a formal term applied to 'fathers' (but never to a person's actual father). It was precisely this vocative that the Western Parakanã employed in their relations with white people throughout much of the 20^th century, contrasting with the affinal terms invariably used to address indigenous enemies.²⁰ This usage dates from the end of the 19^th century, when the Parakanã say that they learnt how to extract industrial goods from the whites peacefully, and was reinforced during the long process of 'pacification' initiated at the end of the 1920s. From the Parakanã perspective, the agents of the State behaved like true fathers-givers and thus subject to indigenous control (Fausto 2001a:469-531, 2002b, 2002c). This is precisely the magic of the Parakanã dreamer: his shamanic potency (ipajé) resides in his capacity to extract a voluntary action from dream enemies; a form of magic similar, for example, to the decorated canoes made by kula traders whose enchantment (Gell 1998) is intended to ensnare the recipient and persuade him to release his most precious objects.

There are other Amazonian contexts in which mastery was also applied to the relation with white people. This is the case of the Paumari, an Arawá-speaking people, for whom, Bonilla writes (2007), all existents (animals, plants, objects) possess a human form conceived as an owner-master. As is usual in Amazonia, the relation between an owner and its species is conceived in the same way as the familiarization of animals and the adoption of children. But there are also asymmetric relations between different species, which are assimilated to the boss-employee relation (Bonilla 2007:199-205).

White people provide a clear representation of this double condition of 'masters' and 'bosses,' to the extent that the Paumari have borrowed a term from the Amazonian Língua Geral to designate whites as a whole: Jara, 'owner.' The whites are masters possessing a predatory power expressed in the capture of Paumari children, kept in boxes like pet animals. The shamans must release these children as they do in other cases of 'soul theft' (Bonilla 2007:87). But though white people appear here as one more figure among the masters populating the non-human world, it is not just the indigenous world that is projected onto the relation with the colonizers: the indigenous model itself is inflected by the historical relations of work and dependency. In the words of a former Paumari shaman: "The shaman is the father of the itavari [auxiliary spirit]. He's like a governor. Whatever the shaman tells him, he must do and obey, like an employee. The itavari are keen to work and follow the shaman's orders since then they will be able to come to the ihinika [ritual]" (Bonilla 2007:355).²¹

As Bonilla suggests, there is yet another twist to this tale since the model of adoption, inflected by the boss-employee relation, was a way of using the asymmetric interaction with whites to control the latter's predatory potential. The Paumari placed themselves in the position of wild pets, trying to convert a predatory relation into care and protection. The strategy of submission also contains a lure, since it is a way of eliciting the action corresponding to the owner position, defined as someone who looks after and feeds his children-pets (Bonilla 2005:58).²² The masters live in a world of abundance — they are bigger, richer and more fertile – and people expect that if they do not behave as predators, they will behave as providers.

The same resonance between historical relations of power and exploration and the indigenous cosmology is found among the Ávila Runa of Ecuador (Kohn 2002, 2007). Here the animal masters express different figures of power and authority from the pre-colonial, colonial and post-colonial past. The world in which they live is described as an urban network with its own hierarchy: the main owners, curagas, live in a kind of 'Quito in the Forest,' located within the Sumaco Volcano, linked by roads to other smaller towns where less important owners live. Another image employed to describe this world is that of the haciendas of the rubber era with their bosses-owners and their domestic animals kept in corrals.²³ Both towns and haciendas are imagined as places of abundance in which the most powerful masters walk around with jaguars by their side like pet dogs (Kohn 2007:109-120).²⁴

The mastery-ownership relation is not only productive in terms of conceptualizing the asymmetries between Indians and whites, or humans and non-humans. It is also a key to understanding the asymmetric relations between indigenous peoples. I suggest that mastery-ownership was an important sociological mechanism in the past, serving to structure hierarchical relations between different indigenous groups, something still observable today in some regional systems. This appears to be the case of the asymmetric relation between the Maku people and the Tukanoan and Arawakan peoples on the Upper Rio Negro (Ramos, Silverwood-Cope & Oliveira 1980), or between the Guaná (Terena and Kinikinau) and the Mbayá-Guaykuru (Kadiwéu), historical evidence of which dates back to Ulrich Schmidel's voyage in the first half of the 16^th century (Cardoso de Oliveira 1976:31-2).²⁵

