Olivier Allard. MPhil SAR. Department of Social Anthropology. University of Cambridge.

EMOTIONS AND RELATIONS: A POINT OF VIEW ON AMAZONIAN KINSHIP.

1st September 2003.

In accordance with Regulation 8 of the General Regulations for the MPhil Degree. I declare that this thesis is substantially my own work. Where reference is made to the works of others the extent to which that work has been used is indicated and duly acknowledged in the text and bibliography.

This length of this work is 18, 500 words, including footnotes but excluding bibliography.

1

Table of Contents.

Preface………………………………………………………………………………………

3

Theoretical Introduction: emotions and kinship…………………………………………….

5

A. The anthropology of emotions………………………………………………………... 1. The dual nature of emotions………………………………………………………… 2. Felt emotions: psycho-biological universality……………………………………… 3. Emotions as cultural meaning………………………………………………………. 4. Is synthesis a solution?………………………………………………………………

5 5 6 9 11

B. The new studies of kinship……………………………………………………………. 1. The Schneiderian critique…………………………………………………………… 2. Some new studies of kinship……………………………………………………….. 3. Consanguinity and the rest…………………………………………………………..

14 14 16 17

I. A world of others………………………………………………………………………… 1. A world of affines……………………………………………………………………... 2. Male others: a world of suspicion…………………………………………………….. 3. Female others: desire………………………………………………………………….. 4. Between humans and non-humans……………………………………………………..

21 21 22 26 27

II. Familiarisation and compassion…………………………………………………………. 1. A cosmological scheme……………………………………………………………….. 2. Seduction and adoption………………………………………………………………... 3. Affective states………………………………………………………………………… 4. The nature of compassion……………………………………………………………...

29 29 31 33 35

III. Shame and restraint: in-law relationships………………………………………………. 1. Real affines cannot be avoided………………………………………………………… 2. Self-control……………………………………………………………………………. 3. Shame and restraint…………………………………………………………………….

39 39 41 43

IV. The production of children as kin………………………………………………………. 1. From biology to ethnobiology…………………………………………………………. 2. The construction of consanguinity: a new point of view on substance………………... 3. From substance to affects……………………………………………………………… 4. Between care and desire………………………………………………………………..

46 46 48 51 55

Conclusion…………………………………………………………………………………..

58

Bibliography………………………………………………………………………………… 60

2

PREFACE.

From March until September 2002, I conducted fieldwork in a Guaraní village located in eastern Paraguay, seeking to understand the difficulties of a project of development that involved the inhabitants. I was interested in all aspects of social life, and sought to gather data about everything that could explain how these people acted. My aim was however more restricted in practice, because my ideas led me to focus on structural aspects of religious power and on kinship relations: I compared the positions of the shaman and of the chief, tried to record genealogies and terminologies, tried to interpret stock-breeding by reference to cosmological ideas about nature. This proved of course to be quite unsatisfactory, probably because I did not stay long enough, probably also because of some limitations of my point of view. I realised afterwards that I had not paid any attention to many phenomena. People talked a lot in terms of other people being “shameless” or “shameful”, in terms of others being “angry”, or making themselves “angry” by their behaviour. I naively assumed – in spite of sometimes being startled by their use of these notions – that it expressed our natural emotional nature, and that it did not have a place in an anthropological account. When in another, much smaller, village, some told me: “we’re all kin here”, I also tried naively to confirm this statement by asking them about genealogies, without imagining that it could mean something else. My aim here is to try to show that these facts or statements can become meaningful anthropological data if one adopts a point of view radically different from that which I held when I conducted this fieldwork. It means trying to show that it is possible to give a very different account of social reality if one pays attention to affective states and emotions, that some of these difficulties disappear while others emerge when the perspective is changed.

3

In order to do so, I will first offer some theoretical guidelines on the anthropology of emotions and some recent studies of kinship, before turning to more concrete Amazonian ethnography. I would like to thank Stephen Hugh-Jones for his friendly encouragements and his helpful suggestions, Matthew Carey and Jacob Copeman for reading and helping me to improve this dissertation.

4

THEORETICAL INTRODUCTION. EMOTIONS AND KINSHIP.

A. The Anthropology of Emotions. 1. The dual nature of emotions. In spite of the recent, numerous works that have focused on emotions in anthropology, it can still appear as rather a liminal field within the discipline. For many years, emotions have been left to the investigation of psychologists or philosophers, and have been taken for grated by anthropologists. At least until the 1970s, emotions were sometimes described or invoked in the course of study, but anthropologists seemed to be able to know what was felt without selfreflection, by an immediate recognition of signs (tears, laughter, etc.), assuming that the link between a bodily expression (public sign) and an emotion (inner feeling) is universal: a shared humanity enabled them to do so (cf. Leavitt 1996: 514, Rosaldo 1984: 137). The situation has however changed, through a double impulse. On the one hand, a theoretical debate has begun about what it means to study what we assume corresponds to expressions of the inner self. On the other hand, more careful attention has been paid to native conceptions of emotions. It soon appeared that “emotion” as a discrete field of phenomena was not universal. Among the Amazonian Candoshi, for instance, as well as in many other groups, emotion, cognition and sensation are not distinguished (Surrallès 2003: 16, 17n.18). I do not wish to tackle the question of whether it is possible to study emotions in a culture where they do not seem to exist. It is however necessary to establish the nature of our subject – without trying to define it precisely, without seeing it as a closed domain. If we follow for instance Wittgenstein’s remarks, it seems that emotions are defined by their double nature, at the same time bodily and mental, which contrasts them with sensations and ideas: “emotions can colour thoughts; bodily pain cannot. [… Belief] is not an emotion. There is no bodily expression 5

typical of belief” (1980: § 153-154). This is a position also taken by Leavitt, who underlines that a mind-body dichotomy is probably absent in many parts of the world, or irrelevant for this subject. But he follows: “what is so ‘recognised’ [as emotion] consists of experiences that involve both meaning and feeling, both mind and body, and that therefore crosscut divisions that continue to mark theoretical thought” (1996: 516). It is of course problematic to try to prove a similar phenomenon in other cultures. Yet the question of translating emotion words “is just part of the general translation problem in anthropology and no more difficult (but not easier either) than translating religious or domestic language whose references are attitudes and circumstances rather than concrete objects and actions. […] One learns to identify the emotional life of a people along with learning everything else” (Solomon 1984: 251): anthropologists tend to recognise something in other cultures, similar to our emotions, that correspond both to feeling and to meaning. It is now possible examine more closely the transformation of this dual nature into a theoretical dualism. My aim is not to provide a subtle discussion of all the possible positions in the debate, but merely to indicate some common theoretical choices, in order to place the ethnographic data analysed later within a wider context. 2. Felt emotions: psycho-biological universality. A path chosen by many anthropologists states that emotions are primarily defined by the bodily feeling that characterizes them. Linked to the founding assumption of anthropology that cultural diversity only hides a common biological nature, it leads to the idea that emotions can be studied in so far as they are universal, panhuman. It is probably most accurately exemplified by the work of Ekman, whose goal is to highlight a set of “basic emotions”, which recalls the basic colours studied by Berlin and Kay (1969), even though Ekman is not interested in “emotion terms”, but in pre-linguistic recognition of universal emotions. What is similar, however, is that Berlin and Kay have shown that the presence of only one word for blue and 6

green, for instance, does not mean that people living in that particular society do not distinguish between blue and green: Ekman wants to show that whatever the words are, basic emotions are the same. In order to justify their hypothesis, Berlin and Kay had used a strong technical methodology to show that the foci of the colour terms were similar in all cultures (contrary to the borders), when people had to link a coloured pill to a colour word. In order to achieve his goal, Ekman also used a point of reference outside language: facial expressions recognised in every society (e.g. 1984). This focus on pre-linguistic data is essential to his position, for as we will see later, language is an unavoidable topic here: do emotion terms, that may vary across cultures, represent emotions felt? According to those who criticise them, anthropologists who defend the psycho-biological universality of emotions have “a naïve understanding of language as merely the encryption of a prior ‘thing.’ Biologically based studies of emotions have assumed that words are purely representational in function and serve as interchangeable labels for a basic experience that is the same for an Inuit, a Spaniard, or an Ilongot” (Tarlow 2000: 716). This does however partly explain the strategic utility of the use of bodily expressions as central data: it reveals what language would hide. The complete absence of any culturally variable element is however undoubtedly an unacceptable option for social anthropologists, and more complex approaches have derived from the assumption that emotions are bodily and universal. A first strategy (that of most classical anthropology) would be to distinguish emotions from the way they are socially and culturally dealt with, which can vary. Solomon, for instance, stresses the variability of the causes of emotions, of their names, of their expressions, and of the metaphors used to describe them (1984: 240-241). Ekman notices similarly that if emotions are universal, their expression, the cultural stimuli that provoke emotions, and the control of natural emotions are cultural (1984): we will see that relationships with kinspeople could be interpreted in Amazonia in terms of the social control of anger (cf. infra, chapter 3). A second strategy consists in the development of a two-layer model of the person, a common path in psychological 7

anthropology. It involves, for instance, the analysis of complex emotions that are produced by the combination of universal basic emotions (Shaver, Wu & Schwarz 1992), cultural variability being the outcome of contingent associations of identical elements. These approaches are more satisfactory, yet they still rest on assumptions about the nature of emotions that create many difficulties. The first problem is that such positions may easily deal with emotions treated as examples in the course of an argument, but often prove to be overly simplified when faced with the complexity of the data gathered by anthropologists. Tarlow, for instance, states that Leff wanted to study a universal state of “sorrow” by showing subjects photographs of mourners at a funeral and asking them to chose among context-free faces. But she underlines that his two assumptions (that sorrow is universal, and that funeral will always produce sorrow) are contradicted by many ethnographies (2000: 715). Assuming the universality of emotions always leads to a reduction of diversity. The second difficulty is methodological. We can have access to language and to facial expressions, but does this reveal anything about what people feel? How do we know that other people’s feelings are authentic? This is of course a major topic of debate in western culture. A famous instance is Wittgenstein’s discussion (1980: § 588sq.). In anthropology these debates have taken a more specific direction, for we cannot put ourselves in other people’s skin, as Geertz reminded us: it is a methodological rule that we should not even try to study inner thoughts, inner feelings, or anything that goes on in one’s soul – if we want to avoid the hermeneutic illusion. This seems to follow on from one of Wittgenstein’s remarks: “Only God sees the most secret thoughts. But why should these be at all important? And need all human beings count them as important?”(1980: § 648). The solution must then be to avoid this question by focusing instead on public (social) life: emotions can be hidden, repressed or faked, but it does not matter for the anthropologist, as long as we can study what is said about emotion – it is a public message. 8

3. Emotions as cultural meaning. Anthropologists who have taken this path focus instead on language as an access to emotions, and their conception of language is more complex than that evoked earlier. Language is not a mere representation of a prior existing thing, but has generative powers, up to the point that “emotion can be said to be created in, rather than shaped by, speech” (AbuLughod & Lutz 1990: 12). There is additionally a great diversity among this trend, but it is legitimate to assert that all of them follow some of Geertz’s rules: their aim is to study sociocultural, public, messages, and not private (by definition inaccessible) feelings: emotion is “about social life rather than internal states” (id.: 1-2). This can lead to an interpretative anthropology, which is probably best exemplified by M. Rosaldo’s work on the Ilongot (1980). In spite of her introductory claims, she mainly restricts herself to a semantic analysis (cf. infra), showing how notions of passion (liget) and knowledge (bêya) play a central role in Ilongot accounts of selfhood, which is a way of describing their practices that focuses on (or restricts itself to) language. Many an analysis of Amazonian kinship relations follows this semantic path, for instance with subtle interpretations of the connotations of lexemes expressing love and compassion (cf. infra, chapter 2). Another notable approach is cognitive. Solomon, for instance, dismisses the question of whether emotions are felt or not by underlining that “the physiological reaction itself is virtually never sufficient to distinguish one emotion from another”. Therefore, for example, “anger is not just a physiological reaction cum sensation plus an interpretation, a cause and certain forms of behaviour. It is essentially an interpretation […] a judgmental emotion” (1984: 250). It clearly shows his belief in the primacy of the mind. Such studies were however criticised, and Geertz’s legacy transformed, not because of their cultural reductionism, but because two other aspects have been put forward: language is not reducible to a set of nouns and some have paid more attention to grammar, intonation, etc. (e.g.

