From the Milk River: Spatial and Temporal Processes in Northwest Amazonia (Cambridge Studies in Social and Cultural Anthropology)

Cambridge Studies in Social Anthropology General Editor: Jack Goody

26 FROM THE MILK RIVER

For a full list of titles in the series, see p. 303

From the MlkRiver\ Spatial and temporalprocesses in Northwest Amazonia CHRISTINE HUGH-JONES

The right of the University of Cambridge to print and sell all manner of books was granted by Henry VIII in 1534. The University has printed and published continuously since 1584.

CAMBRIDGE UNIVERSITY PRESS, CAMBRIDGE NEW YORK NEW ROCHELLE MELBOURNE SYDNEY

CAMBRIDGE UNIVERSITY PRESS Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, Sao Paulo Cambridge University Press The Edinburgh Building, Cambridge CB2 8RU, UK Published in the United States of America by Cambridge University Press, New York www. c ambridge. org Information on this title: www.cambridge.org/9780521225441 © Cambridge University Press 1979 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 1979 First paperback edition 1988 Re-issued in this digitally printed version 2007 A catalogue record for this publication is available from the British Library Library of Congress Cataloguing in Publication data Hugh-Jones, Christine, 1943From the Milk River. (Cambridge studies in social anthropology; 26) Bibliography: p. Includes index. Based on the author's thesis, Cambridge University, 1977. 1. Barasana Indians. 2. Tucano Indians. 3. Macii Indians (Papury River watershed) I. Title. F2270.2.B27H82 980'.004'98 78-73126 ISBN 978-0-521-22544-1 hardback ISBN 978-0-521-35889-7 paperback

FOR LEO AND TOM

But I was so much older then, I'm younger than that now. BOB DYLAN

CONTENTS

List of figures, tables and maps List of myths Preface A cknowledgemen ts Orthography

page

x xii xiii xviii XX

1

Introduction Focus of the study Physical setting History of white influence Changes in traditional life The unit of study

1 1 3 5 9 11

2

Social structure Introduction The units Tukanoans and Maku My use of technical terms for Tukanoan units A note on exogamy and language The model Exogamous Groups The phratry The sib The local descent group The model applied Territory Names for groups The internal organisation ofExogamous Groups Functions of social-structural units Ideology of descent Origin of sib s The longhouse and its inhabitants Composition of the longhouse group The longhouse setting The longhouse

13 13 14 14 15 17 18 19 21 22 22 22 25 26 26 30 33 38 40 40 43 45

Contents The longhouse interior Social and economic organisation of the longhouse community

46 49

3

The set of specialist roles Erosion of the system The specialist roles Politico-economic domain: chiefs-servants Metaphysical domain: dancer/chanters—shamans Domain of competitive intergroup relations: warriors Analysis Analogy with life stages Analogy with external relations

54 54 56 57 60 63 64 65 69

4

Kinship and marriage

76 76 77 83 83 84

5

Introduction The O-generation categories Marriage rules and preferences Negative rules Positive rules Analysis of marriage rules and preferences in relation to specialist roles Marriage practice Obtaining wives Wife-getting methods as a function of social distance General considerations Continuous and symmetrical organisation of the models General value of the models

87 93 93 97 100 100 102

The life-cycle Introduction The end of life Life and death Events after death The elements of the individual separated at death The beginning of life Ante-natal development Post-natal development Birth Naming Menstruation The practice The nature of the menstruating woman Menstruation in myth Menstruation in the life-cycle Male initiation {He wi) The practice The He wi cycle Changes that take place during He wi

107 107 107 107 109 112 114 115 117 123 133 134 134 136 137 139 142 142 145 147

Contents Metaphors of change: contact with the ancestral world Metaphors of change: rebirth Metaphors of change: change through paint The sacred instruments and female rites The natural and social Summary of the life-cycle Perpetuation of Pira-parana society

148 149 149 152 15 5 159 161

6

Production and consumption Introduction The sexual division of labour Manioc Production Analysis of the manioc process Meat: analysis of production Structuring time by production and consumption Secular production and consumption Production and consumption on ritual occasions Reintegration through food consumption Interpretation of foods Starch Cooked meat Pepper pot Cultivated plants and social models

169 169 170 174 174 180 192 200 200 204 213 217 111 222 224 226

7

Concepts of space—time Introduction Horizontal space—time Description of horizontal systems Synthesis of horizontal systems Vertical space—time Synthesis of horizontal and vertical space—time Horizontal to vertical Vertical to horizontal

235 235 238 238 251 257 266 266 269

8

Conclusion

275

Appendix 1: Named groups Appendix 2: Kinship terminology

282 287

Works cited

291

Index

293

FIGURES, TABLES AND MAPS

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

Figures Units of social structure Anaconda journeys Tripartite classification of Exogamous Group ancestors compared with marriage relations between Exogamous Groups The longhouse setting Ground-plan of longhouse interior Relation between concentric and hierarchical arrangements of specialist roles Concentric organisation of specialist roles and longhouse structure Functioning of internal and external aspects of specialist roles O-generation kinship terms Simplified kinship terminology for male ego O-generation kinship relations Symmetrical and continuous organisation of specialist roles and O-generation kinship categories Transference of shamanic activity to body of client Aspects of child development Metaphors of birth Life-cycle of the body Elements of the soul Loss of Romu Kumu's spiritual power Continuity of female generations Differentiation of ritual according to presence of//e; presence of women; patterns of paint on body Body paint in He wi cycle Aspects of alternation of generations Manioc processing: cassava, juice and hiari Processing and products of bitter manioc Cycle diagram of cassava production Alternative interpretations of manioc-separation process Comparison of meat production and reproduction Reverse cycle made in the myth of No-Anus Spirit Coca processing

23 34 37 44 47 56 70 74 78 79 82 103 121 124 127 130 135 139 141 151 153 163 175 176-7 181 191 193 199 202

Figures, tables and maps 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46

Structuring time by food and drug consumption Production and consumption of protein, manioc and coca Sequence of food shamanism following He wi Transformation of female body liquids Old house site planted with pepper and tobacco Comparison of movable and immovable systems of space-time The earth's surface Alternative models of river system of earth Comparison of typical longhouse setting with conceptual longhouse setting Models of house as body and womb Horizontal models of universe as womb Models of horizontal space-time Comparison of linear and concentric orders of horizontal space-time Alternative models of vertical structure of cosmos Relationship between vertical planes of house, longhouse setting and universe Relationship between vertical and east-west horizontal axes of universe Incorporation of Underworld into longhouse setting

2

Tables Comparison of authors' use of terms for Tukanoan socialstructural units Composition of longhouse groups

1 2

Maps The Vaupes region The Pira-parana and surrounding areas

1

205 211 214 225 229 238 240 242 245 247 250 252 255 259 265 268 273

16 42

4 7

LIST OF MYTHS

Note *starred items are reproduced in fuller form in Stephen Hugh-Jones 1979, part V. *Manioc-stick Anaconda and the origin of exogamous marriage Live Woman in the Underworld *Excerpts from myths about Romi Kumu (1) Romi Kumvi's life in the sky (2) Romi Kumu's immortality (3) Romi Kumu steals the sacred Yurupary instruments (4) Fire is stolen from Romi Kumu (5) Warimi steals poison from Romi Kumu's father Frog Wife T h e origin of manioc *The original planting of manioc Dragonfly's daughters *Yeba's penis The poisoning of White Spirit * Yeba in the Vulture's land No-Anus Spirit *Origin of cultivation of coca *Manioc-stick Anaconda and Macaw

xu

88 110 137 137 137 137 137 137 166 182 183 185 185 187 188 197 212 261

PREFACE

The field research for this book was carried out in Colombia between September 1968 and December 1970. Twenty-two months of this time was spent in the field. I took part in a joint project in which Stephen Hugh-Jones and I were to study a group of Tukanoan Indians and Peter Silverwood-Cope was to study a group of semi-nomadic Makii. By careful choice of field location, we hoped to report on each side of the symbiotic relationship between specific groups of Tukanoans and Makii. However, as is the way with fieldwork projects, our plans had to be modified as soon as we had made our first exploratory trip down the Pira-parana. We had chosen this river because most of the Tukanoan population were still living in traditional longhouses, but it was not until we got there that we learnt that it was barely ever visited by Makii and that there were no ongoing Makii—Tukanoan exchanges. Peter Silverwood-Cope left to study the Makii on the Makii-parana, a tributary of the Papuri (see map 1 below and Silverwood-Cope 1972). Although we could not follow our original plan, various ideas that the Pira-parana Indians hold about the Makii are presented here. Throughout our stay in the Pira-parana, we had to weigh up the advantages of making close ties with a single community against the disadvantages of having little comparative data and relying on a mere handful of adult informants. We decided in favour of close ties with a single community. This was partly to avoid living through the difficulties of establishing our position as participant observers more than once. We found that by far the most satisfactory and congenial way of doing our research was to live in a communal longhouse, partake in communal meals and help in productive activities (in as much as our fumbling efforts counted as 'help'). In fact, this seemed the only way, as we had a long period of language learning ahead and we also had to xiii