From the same period comes the first information on the Chiriguano, a people formed by the violent asymmetric fusion of the Guarani and Chané, the latter speakers of an Arawakan language (Combès & Saignes 1991). In the 16^th century, the Chané were literally cannibalized and incorporated in a subordinate position, to the point of being defined as 'slaves' (tapii) to the Guarani, described as 'masters' (iya) (Combès & Lowrey 2006:692).²⁶ However, from the 19^th century onwards, a group of Chané descent, the Izoceños, began to try to reverse this asymmetry, proclaiming themselves, significantly, Iyambae: 'those-without-masters.' This term, initially used as a surname by a dynasty of Izoceños chiefs, has today been converted into a new ethnic marker, providing the names for a territory called the 'land without owner' (Ivi Iyambae) and a homonymous foundation (Combès 2005; Combès & Villar 2005):

]]> For outsiders to Izozog, it suggests freedom and equality; it can evoke […] the 'society against the state' scenario that Pierre Clastres (1982 [1974]), on the basis of Guaraní examples, embedded in the expression. For Izoceño insiders, by contrast, it takes on an establishmentarian cast. In Izozog, to be 'without owner' is to occupy a particularly embedded social position that is materially rewarding; to be 'without master' is to occupy the summit of an Arawakan hierarchy. (Combès & Lowrey 2006:700-701)

There is no space here to discuss the nature of this hierarchy or its association with Arawakan peoples (see Heckenberger 2002, Santos-Granero 2002). Neither is this the place to ask whether, when, where and how these relations of ownership were converted into relations of domination. For the purposes of this article, it is enough to note the productivity and generality of the owner-master idiom.

Conclusion

We have seen how the relation between mastery, conceived as an adoptive filiation, operates at different scales ranging from the microconstitution of the person to the macroconstitution of the cosmos. We have also seen that as a relational schema it defines interactions in highly diverse spheres and between very distinct entities (humans, animals, plants, spirits, artefacts). And, finally, we have observed the existence of a dynamic in which this same schema is inflected by new historical situations. But what, in sum, am I arguing?

In fact, the first paragraph of this text already announced my intentions. It paraphrases a passage from Lévi-Strauss's 1943 article "The Social Use of Kinship Terms among Brazilian Indians," which inaugurated a whole school of Americanist thought on the brother-in-law relationship, of which we are all heirs thanks to the works of various colleagues.²⁷ What I am suggesting, then, is that the relation of mastery operates, like symmetric affinity, as a 'cosmological operator' (Viveiros de Castro 1993). If, as Viveiros de Castro proposes, Amazonian sociocosmologies posit an 'affinity without affines' (intensive and potential) they also posit a type of cosmopolitical and interspecific filiation (a metafiliation) in which adoption rather than the vertical transmission of substances is the crucial element.²⁸ But could the same not be said of other relational modalities, like asymmetric affinity (the father-in-law/son-in-law relationship) or symmetric consanguinity (siblinghood)? Specific contexts aside, these modes of relations do not reach the level of generality of the brother-in-law relation and adoptive filiation. This is because they occupy the polar positions of identity and difference, tending to slip either into the sterile fixity of the same or into uncontrollable cannibal potency.

In Amazonia, siblinghood, particularly same-sex, is often taken to be the core of identity (Fausto 1991:72) and limited to this domain. There is no meta-siblinghood as found in India or universal brotherhood of the Christian kind. In Amazonia, siblinghood only connects wider domains where religious conversions have taken place, especially to the new Evangelism (Vilaça 1996). Whenever siblinghood emerges as a sociocosmic idiom, an asymmetry based on birth order is introduced.²⁹ This is the case of the twin sagas analyzed by Lévi-Strauss (1991), as well as the myths on the origin of the difference between Indians and white people, equated with an inversion of seniority between brothers.³⁰ Birth order can also serve as a sociocosmic ruler for marking hierarchical differences between segments of the same people, as occurs in the Upper Rio Negro system, or between siblings descending from chiefs, as happens in the Upper Xingu. As an identificatory relation, however, siblinghood does not constitute a generalized cosmopolitical idiom, though the sibling group is a fundamental unit in the structuration of the region's political dynamics.

The father-in-law/son-in-law relation is found on the opposite pole to siblinghood, since it is constituted on superimposed differences and asymmetries: the difference between wife-givers and wife-takers and the difference between generations. The relation is overly potent and quickly veers towards figures of power and cannibal voracity. Not by chance, the two pre-eminent images of this relation in the regional literature are the overworked son-in-law and the jaguar father-in-law, as suggested, for example, in Yekuana myths in which the son-in-law has to carry out super-human tasks to avoid being devoured by his cannibalistic father-in-law (Guss 1989:80, 94).