9

Irvine 1990); and it is not language in itself that matters, but “social discourse”, which means it is inseparable from questions of power relations or everyday sociability (e.g. Abu-Lughod & Lutz 1990). It must however be noticed that some of these recent works that claim to be the outcome of the (post-) modern theories of emotions are in fact veiled applications to the field of emotions of much older ideas about language. For instance, White studies how anger – which can be a threat to social order – is transformed by a specific kind of discourse into a form of sadness that can lead to reconciliation: emotion rhetoric is at the same time a justification of people’s behaviour, and the means to reach social peace (1990). It is very close to Wright Mills’ study of the imputation of motives: he clearly stated that “we cannot infer physiological processes from lingual phenomena” (1940: 909), for there is nothing deeper. Motives – and, we could add, emotions in so far as they are a motive for action – are not the subjective cause of action, but a public message directed toward others in order to achieve an agreement (id.: 907-908). It appears, to take an Amazonian example, that “compassion” can be interpreted frequently as a public justification of actions such as adoption of children and pets (cf. infra, chapter 2). This path of study would constitute a major contribution if it could deal satisfactorily with emotions. However, in spite of the insightful analyses that have been conducted by many of these authors, their position cannot be accepted if one does not want to amputate the object of study: as biological reductionism led to simplistic accounts of ethnographic diversity, cultural reductionism seems to make emotions thinner than they are (or leave the reader the freedom of using empathy to go further). As Leavitt pointed out, the use of the disjunctive “rather than” in works pertaining to this trend imply that there must be “a choice between mutually exclusive terms: if emotion is discursive, performative, and social, then it cannot be bodily, expressive, or personal” (1996: 523). But if emotions can colour thoughts, they cannot be reduced to coloured thoughts.

10

4. Is synthesis a solution? The anthropology of emotions has so far been presented in a manner congruent with extreme dualism, recalling nature vs. culture, body vs. mind, etc. This was of course for the sake of presentation, a device also acknowledged by most reviewers of the subject. However, most recent works and most reviews underline that such research was located on a continuum joining both poles, and that synthesis would lie in a balanced acknowledgement of the dual nature of emotions, at the same time bodily and symbolic, universal and particular, etc. (Tarlow 2000: 717-718). Yet this could prove an unsatisfactory solution, for taking a middle position still leaves the opposition between the two terms in place (idem). You can only assume that emotions are both about a bodily feeling and a cultural meaning, but it will not solve the theoretical debate. It is perhaps useful here to highlight one of the difficulties of most research on emotions. Probably as a reaction against former studies that took emotions for granted, more recent works have tried to focus on the emotions themselves. More precisely, they have focused on a limited set of emotions, labelled by western or indigenous words. Psychological anthropologists are explicit examples of such an approach, because they struggle to “define a bounded thing representatively labelled as an ‘emotion’”, achieving a process of “abstraction” and “reification” (Tarlow 2000: 714). This is obvious when reified emotions are based on western concepts, e.g. love or anger. However, a similar process happens when the focus is on “liget” (Rosaldo 1980). As Rosenberg has shown, Knowledge and Passion is mostly a metacommentary on some “emotional terms”, such as liget (anger or passion) or bêya (knowledge), which are mostly nouns, in spite of Rosaldo’s claims to do much more than a simple semantic analysis: these nouns “seem to come to life, first as essences and then as seeming psychological forces, in a species of bilingual ethnographic writing in which the ‘focal terms,’ having been glossed early in the text, become so familiar to the reader that they can be used

11

throughout the remainder of the account […] in sentences in English grammar” (Rosenberg 1990: 164-165). These nouns are abstracted from language and substantialized, in a way that echoes western habits: “The Ilongot rendered in terms of ‘liget’ is a story that is resonant with an old genre in which all the peculiarities that seemed to characterize the French could be woven together under the slogan ‘vanity,’ and just so the Spaniards under ‘pride,’ the Zuni under ‘Apollonian,’ and the Kwakiult under ‘Dionysian’” (id.: 166). The critique of this glossary approach by Rosenberg shows that giving a major role to a set of emotions (or emotion terms), even if they are of course interrelated, can be misleading because we tend to see emotions as things that have a direct effect. It is therefore necessary to achieve a displacement of perspective in order to avoid difficulties stemming from the dualism of theoretical tendencies and from the glossary approach. It is acknowledged that emotions have a relational nature, but we could take this statement the other way round: start with some relations, and study emotions as a quality, as an aspect of these relations – and here restrict ourselves to the emotional character of kinship relations. Focusing on kinship relations that have at the same time a precise cultural meaning, and often constitute the context of social action (i.e. action directed toward someone else) will perhaps enable us to study “relational dispositions” (Taylor 2000: 319) that we could see from the perspective of Bourdieu’s habitus were we to incorporate an emotional element. As the habitus, the dispositions to be studied are “constituted in practice” (Bourdieu 1990: 52) whilst also structuring “perception, thought and action” (id.: 54). They are also essentially relational. But Bourdieu mainly insists, for instance when he studies kinship, on the strategies that aim to ensure the “biological and social reproduction of the group” (id.: 189): it is above all a question of power and symbolic patrimony. If Bourdieu places a lot of importance on the body (e.g. id.: 66-79), he pays little attention to affectivity (only one footnote on “emotion” according to the index of The Logic of Practice). Yet we will try to show precisely that it can be interesting to

12

use habitus with an emotional quality as a tool to analyse social relations, especially in the context of kinship. Certain defined emotions will be pointed to as characteristic of relations in my analysis, while others will be referred to only as emotional qualities; yet these emotions will not exist as “things”, for they will be embedded in relationships. The focus will therefore not be on the theoretical aspect of the anthropology of emotions, with all the subtlety that such a debate would require, thanks to a theoretical displacement that enables us to assume the dual nature of emotion (being confident in anthropologists’ capacity to learn the native emotional life), and to make a shift from body and language towards action. In order to do so, it is first necessary to study how kinship relations are usually conceptualised in anthropology, in order to understand how the emotional or affective quality of these relations can properly be analysed.

13

B. The new studies of kinship. 1. The Schneiderian Critique. Kinship is one of the subjects that results in palpable international academic divisions. In the Anglo-American world, kinship studies in the 1990s turned to a range of problems completely different from their predecessors, whereas in France for instance a mainly Lévi-Straussian approach has continued to be pursued 1. Such a difference probably comes from distinct and different receptions of Schneider’s critique of kinship: on the one side, it was so destabilising that kinship could only reappear in a completely distinct way; on the other, it was only one of many attempts that could not do much to destabilise dominant structuralist studies of kinship – from her (very confident) point of view, Héritier (1999) could say that “the fortress has not been taken”, in an article about The Elementary Structures of Kinship fifty years later. It was perhaps also because Schneider was very much concerned with consanguineal ties, rather than with alliance, which is the key to a Lévi-Straussian perspective. It is necessary to recall some aspects of Schneider’s critique of kinship, in order to understand more accurately the “new” studies of kinship. His main objective, as we will see, is to show that, in spite of all their strategies, anthropologists have not been able to get rid of “biological” ties (i.e. connectedness issued from sexual procreation) to define kinship. In the beginning, at the end of the 19th century, were Morgan and Maine: kinship was a biological relationship, culturally recognised. Hence the Morganian distinction between descriptive and classificatory terminologies – the later, that of the natives, confounding what should stay distinct – which must be understood in relation with the development of marriage. Only a progressive evolution towards monogamous marriage could ensure the knowledge of biological connections, and therefore enable the appearance of a descriptive terminology 1. A whole issue of the journal L’Homme (2000), devoted to kinship at the turn of the century, is still full of terminologies and diagrams, contrary to Carsten’s statement about kinship studies since Schneider (2000: 3).

14

(1870). Maine also stressed clearly that biology was the pivotal reference point: he considered adoption as a “legal fiction”, giving way to the expression of fictive kinship whenever there is no biological connection (1861). Durkheim strongly reacted against this point of view, defending the supremacy of society, and the purely social nature of kinship. In direct opposition to Morgan, for instance, Durkheim provided examples in which the natural child is not kin to his father, whereas the adoptive child is kin to his adopter (cf. Schneider 1984: 100). But, as Schneider points out, kinship can never be “purely social”: why would it be “kinship” (more than anything else) without the biological referent implicit even in Durkheim’s arguments? A third and last position (to over-simplify anthropological history somewhat) would be that of Barnes (1961, 1964) and Scheffler and Lounsbury (1971) who stressed a distinction between western biology, folk biology and social links: what really matters is not biological relations as we understand them, but as they conceptualise them, in relation to the social way of dealing with them. The primacy of biological connections has, in their analysis, been replaced by the primacy of ethnobiological connections. Schneider therefore underlines that (ethno-) biological bonds have been assumed to have the same value everywhere, a belief that is contradicted by, for instance, his own data on Yapese culture (1984: 122). This assumption has made possible the cross-cultural comparative framework that has constituted the principal strength of kinship studies since the 19th century: “the assumption that Blood Is Thicker than Water says that whatever variable elements may be grafted onto kinship relations, all kinship relations are essentially the same and share universal features. Hence the genealogical grid can be used as an etic grid: comparable things are being compared and analyzed by first establishing their identity and universality and holding constant the components that make them comparable” (id.: 174). Comparison, and therefore kinship studies as they were, become impossible if we abandon this assumption.

15

2. Some new studies of kinship. Kinship studies have not disappeared, but have been transformed by the Schneiderian critique. The aim of the new flow of works in the realm of kinship is not anymore to compare systems by relying on biological connections, but instead to question the value and meaning of these connections. What is compared is now precisely the notion of relatedness: how are people connected to each other? What is the place of “biology” in social relations? Certain key themes are of central importance in this new trend. Firstly, kinship is processual: connections are not given by birth, by sexual intercourse, or by sharing of a substance – by blood or genes (cf. Schneider’s distinction between states of being and states of doing, 1984: 165-166). On the contrary, more weight is now given to processes whereby you can progressively turn others into kin – and sometimes this process must apply even to your own babies (Rival 1998, Vilaça 2002). In addition, some people that used to be relatives can progressively lose this quality (Bodenhorn 2000, Carsten 1995: 234). In most cases, this is a slow process because – and this is the second point – kinship relations are built through everyday practices, e.g. caring, feeding, and living together. This appears to be a widespread idea, valid in the indigenous Americas as well as among the Malays (Carsten 1995: 225), and what contradicts western assumptions is the reality of kinship relations produced in such a way. “Real” kin, in the above instances, are those who share a same body, because everyday practices of commensality have made their substance identical (e.g. Gow 1991, Weismantel 1995). The usual process seems inverted: one cannot explain caring by an emotional bond (motherly love, etc.) that would be the natural consequence of a biological tie (according to the adepts of the psycho-biological universality of emotions, cf. supra), precisely because this (ethno-) biological assumption is contested. As I will seek to document, these new conceptualisations of kinship are likely to impact strongly on the study of emotions (cf. infra, esp. chapter 4).

16

These new ideas enable the anthropologist to dispose of the opposition between the social and the biological, especially when it is equated with an opposition between procreation and commensality (Carsten 1995: 236). Both processes are at the same time “biological” – because they transform the body in its substance – and social – because they involve social relations –, but they cannot be opposed to one another. Such a situation is perfectly exemplified in Strathern’s Melanesian analyses (in spite of being prior to this new trend, and of a different theoretical approach). She states abruptly that mothers “do not make babies” in the sense that newborns are not the natural product (to be socialised) of a natural procreative capacity of women, but are in fact the outcome of a system of social relations that they embody, being therefore social themselves. At the same time, consumption is obviously social – because the social relations that produced the food are finally revealed – but it also transforms one’s body accordingly (e.g. 1988: 311-318). 3. Consanguinity and the rest. Most of these “new studies of kinship” however focus on consanguinity and “domestic” kinship. Probably because Schneider himself stressed the question of biological connections, that is to say of links of filiation and siblingship, new studies question the meaning of these relations, and pay very little attention to affinity, except when it is reduced to marriage, that is to say relations of sexual procreation. Or, when affinal relations are noticed, it is usually in a depreciative way. Lambert, for instance, in a discussion of Indian kinship practices, states that consanguineal and optative kinship “are sharply contrasted with affinal relations, which are assumed to lack the natural affective qualities commensurate with consanguineal ties” (2000: 75). What interest these authors seems therefore to be relations that are grounded in affection (in the western common meaning of the word) or in substance (biological relations, or shared substance produced by commensality): in a word, what was considered to be consanguinity before Schneider. This focus – and the implied lack of interest in alliance – could come from 17

the particular meaning of “kinship” in English, that does not cover the whole range of relations included in the French parenté, whose meaning includes affinal relations, as Dumont has shown. Consequently, “filiation becomes the essence of kinship, and other relations mere attributes of this essence. It is possible to understand the success of this point of view, which tends to give first place to the ‘biological’ aspect of things. Unfortunately this is not a comparative view, and when it claims to be such, it introduces socio-centrism” (Dumont 1971: 14, 17). By focusing on consanguinity, and on the everyday practices that produce it, anthropologists working on kinship in lowland South America from this perspective have therefore operated a truncation of their object. This statement does not only derive from the influence of French alliance theorists in the area, but also from the meaning that affinal relationships – and relations with the exterior in general – have in Amazonia. According to Viveiros de Castro (1996: 190), one of the main trends in Amazonian anthropology relies on the idea that symbolic exchanges with the exterior – through hunting, warfare, or shamanism – are essential for the constitution of the inside, that is to say the production of persons and the reproduction of society (he gives his own work and that of Descola as examples of this “symbolic economy of alterity”). More precisely in the context of kinship, he has stressed that, if consanguineal relations are to be constructed, affinity is the unmarked term, the “given” of Amazonian kinship (2001: 19sq.). And consanguinity is not produced out of nothing, as seems to be the case in some studies (a link between previously independent entities), but it is precisely produced out of affinity, through “the cosmological movement of transformation of affinity (alterity) into consanguinity (identity) and back again [with death…]. The construction of kinship is the deconstruction of potential affinity; but the reconstruction of kinship at the end of each life cycle [until the next] through procreation must rely on the affinal givenness of human sociality” (Viveiros 2001: 33-34, cf. also Vilaça 2002). Schneider’s work follows the choices of the authors he criticises (Morgan, Maine, Malinowski, etc.), and this explains his insistence 18