Preface dissociate ourselves from certain aspects of 'the white role' if we were going to understand many facets of Indian culture. Our success in establishing our position within a longhouse depended upon finding a way in which we could reciprocate our hosts' hospitality in a generalised way, rather than exchanging our gifts item for item. Solutions to these problems developed over time and, in spite of never being resolved to everyone's satisfaction, the question of gifts became less nerve-racking over the months. While we remained in a community of Barasana (MeniMasa sib, see appendix I for lists of sibs) located on Cano Colorado (see map 2 below) for most of the time, we also accompanied our hosts on many short visits and spent longer periods in several other communities. Together, we made extended visits to Makuna (Saira) on Cano Komeyaka, Barasana (Kome Masa) on the central Pira-parana and two communities of Bara (Munganyara) on upper Cano Colorado. I also visited a Tatuyo (Jiamoa) community in the Pira-parana headwaters. The Vaupes economy is marked by a strict sexual division of labour. This meant that we spent most of each day apart, each engaged in activities in which it would have been impossible for the other to participate. At times I found the female role irksome and depressing. Quite apart from such indignities as having to eat breakfast after my husband and having to sit out on most formal occasions, I suffered from the conviction that everything important was going on in the men's world and that I was not learning the exciting things about Piraparana society: I was a few years too early to have been armed with a 'raised consciousness'. However, there was no choice but to stay in a deserted longhouse for most of the day or to accompany my companions in their repetitive round of manioc work. Progress seemed slow as I gardened, peeled, grated and sieved. Much of the time the work was too hot and tiring for conversation and, when it was not, the women were often conversing in several different languages. I must have been a dreadful liability, but my friends put up with my technical inefficiency, rude interruptions and foolish questions and, along the way, I learnt a great deal about the domestic round. It is unlikely that I could have learnt the same things in any other way, or even that I would have made the necessary effort to learn them, because I did not realise at the time that they would be an important part of my analysis. In fact, I came to enjoy much of the daily routine for its own sake, but the process of writing up my field material has lead me to appreciate its theoretical significance too. A large part of

Preface the analysis contained in this book is concerned with the structure of secular life and its relation to the other structured domains, such as kinship, myth and ritual, which are usually given preferential treatment in interpretive monographs. I hope this book will be considered a contribution to the ethnography of Vaupes Indians, but this is not its primary object. My aim is to present Pira-parana society as an integrated system. The analysis has grown out of my own attempt to make sense of the data in relation to each other and therefore, although I have obviously been influenced by some types of analysis more than others, in no way is this an exercise in a particular style of interpretation. Originally, I began to write about social structure in the conventional, limited sense of the word. I wished to give an account of kinship groups, the operation of marriage rules, the discrepancy between ideal models and practical behaviour and so on. I abandoned this project because it seemed to me that social structure was not a legitimate isolate and that the most interesting aspects of it could not be understood from within. However, in a sense, a concern with social structure is still basic to this work because much of the analysis is a response to the question of how to present an 'open-ended' society as a system. It is not simply that the boundaries of 'the society' are indistinct or subject to fluctuation; Pira-parana society is part of a wider complex whose distinctive feature is the lack of bounded groups. Instead, there are a great many exogamous patrilineal groups connected by marriage ties in an open-ended network. We are all familiar with segmentary models but these presuppose a primary, all-embracing unit. Once the rule of exogamy is attached to the highest-level units, we are obliged to recognise that a very different type of system exists. Even the well-tried, but misleading, feature of common language must be discarded as a defining feature of 'the society' (or 'a society') because, in this case, languages are attached to exogamous groups. I believe that it is possible to represent a social system without recourse to the notion of 'a society', but it is the analysis of concepts of space and time, of lifecycle development and of other phenomena outside the realm of kinship and marriage, that have convinced me of this. The nature of my enterprise, in attempting to show the interrelations between different aspects of Indian life, has certain implications for the form of this book. It is difficult to represent the interconnectedness of the diverse material in linear narrative form and, therefore, to help the reader there are both a large number of diagrams

Preface and a large number of cross-references. The style of writing is necessarily rather condensed because a wide range of material has had to be fitted into a reasonable space. It is easy to miss essential pieces of information, but the chapter sub-headings are listed in the contents to assist in their retrieval. Throughout, I have made extensive use of myth but, for reasons of length, I have included only much reduced versions and short episodes of myth. I have included a list of the myths given whole or in part in the text, and indicated which of these are reproduced at greater length in Stephen Hugh-Jones (1979). Although more complete versions of the myths to which I have referred would be interesting for their own sakes, they would not be directly relevant to the analysis here. In order to stress the relative autonomy of this work, I have not peppered the text with references to standard theoretical works. Nor, with a few exceptions, have I incorporated other people's data into this book, either as supporting evidence or as comparative material. The data from which I work are very similar to those collected by other anthropologists from groups elsewhere in the Vaupes and, of course, from the Pira-parana itself, but the analysis I attempt is different from other extant works in either type of theoretical approach (broadly speaking, mine is a modified structuralist one) or scope, or both. Also, as suggested above, the notion of a discrete society (or societies) is completely inappropriate throughout the Vaupes: it therefore seems theoretically desirable to preserve the point of view of a small set of communities living on Cano Colorado as far as possible. This reflects the Indian perspective, for each longhouse community is the centre of its own world. Secondly, and following from this, the Vaupes and surrounding areas are a rich field for comparative study but, if material from different groups is fused together, the basis for comparison is destroyed. This work should be seen as a stage prior to comparison — an attempt to understand a single perspective in depth in order to see what can be meaningfully compared. Thirdly, this book is long enough as it is, without additional ethnographic examples. I have, however, made considerable use of my husband's analysis of Barasana initiation ritual (S. Hugh-Jones 1979) which is also based on our joint fieldwork. In some respects my analysis starts where his left off. I have provided the 'general ethnography' within which his work on initial ritual belongs but, in this case, instead of the particular analysis following the general ethnography, it has been the other

Preface way round. Our separate books are intended to be complementary rather than overlapping. Cambridge August 1978

C. H.-J.

ACKNOWLEDGEMENTS

My sincere thanks are due to very many people whose generosity has made this book possible. Of these I can only mention a few, but anyone who has undertaken the same kind of task will know that there are countless more. The research on which this book is based was financed by the British Social Science Research Council. Their grant also lasted through part of the writing-up period, and in addition I received help from Darwin College and the Board of Graduate Studies, University of Cambridge, with the preparation of the manuscript. I am very grateful to all these institutions. Professor Sir Edmund Leach supervised my doctoral thesis and it is due to the understanding, advice and unfailing support that he offered over many years that this book has reached its present form. My warmest thanks are for the people of Bosco's longhouse on Cano Colorado who allowed us to share their life, their food and their knowledge. In particular, I thank Paulina who generously took care of me throughout my stay. The people of Umero's house, Maximilliano's house, Ignacio's house and many more were welcoming hosts and warm friends. Together, the people of the Pira-parana showed us a way of life for which I feel lasting admiration and great nostalgia. I apologise to them for any errors I have made. In Colombia we received invaluable help and hospitality from Professor G. Reichel-Dolmatoff, Dr F. Marquez Yafiez and others of the Instituto Colombiano de Antropologia; from Monsenor Belarmino Correa, Padre Manuel Elorza and others from the Prefectura Apostolica del Vaupes; and from Joel and Nancy Stolte, Richard and Connie Smith, David and Jan Whistler and several skillful pilots — all members of the Summer Institute of Linguistics. Dr Fred Medem, the Bright family, the Bahamon family, Rosnelle Baud and countless xviii

A cknowledgemen ts other people made our stay a great pleasure and gave us many different kinds of help. Both during and since fieldwork, I have benefited from the field notes, advice, information, encouragement and friendship given by anthropologists with experience among South American Indians. Above all, I have appreciated the open and generous spirit with which these have been offered. Among these friends are Bernard Arcand, Kaj Arhem, Patrice Bidou, Irving Goldman, Paul Henley, Jean Jackson, Pierre-Yves Jacopin, Tom Langdon, Howard Reid, Peter Riviere and Peter Silverwood-Cope, but there are many others. I must offer very special thanks to Terry Turner for reading drafts of this work in different stages of elaboration, and providing inspiration through his extensive and illuminating comments. It is impossible for me to do justice to the extent of his help in the text, but let me say now that I have made use of his rare gift of creative criticism throughout. Many people here in Cambridge have treated me with warm sympathy. In particular, my children, Leo and Tom, have put up with a great deal of interference from their paper sibling. Finally, there is no way I can thank Stephen, my husband, enough for his part in this book. I have made extensive use of his field notes, his insights into Barasana culture, his time, his energy and his remarkable domestic skills. The bulk of this book was written after he had completed his own analysis of Barasana ritual and so I have had all the considerable advantages that this entails. To say that I could never have produced this work without him is certainly no cliche.