As Turner (1979, 1991) and Rivière (1984, 1987) show, this is the only kinship relation in Amazonia that involves the substitution of one person's work by another. The son-in-law works for, or in the place of, the father-in-law: he must hunt, fish, clear swiddens, build the house. This obligation derives from the fact of marriage itself and there are few ways of evading it entirely, except by capturing spouses. Even in the Upper Xingu, where bride wealth exists in certain situations, the latter serves only to attenuate, not annul, the services provided to the father-in-law. Rivière and Turner analyze this relation as a mechanism for controlling persons with repercussions on the formation of leaders and on the autonomy of adult men in general. Important variations exist in terms of its structural effects, depending on whether the society is uxorilocal or virilocal, on whether the rule of residence is mechanical or statistical, and on the length of bride-service. Even so, it is likely that had they to identify one kinship relation involving authority and control in indigenous Amazonia, nine out of ten specialists would pick the relation between father-in-law and son-in-law.³¹

Nonetheless, it does not provide a general idiom for schematizing relations as diverse as those between shamans and auxiliary spirits, warriors and victims, captors and captives, masters and pets – despite the fact that, in Amazonia, the son-in-law's position in an uxorilocal system is frequently compared to that of a captive enemy or a pet animal. The lower generality of asymmetric affinity can be traced to the fact that Amazonian masters are double-sided: they are voracious jaguars for other peoples and protective fathers for their own. Fathers-in-law, on the contrary, tend to be all jaguar. This does not mean that the relation cannot operate as a cosmopolitical idiom under certain contexts. Tupi groups, with their cannibal inclination, have always flirted with this possibility. Among the Araweté, for example, asymmetric affinity schematizes the relation between the living and the divinities, pervaded by the same positional ambivalence as other shamanic relations (Viveiros de Castro 1992:218).

By contrast, matrimony is a more productive relation. In mythology it appears as a central mechanism in the passage from one kinship situation to another. Interspecific marriages are numerous in myths and very often eclipse the same-sex affinal relations they necessarily imply. In addition, some indigenous peoples conceive the shaman's relation with his auxiliary spirits as a matrimonial bond, and shamans may constitute spirit families. The examples I know are concentrated in Western Amazonia, found among the Shipibo-Conibo of Peru (D'Anglure & Morin 1998), the Chimane of Bolivia (Daillant 1998, 2003:308-313), the Achuar of Ecuador (Descola 1986:346-48, Taylor 1993:437-439), the Harakmbut of Peru (Califano 1988:117-119), the Wari' of Rondônia (Vilaça 2006:202-203) and the Nambikwara-Mamaindê of Mato Grosso, both in Brazil (Miller 2007:198-200).

Matrimony dynamically expresses a set of kinship relations, since the shaman constitutes actual spirit families: he has a wife and affines, and engenders spirit-children. Even in these cases, however, there seems to be on one hand an emphasis of paternity – a man begins his shamanic life as a husband and ends up as a father, a bond conceived to be more stable and secure (D'Anglure & Morin 1998:67; Daillant 2003:313) ­– and on the other hand, a relative de-emphasis of affinity.³³ It is as though matrimony itself converges towards adoption rather than alliance, in contrast to what occurs, according to Hamayon (1990), in Siberian hunting shamanism.

The Nambikwara-Mamaindê provide us with the most suggestive example of this assimilation of marriage and familiarization (Fausto 2001b). The spirit-wife, a jaguar, is denominated mãindu ('my fosterling' or 'my wild pet') by the shaman-husband. As would be expected, we can also observe here the positional instability that marks the relations between shamans and auxiliary spirits in general: "it is never known for certain who is 'fostering' whom. Although the shaman calls his spirit-wife 'my fosterling,' by sharing food and body decorations with her, the shaman indicates that it is he who is being 'fostered' by her" (Miller 2007:199).

In sum, none of the relations analyzed above seems to have the same degree of generality in Amazonia as symmetric affinity and asymmetric consanguinity. The first combines difference and symmetry; the second, identity and asymmetry. The overlapping of new differences (of gender or generation) is less operative and is limited in terms of both ethnic and spatial distribution, as well as sociocosmic domains. Meta-affinity and meta-filiation are both elective, dispensing with any other prior relation: one can be an enemy/brother-in-law of anyone, just as one can adopt any enemy one wishes. We are not talking about just any adoptive filiation, though, or just any brother-in-law relation. The latter, in its intensive modality (potential affinity), is a figure of enmity, while the former is a figure of ownership, of the asymmetric relation between the owner-master and his children-pets.