on assumed biological ties. Those who have studied kinship in the context of the debate created by his critique have therefore followed this path, but what may be justified in some contexts is certainly not so in others: one cannot avoid the privilege given to affinity in lowland South America. We can therefore suggest the hypothesis that the generality of the idea that kinship is processual, built through everyday practices of feeding and caring, and using an idiom of shared substance, can easily turn into banality – precisely because of this reduction of the focus to consanguinity, with an abstraction of affinity. It can remind us of Lévi-Strauss’ idea that by taking only similarities into account, anthropology “thus loses the means of distinguishing between the general truths to which it aspires and the trivialities with which it must be satisfied” (1968: 14). From this point of view, the ideal work, in order to reinstate diversity, would perhaps consist in highlighting kinship systems in their globality: comparison (within lowland South America) should be done between systems of relations, even when it is the meaning of these relations that is questioned. This task is however out of reach here, partly because it is probably the most difficult work to do, partly because most works on Amazonia that I discuss with are themselves generalising, and partly because of the lack of data on the affective content of kinship relations. There are of course many references to emotions, but they are usually not informed by the debates on the anthropology of emotions (except Surrallès 2003, Taylor 2000). Emotion terms are frequent in the “moral economy of intimacy” school (i.e. Overing and her students according to Viveiros de Castro 1996), but they achieve precisely the focus on the domestic domain that we will try to avoid. On the other hand, emotional elements are virtually absent from most of the works representative of the “symbolic economy of alterity” school. This explains partly why the study of emotions and kinship relations in Amazonia is not easy, and why only some general or common schemata are going to be stressed, in order to attempt to provide some guidelines for a future more concrete and complex study of the subject. We will not focus mainly on consanguinity, and see affinity from 19

a negative point of view, but instead, following Viveiros de Castro, begin with meta-affinity as the general frame for relationships. From this vantage point, it will be possible to follow the tension that tries to produce consanguinity by extruding alterity (for instance through the process of familiarisation), until we finally reach the limit of consanguinity. This will ensure that “emotion” is restricted neither to biological ties, nor to the small world of consanguinity.

20

CHAPTER I. A WORLD OF OTHERS.

1. A world of affines. After having tried unsuccessfully for several decades to apply schema elaborated in Africa or Australia to Amazonian kinship, anthropologists have begun to focus on its specificities. One of these has recently been made explicit by Viveiros de Castro: in lowland South America, affinity “stands as the given dimension of the cosmic relational matrix” (2001: 19), it is the “unmarked category” of relatedness (id.: 26). This observation was made as a reaction to a recent focus on the production of domestic kinship ties, and justifies the attention paid to the most general affinal relations. For among Amazonian people, the idiom of affinity is not restricted to those with whom one exchanges women. On the contrary, it stands as a general idiom used with unrelated individuals, an idea already recognized by Lévi-Strauss (1943: 398). In order to make this specificity clearer, anthropologists have tried to refine kinship models and idioms applied to the area. Viveiros de Castro for instance reflects upon the Dravidian kinship type, especially described by Dumont with reference to South India, and which is very common in South America: the equistatutory relationship between the categories of affinity and consanguinity, which is a defining feature of the Indian system, is replaced among Amazonian people by a hierarchical opposition: affinity encompasses consanguinity (2001: 2021). This concentric form leads to the distinction between real, effective affines (that are often treated as consanguines, cf. infra), and meta-affines with whom reciprocity is not the rule (Taylor [2000: 312] uses the word “meta-affinity” instead of Viveiros’ misleading “potential affinity”). Amazonian concrete kinship relations therefore seem to be located in a world of metaaffinity, which is that of the unknown, of people with whom no relation exists so far – except, of course, warfare. Instead of restricting oneself to the affective content of consanguineal 21

relations, it is possible to begin with what is said to be the context of domestic kinship. We will however need to distinguish distinct types of relationship according to the (real or symbolic) gender of the other – adopting a male point of view (due to a lack of studies considering the other aspect, cf. Taylor 2000: 309n.1) – because it must be noticed that affinity is essentially a same-sex relationship between males, as Rivière partly noticed (contrasting “marriageability”, a cross-sex relationship, with affinity and “affinability” [1984: 56]). Even without adopting Rivière’s terminology (that does not give sufficient attention to meta-affinity), we can follow his distinction between same-sex and cross-sex relationships, that is to say between male and female others. 2. Male others: a world of suspicion. Meta-affines are primarily those with whom war is made: enemies (cf. Tupinamba who used the same word for brother-in-law and enemy, Viveiros 1992: 294). However, it is necessary to notice that Amazonian applications of the kinship idiom extend far beyond the limits of humanity as a species: non-humans can be persons with whom kin relationships are sustained, and this applies especially to animals (Viveiros 2001: 23, Taylor 2000: 309-310). Animals in general, when no particular relationship (hunting or familiarisation, for instance) has been introduced, are conceptualised as males, and the predator/ prey scheme appears in its primary meaning. For the relationship of predation, stemming from the world of hunting, has been highlighted as a key aspect of Amazonian cultures by Viveiros de Castro and others. They have underlined the reversibility of this relationship that relies on the equivalence its terms. This form of relationship is only present when a same-sex (male) relationship is involved, that with enemies or with animals in general (Taylor 2000: 317). This feature probably explains why meta-affinal relationships are always full of suspicion and fear: ego can always happen to be prey rather than predator. Arhem presents the situation of the Colombian Makuna very clearly: “the shaman relates to the Spirit Owners of the animals 22

as to a male affine; it is a relationship of equality and reciprocity but also one of potential danger and violence” (1996: 192). Potential danger is of course characteristic of relationships between unrelated groups. Descola describes the fear of the Jivaro accompanying him to a settlement that was at war with another one: killing without warning was always a possibility (1996: 279), and he underlines how travellers in Jivaro territory always advertise their arrival by noise, in order to show that they are not a silent war-party, and are usually received by men holding their weapons (id.: 282-283). “Fear, suspicion and bravado formed the dominant tone of such interactions” [between Jivaro and White people], and male stories stress the narrator’s fierceness, but also “the fear and confusion felt by the narrator as well as others during the battle” (Taylor n.d.). Basso, when she cites narratives of small groups fleeing from “fierce Indians” and regrouping themselves in the same region, in order to describe the formation of the Xingu area in Brazil, always stresses the possibility of being killed while encountering other people whose nature is unknown: the young man who had fled with his mother and then tries to make contact with the settlement he has just discovered therefore says “I don’t care what happens to me. If someone kills me, so be it” (1995: 216); and another one “stands about clutching his arrows to him, as if he were about to attack the very people he is trying to convince of his peaceful intentions” (id.: 132). It is true that scattered groups unrelated to each other are partly a consequence of the European conquest, and that large scale ensembles used to exist, and sometimes still exist, providing a sphere of peaceful interaction (e.g. RenardCasevitz 1993, Taylor n.d.). But even in this case, encounters with unrelated groups are moments of uncertainty (e.g. Hugh-Jones: pers. com., on circumstances in the Northwest Amazon). And potential danger is of course characteristic of interactions with the non-human world of the forest: the jaguar (or sometimes the anaconda) often symbolically concentrates all the animal dangers (Clastres 1974: 126), and the finality of ritual treatments of the game is often to prevent accidents from happening – the jaguar, for example, can be the figure that

23

revenges some species of prey (for instance among the Aché, Edeb: pers. com.), and Spirit Owners of the animals are keen on punishing misbehaving hunters (e.g. Arhem 1996: 196). What are the consequences of such mixed feelings concerning this world of meta-affines? The first solution consists in the precocious creation of a relationship, and the concomitant diminution of fear: it is notably the case when distinct fleeing groups meet, for instance in the Kalapalo narratives. When an encounter does not lead to immediate killing, a new settlement can be created with the remnants of former groups, and people try to live together. Narratives show “the delicate manoeuvring that overcomes fear of outsiders, and (most important) the way music is performed in order to generate that spirit of happiness conductive to the generation (and perpetuation) of community” (Basso 1995: 199). Progressively, the Kalapalo “come to experience a lessening of that anxious tension which accompanies constant negotiation, dispute, and argument that are the consequence of new encounters” (id.: 195). The fragility of this process, that relies on a progressive growing confidence, shows the uncertainty that defines all relations with strangers: peaceful interactions are not the rule, they are to be achieved through a dangerous process. At the same time, such narratives cannot be reduced to an interpretation in terms of cultural stimuli of natural emotion (cf. supra) – the stimulus being here the encounter of strangers, and the emotional reaction fear –, for they also reveal that happiness and confidence can be the product of an intentional work, for instance through dancing and music (an idea also developed by Rivière [2000: 254-255] with the Trio notion of sasame). This idea that emotions are not only things to be dealt with, but can be socially produced, will be encountered again below. The second solution is that of a destructive relationship, of an actualisation of the position of predator: there another emotion appears, anger. A classical position would be to say that the emotion of anger, suppressed within the group in most Amazonian societies, is let loose outside the group: anger is directed towards the outside, because it is forbidden inside (e.g. Kidd 2000: 122). And often, it is provoked by causes that we can understand, for instance the 24

loss of beloved parents (e.g. Alès 2000: 136). However, such a description of anger as a natural feeling that can be suppressed by social rules – which would be the interpretation of some anthropologists of emotions (cf. supra, Introduction) – is misleading, for it does not do justice to facts. In several Amazonian societies, anger needs to be produced before war and killing: emotion is a precondition to action. For instance, for the Jivaro men, anger is not something that merely exists, but something that they must “cultivate” (Descola 1996: 193). More precisely, we find among the Jivaro a ritual called anemat by the Achuar and koraka by the Candoshi, the second phase of which consists in the division of the war party into two groups, and of the “simulated aggression” of one by the other: by “arousing the men both physically and mentally”, it prepares them for battle (Descola 1996: 390, cf. also Surrallès 2003: 192). Similarly, among the Brazilian Parakanã, the “will to kill an enemy” is a pre-requisite to war, that can be intentionally exacerbated (Fausto 2001: 271). We apparently face a phenomenon very different from that analysed in existing anthropological studies of emotions. Emotions are not merely something “felt”, for this expression implicitly states the passivity of the person who is affected by an inner force. But it cannot be either a mere cognitive or linguistic phenomenon. It is impossible to state that anger is “essentially an interpretation, a view of its cause (more accurately, its ‘object’) and logically consequent forms of behaviour” (Solomon 1984: 249). Nor can it be reduced to a public message, a “socially contested evaluation of the world phrased in an emotional behaviour” (Abu-Lughod & Lutz 1990: 11). For what matters really is not what people say, but what they do. And they do it because they have action as a target: it is perhaps this lack of interest in (social) action that has led many anthropologists to focus either on bodily feeling or on cultural meaning – they have not paid attention to the fact that emotion can be about something else. In some parts of lowland South America, it is therefore possible to view anger as a disposition towards action, a precondition of war, that can and must be socially produced.

25

3. Female others: desire. Meta-affines are conceptualised as prey to be killed or as dangerous predators only in so far as they are males. For when the relationship is cross-sex – that is to say when the other is female, if we keep our male point of view –, it is irreversible, an “unambiguous relationship of prey to predator” (Taylor 2000: 317). We can probably say that such a change transforms greatly the affective nature of the relationship: it is shown by the partial equivalence that we can establish between relationships with prey and with women. It has already been said that animals as a whole are conceptualised as affines – potential affines in the case of game, and meta-affines in the case of dangerous predators. However, when game animals enter concrete relationships of hunting, with specific men, this elicits their female gender. In many an Amerindian or Siberian culture, hunters have to seduce their prey, as they would seduce a woman. There is of course plenty of evidence of this phenomenon: for instance, hunting magic attracts women (Gow 1991: 125), men must avoid having sexual relations before hunting, often because of possible jealousy, and “the hunter is explicitly said to attract and seduce his prey” (Arhem 1996: 192). However, as we have seen previously, game animals are not unambiguously female, the hunters can on occasion treat game as male affine. We can thus assert, following Taylor (2000: 316), herself inspired by Strathern (1988), that gender is a result of the relationship – game is female because it is seduced. Several key aspects of the topic are revealed when we look at the other term of the equivalence – women, in so far as they are not possessed yet, which makes them stand as others. Jivaroan magical songs (anent) are here useful, because women usually use animal figures to induce specific affective states in other men’s hearts. When they try to inspire compassion or tenderness, they will speak as a helpless pet talking to his master, when they try to express their wrath, as a dangerous predator, when they try to create desire, they will claim to be out-of-reach prey, tracked by a hunter (Taylor 2000: 314-315). Although it may seem