ORTHOGRAPHY

The Barasana orthography used in this book follows that developed by Richard Smith (n.d.) of the Summer Institute of Linguistics. This orthography uses symbols chosen to conform to that of Colombian Spanish. For English readers I have substituted the symbols 4h' and 'ny' for f and 'n'; I have also not used the symbol 'q\ as it has the same value as 'k' which I use instead of 'c\ Vowels Un-nasalised Nasalised a as in m#sk a e as in egg e i as in mk I o as in orange 6 u as in scoop u & similar to the German li £ Consonants b similar to buy but with prenasalisation (mb) k as in kite d prenasalised as in and g as in go but with prenasalisation (ng) h as in house m as in man (phonologically a variant of b, conditioned by a contiguous nasalised vowel) n as in nose (phonologically a variant of d, conditioned by a contiguous nasalised vowel) ng as in tongue (phonologically a variant of g, conditioned by a contiguous nasalised vowel) ny as in Spanish ma/iana (phonologically a variant of y, conditioned by a contiguous nasalised vowel) p as in pen r between r and 1 in English s similar to English ts as in boata t as in rime w as in wine y as in .yam xx

1 Introduction

Focus of the study Pira-parana Indians see themselves as existing within an ordered cosmos created in the ancestral past. The world of their present-day experience is a residue or product of the ancestral doings related in myth, ritual chants and shamanic spells. From their own point of view, this cosmos and the mythical deeds associated with it control their contemporary social life and provide a moral framework for present-day action. Here, I work the other way round — from the inside out. Instead of starting with the cosmology, I start with the building of basic units of social structure, families and patrilineal groups, through marriage and procreation. I begin by showing how different phases of that temporal processes are associated with different spaces within and around the longhouse and end by showing that the very same 'space-time' principles underlie the structure of the cosmos. I argue as if the basic principles of social structure were primary and the cosmos was a reflection of these but I do not mean to imply that there is a simple relation of cause and effect between the two or that they should be seen as 'infrastructure' and 'superstructure'. The anthropologist must regard the ancestral cosmos as an imaginary projection of present experience, but at the same time it is a projection which both controls present experience and forms an integral part of it. There is therefore a sense in which each world - the ancestral one and the present-day secular one — regulates the other. This interdependence is reflected throughout my study, because, although I work outwards towards the ancestral cosmos, I am forced to refer to my goal throughout. Without recourse to myth and exegesis I would be quite unable to construct any model of the present-day situation, because I would be unable to 'see' it. For instance, it is quite 1

Introduction impossible to understand the simplest fact about Pira-parana social structure without first understanding 'imaginary' descent from the founding anaconda ancestors. The theme of the book is more complex than the isolation of an inner structural level and an outer cosmological one would suggest, for I show that there is a compatibility and coherence between the various domains of experience which typically form the chapter headings of the social anthropologist's monograph. I demonstrate that 'social structure', 'kinship and marriage', 'the life-cycle', 'polities', 'economics' and 'religion' are ideologically integrated just as they are also inextricably bound together in concrete behaviour. In this sense the book is an attempt to overcome the distortions and limitations resulting from the establishment of those anthropological subdisciplines which separate off categories of data in a manner reflecting anthropologists' professional interests rather than the nature of the societies they study. In the case of the Vaupes Indians, there is special reason to avoid such a rigid classification of data, because questions about almost any aspect of life are answered by the telling of mythical episodes, and in myth there is no convenient separation into discrete institutions. People actually spend their time living, taking decisions, experiencing biological changes and promoting changes through their own activity. Even though anthropologists may choose to regard them as 'actors' who are continually 'expressing social relationships', this does not seem to be the way they regard themselves. At the same time, there is obviously a sense in which their culture and its institutions may be maintained in spite of the changes that are occurring on all fronts. I therefore describe Pira-parana society in terms of dynamic processes which take the form of repeated cycles. Even the systems of classification which seem to provide a fixed, 'static' framework for positive action can be conceived of in terms of dynamic processes. Thus the structure of descent groups, of the cosmos, of the life-cycle, of a meal and so on are all created by movements of people, ancestral beings or other elements through space and time. It is because these structures are created in this way that they possess the power to engender further creative change — this, at least, is the way in which I believe Indians see things, and it is also instructive for the model-building anthropologist. The book contains a series of discussions of these separate processes, all of which contribute to the composite phenomenon of 2

Physical setting social reproduction - by which I simply mean the continuity of a population with a recognisable culture and set of social institutions. In the following chapter, I set out a model of Pira-parana descentgroup structure and discuss its relation to the data from which it is drawn. Then I describe the process of reproduction of the local community through formation and dispersal of family units. In chapter 3,1 take the specialist-role system associated with the internal classification of descent groups and show that its structure is related to the formation and dispersal of family units and also to the range of different types of communication with the outside groups that are potential sources of wives. In chapter 4,1 describe certain aspects of kinship and marriage, concentrating on the role of marriage in creating close relations between opposed exogamous groups, which thus promotes changes in the network of intergroup relations over time. In chapter 5,1 describe the life-cycle rituals in order to elucidate Indian concepts of the life process with its interwoven physical and social aspects. Chapter 6 is devoted to processes of production and consumption, particularly of food and 'drugs'. I show how these are integrated with the ritually elaborated life processes outlined in the previous chapter. Finally, in chapter 7,1 discuss the concepts of space-time which unite all these different processes, and develop a general model of space-time systems which accounts for the relationship between the world of present-day Indian experience and the ancestral past. Physical setting The Pira-parana and its tributaries lie between 70 and 71 degrees W. and the main stream crosses the Equator roughly halfway along its length (see map I). This part of the Northwest Amazon area is just inside the Colombian border with Brazil and lies in the administrative district of the Comisaria del Vaupes. Geologically the area is the southern fringe of the Guiana shield. Most of the land lies around 700 feet, with isolated hills and mountains standing above the surrounding forest. Apart from a few open, sandy areas with tough and sparse vegetation and swamps with miriti palms (Mauritia flexuosa), there is a general forest cover. From the air this is dense green with an occasional flowering treetop; from the river it is an impenetrable and richly varied tangle of trees and creepers, but once in the forest, shaded by the canopy far above, it is 3

Map 1 The Vaupes region

History of white influence damp and sombre and surprisingly open. It is only on the river, where a huge tree has fallen, or in a man-made clearing, that the colours are light and bright and it is possible to feel a sense of space and distance. The average rainfall for the Vaupes is around 3500 mm but there is great local and annual variation (Instituto Geograflco Agustin Codazzi 1969). There are two rainy seasons and two dry ones: a long dry season lasts from December to March; heavy rains fall until August or September and then a short dry season is followed by more rains. The temperature varies between 20 and 35 °C throughout most of the year, except during the am or friagem, a cold spell characterised by fine drizzle, when it may drop to as low as 10 °C. Even in the dry seasons there is frequent rain, but the relative lack of it can be seen in the height of the rivers. At the wettest times of year the rivers flood the adjacent forest and all the lower ground becomes waterlogged; at the driest times, sandy banks and the rocks and dead branches lodged in the river beds are exposed. The people of this area are scattered in small longhouse communities situated by rivers and separated by anything from half-an-hour's to a day's journey. They prefer to travel by river, even if it means a longer journey, but there are also many paths in use, particularly in headwater areas. The Pira-parana and some of its tributaries are full of rapids and waterfalls which are rightly considered treacherous by both Indians and white travellers. The relative isolation of the Pira-parana from the encroaching white society and culture can be mainly attributed to the problems of travel. The establishment of mission airstrips has naturally been a major factor in opening up the area. Because the Pira-parana is an affluent of the Apaporis to the south, but is also accessible from the Vaupes system, its inhabitants have been influenced by whites from both directions. History of white influence This book is partly a salvage operation: because of the nature of the analysis, I emphasise the traditional aspects of Indian life at the expense of the features resulting from deculturation. Although I do incorporate processes of change into my analysis, these are changes within a traditional institutional framework and not changes of that framework itself. However, I do not want to give any false impressions about Pira-parana people as I knew them in 1970, and the following 5

Introduction account of their recent history is intended to set the record straight. When I use the 'ethnographic present', I refer to the period of my field work — no doubt things are different now. The first mention of the Pira-parana dates from the mid-eighteenth century (Briizzi 1962 : 22). However, it was not until 1965 that the first mission outpost was established by North American Protestant missionaries belonging to the Summer Institute of Linguistics. This was closely followed by other SIL posts, so that in 1970 the Piraparana had four, corresponding to the four main languages spoken along its course. The Catholic mission was established in 1968. The sudden influx of missionaries was accompanied by the appearance of many anthropologists, adventurers and students of this and that so that, rubber gatherers aside, Indians saw more non-Indian strangers in several years than they had done in as many centuries. However, the shock of change was not quite as great as this sudden influx would imply, because for many decades the Pira-parana had been surrounded on all sides by peoples whose ways of life had been more severely interfered with by missionaries and rubber gatherers than had their own. As an index of this, by the time of our fieldwork, the Piraparana, adjacent parts of the upper Tiquie, a few isolated areas on the Papuri system (such as Cano Inambii where Jean Jackson did her research) and, to the south, parts of the Apaporis system, were the only areas where Tukanoan groups were still living in longhouses. In most cases these were not built to the traditional size but they did, at least, conform to the traditional structure. In the Pira-parana area, the influence of the rubber trade was first felt in the latter half of the nineteenth century and the industry reached its height in the first decade of this century. After that it went into a gradual decline until the Second World War brought a sudden increase in demand. By 1970 there were rumours that the market was approaching total collapse and that government support for the industry was to be withdrawn. The rubber gatherers (caucheros) entered the Pira-parana from the Vaupes and Apaporis and, in most cases, carried off the able men by force, often killing others and raping women. Many Indians tell how the reprisals against the brutal intruders were also extended to fellow Indians who had directed them to hidden longhouses. The location of longhouses was an indication of white activity, because Indians fled to the headwaters and concealed approaches to their longhouses when times were bad, and then moved back to the larger rivers as white people withdrew. 6