There is a final point I wish to make: adoption is, so to speak, an incomplete filiation. It does not produce full identity but an ambivalent relation in which the substrate of enmity is obviated, yet not entirely neutralized. Hence my emphasis on the persistence of the other's perspective in the shaman-auxiliary spirit or killer-victim relationship. This may explain why captives, orphans and pet animals often receive treatment that veers between care and cruelty. The master's double face is matched by the pet's wildness: the latter is an other and will never cease to be so entirely.

To finish, I once more paraphrase Claude Lévi-Strauss (1943:409): a sufficient number of convergent indications have been recorded so that we may consider the outstanding character of the mastery relationship a specific feature of Amerindian sociocosmology, configuring a world of owners and enemies, but not necessarily one of domination and private ownership.

Notes

¹ I have been writing this text in my head for years. Some of its ideas have been presented on my courses in Brazil, as well as in a seminar at EHESS, in 2005, run by Carlo Severi, whom I thank for the invitation. I also thank Marina Velasco and Federico Neiburg for the opportunity of studying Locke on a course that we gave together in 2004. My thanks to Marc Brightman, Vanessa Grotti and Olga Ulturgasheva for inviting me present this text in the conference "Humains, animaux, plantes et choses: la notion de personne en Amazonie et Sibérie contemporaines," at the Musée du Quai Branly, in 2008. Finally, I am grateful for the reading and comments of Aparecida Vilaça, Cesar Gordon, Federico Neiburg, Marina Velasco, Marnio Teixeira-Pinto and in particular Luiz Costa, with whom I have maintained an intense dialogue on the theme.

² Yawalapiti is a southern Arawak language. The cognates of wököti in the region are the Mehinaku wekehe (Gregor 1977) and the Wauja wekehö (Barcelos Neto 2004). For a discussion of owners and ownership among the Arawak of the Upper Xingu, see Ball (2007).

⁴ I employ the concept of singularity to designate an internally multiple and non-self-identical unity, following its contemporary usage, inspired by Deleuze (1968). Sometimes I also use the composite expression 'plural singularity.' As Viveiros de Castro points out (2007), in anthropology the concept has resonances with the proposals of Strathern (1988, 1992) and Wagner (1991) for redefining the relation between part/whole, particular/collective at different scales from the microconstitution of the person to the macroconstitution of the social. Though I cannot develop the point here, it is important to note that the type of sovereignty implied by the Amazonian notion of 'owner' differs from that implied by our own concept of political body; in other words, the chief-owner-body is not a Leviathan.

⁵ See, especially, Fausto (2001), as well as the articles by Erikson (1987, 2000) and Descola (1994) on which this argument is based. In relation to shamanism, see Albert on the Yanomami (1985:316), Bonilla on the Paumari (2007:355), Briggs on the Warao (1994:141-142), Chaumeil on the Yágua (1983:120), Costa on the Kanamari (2007:49), Henry on the southern Gê (1964:73) and Wagley on the Tapirapé (1976:242). In relation to warfare and trophy rituals, see Menget on the Ikpeng (1988:67), Santos-Granero on the Conibo (2005:156-157), Sterpin on the Nivakle (1993:42) and Taylor on the Jívaro (1994). In relation to ceremonial objects, see Staden on Tupinambá maraca rattles (2008[1557]:125), Menget on Ikpeng flutes (1988), Hugh-Jones on Barasana body adornments and musical instruments (1996:141) and Maia Figueiredo (2009) on the Jurupari flutes made by the Baré.

⁶ I am unable to discuss here the quality of this 'care,' a central theme in the works of Overing Kaplan (1999), inspired by the moral philosophy of virtues (MacIntyre 1981, Larrabee 1993, Baier 1994).

⁷ The term iamït applies to the animals controlled by masters, to the pets raised by humans and to adopted orphans (Teixeira-Pinto 1997:314).

⁸ Analyzing ayahuasca visions and the foetus-placenta relation among the Piro, Gow (1999:237) likens this typology to that of a Klein bottle in which the inside is simultaneously the outside of the recipient.

⁹ The reification of the chief-form, which makes him the singular image of an owner-master, also makes him something that belongs to the community: "Persons are owned as things through a political-ritual fabrication that presents the person being claimed by another as singular, entire and whole [...]" (Strathern 2005:120). Entire and whole, but simultaneously singular and plural, since here the individual is not opposed to the collectivity: "for whether we see a man or a clan is in one sense irrelevant: collective action aggrandises each man's performance but is no different in kind from his own aggrandisement as a single person" (Strathern 1999:37).