26

disappointingly banal, and Proust said it with much more elegance, these magical songs probably show that desire, among the Jivaros, targets only something inaccessible, or at least not accessed yet (cf. also Descola 1996: 195). There is no desire for a pet, but only compassion, as we will see later, for it is already under its master’s protection and dependence. The animal that the hunter desires and tries to seduce is the prey that has not been caught yet. Both women and prey, that is to say, others under female form, are desired in apparently similar ways. However, there is a major difference that will be explored later: prey are killed by hunters, whereas women become wives. 4. Between humans and non-humans. We have seen that in most Amazonian anthropology, animals are treated as persons, because animist ontologies “use the elementary categories structuring social life to organise, in conceptual terms, the relations between human beings and natural species” (Descola 1996b: 88) – in lowland South America this often means that a kinship idiom is used for human/ nonhumans relationships. What happens in practice though? In most cases, only shamans can achieve in reality the use of this idiom, by having sexual relations with animals (McCallum 2001: 75, Taylor 2000: 325n.20), “by the establishment of kinship relations with an animal species” (Vilaça 2002: 358), etc. But ordinary men cannot cross the boundaries of humanity, and there is therefore an ultimate distinction between human relationships and human/ non-human relationships: “seduction, in human relations, leads to affection and procreation; in relations between men and prey, it leads to murder and death” (McCallum 2001: 55). By not killing his prey, the hunter would be tricked by his own work of seduction, he would engage in a relationship with a (female) prey, and could not come back to his state of humanity (Willerslev 2003). However, if humans and animals are conceptually equivalent, they are not indistinct: the boundaries of humanity and animality are frequently marked by the presence of inversions, of a 27

mirror that deflects the nature of relationships. Thus, seduction of women leads to sexual relations, and seduction of prey to deaths, but both events are equivalent even though the crossing of a boundary makes them differ – they are part of a process that goes from seduction to reproduction, and are even the conditions of reproduction. For if procreation is essential to human reproduction, it has been stressed that (at least in some parts of Amazonia) death at the hands of the hunter is essential to animal reproduction. “An animal must be killed for another to be reborn […]. The perpetuation of cosmic order […] requires ‘male’ predation as well as ‘female’ fertility” (Arhem 1996: 192). In some groups, this happens explicitly because each animal species has a limited number of “forms” or “silhouettes”: “by killing his prey and sending back its form to the “mother of the game”, the hunter fecundates animals, and this is one reason for the feminisation of game instituted by the relationship of hunting” (Taylor 2000: 325). Humans/ non-human relationships therefore seem to be marked by a specific inversion when reproduction is concerned. If it is true that hunters can treat game as male affines (in general) or as women to be seduced (in specific relationships), it remains the case that reproduction happens because of sexuality in human relations, and because of death in human/ animal relations. This argumentation shows that there can be an equivalence between humans and non-humans from a cosmological point of view (production of persons and reproduction of animals, predation of forms and of identities, etc.), but much less when we try to look at the affective content of relationships: seduction of prey and seduction of women are similar because they lead to actions that are cosmologically equivalent (sex and death), but not emotionally so. Hunters can seduce non-humans (animals), but only shamans can love them.

28

CHAPTER II. FAMILIARISATION AND COMPASSION.

1. A cosmological scheme. Beginning with a focus on meta-affinity, I stressed the fact that most recent Amazonian anthropology has paid much attention to shamanism, hunting or warfare as examples of the scheme of predation – Amazonian people seem governed by the necessity of acquiring without reciprocity what they do not have, and the production of persons as well as the reproduction of society is dependant on the exterior. However, some anthropologists have tried to underline that the relationship with the exterior is not always destructive, that consumption can be productive, and so on. Fausto gave to these ideas a central importance in understanding Amazonian cosmologies and practices, through the notion of familiarising predation (1999: 936-938, 2001). One of his most significant insights stems from the attention he pays to processes: his work is not so much focused on oppositions between consanguines and affines (predation being excluded between consanguines and achieved with meta-affines), for instance, than on the processes on consanguinisation (as a familiarising predation of affines). The power of his analysis comes mainly from the generality of the scheme of familiarisation, “schematized in the Amerindian context as the protective relationship between father and child or between master and wild pet” (Fausto 1999: 938). It concerns firstly the taming of the young of animals killed while hunting, a phenomenon that had already been noticed, although its general application had not been really exploited (e.g. Erikson 1987), and that can find its counterpart in the idea that non-humans can also capture humans and turn them into kin (Fausto 2002: 14-15). Secondly, we have “the shaman’s relationship with his auxiliary spirits” (Fausto 1999: 938): it is an “ambivalent” adoption, because of the powerful agency of the entities domesticated, but the idiom of fatherhood is very widespread in Amazonia (e.g. Chaumeil 1983: 120). Thirdly, Fausto notes that “captive and abducted children are often 29

conceptualized as pets” (1999: 939) – war seems comparable to hunting, and there is often a single expression for adoption and domestication (e.g. Surrallès 2003: 67). Finally, Fausto exposes the relationship that he privileges in his own work: “the binding control established by the killer over his victim’s subjectivity should be compared to the relation between shamans and familiar spirits. Both are conceived of as an adoption, as the transformation of a relationship of predation (real or virtual) into control and protection, modelled as the passage from affinity to consanguinity” (1999: 939). The relationship with the enemy’s spirit, which is very similar to that of many other groups (e.g. Wari’, Nivacle, Tupinamba) gives to Fausto’s analysis its specificity, as will be made explicit. His interpretation of familiarisation focuses on a cosmological idiom of control, stemming from the fact that the hierarchy of beings in lowland South America is often expressed in terms of more or less powerful (predating) subjectivities (e.g. Fausto 2001: 537-538, Surrallès 2003: 44). Familiarising another being then means to control its subjectivity: “the pet is that which loses potency or, shifting ahead a little, that which loses self-consciousness and lets itself be dominated by the other’s perspective” (Fausto 1999: 940). This interpretation explains for instance why rituals involving the killing of domesticated entities often contain a step of “estrangement”: a completely domesticated being would have no ritual use (Fausto 2001: 540), and that is why the Tupinamba pet-prisoner or the Parakanã spirit of the enemy had to be “reenemised” before being killed (Viveiros 1992: 291, Fausto 2001: 440). However, this idiom does not concern only the process in itself, but also justifies its aims: “the principal themes are ontological: acquisition of souls and virtualities of persons, names and existence, development of the full capacities of persons and maturation, control of deathly processes and longevity, conquest over death and immortality” (2001: 418). It is however possible to argue that this is a reduction of a wider process to the single case of familiarisation of an enemy’s spirit: it is only by paying so little attention to the familiarisation of other entities, such as children or pets – in spite of showing the similarity of the processes –, 30

that Fausto can give so much importance to cosmological themes. We can here achieve a slight displacement, and at the same time try to introduce an element of affect: for familiarisation is not only a control, but the creation in the domesticated animal or spirit of a disposition of submission. If we introduce such notions, both as a cause and as a result of the process, a quite different picture can be drawn, stressing differences between distinct types of familiarising processes. 2. Seduction and Adoption. Fausto, following the kinship idiom usually used, sees the relationship between master and pet as a purely filial relationship, with control on the one side, and submission on the other. It is however possible here to use Taylor’s insight (2000: 313) which recognises the ambivalence between adoption and seduction in the familiarising process. Fausto’s focus on the idiom of filiation apparently comes from the fact that he does not fully acknowledge that gender is a key aspect of the question – the entities he speaks of seem to be of undefined gender, to be already under the control of the master (as a young animal captured when its parents have been killed): the process is reduced to the creation of a disposition of submission to the master. Fausto’s description seems to be adequate in the case of animal pets, of spirits and of children of enemies. In these cases there is no seduction, and the gender of the entities matters little (although the spirit of the enemy is obviously male). We could probably say that there is no need to attract something that is already under control. The situation is however quite different in the case of women and of shamanic auxiliary spirits. Fausto had not really envisaged the example of women, probably because among the Parakaña, even captured women are not “familiarised” but “pacified”, only captured children being familiarised and adopted (2001: 300). However, among the Jivaro for instance, even women who are not captured in war need to be domesticated by their husbands. This appears both in the male discourse of domestication through sexual relations (e.g. Descola 1996: 185), 31

and in the female magical songs, where the husband is called “little father”, and where the wife is compared to a pet longing for its master (Taylor 2000: 316). A comparable situation exists among the Tukanoan Barasana (Hugh-Jones, pers. com.) where most marriages have a ritual element of capture, and where men have to domesticate (i.e. “feed”) their wife, before being fed by them (through food and sex). But then it must be noticed that such an idiom of filiation and of domestication only happens after women have been seduced, as we have seen earlier, and the idiom of seduction is that of hunting, not of familiarisation (at least among the Jivaro). Seduction also appears when shamanic spirits are involved, even though here the situation seems more complex. In some cases, they are powerful beings only pacified through gifts of tobacco (e.g. Chaumeil 1983: 206-208), but in other cases, they must be attracted. Chaumeil for instance notes their “seductive” aspect, when they appear to the apprentice under the aspect of beautiful women (1983: 117). The major difference with game or women is that shamanic spirits cannot unambiguously be conceptualised as female prey, otherwise they would be useless, lacking power. For instance, in the case of the Colombian Embera, shamanic spirits are bisexual, and the shaman who tries to familiarise them first needs to seduce them by playing on his gender: he will sing alternatively with male and female tones in order to attract the spirits, and will use smelling plants called “the spirits’ seductress” (Losonczy 1990: 80). Once they have entered the process, it seems that all are treated and conceptualised similarly: faithful, helpless, obedient beings that need the protection of a master. What differs, when we introduce the notion of seduction, is the process to gain control over the domesticated beings. In the case of wild pets, of children captured in war, or of the enemy’s spirit, we could assume that the mere act of killing their parents or the enemy himself seems to place them in a condition conducive to domestication. On the contrary, to use Fausto’s terms, it seems that adult women and shamanic spirits have too strong a potency or self-consciousness: they must be attracted, seduced, that is to say probably tricked into a state of dependency. But this is only the beginning of the process. 32

3. Affective states. We have introduced the notion of seduction into the analysis of familiarisation, but have kept an idiom of subjectivities in confrontation with each other: is this the only possible description? Indeed, the relationship is not only cosmological or ontological, not only about agency, subjectivity or substances, for the goal of the process of domestication is above all the creation of a specific disposition in each being, a relational affect that would conduct their behaviour towards each other. By domesticating or adopting a child, a pet, a spirit, a wife, etc., one tries to instate a relationship marked by “faithfulness, obedience and assistance” towards the master (Chaumeil 1983: 120). Texts say very little about this submission that seems to appear in all cases. What about the other side of the relationship though? What do anthropologists say about the relational disposition of the master/ owner/ father? From this point of view, we face once again a distinction between different kinds of familiarisation. When spirits are involved – whether shamanic spirits or the spirits of enemies – there only seems to be a will to control powerful beings. They must be put into a relationship of dependency and control, the shaman for instance causing the spirits to submit by controlling their feeding, when he feels their thirst (Losonczy 1990: 80-82). The relationship seems however to differ substantially when it concerns wives, children or pets: in such cases there appears the notion of compassion, or compassionate love. This is for instance the affective state that a Jivaroan woman tries to induce in her husband, thanks to magical songs, when she compares herself to a helpless pitiful pet that needs its master (Taylor 2000: 314-315). Among the Candoshi, orphans compare themselves to featherless young birds, asking the hunter to bring them home with him (Surrallès 2003: 67), trying to induce compassion by stressing their helplessness once they are deprived of the protection of their father (and compassion is also given as a justification for adoption among the Canela, Crocker 2002: 99). And the recurrence