Map 2 The Pira-parana and surrounding areas

Introduction Besides reducing the population and disrupting social life in these ways, the rubber trade introduced new diseases on a large scale. It also changed Indian culture and aspirations by creating new needs for white men's merchandise. But the caucheros did not purposefully attempt to change Indian culture from within by altering patterns of social and domestic life. Their very lack of interest in Indians as anything beyond a workforce to be ensnared in an eternal credit system meant that those who escaped their recruiting drives continued to live in a traditional way. By the period of our fieldwork, the methods of caucheros were less violent, but nevertheless very variable. Many Indians born in the upper Pira-parana were permanently living in rubber camps on the Vaupes or beyond, while others in the lower Pira-parana worked regularly on the Apaporis for a part of each year. The missionaries played a complementary role to the caucheros: proclaiming themselves to be against the economic exploitation of Indians, they set out to convert and 'civilise'. The first missionaries to enter the Colombian Vaupes area were Montfortians, who settled along the Papuri following an exploratory voyage in 1914. They destroyed longhouses and burnt ritual goods, and forced the inhabitants to build wattle-and-daub villages, each with its own church. The manner in which they waged their war against sin softened over the years, especially with the replacement of the Montfortians by Colombian Javerians in 1949, but the aims remained basically the same until around 1970 when there came a wave of radical questioning of missionary policy from within. As the methods of persuasion became more gentle, the bond between Indian and missionary strengthened in economic content. The labour required to build airstrips and to feed and maintain mission headquarters and boarding schools was bought with manufactured goods. As these became indispensable, Indians found themselves dependent on mission centres and so the inflationary spiral of demand for manufactured goods began. Such demands are met by working for the missions and this both accelerates the acquisition of mission values, and increases mission power by augmenting capital resources such as airstrips, buildings and so on. In this way, the act of earning manufactured goods strengthens the acculturative force and so creates the desire for more goods. Many Pira-parana Indians had visited mission centres outside the Pira-parana area long before missionaries visited them and all were familiar with the elements of Christianity. The Vaupes marriage system, which requires each man to find a wife from beyond the area

Changes in traditional life occupied by his own linguistic group, creates a network of alliances which facilitates the spread of missionary culture well beyond the limits of missionary activity. However, at the time of my fieldwork, the disparity between longhouse life in the Pira-parana and missionary village life in the surrounding areas seemed to have dampened the exchange of women across the divide, so that Pira-parana society taken as a whole was relatively endogamous and culturally distinct (in terms of existing culture rather than traditional culture) from the rest of the Vaupes. The influence of the four North American SIL teams working on the Pira-parana during the time of our fieldwork was almost purely economic. Their policy was to sell merchandise in exchange for labour, food and artifacts and to offer limited medical aid. Meanwhile they would concentrate on a few suitable candidates for conversion, but by far the greater part of their energy was devoted to language learning. Each team (a married couple in all cases on the Piraparand) would spend stretches in a small house built alongside a longhouse. These were interspersed with periods at a large residential base, outside the Vaupes area, where work among Colombian Indians was administered and co-ordinated. The second stage of true evangelising and education had to await New Testament translations and had hardly got under way in the oldest-established Pira-parana post before we left. Changes in traditional life As a result of white activity, Pira-parana Indians had steel axes, machetes and knives, aluminium cooking pots, some shotguns, fishing nylon and hooks, cotton hammocks, clothing, matches, beads and so on. They still used plenty of ceramic pots and some string hammocks and blowpipes, all of which they made themselves. Women wore home-made skirts or full dresses made from manufactured cotton textiles, while men were sometimes clad in G-strings alone and sometimes in ready-made trousers and shirts. Stone axes were never used, but the older men could remember when there were only very few steel axes which were circulated from house to house for felling cultivation sites. These white goods must have produced considerable changes in the Indian economy. The time required to fell trees, build houses, collect firewood and peel manioc is so much reduced by the use of steel that there can be no doubt that potential productive capacity 9

Introduction has increased. We know from descriptions of travels in the past that longhouse size has actually decreased, but it is reasonable to assume that cultivation size has increased. In fact it is impossible to imagine otherwise, because even now, with the help of steel, the annual felling lasts several weeks and the women's daily round of cultivating and processing manioc and gathering firewood absorbs very nearly all the daylight hours. It would be impossible to achieve the same output with instruments of wood, stone, tooth and bone. The older women describe a number of edible, wild starchy roots which are never collected now; this also suggests that manioc production used to be on a smaller scale. The use of shotguns and fish-hooks has changed the entire style of hunting and fishing, particularly in its social aspects. These are now mostly individual pursuits: the individual puts in a great deal of time searching for game or waiting for fish to bite and brings in a relatively small catch at the end of the day. In the past much more time was spent in preparation for capturing fish and game: traps, blowpipe darts, arrows and other devices were made with considerable expense of time and effort — often communal effort. My impression is that the result of these indigenous methods was larger but more infrequent catches, corresponding to the larger scale of each enterprise. Shotguns and fish-hooks and line do not lend themselves to group expeditions, and they may well have been additional factors in the breakdown of large longhouse groups in many regions. Lastly, there is no doubt in Indian minds that shotguns are to blame for the present poverty of game in most of the Pira-parana area. Hand in hand with the economic changes has come the decline in intergroup hostilities. This is one of the most difficult of the processes of change to assess because there was evidently an intermediary phase when much of the killing among Indians was directly related to white activity. Indians give two reasons for the cessation of hostilities among themselves: one is the general upheaval, dispersal of groups and terror caused by the arrival of whites; the other is the order to stop killing issued by the priests. I think it is safe to assume that whatever the direct or indirect effects of the missionaries on fighting, the decline was under way long before they arrived. It is likely that the scale of fighting decreased even before the first caucheros arrived in the Pira-parana, because many traders must have already used this known thoroughfare linking the Vaupes and Apaporis. 10

The unit of study There were Indians alive in 1970 who had been involved in isolated killings but not in the full-scale raids made by their forbears. Evidence that the Pira-parana groups were once warlike comes from their own stories as well as from the content and explicit purpose of ritual which is designed to make men fierce. Broadly speaking, most people know and agree about famous intergroup encounters and can plot many of the consequent migrations. Besides the fighting and shifting alliances which occurred between resident Tukanoan groups, Indians often mention a time when cannibal tribes swept through the Pira-parana; they even point out a cave with smoke-blackened walls where they claim their ancestors lived in hiding. The cannibals are commonly said to have been Barea gawa, probably Bare Indians from the northeast, close to the Casiquiare canal, and Hode gawa, Carijona, from the west. The unit of study In this book the categories 'Vaupes Indians', 'Pira-parana Indians' and 'Barasana' are used. Whenever Indian words are included, these are in the Barasana language unless otherwise stated. The rather confusing use of terms requires some explanation. The river Pira-parana and its affluents are inhabited by a number of separate peoples, including the Barasana. The area covered by the Pira-parana river system forms the southwestern part of the Vaupes region, which overlaps from Colombia into Brazil. The other rivers of the Vaupes region are part of the Vaupes river system, whereas the Pira-parana flows into the Apaporis to the south (see map 1). The Tukanoan groups of the Vaupes region are related by common cultural patterns, by their use of Eastern Tukanoan languages and by their participation in an open-ended social system based on the exchange of women between exogamous patrilineal descent groups. These exogamous groups are the peoples such as the Barasana, Tatuyo, Desana etc. Generally speaking, each of these exogamous units has its own language, but, in the exceptional cases where language and descent boundaries do not coincide, there is enormous confusion in the use of these two criteria for distinguishing groups. The confusion is made worse by the use of group designations in Spanish, Portuguese, Lingua Geral and a great number of Tukanoan languages. The terms are applied by missionaries, linguists, anthropo11

Introduction logists, Indians and others — each having their own reasons for wishing to stress particular defining characteristics.1 'Barasana' is a case in point: it applies to both a language which is also spoken by the Taiwano (regarded as a descent category) and a descent category which also includes speakers of Makuna (regarded as a language category). However, I use 'Barasana' to refer to a descent category, except where it is quite clear that I mean a language. Barasana is a word of unknown origin, current among whites but not used as a self-name except in dealing with whites. Since the Barasana are obliged to marry out, they in no way constitute an autonomous social unit; nor do 'the people of the Piraparana', since these include members of descent groups which are mainly based on other rivers, as well as women who have married in from outside and sisters who will marry out of the area. However, all my data come from people living on the Pira-parana, and the population does at least cover several intermarrying exogamous units. Also, as mentioned above, there is a tendency towards endogamy in the Pira-parana area taken as a whole. In spite of all these reasons for regarding the Pira-parana as the unit of study, inhabitants of each geographical area and each local community consider themselves to be at the centre of a spreading system of intermarrying groups which theoretically has no geographical or social limits. This means that it is most important to recognise the incorporation of Pira-parana dwellers into the greater Vaupes system, and also their links with the culturally slightly different groups to the south, such as the Yukuna, Tanimuka, Letuama and Matapi. 1 The Summer Institute of Linguistics have made matters worse by calling the Bara language 'Northern Barasana' and the Barasana language 'Southern Barasana'. This is reputed to be because they regard 'Bard' as an unsuitable name: bara means 'love-charm' or 'aphrodisiac' in some Tukanoan languages.