¹⁰ As I tried to show, this prey/predator split is an essential element in the constitution of the person in Amazonia (Fausto 2002a, 2007; Taylor 2000). For a different but consonant formulation, see Gordon's analysis (2006:217-218) of the Xikrin categories àkrê (described as a capacity for self-subjectification and other-objectification) and uabô (described as a capacity for self-objectification and other-subjectification).

¹¹ Even in the case of Locke, I focus only on the Second Treatise of Government and chapter XXVII, Book 2, of An Essay Concerning Human Understanding. Locke's work contains another model of the person, described as a passive repository of capacities, that pervades his ideas on education, especially of labourers (Tully 1993a:88). It should also be noted that Locke's notion of the individual as an owner of his/her self (and, therefore, as a non-slave) can be traced back to a long tradition of Roman law, which is taken up by Grotius, Hobbes and the Levellers, before Locke (Tully 1993b).

¹² Unsurprisingly the argument has baffled some commentators. Lloyd Thomas, for example, considers Locke's premise absurd, since it is based on the idea that the mixture of bodily secretions with things adds something to the latter that nature did not provide them: " 'labour' cannot be mixed with the substances things are made of" (Lloyd Thomas 1995:109). From an anthropological perspective, on the contrary, the argument encounters numerous ethnographic resonances.

¹⁴ This expression was used to explain to me why the old school had been practically abandoned by 1998, even though the Kuikuro consider school education to be extremely important. Any researcher who has tried to give something 'in common' to an indigenous people must have faced an impasse, since everything always has an owner. Today among the Kuikuro with the proliferation of commercially valuable objects, a distinction is marked between what belongs to the comunidade ('community') and what is 'particular' – two Portuguese terms that are interspersed in Kuikuro phrases. 'Community' objects are those belonging to the chief without being his private property.

¹⁵ From which Balibar concludes, with a certain irony, that the notion of identity as that which 'differs from the different' was not a discovery of the Hegelian dialectic.

¹⁶ A more rigorous exposition of the differences between our philosophical alternatives and indigenous cosmologies eludes the objectives of this text. It would require us to consider other Western philosopical models of personal identity that, in contrast to Locke, are not based on the equation between ipse (self) and idem (same).

¹⁷ I refer above all to Society Against the State in which 'primitive society' appears as a kind of Gulagian nightmare (pictured as a dream) "from which nothing escapes […] since all exits are closed." (Clastres 1978:147-148).

¹⁸ Among the Yanomami, warriors about to leave on a war expedition underwent a ritual to absorb a vulture-image, necessary for digestion of the future victim (Albert 1985:363), while Wayana men were scarified with jaguar or bird-of-prey patterns to instil them with the predatory disposition needed for success (van Velthem 2003:354).

¹⁹ The mapping of the Kanamari structure onto the region's hydrography – in which the largest river is the -warah, which contains its affluents, which, in turn, contain their own affluents, and so on – also echoes the hierarchical structure of the rubber economy. On the relationship between this cartography and shamanism, see Gow (1994) and Carneiro da Cunha (1998).

²⁰ Following the submission to State administration in the 1980s, the Parakanã began to address white people by the vocative wepajé, 'my ritual-friend,' a term with connotations of affinity and enmity. On this term, see Fausto (1995:75-78; 2001:285-297).

²¹ Ihinika is a ritual in which children are captured by the human-part of foods and later rescued by shamans (Bonilla 2007:14).

²² This strategy, in which a subject looks to place him or herself under the protection of a master and extract an action from him, also characterizes the affective language of interpersonal relations among the Candoshi, for whom "the paradigm of familiarization comprises [...] the conceptual basis of all affective relations within the family" (Surrallès 2003:69). Among the Toba of the Chaco, the language of submission is designed to elicit the compassion and generosity of the spirits-masters (Tola 2006).

²⁴ The association between dog and jaguar, including at a lexical level, occurs in some parts of the Americas and is not merely derived from their morphological and behavioural similarity (many of the dogs of the Conquistadors were large hunting dogs): it also reflects the status of canines as a ferocious domesticated animal under the control of an owner, which enabled them to be associated with the (invisible) jaguars familiarized by shamans and warriors.

²⁵ The relation of dependency and protection between these peoples contrasts with the violence of the Guaikuru against the Guaxi, Guató and Chamacoco, indicating that the Guaná tried to control predation through submission, just as the Paumari did in relation to the whites.

²⁶ The translation of tapii as 'slave' should be considered carefully (see Combès 2005:60-68). Among the coastal Tupi, the term designated the non-Tupi Indians, but did not indicate a relation of submission, as appears to have happened in the Chiriguano case. For a survey of the theme of slavery and other forms of subordination among indigenous peoples, see Santos-Granero (2005; 2008).