33

of the metaphor of the young animal to be tamed shows that it is one of the reasons to save them (this compassion can also be interpreted as a kind of compensation for having killed the adult game, Erikson 1987). However, it seems that compassion has a much wider application than simply relationships of familiarisation: it characterises apparently in lowland South America most asymmetric relations, at least when one term is defenceless and dependent on the other. It is for instance the emotion that babies are supposed to elicit among the Peruvian Piro: they elicit compassion in others because they are helpless, and they are helpless because they are alone, not having kinspeople yet – being related to others being a progressive achievement (Gow 2000: 47). It is also out of compassion that the Enxet helped the Paraguayan deserters during the Chaco war, instead of shooting at them (Kidd 2000: 119), and it is compassion that the Miskitu and the Wari’ feels for mourners (Jamieson 2000: 84, Conklin 2001: 96-97). Another example is that of Peruvian Amuesha, where compassion is the characteristic of asymmetrical love in general, that of parents for children, of divinities for humans, of priests for their followers, etc. (SantosGranero 1991: 203). Or, once again among the Piro, shamans cure because they feel compassion for their patients (Gow 2000: 58). Compassion (or the various native terms so translated in English) therefore seems to appear in front of someone (or something) who is helpless, and who depends on a more powerful being for his well-being and even his life. An important motive for action among Amazonian people, it also keeps a central role in the process of domestication. Except in the case of spirits, compassion would appear to be the justification for familiarisation, especially in the sense that not all orphans or young animals are saved, but only those that are able to elicit this particular feeling. A distinction then appears between human adoption, and the domestication of pets and wives: this distinction concerns reversibility. For wives and pets (and, if we go further, humans compared to divinities, laymen compared to shamans, etc.) will always stay in a state of dependency: they will never be in a position to experience compassion towards their current 34

masters, because the asymmetry is founded on characteristics that cannot be changed (divinity, gender, species, etc.). Therefore, their helplessness will never disappear, they cannot be freed, and this is what enables Jivaro women to appeal to the compassion of their masters/ husbands. On the contrary, children will not stay in this situation, because asymmetry is only founded on age difference (even though adult orphans sometimes underline that their pitiful youth has marked all their life, e.g. Descola 1996: 295): the relationship is reversible. And this seems to be a specificity of human relationships: relations with non-humans can be ambivalent in general (e.g. game can be conceptualised as male or female), but not in particular instances – once relationships are established, they cannot evolve. Pets do not procreate, and this is what enables their relation of pure consanguinity with their masters to be maintained – and, among the Jivaro, “consanguineal ties remain non-relations as long as they are not associated with relationships of affinity” (Taylor 200: 322). Only humans really go through the cycle of kinship, for adoptive children will become adults, marry and themselves have children (or capture some in war). Not restricting ourselves to the idiom of adoption and general cosmological or ontological schema therefore enables us to make several distinctions. We have already seen that some cases were rendered more complex by the presence of a seductive component. Now it appears that the notion of compassion, and the possible reversibility of this affect, show that schema that seem equivalent from a cosmological point of view often differ with a different approach. 4. The nature of compassion. We have so far used the word “compassion” without reflection, but what is the nature of this disposition that is so strongly related to familiarisation? Here we find again the ambiguities of the anthropology of emotions. For on the one hand, we could present compassion as a linguistic phenomenon. It is a “motive” for action, publicly avowed or acknowledge in order to justify one’s behaviour, a public message: there are various anthropological words that 35

anthropologists have translated as compassion, which are assumed to explain such actions as adoption of an orphan or of a pet, attention given to suffering others (Kidd 2000: 119), an animal spared by a hunter (Rival 1996: 150), etc. To follow Wright Mills (1940), compassion would not appear as the psychological cause of action, but as a justification given to others – a public message. It is even possible to go further than this glossary approach, in order not to look only at nouns, but also at more complex linguistic phenomena: in order to express their compassion towards mourners, Miskitu people lament the dead “through the strategic use of particular affixes” that connote intimacy, instead of the more distant class of affixes that they would usually have used (Jamieson 2000: 88). This approach helpfully prevents us from falling under Rosenberg’s criticisms (1990), for studying these linguistics phenomena would enable us not to essentialise compassion (cf. supra, Introduction). However, we could also adopt the other path – emotion as a bodily feeling –, and refuse to consider emotion as a purely linguistic phenomenon. Relying on the anthropologist’s empathy would probably be too old-fashioned, but the native exegesis itself shows that compassion is something felt. For several studies stress that the words or concepts so translated contain the notion of pain: it is explicitly said that Amuesha divinities’ suffering “results from an intense feeling of love/ compassion” (Santos-Granero 1991: 203), there is often a linguistic association between compassion, love and pain (e.g. Jamieson 2000: 84), etc. And pain is something felt, a bodily experience, which gains here a cultural meaning through its association with compassion: excluding it from our study would amputate the notion of compassion, and limit our understanding of it. People act a certain way because they feel this specific kind of pain caused by compassion. Without taking this aspect into account, we can explain social discourse, but we cannot account for action: seeing emotion as a disposition for action makes it necessary to find a place for bodily feelings in the analysis. As we have already underlined, people act out of compassion as they act out of anger (cf. supra).

36

However, the dualist approach to emotions cannot escape its contradictions. For, it would be easy to state that people say they feel something, but we only know what they say. And even if the question of authenticity of emotions is pointless in some indigenous cultures (e.g. Gow 2000: 51), it is too pregnant in the Western world to be dismissed. The choice of one or the other path can only be an act of faith, and therefore we shall try to avoid both of them. A return to the Jivaroan magical songs can help us resolve this matter. Their purpose is to create affective states imputed to others according to cultural models of relationship, and to the position of ego in this relationship (Taylor 2000: 316). Women magically create compassion in their husbands by adopting the position of a pitiful pet, for compassion is the emotion that a master should have towards his pet. This case is of course highly specific, because it concerns private silent songs, and not public messages. It is however possible to view this as revealing. People make a strategic use of the emotions they impute to others: humans try to provoke divinities’ compassion (Santos-Granero 1991), hunters can try to provoke the compassion of game towards them, for “animals may ‘pity’ the hunters who have need of their flesh” (Brightman 1993: 187, this is especially true of North America), etc. And the same could probably be said of other relational dispositions that go in pairs. The issue must therefore not only be that of meaning or of bodily feeling, but that of action. People try to force others to act 2 by generating certain feelings (here pity or compassion) that correspond to the disposition complementary to that which they express (here helplessness) – and a reciprocal phenomenon happens when actions must be justified (one is more likely to say “this animal/ baby is so small, it will never survive alone”, than “I feel compassion”). Of course, we do not now what happens under people’s skin, we do not know whether it is a production of “real”, “authentic” emotions, or only an imputation (by merely expressing some feelings, one imputes the complementary feeling to the other). Yet we are not limited to what people say, because we

2. I do not grant agency to babies and animals – this expression concern essentially children and adults, who use the similarity between their position and that of babies and animals.

37

know that they act and how they act: what matters here is to gain an understanding of how people use such imputations to make others act.

38

CHAPTER III. SHAME AND RESTRAINT: IN-LAW RELATIONSHIPS.

1. Real affines cannot be avoided. As we have seen, following much of contemporary Amazonian anthropology, affinity is the unmarked term, an idiom that goes much further than a simple brother-in-law relationship. It is however mostly of meta-affinity that we have spoken so far, and not of real affinity, not of males connected through close alliance to each other. For in spite – or because – of the general idiom, real affinity seems to create a difficulty for Amazonian peoples. Some anthropologists, referring to myths or interpreting rituals, have alluded to the dream of a world without affines, where one does not have to deal with one’s brothers-in-law and father-in-law. As a general situation, it is of course a dream, but it can sometimes be achieved: when women are captured and not properly exchanged (an inverse practice – which expresses the same dream – is that of the ancient Tupinamba, who captured enemies, transformed them into brothers-in-law by giving them wives, and then killed and devoured them, cf. Clastres 1972). Thus captured women need more than others to be familiarised, for they often try to escape, but there is no need to deal with in-laws, and it is possible to avoid brideservice. Such practices, although not universal in lowland South America, are quite common, for instance among the Tupi (Fausto 2001: 194-195), the Jivaro (at least one kidnapped woman in each local group, according to Surrallès [2003: 160], cf. also Descola 1996: 176), or the Txicao (Menget 1982: 196). However, this is not always possible, and there is another solution to avoid having to live and deal with strangers: marrying a closely related kinsperson. Both practices are sometimes related, precisely because they have the same goal: among the Jivaroan Candoshi, two groups of siblings usually exchange their sisters, soon creating a very strong link that renders most tensions impossible. However, this close alliance is perhaps before anything else the means to get strong allies to capture women in more distant groups (Surrallès 2003: 156-160). About the 39

Tupian Parakanã, Fausto also notes that kidnapping “is a way – as marriage with the niece at the other extreme – of escaping brideservice, […] of asserting one’s own autonomy” (2001: 303). Of course, many societies have chosen to content themselves with the second solution, by stressing local endogamy and alliance with close kin (e.g. direct cross-cousins). This, for instance, is the case in the societies of Guiana: “within the ideal settlement affinity does not exist” (Rivière 1984: 70). Why such a situation? Before being the male consanguines of one’s spouse, affines are others, disagreements are always more likely to happen with them (Journet 1995), and the father-in-law/son-in-law relationship is often the central relationship of control in Amazonian societies, an uncomfortable relationship indeed (Rivière 1984: 77). But not all women can be seized by force, and marriage with close kin is not always possible (or allowed): real affines are not absent from Amazonian life. How are they dealt with? In most cases (with some exception, e.g. the Peruvian Piro), the concentric Dravidian systems are such that real affines are attracted toward consanguinity, i.e., up to a certain point, they are consanguinised. There is firstly a very widespread avoidance of affinal terms to refer to or address real affines: preference goes to “cross-kin alternatives or […] teknonyms expressing co-consanguinity” (Viveiros 2001: 24). But it usually goes further, with an ideal of behaviour: it is explicitly stated to anthropologists that people should treat their brothers-in-law as brothers, their sons-in-law as sons, etc. (Alès 2000: 141, Journet 1995). It is, of course, a moral ideal, and Alès even speaks of “obligations of demonstration of affection” (id.), thus implicitly revealing doubts concerning the “authenticity” of the emotion. Finally, incoming affines are consanguinised because they are coresidents: commensality, exchange, and even the fact of living together, are essential parts of the transformation of an other into a relative (this idea has already been introduced in the introduction, cf. supra, and will be examined again, cf. infra, chapter 4). But such a process, in the case of male affines has to remain incomplete. For they are not women that can be domesticated by sexual relations (e.g. Descola 1996: 185, Viveiros 2001: 24), and they are not children that can be familiarised as pets: because they are male 40

adults, coresident affines will keep their alterity. Often, it is only in specific situations – for instance war – that consanguinisation of the local group is possible: “Fathers-in-law, brothersin-law and sons-in-law disappear from the field of social references to be transformed into chosen blood relatives, while affinity, now lacking an effective basis, tends to be converted into an abstract relationship, making it possible to characterize enemies as givers of women” (Descola 1996: 292) – a process that refers to the meta-affinity discussed earlier. Moreover, sociality is not restricted to the local group: interaction and cooperation must often take place with affines that are not coresident, and that cannot therefore be consanguinised. What methods do Amazonian people employ to deal with them? 2. Self-control. The situation is of course diverse, because kinship systems differ greatly from each other. One specific case would for instance be the Jivaro, where young brothers-in-law joke freely with each other, and can even make physical contact in public, before – it is true – evolving towards greater formality when they progressively turn into allied warriors (Taylor 2000: 311). Usually, however, relationships with affines are marked by self-control, etiquette and formality. Jokes in general, and especially sexual jokes, are often forbidden with in-laws. Often, there is not even direct speech, especially when spouses were unrelated before marriage: “even when in close proximity, a man will ask his daughter to ask her husband to lend him his knife; the woman then asks the husband, who hands her the knife which she passes to her father” (Rivière 2000: 259). All these forms of avoidance, the interdiction of familiarity, etc., are very frequent in Amazonian cultures, and more accentuated between real affines of different generations (i.e. father-in-law/son-in-law). It is true that the area of application of such behavioural rules varies a lot. In some cases the brother-in-law relationship is the exception, “not as comfortable as other relationships”, because “men do not banter with their brothers-in-law about the earthy matters usual to their humour” (Overing 2000: 68-69). In 41

other groups though, restraint between most adult males seems to be the rule, for instance among the Jivaros (Descola 1996: 58), and formal etiquette is compulsory with men of neighbouring villages, for instance in the case of the Brazilian Kalapalo: the notion of respect is fundamental in the constitution of the Xingu as a multi-ethnic peaceful area (Basso 1995: 271). We could also add that such practices are not restricted to real human in-laws, for relationships are not free with game affines, when they take the aspect of masters of game – powerful and dangerous fathers-in-law that give their daughters to the hunters (e.g. Jivaroan magical songs, where hunters ask animal in-laws to give them their children and their sisters, Descola 1994: 261-262). The notion of respect, both in the treatment of the prey and in the ways people address masters of game, has been stressed by many anthropologists in all America: among the Crees, people do not call animals by their names, the same way one avoids addressing others of the same age or older with personal names, because of the “respect” due to them (Brightman 1993: 115); while among the Brazilian Juruna there is an imperative of “verbal moderation” (Fausto 2002: 14). Among the Makuna, the Spirit Owners of the animals give their daughters to the hunters, and humans need to reciprocate – otherwise revenge will be taken (Arhem 1996: 196). How do these practices appear, from the point of view of the anthropology of emotions? If we render explicit most descriptions, the idiom seems to be that of social control over natural emotions. These are a force within the individual, and their free expression could be a threat to social peace, especially when affines are in the picture. People do not joke with their affines, as this could be interpreted as a serious insult (Overing 2000: 69), they behave according to etiquette when they visit other groups, because foreigners are always assumed to try to seduce women (Basso 1995: 271, Descola 1996: 199), they should not express their anger, because it could easily lead to war or at least to scission (e.g. Kidd 2000: 114), especially in a context where suspicions of sorcery are often directed towards affines (Rivière 1984: 80). Affinal 42

relationships have political significance, and therefore “respect” is more important than “affectivity” (Surrallès 2003: 33). We thus encounter a social law that appears to be most reasonable, corresponding with classical western oppositions: collective vs. individual, social vs. biological, reason vs. emotion, expression vs. feeling, etc. Most recent works on the anthropology of emotions have of course criticized these oppositions, treating them as products of a misleading ethnocentrism (e.g. Leavitt 1996, Lutz & Abu-Lughod 1990). However, part of the problem is that such an idiom of social control seems to be congruent with Amazonian natives’ point of view. It is particularly obvious when anthropologists tackle the subject of drunkenness – a usual reason for fighting. When they are drunk, people “begin to lose their self-control and behave with less restraint, […] it can end in the ultimate disrespect of anger” (Kidd 2000: 126). “Public outbreaks of anger [are] explained as a consequence of alcohol’s erosive effect on people’s control of their emotions, which is probably the physical effect in most cases. However, drunkenness is also seen as a state in which people can come in contact with their own true selves (noseire) which means that the social restraints set by the generally shared sense of community may be discarded in order to let the suppressed self take over and dominate” (Rosengren 2000: 226). These quotes could be very satisfying: native folk models seem to be similar to western folk models about emotion and self-control. Yet is it enough to accept these metaphors as accurate descriptions and models (Solomon 1984: 241)? Can we accept the idea that emotions are inner forces to be controlled (especially with reference to affines) because some Amazonian and Western metaphors confirm this view? 3. Shame and restraint. We have seen that anger is a disposition that can be cultivated (cf. supra), and now it appears as a thing to be controlled, in the context of a restraint of all affectivity. But this last point of view is misleading because it sees emotion as a “thing”, a limited reification already alluded to (Tarlow 2000: 714). For if we take another look at all these situations, it appears that 43