12

2 Social structure

Introduction The character of Vaupes social structure is such that no model can come close to the 'facts' as revealed by field research. The anthropologist's social structure must be pieced together from a muddling mass of statements that Indians make about kinship connections, group names, ancestral derivations, linguistic affiliations, geographical sites and so on. These statements employ named descent groups, named languages and named geographical features, and the categories based on these different criteria do not necessarily coincide with one another. Besides, changing identity, conflicting ideas about group membership and group status, and also the fission and fusion of groups are all in the very nature of the traditional system. It may well be that those groups living in the Pira-parana region, on the edge of the Vaupes culture area, are less stable than those living on the River Vaupes itself or on its principal affluents. Of course, similar reports of uncertainty and flux in indigenous classification of groups are far from uncommon in anthropological literature, and it has often been possible to represent these potentialities for change by means of simple segmentary models of social structure. However here, as we shall see, this would not be a very satisfactory course because, at an ideal level, the internal organisation of the most important groups is fixed: there are divisions into five sub-units each of which possesses a specialist occupation. This set of specialist roles provides an explicit indigenous model which prohibits the proliferation and confusion of sub-units but, nevertheless, the changes still occur in practice, with the result that the existing situation differs from the model in many ways. I must emphasise that in this book the focus is on the properties 13

Social structure of the model. This means that a mere description of the constant features of social structure, with details of existing groups (as, for example, in Goldman 1963), is not enough. However, the question of the relation of my more formal model to contemporary practice cannot be ignored and will be treated in some detail. As the analysis progresses in the following chapters, it will become clear that, although the model assumes a fixed internal structure for major groups, this same fixed structure embodies dynamic features especially appropriate to a situation of changing intergroup relations. However, the dynamic aspects can only be appreciated by enquiring beyond the conventional limits of social structure and this cannot be done until the model has been set out. To make the presentation as clear as possible, the model is described first and the extent to which it is an accurate reflection of the social groupings which people recognise in practice is discussed afterwards. This treatment of the larger structural units is followed by an account of local longhouse communities and of how these are reproduced over time. Before proceeding to the model some preliminary points must be made about the relation of Pira-parana Indians to other Vaupes Indians; the use of technical terms for social-structural units; and the relation of these units to patterns of language affiliation. Let me say at the outset that I am well aware that much of the material in the remainder of this chapter is not easy to understand, but I do not believe it can be simplified without distorting the data. The units Tukanoans and Maku The Indians of the Vaupes fall into two major categories: Tukanoan and Maku. Tukanoan' is essentially a linguistic category referring to affiliation to one or another of a group of languages known as 'Eastern Tukanoan' and exclusively represented in the Vaupes area. Sometimes the Tukanoans are simply called 'Tukano' but, since this term is also the name of one particular Eastern Tukanoan language, I shall not use it in the more general sense here. The culture of all these Tukanoan groups is based on extensive cultivation of bitter manioc and the people live, or used to live, by rivers in large communal longhouses. Groups speaking Arawakan languages and living on the fringes 14

The units of the Tukanoan area are similar in culture to the Tukanoans rather than the Maku. Maku are differentiated from Tukanoans by both language and lifestyle. There are two distinct Maku languages which have so far proved difficult to classify in terms of major South American language groups. Maku are found in interfluvial areas where they lead a semi-nomadic life alternating between village settlements and hunting treks. They cultivate some of their own manioc but rely heavily on forest products, particularly game. They still maintain their traditional exchanges with riverine Tukanoan groups in which meat, curare (poison for blowpipe darts) and labour are given in exchange for cultivated products (Silverwood-Cope 1972, Jackson 1973b). It is not entirely clear to what extent Maku have been gradually adopting the life-style of Tukanoan groups over the last decades, for their present life-style is certainly closer to the Tukanoan one than earlier accounts suggest. Although Tukanoan groups of the upper and middle Pira-parana, from where my data come, had no relations with the Maku, their ideas about Maku culture are an integral part of their system of social classification. Where I mention Maku here, I refer to a conception of Maku held by certain Tukanoan groups. Many aspects of this characterisation bear little resemblance to the actual Maku or to the Makii's ideas about themselves, as described by Peter Silverwood-Cope (1972) for the Colombian Maku and Howard Reid (personal communication) for the Brazilian Makii. The discrepancy might not have been so great in the past, but it is likely that it always existed to some degree. My use of technical terms for Tukanoan units The Tukanoans are divided into patrilineal, exogamous descent groups with further internal divisions according to patrilineal descent. In the two serious attempts to describe Tukanoan social structure, there is only partial agreement about the technical terms for different levels of grouping (Goldman 1963, Jackson 1972). Goldman writes of 'tribe', 'phratry' and 'sib' among the Cubeo, while Jackson writes of 'phratry', 'language-aggregate' and 'sib' among the Bara. For simplicity of comparison I also use the terms 'sib' and 'phratry' in Jackson's sense, but I use two terms, 'Compound Exogamous Group' and 'Simple Exogamous Group', in place of her 'language-aggregate' as indicated in table 1. 15

Social structure Table 1 Comparison of authors* use of terms for Tukanoan socialstructural units Exogamous Units Goldman (Cubeo) Jackson (Bara and neighbours) C. Hugh Jones (Barasana and neighbours)

tribe*

-

phratry

sib

—

phratry

languageaggregate0

sib

-

phratry

Compound or Simple Exogamous Group

sib

indicates the author regards language as defining feature.

The application of the term 'phratry' to units of different order by these two authors reflects a significant difference in the social structure of the Cubeo and Bara for, although each is a language-bearing unit, the Bara are also a single exogamous unit while the Cubeo are composed of three intermarrying exogamous units. However, I believe that by concentrating on descent and exogamy rather than language in defining units of social structure, the two groups can be fitted into a basic pattern common to all Tukanoan peoples. Although I do not consider language to be a useful defining feature of groups at any particular structural level, I do not mean to underestimate its extreme importance nor to imply that differences between the Cubeo and Bara in this respect should be overlooked. I have simply tried to set out a model which emphasises the common ground rather than the differences between Tukanoan groups. The four types of grouping to which I refer are all patrilineal exogamous descent groups, and so I use the terms 'Compound Exogamous Group', 'Simple Exogamous Group' or just 'Exogamous Group', written thus with capital initial letters, to refer to specific Vaupes institutions. I make the rather awkward distinction between 'Compound Exogamous Group' and 'Simple Exogamous Group' to account for a type of internal structure among groups of sibs which has not been reported in the Cubeo or Bara ethnography (the internal structure of Exogamous Groups is the subject of chapter 3 ) This distinction will not be relevant in the following chapters, for I shall 16

The units just refer to these as 'Exogamous Groups' in order to avoid the use of the more cumbersome term. However, it is important to make the distinction here in order to show how my model society corresponds to the groups that are recognised in practice. A note on exogamy and language The unit I have called a Compound Exogamous Group is ideally a language-bearing unit; the same applies to a Simple Exogamous Group in cases where this exists on its own and is not organised into a Compound Exogamous Group with other like units. In practice the boundaries of units at this level are usually co-terminous, or almost so, with linguistic boundaries. This means that in the vast majority of Vaupes marriages, the partners have different first languages and there are sometimes four or more different language groups represented in a single longhouse. Men and women speak the language of their own descent group throughout their lives and another language is only used under the following circumstances: 1 — if the hearer does not understand the speaker's patrilineal language — a relatively rare event unless the two come from widely separated groups. - when imitating or reporting speech uttered in another language. — if the person has been brought up in a foreign patrilineal community of different language affiliation to his or her own. This is also rare because the child's agnates make every effort to prevent it. - if the speaker is an old woman who has lived her married life isolated from speakers of her own language. In this case she may make a partial or total switch to her husband's language but, on the other hand, provided her own language is generally understood, she may well continue to use it exclusively. From the basic rule that patrilineally inherited languages are spoken at all times, it follows that conversations between individuals speaking different languages are commonplace. Conversely, the case of husband and wife speaking the same language is regarded as anomalous. Indians accept that affinal groups sometimes share a single 1 This does not create any problems in mutual comprehension unless the parties come from widely separated geographical areas. Each person is brought up in a longhouse with members of several different language-bearing exogamous groups and is continuously exposed to all the languages represented in his or her field of social interaction. Since it is clear that children cannot be learning language from their mothers (who speak the languages of foreign descent groups), nor from adult descent-group men, as they spend most of their time apart from men, it follows that they must learn from the older, co-resident children.

17

Social structure language but they regard this as 'wrong' and try to account for the missing language in terms of battles and migrations. The Cubeo appear to be an exception here since there is no evidence that they regard language exogamy as the ideal state of affairs. The ideal coincidence of language unit and Exogamous Group boundaries leads Indians to use language as a way of talking about descent. They often indicate Exogamous Group membership by saying an individual 'says . . . ' repeating a stock phrase in that person's patrilineal language.2 Minute variations of pronunciation, vocabulary and even speaking style are used to support claims about the internal organisation among the sibs of an Exogamous Group. For example, I was assured by a member of a middle-ranking sib that the most senior sib of his group speak with an exaggerated, forceful delivery while the lowest-ranking group mumble and blur their words. Jackson's research among Papuri groups, which are nearer to the centre of the Tukanoan area than Pira-parana groups, shows that the ideal coincidence of linguistic and descent boundaries is almost perfectly matched in practice. The number of exceptions among Piraparana groups probably reflects their marginal position at the southwestern extreme of the Tukanoan culture area in the same way that the exceptional language situation among the Cubeo probably reflects their extreme northerly position. However, the empirical exceptions to the rule that language- and descent-group boundaries coincide are so numerous among groups which share the common features of descent-group exogamy and ranked sibs with the central Tukanoans, that I feel justified in discarding language as a distinctive feature of Exogamous Groups. I think this decision might, in the end, lead to a better understanding of the social role of multilingualism in the Vaupes because it allows language exogamy to be treated as a variable instead of an invariant feature of the Vaupes system (see Sorenson 1967, Jackson 1974). The model The clearest way of presenting my model of Vaupes social structure is to start with Exogamous Groups and, after discussing their internal organisation, to move on first to larger units and then to smaller units. This is because the Exogamous Group with its five-fold internal 2 The exemplary phrases most often used may be glossed as, 'that is the way it is', thus tobahiro bahiaha is the phrase chosen to indicate Barasana language-group membership.