²⁷ See, above all, the innovative synthesis produced by Viveiros de Castro (1993), built on the works of Rivière (1969, 1984), Overing Kaplan (1975, 1984), Basso (1975), Menget (1977), Carneiro da Cunha (1978), Taylor (1983, 1985, 1989), Albert (1985) and Erikson (1986), among others. Also see the subsequent works of Descola (1993, 2001) and Taylor (2000). For a discussion of the inaugural nature of Lévi-Strauss's article, see Coelho de Souza & Fausto (2004).

²⁸ I use the prefix meta- in the sense given by Taylor (2000:312) who, in turn, takes the expression from Jamous (1991) on meta-siblinghood in India (see also Dumont 1975).

²⁹ The only exception that comes to mind is the relation of the Guajá person with his or her homonym (animal, plant, artefact), a relation conceived as siblinghood (Cormier 2003:91).

³⁰ See, for instance, the Barasana myth analyzed by Hugh-Jones (1988:143-44), or the 17^th century Tupinambá version recounted by Abbeville (1975 [1614]:251-2).

³¹ For a re-reading of the theme of control and leadership in the Guianese case, including a discussion of the notion of ownership, see Brightman (2007). For a general hypothesis on the structural effects of marriage in bride-service societies, see Collier & Rosaldo (1981). For a critical analysis of this hypothesis in the Melanesian context, see Kelly (1993:415-525), and in the Amazonian context, Fausto (2001a:201-210).

³² Some caution is needed in relation to the translations. For example, the 'mother of the peccary' among the Munduruku (Murphy 1958), or the 'mothers of the game' among the Achuar (Descola 1986:317) are effectively called 'mother' in the indigenous language. However, the regional Spanish translation of the Yágua term hamwo or the Arakambut term wachipai as 'madre' is equivocal, since these terms have another meaning in these languages (Chaumeil 1983:74; Gray 1997:53).

Bibliography

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). Tese de Doutorado, Université de Paris-X (Nanterre).

ANDRADE, Lucia M. M. de. 1992. O Corpo e o Cosmos: Relações de Gênero e o Sobrenatural entre os Asurini do Tocantins. Dissertação de Mestrado, Universidade de São Paulo.

ÅRHEM, Kaj. 1996. "The Cosmic Food Web: Human-Nature Relatedness in the Northwest Amazon". In: P. Descola e G. Pálsson (orgs.), Nature and Society: Anthropological Perspectives. London: Routledge. pp. 185-204.

BELAUNDE, Luisa Elvira. 2001. Viviendo Bien: Género y Fertilidad entre los Airo-Pai de la Amazonía Peruana. Lima: CAAP.

BIRD-DAVID, Nurit. 1990. "The Giving Environment: Another Perspective on the Economic System of Gatherer-Hunters". Current Anthropology, 31:189-196.

BRIGHTMAN, Robert A. 1993. Grateful Prey: Rock Cree Human-Animal Relationship. Berkeley: University of California Press.

CALAVIA SÁEZ, Oscar. 2001. "El Rastro de los Pecaríes: Variaciones Míticas, Variaciones Cosmológicas y Indentidades Étnicas en la Etnología Pano". Journal de la Société des Américanistes, 87:161-176.

CHAUMEIL, Jean-Pierre. 1983. Voir, Savoir, Pouvoir: Le Chamanisme chez les Yagua du Nord-Est Péruvien. Paris: École des Hautes Études en Sciences Sociales.

CLASTRES, Hélène. 1975. La Terre Sans Mal: Le Prophétisme Tupi-Guarani. Paris: Seuil.

CLASTRES, Pierre. 1968. "Ethnographie des Indiens Guayaky". Journal de la Société des Américanistes, LVII:8-61.

___ . 1982. Arqueologia da Violência: Ensaios de Antropologia Política. São Paulo: Brasiliense.

CONKLIN, Beth. 1995. "'Thus Are Our Bodies, thus Was Our Custom': Mortuary Cannibalism in an Amazonian Society". American Ethnologist, 22 (1):75-101.

___ . 2001. Consuming Grief: Compassionate Cannibalism in an Amazonian Society. Austin: University of Texas Press.

DESCOLA, Philippe. 1986. La Nature Domestique: Symbolisme et Praxis dans l'Écologie des Achuar. Paris: Maison des Sciences de l'Homme.

___ . 1992. "Societies of Nature and the Nature of Society". In: A. Kuper (org.), Conceptualizing Society. London: Routledge. pp. 107-126.