interpreting them in terms of self-control can only be a partial description. In many cases, native Amazonians instead speak of their “shame”. It can happen when they realise that they have behaved inappropriately, and shame can force them to leave the community (Kidd 2000: 127, Rivière 2000: 259). Rosaldo (1984: 144) made a similar comment about the Ilongot: a drunken fight does not reveal anger as a deep truth, for it is not followed by a lasting disagreement, but only by shame. In the Amazonian region, shame can express people’s feeling in front of someone who behaves without restraint, their shame in front of someone who is precisely shameless, who expresses his anger or asks for food or money all the time (Belaunde 2000: 210, McCallum 2001: 97, pers. obs. among the Guaraní). It can be a quality of a relationship, more than a general state of the individual: shame “should obtain between a man and his affines. A person with intense shame toward another does not look straight at him or her, nor speak to him directly, nor use his name, share cooked food with him, engage in sexual relations with him, or joke with him in any way” (Seeger 1981: 125). Similarly, a Kulina man is full of shame towards his mother-in-law (Pollock 2002: 49). In many cases, it therefore appears that it is not reason that should control an inappropriate emotional behaviour (anger, desire, etc.) but shame, that is to say another notion of affective nature. What about the old opposition between emotion and reason? Has it been replaced by the more surprising opposition, emotion vs. emotion? It is more interesting, however, to follow another path, and to think of emotion not as a thing opposed to others, but as a quality of the world (i.e. of relationships, of actions, of persons, etc.). This would explain why the same notion could for instance be summoned to explain processes as distinct as the transmission of cultural models (e.g. D’Andrade & Strauss (eds) 1992) and action (cf. supra). Both an inappropriate behaviour (anger) and its control (shame) can be understood by referring to emotion: it is only a aspect of the dispositions that generate people’s actions and justifications, a quality of the habitus. The use of this concept therefore enables us to avoid tendencies that essentialise emotions (cf. supra, Rosenberg commenting on Rosaldo): the habitus is not a 44

thing, because it has essentially a relational nature that explains the possible existence of tensions – here between anger and shame, in Bourdieu’s analysis of Arab kinship (1990: 187) between a centripetal drive (desire of agnatic security) and a centrifugal drive (desire of a prestigious alliance).

45

CHAPTER IV. THE PRODUCTION OF CHILDREN AS KIN.

1. From biology to ethnobiology. We began by tackling the issue of the production of consanguineal ties with the notion of familiarisation (pet-children and pet-animals) and with the consanguinisation of real affines. All the situations so far considered excluded by definition any procreative tie: kinship could only be of a social nature – that is to say “fictive”. For in most (Amazonian) studies of kinship, until fairly recently at least, the primacy of “biological” connections has been assumed, whether when drawing genealogies in order to understand alliance systems, or when trying to understand practices and conduct. What was understood by biological connections was mostly transmission of genetic substance, following a western model, – that is to say relations of procreation. Anomalies were noticed (e.g. partible paternity), but they were reduced as much as possible to the basic model (e.g. distinction between primary father and secondary fathers – the latter being forgotten when studying kinship). Such analyses followed at least implicitly Malinowski’s dogma that “statistically speaking, the biological ties are almost invariably merely reinforced, redetermined, and remoulded by the cultural ones” (1930: 137). However, it has proved necessary to take into account some native specificities. Many anthropologists have found it necessary not to rely on the opposition between the biological and the social, and have instead developed an interest in folk biology. In anthropology in general, it began with Barnes’ tripartite division between genetic father (i.e. father according to western science), genitor (i.e. father according to folk biology) and pater or social father (1961, 1964). These distinctions have been followed by Scheffler and Lounsbury, who stressed that what really matters for anthropologists is the opposition between the second and the third term. These anthropologists defined “genealogical connections” as “those culturally posited forms of interpersonal connectedness that are held to be direct consequences of the processes of 46

engendering and bearing children and that have the property of indissolubility” (1971: 37-38). In Amazonian anthropology, this approach has led until now to an interest in native conceptions of conception, with always a focus on the transmission of substances that could be assumed to contain what we would see as “genetic substance”, essentially semen in our case. Firstly, we have partible paternity or “multiple ethnobiological fatherhood – the belief that more than one man is the biological father of a child” (Pollock 2002: 42). The fact that this is held to be “ethnobiological” comes from Amazonian ideas about conception (when there are some): the foetus is formed by the accumulation of semen in the womb of the woman, and therefore all the men who have had sexual relations with the mother have contributed to the foetus. This is false biology according to us, and therefore is ethno-biology. Yet this is a belief held by many in Amazonia (in spite of not being universal, it is very common throughout lowland South America), and a belief may have a reasonable cause and important effects. “In the absence of DNA analysis nonmarital and extramarital sex can create a problem in establishing the social identity of any child born to a woman who has had multiple sex partners at the time she becomes pregnant” (Kensinger 2002: 24): according to some authors, the native ideology of partible paternity is an answer to such an uncertainty – a practical belief in a situation where truth is unavailable. There is a second frequent ethnobiological conception: women are mere pots that men fill with semen. “A foetus is formed entirely of paternal semen, and the pregnant woman makes no material ‘biological’ contribution to foetal development during this gestational stage. However, at birth […] mother’s milk becomes, at least conceptually, the exclusive substance nourishing the child” (Pollock 2002: 47). This is sometimes contrasted with the Trobriand claim (reported by Malinowski) that fathers do not make any material – seminal – contribution to the foetus (e.g. Alès 2002: 65, Pollock 2002: 42): a substantial link between mother and child only appears later, when she feeds it with her milk. These conceptions about the substantial formation of the baby are what constitute ethnobiology: native ideas about what is, according 47

to us, the foundation of kinship. There can be several genitors, all those who have contributed to the foetus by their semen (and sometimes several genitrix, who have breast-fed the baby), but they are all opposed to social fathers, that is say to men who call the baby “son” or “daughter” (and are called “father”), treat it accordingly, and are usually the mother’s husband (e.g. Alès 2002: 68). We face here the consequences of the fundamental assumption that “blood is thicker than water” – the implicit idea highlighted by Schneider that kinship consists of compelling bonds that are “states of being” (and not of doing), because of their (ethno-) biological nature (1984: 165-166): kinship as a social construct has always to deal with them. 2. The construction of consanguinity: a new point of view on substance. Although it is still present in some of the most recent works on Amazonian parenthood, such a perspective has of course been criticised by many. It is meaningless to oppose the (ethno-) biological facts of procreation to posterior social practices. For, as Viveiros de Castro stated, in Amazonia, consanguinity is not (constructed as) given – by biology –, but (given as) constructed – by practices that involve at the same time the body and social relations (Viveiros 2001: 19). We can therefore criticise the focus on the transmission of some substances that are given a privileged status because of the symbolic importance they have in our own conceptions of kinship: food for instance seems to be nothing compared to blood and semen – but is this the case for Amazonians? This question will be dealt with later, for it seems necessary to begin with an “ethnobiological” fact apparently characteristic of lowland South America: the absence of any contribution from the mother to the foetus. The focus on material substance seems as misleading here as it is in Melanesia, for if we follow Strathern: “The mother’s body is an aesthetic construction: what matters is its form. […] It is a consequence of a relationship between mother and child but one that need not be mediated by feeding or any other transmission of ‘substance’. The relationship consists of form itself: a contained entity is 48

enveloped by a container” (1988: 329). My point is of course not to say that what is true of Melanesia is also valid in native America, but western assumptions have sometimes taken a similar form in both cases: a focus on substance. The first implication that we can draw from Strathern’s warning is that being a mere container is not an irrelevant contribution. This point has been noted by Alès, who underlines that, among the Yanomami, “only uterine affiliation can confer ‘true’ consanguineal siblingship” (2002: 67); and by Viveiros de Castro, according to whom: “As the site of transformation or as the transformer herself, the mother exercises a metonymical causality 3 over the material she receives in her depths. The idea of an exclusive masculine genesis must be viewed with caution” (1992: 179). It therefore seems that the seminal contribution, and the connection that is assumed to result from it, is not the ultimate ground of relatedness: we will explore this issue further. Most Amazonian ethnography underlines the difference between “real” kin and kin in general. But contrary to the assumptions of western conceptions, “real” does not mean biological here. “Real kin” are in fact usually coresident kin, that is to say, those with whom one eats, shares, works, lives, etc. One of Gow’s informants told him that they were all kin here, that all her kin were here – a statement that excluded several related persons, and included several unrelated co-villagers. But “to be real, one might say, a relationship must be enacted […]. Coresidence, living together in the same village, is a demonstration of caring” (Gow 1991: 193-194). This question of the “reality” of kin can be explored further by a study of adoption. We discussed earlier the cases in which the children of enemies are captured and familiarised as pets, but adoption can be a much wider phenomenon in Amazonian societies: many children are adopted, not by foreigners or enemies, but by related people. A cousin, a grandmother or a co-villager can be turned into a mother (e.g. Alès 2002: 72, Peluso & Boster 2002: 142).

3. A comparable metonymic causality appears when women are conceptualised as “mothers of crops”, and make tubers grow, among the Barasana (Hugh-Jones, pers. com.).

49

Adoption is also a long process (McCallum 2001: 67), and not the result of a single performative legal act (as in western culture): only through constant feeding, caring, living together, sharing, etc., is it possible to turn a child into one’s own child. This insistence on food shows the link with domestication: in Barasana, “to domesticate” means “to give food to” – a dog is “a jaguar you feed”, a wild animal is “an animal that is not fed” (Hugh-Jones, pers. com.). An example from Ecuador is particularly apposite: a man was feeding a young boy, and said proudly that the boy was an orphan that he had adopted. A nurse coming from the capital, who was there with the anthropologist, was horrified, and told the man that he should not talk like that in front of the boy: “if he’s lucky, he’ll forget about his own parents and grow up believing you’re his real father.” The man answered indignantly: “I am going to be his father. Aren’t I feeding him right now?” (Weismantel 1995: 690). He did not say that he was the father, for the process had hardly begun, but he would be, one day. And he would of course be the true father. What has interested many recent discussions of Amazonian kinship is that the idiom expressing this true kinship is that of substance. “Instead of the concern with the relationship between the biological and the social, authors aimed to show the complexity of the ‘biological.’ The emphasis was thus on native notions of body and consubstantiality; rather than being seen as natural givens, these came to be understood as products of society and culture. Consubstantiality, located in this new body, was no longer a relation determined by birth, but a condition being continuously produced through acts of sharing, particularly of food and mutual care” (Vilaça 2002: 348). This idiom of consubstantiality has been recognised by anthropologists such as Rival who states that the result of such practices is being “of one same flesh” (1998: 621), Gow who talks of “shared substance” (1991: 161), or Fausto who notes that commensality consists in the production of “bodies related as kin” (corpos aparentados, 2002: 8). These interpretations were probably privileged because they deal adequately with the oldfashioned opposition between the biological and the social. What we would consider as 50

biology (i.e. genealogical connectedness) does not really matter, because in a certain way there is no difference between adopted children and “real” children: all must be made one’s own children, because newborns are “guests” (Rival 1998), or “others” that can easily be turned into animals or spirits (Vilaça 2000). What matters in fact is the indistinctively social and biological process of fabricating bodies, that is to say of making them consubstantial in order to ground kinship ties in the flesh. 3. From substance to affects. Many anthropologists stress this idiom of consubstantiality. Some, however, underline its coexistence with “an idiom of caring and its reciprocation in the use of kin terms” (Gow 1991: 161). Even if “substance” is not absent from Gow’s description of the Piro, he gives more importance to the relationships of caring and feeding, in so far as they are remembered, and often determine the strength of kin ties between adults (id.). The reader, however, may wish to question his focus on food and substance. Once again, we should recall what Strathern said about Melanesia: “For a further Eurocentric image to be discarded lies in the idea, often imputed to Melanesian cultural practice, that food creates identity, with the implication that food contributes directly to the substance of persons, and that substance is always an inner condition. […] It is not the food as such that must be analyzed, but the feeding relationship” (1988: 251). This feeding relationship is not irrelevant: Gow for instance notes that the relation with the parents changes once the child eats “real food” (i.e. game meat and garden crops), that is to say receives gifts of food from his parents: these gifts, “given out of love for the child, evoke the child’s love for its parents” (1991: 157). Therefore, the feeding relationship indeed contributes to the growth of the child, but not only that: it is also central in the constitution of a memory of past feeding and of a relationship of care, a consideration that leads to Taylor’s definition of paternity as a “cumulative and shared relational disposition” (2000: 319, cf. also Vilaça 2002: 352). 51