18

The model structure is the most significant portion of the model, and the more inclusive phratric groupings are both ill-defined and awkward to understand without any knowledge of the nature of the constituent Exogamous Groups. The entire account is difficult to grasp without some preliminary population sizes but, unfortunately, there are no reliable sources known to me and only rough indications can be given. An official population figure of 5280 given for the entire Vaupes region (Rodriguez 1962) was unfortunately a gross underestimate. Figures reproduced in Dostal (1972 : 393—6) suggest a population of over ten thousand, divided into 'social and linguistic groups' which roughly correspond to my category of Compound Exogamous Groups (or Simple ones where these exist on their own). With the help of this breakdown we may estimate figures of between two hundred and one thousand members for such Exogamous Groups, with the central Vaupes groups at the higher extreme and the marginal ones at the lower extreme. The Tukano themselves are more numerous than this but it is difficult to say to what extent, because many groups which are not considered Tukano by descent speak the Tukano language and therefore get included in estimates. The size of a sib is from one to fifty or so members in the Pira-parana area but there may be larger sibs elsewhere in the Vaupes. Exogamous Groups Exogamous Groups are collections of sibs arranged in a hierarchical order, modelled on the birth sequence of an agnatic male sibling group. This principle of birth order, which I shall call 'seniority', is found at every level of organisation within an exogamous unit so that each individual, each sub-unit within each sib and each sib itself has a unique position in the order composed of similar units. Among sibs of an Exogamous Group there is a 'first born', a 'next born', a 'next born' and so on to the 'last born'. Each sib is either an 'older brother' or a 'younger brother' from the point of view of any other sib. The distinction I have made between Simple and Compound Exogamous Groups is also a matter of internal organisation. There are five hierarchically organised specialist roles, each of which is allocated to a single sib, so that five sibs belong together in a functionally integrated unit. These roles, in descending hierarchical order, are chief, dancer/chanter, warrior, shaman and servant. A Simple Exogamous Group is a unit containing one set of these roles while a Compound 19

Social structure Exogamous Group contains more than one set. A Simple Exogamous Group may exist on its own or it may be part of a Compound Exogamous Group composed or two or more Simple Exogamous Groups arranged in accordance with the seniority principle from first to last born. From the descriptions of ideal behaviour associated with sibspecialist roles the internal organisation of the set of five roles can be shown thus: chief and servant: reciprocal roles in the politico-economic domain dancer/chanter and shaman: complementary roles in the metaphysical domain warrior: single role in the domain of'natural' intergroup relations, directed outwards and opposed to the same role within the internal organisation of affinal Exogamous Groups This can be represented as follows: politico-economic relations: chief ritual relations: dancer/chanter external relations: warrior (birth order 1 2 3

servant shaman 4

5)

The specialist roles thus give the order among a set of sibs an additional feature of symmetry about the central position. This symmetry depends on the recognition of three domains of social activity each with a characteristic mode of social interaction. In this way a simple hierarchy based on time intervals (between serial births) is converted into a complex system of relationships linked to a comprehensive model of social and cosmological organisation. I believe this model has special significance among the dispersed and semi-autonomous longhouse groups into which the Tukanoan population of the Vaupes is organised. It is not just a theory about a certain kind of social structure, it also contains the more comprehensive and fundamental idea that there is a structured organisation of Indian society as a whole (see ch. 3). An Exogamous Group, Compound or Simple, occupies a continuous territory described in terms of the rivers that structure it. The sibs of each Simple Exogamous Group are arranged along rivers so that the seniority order from first born to last born reflects an upward journey from river mouth to headwaters, and the Simple Exogamous Groups of a Compound Exogamous Group are arranged in the same way. 20

The model The phratry The phratry is an association of Exogamous Groups united by the rule of exogamy but not occupying a continuous area. Thus there are relations of 'brotherhood' between certain geographically separated Exogamous Groups. These relations may be expressed in either a strong or weak form. The strong form is hierarchical: the separate Exogamous Groups are arranged as a set of agnatic male siblings in exactly the same way as described for lower-order units. The only formal difference between the arrangement of more than one Simple Exogamous Group into a Compound Exogamous Group and the arrangement of Compound or Simple Exogamous Groups into phratries is in the geographical separation of the units in the latter case. This formal distinction is backed up in practice by the sharing of language and strong sense of common identity, which exist at the level of Compound or Simple Exogamous Group, but not at the level of phratry. The weak form of phratry structure is non-hierarchical, with the separate Exogamous Groups arranged as uterine siblings amongst whom there is formal equality. Any given group may regard some of their uterine sibling groups as 'people waking-up mother's children' (masa yuhiri hako rid), and others as hako ria, 'because they marry our mother's people'. These last might be called 'de facto hako ria\ people who are unmanageable because actual marriage practice proves them to be 'affines of our affines'. From the variability of relationship between the constituent Exogamous Groups, it is evident that phratric organisation is not at all clear cut. It should be considered an uppermost limit of exogamic organisation at which the consistency between the viewpoints of separate groups breaks down in a fundamental way. Because it is difficult to maintain a meaningful distinction between the model and its application with respect to the phratry, some further points about the practical situation may be included here. While the ties between some major Exogamous Groups, such as the Tukano and the Bara, are universally known and rigorously respected in the observance of exogamy, it is fruitless to look for an overall organisation of Vaupes Exogamous Groups into discrete phratries. The brothers or mother's children of a given Exogamous Group may intermarry with separate groups which are also classified as brothers or mother's children. Besides, the very existence of three types of link between 21

Social structure Exogamous Groups - agnatic sibling, original uterine sibling and de facto uterine sibling - suggests that this is a weak and variable form of organisation. It seems likely that in practice this phratric level is emphasised when there is minimum territorial flux, allowing marriage patterns between Exogamous Groups to become stabilised. Conversely, it seems to break down as Exogamous Groups move and are forced to look to new neighbours for wives. The sib The internal organisation of the sib is hierarchical, with small descent groups, founded by a living or recently dead ancestor, ordered from first born to last born. The identification of separate units within this seniority continuum depends upon the size of the genealogical gap between one descent group and another. These sub-units of a sib are simply shallow un-named descent groups whose size and number varies with the size of the sib and degree of differentiation required by the context. Either a whole sib or a sub-unit of a sib gives rise to a local descent group. The local descent group As I use this term here it refers to a group of close agnates who form the core of a longhouse population. It differs from the preceding categories because descent and residence operate together in determining local descent-group membership. Also, there are many people who do not technically belong to any local descent group; among these are all married women who live among their husband's descent-group members. By contrast, all the previous categories based on descent criteria are comprehensive in that every individual belongs to a category at each particular level - a sib, an Exogamous Group and a phratry: membership of such groups is never forfeited on marriage or because of change of residence, as it is in local descent groups. The various social-structural units are represented in fig. 1. The model applied The problem inherent in all attempts to define social structural units in the Vaupes is that Indians are more concerned with the types of relationship (hierarchy etc.) outlined in the model above than with 22

PHRATKY 1 COMPOUND EXOGAMOUS GROUP SIMPLE EXOOAMOU5 CROUP 1 SIB 1

CHIEFS

h

2

3

1 2,3

Fig. 1 Units of social structure 23

4-

DANCER/ WARRIORS SHAMANS SERVANTS CHANTERS

2,3 etc. 1

2,3 etc.

1, 3 etc.

Social structure the precise definition of social boundaries. They do not have any neat named categories equivalent to the ones I have used to describe my model; instead they use expressions like 'one people' (koho masa), 'one pile' (koho tubua), 'children of one man' (singuna) or 'other people' (ngaherd) to describe relations between a great number of named groups. Sometimes the named groups overlap in membership, and sometimes they act as 'sliding' categories covering a more or less extensive population, depending on context. Naturally, the group membership and location of residence of the informants influence statements about individual or group identity, and there are frequent assertions, backed up by myths, that particular claims to membership of certain groups are false. In spite of the difficulties, there is a fairly high degree of consistency about the broad outlines of intergroup relations, and it is mainly the details of interrelations between sibs that differ. This may be observed in other ethnographers' attempts to set out the existing situation (e.g. Briizzi 1962, Goldman 1963.) For the benefit of other Vaupes specialists, I include my own attempt for the Pira-parana region in appendix 1. As mentioned at the beginning of this chapter, social change is at the heart of any question about the relation of model to reality in the Vaupes. The model I have described is of a sharply defined and fixed system which only crumbles at the upper limit of exogamic organisation. I have emphasised hierarchy and the fixed number of specialist sibs per Simple Exogamous Group, yet, in practice, there is evidence of continual change at all levels both within the traditional system and as a result of white influence. Raiding, typical of the traditional life-style, often resulted in the massacre or migration of substantial groups. Besides this, the small size of sibs, even granted that they were larger in the past, must have made them extremely vulnerable to normal population fluctuation. As for the influence of white society, there is no doubt it has led to vast depopulation and rearrangement of groups on the ground. The consequent deculturation has eradicated many of the traditional functions of social-structural institutions and loosened the cohesion of social units at all levels. All this can be summed up by saying that we know there have always been changes of fusion and fission and changing identity of groups within the total system, but it is only for the last one hundred years that we can be sure of the broad direction in which the total system has altered. This has been towards smaller communities, a minimum of physical violence and a decrease in ritual elaboration. Inevitably, my 24