___ . 1993. Les Lances du Crépuscule: Relation Jivaros. Haute Amazonie. Paris: Plon.

___ . 1998. "Estrutura ou Sentimento: A Relação com o Animal na Amazônia". Mana, 4(1):23-45.

DOLE, Gertrude. 1962. "Endocannibalism among the Amahuaca Indians". Transactions of the New York Academy of Sciences, 24(5):567-573.

ERIKSON, Philippe. 1986. "Alterité, Tatouage et Anthropophagie chez les Pano: La Belliqueuse Quête de Soi". Journal de la Société des Américanistes, LXII:185-210.

FAUSTO, Carlos. 1999a. "Of Enemies and Pets: Warfare and Shamanism in Amazonia". American Ethnologist, 26(4):933-956.

___ . 1999b. "Da Inimizade: Forma e Simbolismo da Guerra Indígena". In: A. Novaes (ed.), A Outra Margem do Ocidente. São Paulo: Companhia das Letras. pp. 251-282.

___ . 2001. Inimigos Fiéis. História, Guerra e Xamanismo na Amazônia. São Paulo: EDUSP.

___ . no prelo. "A Blend of Blood and Tobacco: Shamans and Jaguars among the Parakanã of Eastern Amazonia". In: N. Whitehead e R. Wright (eds.), Dark Shamans: The Anthropology of Assault Sorcery in Amazonia. Chapell Hill, NC: Duke University Press.

GIRALDO-FIGUEROA, Alba Lucy. 1997. Guerriers de l'Écriture et Commerçants du Monde Enchanté: Histoire, Identité et Traitment du Mal chez les Sateré-Mawé (Amazonie Centrale, Brésil). Tese de Doutorado, École des Hautes Études en Sciences Sociales.

GONÇALVES, Marco Antonio. 2001. O Mundo Inacabado: Ação e Criação em uma Cosmologia Amazônica. Etnografia Pirahã. Rio de Janeiro: Editora da UFRJ.

GOULARD, Jean-Pierre. 1998. Les Genres du Corps: Conceptions de la Personne chez les Ticuna de la Haute Amazonie. Tese de Doutorado, École des Hautes Études en Sciences Sociales.

GOW, Peter. 1991. Of Mixed Blood: Kinship and History in Peruvian Amazonia. Oxford: Clarendon Press.

___ . 1997. "O Parentesco como Consciência Humana". Mana, 3(2):39-65.

___ . 2001. An Amazonian Myth and its History. Oxford: Oxford University Press.

GRAY, Andrew. 1996. The Arakmbut: Mythology, Spirituality, and History in an Amazonian Community. Providence, RI: Bergham Books.

___ . 1997. The Last Shaman: Change in an Amazonian Community. Providence, RI: Bergham Books.

GREGOR, Thomas. 1990. "Uneasy Peace: Intertribal Relations in Brazil's Upper Xingu". In: J. Haas (ed.), The Anthropology of War. Cambridge: Cambridge University Press. pp. 105-124.

HARNER, Michael. 1973. Hallucionogens and Shamanism. Oxford: Oxford University Press.

___ . 1978. Shuar: Pueblo de las Cascadas Sagrada. Quito: Ediciones Mundo Shuar.

HUGH-JONES, Stephen. 1996. "Bonnes Raisons ou Mauvaise Conscience? De l'Ambivalence de Certains Amazoniens envers la Consommation de Viande". Terrains, 26:123-148.

INGOLD, Tim. 2000. The Perception of the Environment: Essays on Livelihood, Dwelling and Skill. London: Routledge.

KARADIMAS, Dimitri. 1997. Le Corps Sauvage. Idéologie du Corps et Représentations de l'Environnement chez les Miraña d'Amazonie Colombienne. Tese de Doutorado, Université de Paris X (Nanterre).

KELLY LUCIANI, José Antonio. 2001. "Fractalidade e Troca de Perspectivas". Mana, 7(2):95-132.

KENSINGER, Kenneth. 1975. The Cashinahua of Eastern Peru. Providence, RI:The Haeffenreffer Museum of Anthropology, Brown University.

LAGROU, Elsje. 1998. Cashinahua Cosmovision: A Perspectival Approach to Identity and Alterity. Tese de Doutorado, University of St. Andrews.

LÉVI-STRAUSS, Claude. 1964. Le Cru et le Cuit. Mythologiques I. Paris: Plon.

___ . 1965. "Le Triangle Culinaire". L'Arc, 26:19-29.

LIMA, Tânia Stolze. 1995. A Parte do Cauim. Etnografia Juruna. Tese de Doutorado, PPGAS/Museu Nacional/ UFRJ.