Such an ambivalence between two idioms that define kinship ties – shared substance and caring relationships – must lead to a questioning of the Amazonian body, for much of the insistence on bodily substance probably came from the idea that the body (as a construct) is the central idiom for the understanding of social processes in lowland South America – an idea that has been dominant for the last 25 years (Seeger, Da Matta, Viveiros de Castro 1979). However, Surrallès recently made the surprising statement that there is a general uniformity of bodies among the Jivaroan Candoshi, and in Amazonia in general. He brings this case especially to the question of human/non-human relationships: what differentiates animal species and human groups are not their bodies, but their ethology (i.e. behaviour) and even their psychology (behaviour expresses an intentionality): the Candoshi pay little attention to morphology and physiology (2003: 37). In spite of going against the trend of much Amazonian anthropology, the difference between Surrallès’ analysis and that of Viveiros de Castro lies perhaps more in words than in concepts: “what I call ‘body’ is not a synonym for distinctive substance or fixed shape; it is an assemblage of affects or ways of being that constitute a habitus” (Viveiros 1998: 478). But is this a ‘body’ anymore? From this perspective, for instance, the question of metamorphosis among Amazonian people is misleading: shamans or men do not transform their bodies in the body of a jaguar, for “‘jaguar’ designate only a way of acting without any ontic content and unrelated to any somatic configuration” (Surrallès 2003: 201) – the jaguar skins warriors put on their back is only an “emblem” or “flag”, a symbol of their intentionality (id., cf. also Hugh-Jones [1996: 140] on animal fur and feathers as corresponding to the heraldic concept of “arms”). All this might seem far from our subject. However, it serves the purpose of revealing that the body is not the universal idiom of Amazonian cultures, or at least, that it should not be reduced to the bodily substance, but envisaged as an embodied habitus, and, as we have already tried to show, this habitus can have an affective quality, it cannot be reduced to ethology or intentionality. Surrallès and Viveiros de Castro have noted this, but have not really exploited the point, or explored its implications. Consequently, we could say 52

that emotion here is bodily, but only according to the Amazonian conception of the body. From a certain point of view, this recalls an idea put forward about Inuit kinship: one can choose one’s kinspeople, but not the way of behaving towards them (Nuttal 2000: 45). What defines kinship is the way people behave towards one another (ethology), which is a product of their affective dispositions (psychology), not a shared substance. Surrallès’ idea is also useful in interpreting another aspect of commensality. It is sometimes stated that to eat together is a factor of identification (Fausto 2002). Often though, the idea is precisely that criticised by Strathern: the ingestion of the same food produces consubstantial bodies, because food is what constitutes flesh (e.g. Rival 1998). It is however possible to take another direction, especially when what is at stake is not concrete food, but an alimentary diet, for instance the observance of the same restrictions (e.g. Vilaça 2002: 352): eating expresses intentionality and social values. This reminds us of Basso’s statement about the Kalapalo that dietary practices (the avoidance of the same big animals) define shared values (i.e. not being a predator), and therefore the limits of the group with whom it is possible to coexist (1995: 200). Here we could say that eating the same food also means having the same alimentary ethos (though the former cannot be reduced to the latter) – and this is a major element in defining species. The body does not seem to be a substance, but a disposition that could be defined as Bourdieu’s habitus – a scheme that determines ways of seeing (i.e. Amerindian perspectivism), feeling (i.e. affects) and acting (i.e. ethology) –, and that is a product of social practices, and especially of relationships of commensality. Yet such a definition meets two difficulties: firstly the couvade, and secondly the idiom of shared substance. For in spite of the little attention we have paid to ethnobiological connections, we cannot avoid the fact that all the men who have contributed their semen to the foetus (and the mother as well) must observe certain restrictions – otherwise the newborn could become ill, die, or turn into an animal: all authors state explicitly that ethnobiological fathers are involved, and not classificatory, adoptive or step-fathers (e.g. Crocker 2002: 88). The 53

importance of these connections, that we had dismissed earlier, seems thus to be restored. But does this mean that they are meaningful from all points of view, for the whole of life? Does this mean that they are “indissoluble” (Scheffler & Lounsbury 1971: 38)? As Schneider puts it, “it is assumed that this relationship, of birth or blood or biological relatedness, or arising from reproduction, is self-evidently of high value, that it is of great and grave significance – that it is, in short, privileged”, whereas we should instead question the value and meaning of such ethnobiological relatedness (1984: 123). What is called the couvade is usually defined as a certain number of restrictions imposed on the parents of a newborn (not eating several kinds of meat, often not hunting, not practicing shamanism, etc.) for fear of what could happen to the baby. It seems to result from two concomitant facts: at the same time, the body of the newborn is not independent from the bodies of its parents (who are “blended with the child”, according to Vilaça [2002: 355], while Gow speaks of “corporeal identity” [1991: 154-155]), and it is still mutable (the child could assume the characteristics and dispositions of the animal eaten by his parents, Vilaça 2002: 357). Therefore, after expiry of the restrictions that have a preventive aim, a double process takes place: the parents must attach the child to them, making his “body” akin to theirs (Vilaça 2002: 357) in order to avoid the possibility of his turning into an animal; but the child is at the same time progressively detached from them, so that their behaviour does not have direct consequences on himself. As Menget underlines, “the progressive relaxing of the couvades is thus the sign of a gradual differentiation of substance, and the child’s eventual acquisition of a separate identity” (1982: 202). It is thus possible to postulate the hypothesis that we face here something similar to a shift from an “unmediated exchange” (semen, blood, milk) to a “mediated exchange” (gifts of “real food”) that has the effect detaching people from people, differentiating them (e.g. Strathern 1988: 191-192). In this light it is apparent of course that ethnobiological connections are of major importance for a newborn. But soon, this form of relatedness fades away, when the peril disappears, and another form replaces it – founded not

54

any more on the transmission of semen, but on the constitution of a bodily disposition through feeding and caring. The second difficulty that we face is that of the idiom of shared bodily substance. This idiom is indeed a reality of Amazonian societies, but how can it properly be explained by our interpretation of the body as a habitus, and our idea that kinship relies on such relational dispositions? In order to do so, we must leave the idea that the kinship tie could derive directly from the shared substance – an interpretation that has been expressed in other contexts, for instance by Carsten about the Malays: “sharing of rice meals cooked in the same hearth […] thus also implies shared substance” and “shared substance gives emotions and words a special effectiveness” (1995: 228-229). On the contrary, in Amazonia at least, it is possible to interpret this differently, following the Araweté idea that “what remembers is the flesh” – a memory of past relationships (Viveiros 1992: 207). The bodily substance, the flesh, stands therefore as a record of these relationships of caring and feeding that generate the habitus of kinspeople: it is not the cause but the sign of this habitus. In this sense, it is a symbol standing for something else: as in the fatness of the Melanesian pig being evidence of the woman’s work (Strathern 1988) and the shared flesh of Amazonian kinsfolk being a sign of the relationships that transformed them into kin (J. Kelly: pers. com.). 4. Between care and desire. We need however to define more precisely the nature of this relational disposition that we have been speaking of. We could say, broadly, that it reveals an emotional quality – or content – of relationships, that we could call affectionate love, and that this can take the form of many emotions. Kinspeople, i.e. coresident people related by commensality and caring, pay attention to each other: they “look after” each other, take care of each other during periods of distress (e.g. Gow 2000: 50). It is also often stressed that they feel nostalgia for each other when they do not live together. This serves to explain endogamy or the choice of locality, more than any 55

rule (McCallum 2001: 30, Gow 1991: 121). Here appears another notion: the desire to live with one’s real (and not biological) kin. We have already evoked the question of desire when talking about seduction, and now it must come back into the picture. For a difficulty arises now that kinship ties are not grounded on biology anymore. According to the assumption that “blood is thicker than water”, the compelling nature of biologically based kinship ties should be the ground for social ties. But we have seen that such conceptions were unsustainable in a world where adoption is not an exception, and where “true” kinspeople are those who have cared and still care for each other. The affective disposition of parents and children, or of consanguines in general, generates a desire to live with each other, to take care of each other, to treat each other well. But where does this disposition come from once we cannot rely anymore on Fortes’ “axiom of amity” (1969: 232-234)? We have said that it was produced by past caring relationships, but why did these relationships take place if there were no biological connections to justify them? Why do people desire to be kin in the first place? Couples who have no children anymore (because they are too old, or because they are sterile) or single individuals can desire to raise children because without them a household is incomplete and sad, because old people need help, because they do not want to be alone (Peluso & Boster 2002: 142). Sometimes, economic factors are put forward (as a workforce, for instance among the Paraguayan Guarani, Boidin: pers. com.), but this is not the general case in societies where the agency of children is recognised: if adults can desire to have children to live with them, children can also desire to live with some people more than others, and sometimes their parents do not even think of going against their decision (e.g. Bodenhorn 2000: 147n.11). Among the Barasana, parents often borrow each other’s children for a couple of months, but some children decide not to come back (Hugh-Jones, pers. com.). The contrast is obvious with non-Amerindian societies, for instance, for the Malays, commensality creates kinship, but parents teach their children always to come back home for meals (Carsten 1995: 225). In Amerindian societies, children 56

are not a thing or a possession of parents, for they have subjectivity: desires cannot be unilateral, among adults or between adults and children, when the issue at stake is for them to become kin. Perhaps we reach here the limits of anthropological analysis, yet it is possible to assert that we face a desire here very different from that with which we began. Sexual or hunting (cross-sex) desire targeted a fleeing prey, or a disdainful woman. The (same-sex) desire to live with kinspeople, with parents and with children concerns people who reciprocate this desire – we have indeed seen that same-sex predation was reversible. This is the desire to live with people who also want to live with you, who want to care for you, and who will long for you. Becoming kin with someone else begins with the desire to do so, which is expressed by acting as a proper kinsperson (McCallum 2001: 76, Vilaça 2002: 352) – not through connections of substance.

57

CONCLUSION.

This overview of Amazonian kinship and emotions was an attempt to get out of the impasses of current debates. Amazonian kinship has already been widely studied, and has moved away from models established for other cultural areas: the concepts used by anthropologists are nowadays culturally relative. However, its treatment of emotions is probably not satisfactory. There is too much “love” and “conviviality” on the one hand, which reveals an uncritical use of emotion by the “moral economy of intimacy” school; and on the other hand there are too few affective elements in the analyses of the exponents of the “symbolic economy of alterity” school which remains centred on cosmological and ontological processes. We have therefore tried to push this debate further by focusing on emotional aspects while respecting the primacy of affinal relations characteristic of lowland South America. At the same time, it was necessary to enter concretely into a discussion of consanguinity, in order to avoid jumping from one essentialism to another through an idiom of substance. Could such a handling of emotions, that starts from kinship relations and gives priority to action, have an effect on the debate on emotions? We tried to show, with the introduction of the notion of habitus, that it could. The significance of this introduction came from a displacement of the focus. Rather than a focus on what emotions are (a bodily feeling, an interpretation, etc.), it is more productive to focus on what they do – action. It is therefore useful to introduce a notion that “is always oriented towards practical functions” (Bourdieu 1990: 52). This introduction had another aim, which is to avoid the risks of essentialism inherent in any interest in emotion words – they come to act as “things” that have a psychological effect in themselves (cf. supra, Introduction, Rosenberg 1990). On the contrary, the habitus is a relational disposition, a scheme. We have, for instance, encountered the notion of compassion, but tried not to see it as a thing. We therefore contextualised it within the scheme of familiarisation, that is to say a relation (between compassionate master and helpless 58

and submitted pet), that is always evolving (from seduction or war/ hunting to marriage and adulthood of adopted children). As a scheme that is both socially structured (through experiences, and not only “culture”), and structuring of social experience and practice, the habitus – when its emotional component is elicited – is a way of reintroducing emotions into the field of social anthropology.