The model applied account of Vaupes society and ideology hovers between the observed present and Indians' idealised version of the past with which they give meaning to the present. Without confining myself to simple ethnographic description, I can see no way round this. The best I can do is give some general idea about the applicability of my model, which belongs to this imaginary past-and-present, to the situation on the ground during my field work. Territory While occupation of a continuous territory is an ideal feature of each type of descent category up to, but not including, the phratry, in practice there is considerable overlap between descent-group territories. Indian accounts of group history suggest that there has been a strong tendency for sibs or sub-units of sibs to move back to their original position within their Exogamous Group territory after a period of exile, but they have by no means always done so. In the long run, descent ideology must adjust to sever the close ties between geographically distant groups. Indians are able to describe recent instances of this, but descent-group identity dies hard and there is never a neat pattern of continuous, exclusive descent-group territories at any one time. Although main river and downstream locations are definitely regarded as more prestigious than sidestream and headwater ones, there is no exact correlation among Pira-parana groups between descent seniority and location. However, there is a general correlation for the units of the large Compound Exogamous Group to which our home community belonged, and Patrice Bidou has commented on the strictly maintained correlation amongst the sub-groups of the Tatuyo sib with which he worked (1976). 3 Goldman describes a clear practical association of the same kind among the Cubeo, but it remains to be seen whether residence among central Tukanoan groups bears out this association. 3 It does seem that the groups more culturally and geographically oriented towards the Apaporfs, into which the Pira-parana flows, are closer to the ideal pattern than those oriented towards the rivers of the Vaup6s system. The Pira-parana, flows away from the Vaupes affluents and from the centre of Vaupes culture. It may well be that this geographical feature presents those groups having strong affinal ties to the people living on the Vaupes river system with conflicting ideals. Part of the prestige of downstream residence is associated with increased social interaction and yet, for these upper Pird-parana groups, downstream residence represents increased social isolation.

25

Social structure Names for groups In general, groups at the phratric level do not have names; instead, named Exogamous Groups are listed and described as 'one people' (koho huna or koho masa). Compound Exogamous Groups, or Simple ones where these exist alone, can usually be identified by name from any given point of reference, but these names are often rejected as self-names. Descriptive names of the form 'children of Such-and-Such Anaconda' are sometimes used at both Exogamous Group and phratry level. Sibs are named groups in a way that more inclusive units are not. In fact, the identity of a sib is so intimately bound up with the name that, in a sense, the name is the sib. Disputes about sib membership take the form of insisting that a person or descent group belongs to one named sib and not another, instead of argument over genealogical details. In their absence, people are often referred to in terms of sib membership to avoid the use of personal names. However, there is also a reticence about mentioning a person's sib-name to his or her face, and individuals often substitute a more inclusive descriptive name for their own sib-name in conversation. Thus, Indians often made statements of the form, 'We are not Wamutanyard [sib-name in Bara language], we are Fish Anaconda descendants [Exogamous Group name].' This slight tendency to avoid direct personal use of sib-names must be seen in relation to the much stronger avoidance of individual personal names in address. Sib-names have much in common with personal names for there are joking and derogatory sibnames in the same way that there are joke-names paired with male personal names. The internal organisation of Exogamous Groups There is a certain amount of confusion and disagreement about the exact ordering of sibs within a Compound or Simple Exogamous Group, but there is broad consistency over which portion of the seniority hierarchy any particular sib belongs in and, usually, complete agreement over the first and last sibs. Even when he does not really know the ordering of a group of sibs united by observance of exogamy, an Indian will stick to the convention of listing them from first born to last born, thus demonstrating the force of the hierarchical principle. 26

The model applied In practice, it is difficult to find neat examples of the allocation of the five specialist roles to a simple cohesive set of five sibs, although the Meni Masa, our hosts, belonged to one such set. More often, a set of sibs is recognised as forming a cohesive hierarchical unit and yet only some of the specialist role identifications are known. Such a set is likely to contain more than five sibs existing at the present day and, in addition, a variety of extinct sibs whose hierarchical status is not clear. However, data from the Pira-parana show that even when all five roles cannot be clearly identified among a cohesive set of sibs, the notion of the functionally interdependent set of roles provides an important model of cohesion. For instance, a particular sib will be 'their shamans' rather than 'our shamans', and this implies a separate functioning of 'them' as opposed to 'us', even though the remaining roles are missing amongst 'their' sibs. The reality corresponding to my Simple Exogamous Groups consists of such bounded units. These are usually localised and arranged hierarchically within a more extensive hierarchy of sibs, which corresponds to my Compound Exogamous Group. The importance of the specialist role model in creating smaller, bounded, localised units within large Exogamous Groups with a wide geographical spread and high population leads me to expect the same principle to operate throughout the Vaupes. Stephen Hugh-Jones mentions fragments of evidence for similar systems among the Baniwa (1979 : 149); Goldman reports that at least some elements of the system exist among the Cubeo (personal communication) and Howard Reid, recently among the Makii, was told that both the Desana and the Tukano divide into a number of localised sub-units, each containing a set of sibs differentiated by their specialist roles. It therefore seems likely that this type of sub-division is found throughout the Vaupes and that central groups, such as the Tukano (fifty sibs reported by Fulop 1955) and the Desana (over thirty sibs reported by Reichel-Dolmatoff 1971), fit the category of Compound Exogamous Groups as used here.4 The function of the specialist role system is more problematic. It is an ideological system and can be analysed as such whatever its practical functioning: it is this approach that the present-day anthropologist is bound to adopt. The content of the role categories can be 4 H. Reid says that an informant told him these groups were divided into a number of localised sub-units each consisting of a set of specialised sibs. The set of specialist roles he gave was the same as the one I describe, except that the warrior role was missing, making a total of four roles. This may have been omitted on purpose, since the more acculturated Desana and Tukano are reluctant to admit their warlike past. 27

Social structure drawn from the specialist individual roles existing in present-day society - there are dancers, chanters and shamans — and from idealised descriptions of the other roles of chief, warrior and servant. However, speculation about the past is inevitable. Indians of the Meni Masa warrior sib say that in the past their sib occupied a single longhouse and each of the four other specialist sibs of the set also occupied a single longhouse on the same site. Each performed the appropriate role for the benefit of the others. They support this image of truly specialist sibs by saying that they themselves did not know how to dance because the Rasegana, their immediate seniors and 'dancers and chanters', performed this role on behalf of the whole Simple Exogamous Group. Only when the Meni Masa decided to stop raiding did they learn how to hold their own dance rituals from the Rasegana. At present, there is occasional reference to such a system: members of a 'dancers-and-chanters' sib may be invited to perform their traditional function at a ritual staged by another longhouse group; the 'servant' sibs are treated with less respect than others. But in general, the specialist roles essential to a functioning of present-day society are performed in an individual capacity. Individual specialists are not confined to sibs of the same speciality — indeed, such a system would be unworkable because the sibs of a Simple Exogamous Group tend to be too spread out for individuals from each sib to gather together in a single place and benefit from each other's specialist skills. At present, possession of one ritual role does not exclude an individual from possessing another, and for any specialist role there are degrees of competence and recognition. Within a large longhouse there is usually at least one individual with some degree of competence in each of the roles of chief, dancer/chanter and shaman. It is thought especially appropriate if, within the group of brothers forming the elder generation of the core descent group, the distribution of specialist roles follows the seniority order. The eldest brother should be the headman, the youngest a shaman and the middle brother either a chanter or a dancer. In a set of longhouse groups which meet regularly for ritual purposes, there will be at least one important specialist shaman, one chanter and one dancer of wide renown; thus the distribution of the ritual roles depends largely on the fact that they are indispensable for the functioning of a ritual community. Among my data there are two different principles at work regarding the functioning of the specialist-role system. According to Indians' accounts of the past, the specialist roles operate exclusively among 28

The model applied the five sibs of a Simple Exogamous Group. Although the relations between these are modelled on the relations between individual males of an agnatic sibling group, an actual agnatic sibling group belongs to one or other particular sib and therefore cannot be internally differentiated by specialist role. However, at the present day, such of the specialist-role system as remains appears to operate according to a segmentary principle, because the set of roles appears both among sets of sibs and sets of siblings. There is some evidence that the specialist roles also appear among sub-sections of a sib and this would be entirely consistent with a segmentary principle. By its very nature, a segmentary specialist-role system precludes a single consistent specialist role for any person apart from the very first born and the very last born within the entire Exogamous Group. Only for these two individuals could sib role, sib sub-section role and sibling role coincide. The general lack of consistent role identification for individuals, together with the need to fill the roles, results in a system of specialisation resting more upon relative personal aptitudes and ambitions than social-structural status. I would like to suggest a hypothesis (which, unfortunately, can never be proved in view of the progressive disintegration of traditional social systems) that the 'sib-only' principle and the 'segmentary' principle represent adaptations to different degrees of geographical dispersion of a Simple Exogamous Group. According to this hypothesis, the 'sib-only' principle is the ideal one, associated with the idealised past and also with a concentration of the entire Simple Exogamous Group on one site; the 'segmentary' principle starts to operate whenever the Simple Exogamous Group spreads out. It may be that there has never actually been such a concentration of Simple Exogamous Groups either on one site or along one stretch of river, and that the functional integration of sets of five sibs has always been an ideal, rather than a practical reality. After all, the sibs of a Simple Exogamous Group lend themselves better to expressing such an ideal than do units at lower levels of social differentiation, because sibs are named and change more slowly than smaller units. The set of five roles would seem, in fact, to stabilise not only the identity of sibs but the structure of the Simple Exogamous Group itself. It is a 'conservative' model which makes incorporation of new sibs from outside or fission of sibs within more difficult. When it is considered without the associated specialist roles, the hierarchical system amongst the sibs of a Simple Exogamous Group is much weaker 29