___ . 1996. "O Dois e seu Múltiplo: Reflexões sobre o Perspectivismo em uma Cosmologia Tupi". Mana, 2(2): 21-47.

MCCALLUM, Cecilia. 1996. "Morte e Pessoa entre os Kaxinawá". Mana, 2(2):49-84.

___ . 2001. Gender and Sociality in Amazonia: How Real People Are Made. Oxford: Berg.

OAKDALE, Suzanne. 1998. The Power of Experience: Agency and Identity in Kayabi Healing and Political Process in the Xingu Indigenous Park. Tese de Doutorado, University of Chicago.

OVERING (KAPLAN), Joanna. 1975. The Piaroa: A People of the Orinoco Basin. Oxford: Clarendon Press.

___ . 1991. "A Estética da Produção: O Senso de Comunidade entre os Cubeo e os Piaroa". Revista de Antropologia, 34:7-33.

___ . 1999. "Elogio do Cotidiano: A Confiança e a Arte da Vida Social em uma Comunidade Amazônica". Mana, 5(1):81-107.

REICHEL-DOLMATOFF, Gerardo, 1973. Desana: Le Symbolism Universel des Indiens Tukano du Vaupés. Paris: Gallimard.

RIVAL, Laura. 1996. "Blowpipes and Spears: The Social Significance of Huaorani Technological Choices". In: P. Descola e G. Pálsson (orgs.), Nature and Society: Anthropological Perspectives. London: Routledge. pp. 145-164.

___ . 1999. "Prey at the Center: Resistance and Marginality in Amazonia". In: S. Day, E. Papataxiarchies e M. Stewart (eds.), Lilies of the Field: Marginal People who Live for the Moment. Boulder, CO: Westview Press. pp. 61-79.

SANTOS-GRANERO, Fernando. 1991. The Power of Love: The Moral Use of Knowledge amongst the Amuesha of Central Peru. London: Athlone.

SCHADEN, Egon. 1954. "Aspectos Fundamentais da Cultura Guaraní". Boletim da FFCL, 188 (Antropologia 4), Universidade de São Paulo.

SISKIND, Janet. 1973. To Hunt in the Morning. Oxford: Oxford University Press.

SMITH, David M. 1998. "An Athapaskan Way of Knowing: Chipewyan Ontology". American Ethnologist, 25(3): 412-432.

STERPIN, Adriana. 1993. "La Chasse aux Scalps chez les Nivacle du Gran Chaco". Journal de la Société des Américanistes, LXXIX:33-66.

STRATHERN, Marilyn. 1988. The Gender of the Gift: Problems with Women and Problems with Society in Melanesia. Berkeley: University of California Press.

TANNER, A. 1979. Bringing Home Animals: Religious Ideology and Mode of Production of the Mistassini Cree Hunters. London: Hurst.

TAYLOR, Anne-Christine. 2000. "Le Sexe de la Proie. Répresentation Jivaro du Lien de Parenté". L'Homme, 154-155:309-334.

___ . s/d. "Sick of History. Contrasting Regimes of Historicity in the Upper Amazon". In: C. Fausto e M. Heckenberger (orgs.), Time Matters: History and Historicity in Amazonia (em preparação).

VENGOECHEA, Consuelo. 2001. Classificación Nominal el la Lengua Muinane. Comunicação apresentada no Centre d'Études des Langues Indigènes Amerindiennes (CELIA), Paris, novembro.

VILAÇA, Aparecida. 1992. Comendo como Gente: Formas do Canibalismo Wari'. Rio de Janeiro: Editora da UFRJ.

___ . 1999. "Devenir Autre: Chamanisme et Contact Interethnique en Amazonie Brésilienne". Journal de la Société des Américanistes, 85: 239-260.

___ . 2000. "Relations between Funerary Cannibalism and Warfare Cannibalism: The Question of Predation". Ethnos, 65(1):84-106.

___ . 2002. "Making Kin Out of Others". Journal of the Royal Anthropological Institute, 8(2):347-365.

VIVEIROS DE CASTRO, Eduardo. 1992. From the Enemy's Point of View: Humanity and Divinity in an Amazonian Society. Chicago: Chicago University Press.

___ . 1996. "Os Pronomes Cosmológicos e o Perspectivismo Ameríndio". Mana, 2(2):115-144.

___ . 2001. "GUT Fellings about Amazonian: Potential Affinity and the Construction of Sociality". In. L. Rival e N. Whitehead (orgs.), Beyond the Visible and the Material. Oxford: Oxford University Press. pp. 19-44.

For Lévi-Strauss, in celebration of his hundredth birthday