59

BIBLIOGRAPHY ABU-LUGHOD (Lila) & LUTZ (Catherine A.) (eds). 1990 Language and the Politics of Emotion. Cambridge: Cambridge UP, Paris: Editions MSH. ABU-LUGHOD (Lila) & LUTZ (Catherine A.). 1990. Introduction: emotion, discourse, and the politics of everyday life. In Abu-Lughod (Lila) & Lutz (Catherine A.). Op. cit. 1-23. ALÈS (Catherine). 2000. Anger as a Marker of Love. The ethic of conviviality among the Yanomami. In Overing (Joanna) & Passes (Alan) (eds). Op. cit. 133-151. — 2002. A Story of Unspontaneous Generation. Yanomami male co-procreation and the theory of substances. In Beckermann (Stephen) & Valentine (Paul) (eds). Op. cit. 62-85. ARHEM (Kaj). 1996. The Cosmic Food-Web: human-nature relatedness in the Northwest Amazon. In Descola (Philippe) & Pálsson (Gíslí). Op. cit. 185-204. BARNES (J.A.). 1961. Physical and Social Kinship. Philosophy of Science 28: 296-299. — 1964. Physical and Social Facts in Anthropology. Philosophy of Science 31: 294-297. BASSO (Helen). 1995. The Last Cannibals. A South American Oral History. Austin: University of Texas Press. BECKERMANN (Stephen) & VALENTINE (Paul) (eds). 2002. Cultures of Multipe Fathers. The Theory and Practice of Partible Paternity in Lowland South America. Gainesville: University Press of Florida. BELAUNDE (Luisa Elvira). 2000. The convivial self and the fear of anger amongst the Airo-Pai of Amazonian Peru. In Overing (Joanna) & Passes (Alan) (eds). Op. cit. 209-220. BERLIN (Brent) & KAY (Paul). 1969. Basic Colour Terms: their universality and evolution. Berkeley & Los Angeles: University of California Press. BODENHORN (Barbara). 2000. ‘He used to be my relative’: exploring the bases of relatedness among Iñupiat of northern Alaska. In Carsten (Janet) (ed). Op. cit. 128-148. BOURDIEU (Pierre). 1990. The Logic of Practice. Cambridge: Polity Press. BRIGHTMAN (Robert A.). 1993. Grateful Prey: Rock Cree Human-Animal Relationship. Berkeley: University of California Press. CARSTEN (Janet) (ed). 2000. Cultures of Relatedness. New Approaches to the Study of Kinship. Cambridge: Cambridge UP. CARSTEN (Janet). 1995. The Substance of Kinship and the Heat of the Hearth: feeding, personhood, and relatedness among Malays in Pulan Langkawi. American Ethnologist 22(2): 223-241. — 2000. Introduction: cultures of relatedness. In Carsten (Janet) (ed). Op. cit. 1-36. 60

CHAUMEIL (Jean-Pierre). 1983. Voir, Savoir, Pouvoir. Le chamanisme chez les Yagua du NordEst Péruvien. Paris: EHESS éditions. CLASTRES (Hélène). 1972. Les Beaux-Frères Ennemis: à propos du cannibalisme tupinamba. Nouvelle Revue de Psychanalyse 6: 71-82. Clastres (Pierre). 1974. La Société contre l’Etat. Paris: Minuit. CONKLIN (Beth). 2001. Consuming Grief: compassionate cannibalism in an Amazonian society. Austin: University of Texas Press. CROCKER (William H.). 2002. Canela ‘Other Fathers.’ Partible paternity and its changing practices. In Beckermann (Stephen) & Valentine (Paul) (eds). Op. cit. 86-104. D’ANDRADE (Roy) & STRAUSS (Claudia) (eds). Human Motives and Cultural Models. Cambridge: Cambride UP. DESCOLA (Philippe) & PÁLSSON (Gíslí) (eds). 1996. Nature and Society. Anthropological Perspectives. London & New York: Routledge. DESCOLA (Philippe). 1994. In the Society of Nature. Cambridge: Cambridge UP. — 1996a. The Spears of Twilight. Life and death in the Amazonian jungle. London: HarperCollins. — 1996b. Constructing Natures. Symbolic ecology and social practice. In Descola (Philippe) & Pálsson (Gíslí). Op. cit. 82-102. DUMONT (Louis). 1971. Introduction à Deux Theories d’Anthropologie Sociale. Groupes de filiation et alliance de mariage. Paris-La Haye: Mouton. EKMAN (Paul). 1984. Expression and the Nature of Emotion. In Scherer (Klaus R.) & Ekman (Paul) (eds). Approaches to Emotion. Hillsdale (N.J.): Erlbaum. 319-343. ERIKSON (Philippe). 1987. De l’Apprivoisement à l’Approvisionnement. Chasse, Alliance et familiarisation en Amazonien amérindienne. Techniques et Cultures 9: 105-140. — 2002. Several Fathers in One’s Cap. Polyandrous Conception among the Panoan Matis (Amazonas, Brazil). In Beckermann (Stephen) & Valentine (Paul) (eds). Op. cit. 123-136 FAUSTO (Carlos). 1999. Of Enemies and Pets: warfare and shamanism in Amazonia. American Ethnologist 16(4): 933-956. — 2001. Inimigos Fiéis: história, Guerra e xamanismo na Amazônia. São Paulo: Edusp. — 2002. Banquete de Gente: comensalidade e canibalismo na Amazônia. Mana 8(2): 7-44. FORTES (Meyer). 1969. Kinship and the Social Order. London: Routledge and Kegan Paul. GOW (Peter). 1991. Of Mixed Blood. Kinship and History in Peruvian Amazonia. Oxford: Clarendon Press.

61

— 2000. Helpless. The affective preconditions of Piro social life. In Overing (Joanna) & Passes (Alan) (eds). Op. cit. 46-63. HÉRITIER (Françoise). 1999. La Citadelle Imprenable. Critique 620-621: 61-83. HUGH-JONES (Stephen). 1996. Bonnes Raisons ou Mauvaise Conscience? De l’ambivalence de certains Amazoniens devant la consommation de viande. Terrains 26: 123-148. IRVINE (Judith T.). 1990. Registering Affect: heteroglossia in the linguistic expression of emotion. In Abu-Lughod (Lila) & Lutz (Catherine A.). Op. cit. 126-161. JAMIESON (Mark). 2000. Compassion, Anger and Broken Hearts. Ontology and the role of language in the Miskitu lament. In Overing (Joanna) & Passes (Alan) (eds). Op. cit. 82-96. JOURNET (Nicolas). 1995. La Paix des Jardins. Structures sociales des Indiens curripaco du haut Rio Negro (Colombie). Paris: Institut d’Ethnologie. KENSINGER (Kenneth M.). 2002. The Dilemmas of Co-Paternity in Cashinahua Society. In Beckermann (Stephen) & Valentine (Paul) (eds). Op. cit. 14-26. KIDD (Stephen W.). 2000. Knowledge and the Practice of Love and Hate among the Enxet of Paraguay. In Overing (Joanna) & Passes (Alan) (eds). Op. cit. 114-132. LAMBERT (Helen). 2000. Sentiment and substance in North Indian forms of relatedness. In Carsten (Janet) (ed). Op. cit. 73-89. LEAVITT (John). 1996. Meaning and Feeling in the Anthropology of Emotions. American Ethnologist 23(3): 514-539. LÉVI-STRAUSS (Claude). 1943. Guerre et commerce chez les Indiens d’Amérique du Sud. Renaissance 1(1-2): 122-139. — 1968. Structural Anthropology. London: Penguin Press. LOSONCZY (Anne-Marie). 1990. La Maîtrise du multiple. Corps et espace dans le chamanism embera du Choco (Colombie). L’Homme 114: 75-100. MAINE (Henri S.). 1861. Ancient Law. London: J. Murray. MALINOWSKI (Bronislaw). 1930. Parenthood. The basis of social structure. In Calverton (V.) & Schmalhausen (S.D.) (eds) The New Generation. New-Yord: Macaulay. 112-168. MCCALLUM (Cecilia). 2001. Gender and Sociality in Amazonia. How Real People Are Made. London: Berg. MENGET (Patrick). 1982. Time of Birth, Time of Being. The Couvade. In Izard (Michel) & Smith (Pierre) (eds) Between Belief and Transgression. Chicago: The University of Chicago Press. 193-209. MORGAN (L.H.). 1870. Systems of Consanguinity and Affinity of the Human Family. Washington, D.C.: Smithsonian Institution. 62

NUTTAL (Mark). 2000. Choosing Kin. Sharing and subsistence in a Greenlandic hunting community. In Schweitzer (Peter P.) (ed). Dividends of Kinship. Meanings and uses of social relatedness. London & New York: Routledge. 33-60. OVERING (Joanna) & PASSES (Alan) (eds). 2000. The Anthropology of Love and Anger. The aesthetics of conviviality in Native Amazonia. London & New York: Routledge. OVERING (Joanna). 2000. The Efficacy of Laughter. The ludic side of magic within Amazonian sociality. In Overing (Joanna) & Passes (Alan) (eds). Op. cit. 64-81. PELUSO (Danelia M.) & BOSTER (James S.). 2002. Partible Parentage and Social Networks among the Ese Eja. In Beckermann (Stephen) & Valentine (Paul) (eds). Op. cit. 137-159. POLLOCK (Donald). 2002. Partible Paternity and Multiple Maternity among the Kulina. In Beckermann (Stephen) & Valentine (Paul) (eds). Op. cit. 42-61. RENARD-CASEVITZ (France-Marie). 1993. Guerriers du sel, sauniers de la paix. L’Homme 126128: 25-43. RIVAL (Laura). 1996. Blowpipes and Spears: the social significance of Huaorani technological choices. In Descola (Philippe) & Pálsson (Gíslí). Op. cit. 1145-164. RIVIÈRE (Peter). 1984. Individual and Society in Guiana. A Comparative Study of Amerindian Social Organization. Cambridge: Cambridge UP. — 2000. ‘The more we are together…’ In Overing (Joanna) & Passes (Alan) (eds). Op. cit. 252-267. ROSALDO (Michelle Z.). 1980. Knowledge and Passion. Ilongot notions of self and social life. Cambridge: Cambridge UP. — 1984. Toward an anthropology of self and feeling. In Shweder (Richard A.) & LeVine (Robert A.) (eds). Op. cit. 137-157. ROSENBERG (Daniel V.). 1990. Language in the Discourse of the Emotion. In Abu-Lughod (Lila) & Lutz (Catherine A.). Op. cit. 162-185. ROSENGREN (Dan). 2000. The Delicacy of Community. On kisagantsi in Matsigenka narrative discourse. In Overing (Joanna) & Passes (Alan) (eds). Op. cit. 221-234. SANTOS-GRANERO (Fernando). 1991. The Power of Love. The moral use of knowledge amongst the Amuesha of Central Peru. London: The Athlone Press. SCHEFFLER (H.W.) & LOUNSBURY (F.G.). 1971. A Study in Structural Semantics: The Siriono kinship system. Englewood Cliffs (N.J.): Prentice-Hall. SCHNEIDER (David M.). 1984. A Critique of the Study of Kinship. Ann Arbor: University of Michigan Press.

63

SEEGER (Anthony), VIVEIROS DE CASTRO (Eduardo), DA MATTA (Roberto). 1979. A construção de pessoa nas sociedades indigenas Brasileiras. Boletim do Museu Nacional 32: 2-19. SEEGER (Anthony). 1981. Nature and Society in Central Brazil. The Suya Indians of Mato Grosso. Cambridge (Mass.): Harvard University Press. SHAVER (Phillip R.), WU (Shelley) & SCHWARZ (Judith C.). 1992. Cross-Cultural Similarities and Differences in Emotion and its Representation: A prototype approach. Review of Personality and Social Psychology 13: 175-212. SHWEDER (Richard A.) & LEVINE (Robert A.) (eds). 1984. Culture Theory: essays on mind, self, and emotion. Cambridge: Cambridge UP. SOLOMON (Robert C.). 1984. Getting Angry. The Jamesian theory of emotion in Anthropology. In. Shweder (Richard A.) & LeVine (Robert A.) (eds). Op. cit. 238-254. STRATHERN (Marilyn). 1988. The Gender of the Gift. Problems with Women and Problems with Society in Melanesia. Berkeley & Los Angeles: University of California Press. SURRALLÈS (Alexandre). 2003. Au cœur du sens. Perception, affectivité, action chez les Candoshi. Paris : CNRS-MSH. TARLOW (Sarah). 2000. Emotion in Archaeology. Current Anthropology 41(5): 713-746. TAYLOR (Anne Christine). 2000. Le Sexe de la Proie. Représentations jivaro du lien de parenté. L’Homme 154-155: 309-334. — n.d. Sick of History. Contrasting regimes of historicity in the Upper Amazon. VILAÇA (Aparecida). 2002. Making Kin out of Others in Amazonia. Journal of the Royal Anthropological Institute (N.S.) 8: 347-365. VIVEIROS DE CASTRO (Eduardo). 1992. From the enemy’s point of view: humanity and divinity in an Amazonian society. Chicago: The University of Chicago Press. — 1996. Images of Nature and Society in Amazonian Ethnology. Annual Review of Anthropology 25: 179-200. — 1998. Cosmological Deixis and Amerindian Perspectivism. Journal of the Royal Anthropological Institute (N.S.) 4(3): 469-488. — 2001. GUT Feelings about Amazonia: Potential Affinity and the Construction of Sociality. In Rival (Laura H.) & Whitehead (Neil L.) (eds) Beyond the Visible and the Material. The Amerindianization of society in the work of Peter Rivière. Oxford: Oxford UP. 19-43. WEISMANTEL (Mary). 1995. Making Kin: kinship theory and Zumbagua practice. American Ethnologist 22(4): 685-709. WHITE (Geoffrey M.). 1990. Moral discourse and the rhetoric of emotions. In Abu-Lughod (Lila) & Lutz (Catherine A.). Op. cit. 46-68. 64

WILLERSLEV (Rane). 2003. Not animal, not not animal: a study of hunting, imitation and personhood among the Upper Kolmya Yukaghirs of north-eastern Siberia. Senior Seminar (2nd May 2003). Department of Social Anthropology. University of Cambridge. WITTGENSTEIN (Ludwig). 1980. Remarks on the Philosophy of Psychology. Vol. II. Oxford: Basil Blackwell. WRIGHT MILLS (Cecil). 1940. Situated Actions and Vocabularies of Motive. American Sociological Review 5(6): 904-913.

65