Social structure as a stabilising force. The hierarchical principle readily enables the Exogamous Group to accommodate new outside groups (though more easily at the bottom of the hierarchy) as well as splits within. Besides, by itself, the hierarchical system does nothing to determine the interdependence between brother sibs and nothing to divide the large Compound Exogamous Groups into smaller units with a high degree of internal interaction. The fixed nature of the five-role system means that, on the one hand, change in the composition of Simple Exogamous Groups is inhibited but, on the other, when migrations and population fluctuations force such change, the system of specialist roles loses coherence. There is constant tension between the ideal system and the practical relations between groups on the ground. Even so, however great the divergence between the ideal and practice, the practical situation is described in terms of the five specialist roles as far as possible, because these represent order and interdependence within a local exogamous unit. Functions of social-structural units At all levels up to, and including, the phratry, exogamy is a fundamental function of social units based on patrilineal descent. The way in which marriage alliances involve units at different levels is discussed briefly below (pp. 3If, 85). Apart from exogamy and from the internal organisation of Simple Exogamous Groups according to specialist sib roles, the functions of the larger exogamous units are mainly expressed in rights over ritual property. The rights reflect descent structure, so that the Simple Exogamous Groups making up a Compound Exogamous Group may own ritual property of very similar, but not quite identical, form. In the same way, the geographically separated Exogamous Groups making up a phratry have one or two rights in common, and these are frequently cited as evidence that they should not intermarry. At the lower level, the sibs of a Simple Exogamous Group share virtually identical rights over ritual property and any substantial differences are regarded as evidence of illegitimate status within this cohesive set of sibs. The basic principle operating in the distribution of these rights is clear: exogamy is the corollary of shared descent, and shared descent receives its significance from ancestral power which is transmitted along patrilineal descent lines. The transmission of such power, as we shall see, is intimately connected with male-dominated ritual, and so it is quite 30

The model applied appropriate that variation in rights to ritual items should mirror descent ties between groups. In fact, from an Indian point of view, it would be more accurate to say that ritual property is part of the content of descent ties. 'Ritual property' is a loose designation covering sets of names for sacred musical instruments (Yurupary), sets of personal names, sets of dances, sets of chants about ancestral origins, traditions of origin expressed in myth, various ritual ornaments, varieties of cultivated plants and other ritual paraphernalia. Language could also be included in the list for, although it is not strictly speaking 'ritual property', it is ideally descent-determined and is the medium of expression of many of the items above. Non-ritual items and techniques are also associated with descent groups, so that particular weaving styles, species of roofing leaf, cooked dishes and so on are described as belonging to one or another group. However, although no hard and fast distinction can be made, these are regarded less as part of the inherent nature of the groups who own them and more as random idiosyncracies. Nowadays, the members of a phratry, Compound Exogamous Group or Simple Exogamous Group never collect together for any purpose. To what extent this is due to geographical dispersion in the case of Simple Exogamous Groups is not clear (see p. 28). In practice, it is the longhouse group that is a virtually autonomous, cooperative and functionally integrated unit for the purposes of economic production, consumption and the rearing of children. While this autonomy is maintained in day-to-day secular life, the longhouse community depends upon other communities for marriage arrangements, attendance of rituals, incorporation into a wider political community, particular kinds of communal food-getting expedition and certain articles of trade. Each large longhouse community has a social field or neighbourhood extending around it in such a way that the social fields of nearby longhouses overlap considerably. These fields consist of a number of communities who regularly exchange ritual invitations with the home community, and whose members may join the home community on camping expeditions to fish or exploit seasonal forest products. The communities making up the social field of a particular longhouse community may be affiliated with the same sib, a different sib of the same Exogamous Group or else with a different, affinal sib: typically there are all three kinds. The social fields are not precisely 31

Social structure defined; geographical closeness, common descent and existing affinal ties are the three most important factors in determining the attendance of rituals, but often quarrels develop between particular communities related in these ways and frequency of interaction drops. It is clear from this pattern that, whatever else they may be, the descent units are not corporate groups except at the level of the local descent group. Part of the conceptual difficulty in describing Vaupes social structure, at least as I found it in the Pira-parana area, is the overlap of a comprehensive system based on descent and exogamy with a practical local organisation into longhouse communities, each of which perceives other communities as more-or-less distant outsiders. It is difficult to isolate wider political communities because raiding and killing have ceased and there is no pressing need for military solidarity. In ideal terms political relations would, of course, have been dominated by the internal structure of the Simple Exogamous Group and the external relations with affinal units of the same type. Political relations are now bound up with marriage relations, shamanism and prestige gained through ritual activity. The hierarchical structure of descent groups plays a negative part by preventing those from very low positions in the seniority hierarchy of a Simple Exogamous Group or sib from becoming renowned leaders. Doubtless, political relations always had these same components but, according to Indian accounts, previously, the resort to physical violence led to political alliances of numbers of longhouses and the emergence of powerful leaders. Now there is strong resentment of individuals who seek power over people outside their own longhouse community; threats of violence are relatively empty and differences in political prestige are achieved through the staging of rituals or through shamanic reputation. The social field composed of longhouse communities that regularly attend the same rituals is clearly also the field of operation of practical political prestige. The recognition of important individuals tends to draw together the social fields of the longhouse communities of a geographical area so that they overlap more with each other than they do with those of communities outside the influence of the important man. Sometimes, the headman of a house which holds frequent rituals is also a shaman (in spite of the ideal separation of the roles of chief and shaman), and such people build up a local reputation which is soon reflected outwards as they become the butts for accusations of evil shamanism made by more distant groups. The general picture is one 32

The model applied of a social field extending around each longhouse, the fields overlapping in such a way that relatively dense areas of interaction are created. These areas show a relatively high rate of endogamy, relatively friendly and stable relations between the communities and a relatively autonomous ritual life. They cannot contain two longhouse leaders of wide renown without their being competition between them, which is likely to weaken this social cohesion. The areas tend to be geographically, as well as socially, defined although ease of communication is naturally more important than distance 'as the crow flies'. Ideology of descent In the Pira-parana area each exogamous group above sib level is derived from an anaconda ancestor. In some cases these anacondas represent phratries but in most they represent a Simple Exogamous Group or a Compound Exogamous Group. In any case, there are different versions of the origin of any given group and the best I can do here is to give a general outline of the nature of ancestral anacondas. They originated in the east beyond the Water Door (Oko Sohe) which lies at the mouth of the Milk River (Ohekoa Riaga). The Milk River is the main river of this earth into which all other rivers flow; Indians who have travelled far downstream with whites identify the Milk River with the mainstream Amazon. The ancestral anacondas swam up the Milk River, stopping at numerous sites distinguished by rocks, rapids and other landmarks. These sites, at which the ancestors danced, are associated with mythical events which occurred on the ancestral journeys. The theme of a great number of these is the original obtaining of ritual objects and plants which are group-specific and which are used by the living to recreate mythical times. Some of the sites, particularly those towards the mouth of the Milk River, are mentioned in all the anaconda journeys but others are particularly associated with the journey of one anaconda. The separation of the journeys becomes more marked towards the end, for the anacondas separated and travelled up different rivers and, finally, the sibs derived from each anaconda were deposited in turn along the route. In the last stages of their journeys — those covering the rivers just beyond the territory of the present-day descendants — the anacondas swam up to the headwaters of various affluents and down again. They were seeking good settling places, but many sites had to be rejected because of the mystical dangers located there. 33

Social structure The process of populating the earth with people is called 'people waking-up' {masa yuhi-) and each sib recognises a string of sites as its own 'people waking-up houses'. At these sites, the anacondas emerged from the water, became groups of ancestral people and danced. The sites represent serial stages of differentiation in the course of the journey: the Water Door is the population waking-up door for all the different peoples, then there are people waking-up houses common to all the sibs embodied in a single anaconda and, finally, special ones for each sib (fig. 2A). There are various Indian theories about the precise nature of the ancestral anacondas and the means of 'birth' of the sibs. It is generally acknowledged that the sibs are represented by sections of the anaconda's body in such a way that the head or tongue is destined to be the first-born sib (chiefs), and so on in descending hierarchical order until the tail, which is destined to be the last-born sib (servants). According to different versions, the sibs are separated, 'born' or created in the appropriate hierarchical order by serial vomiting, by A

ANACONDA JOURNEYS AS CONTINUOUS PROCESSES

LAND: DESTINATION

mcreasim differentiation

lA/ATPD • W*TEKDOOR [EAST]

&

ANACONDA JOURNEYS AS REPEATED 2-PHASE CYCLES

OM°LAND

LAND

WATER

Fig. 2 Anaconda journeys 34

(DANCING ROUND)

(SWIMMING ALONG)

ANACONDA IN WATER [EAST]

The model applied the ancestral body breaking up or by disembarkation as from a canoe. Once deposited at their traditional sites they are conceived of sometimes as groups and sometimes as single ancestors. The nature of the anaconda ancestors turns upon the idea that a single entity may be perceived or conceived of in different ways depending upon the point of view of the experiencing subject. A common view is that the ancestral anacondas were anacondas in the water but that whenever they came on land at the waking-up houses, they were transformed into groups of people who danced and performed ritual. This view implies that the final depositing or dispersal of the sibs on land is seen as the irreversible end of a gradual transformation achieved by leaving the water and re-entering it many times (see fig. 2B). The transformation from water to land is explained by equating the anaconda with a canoe from which people disembark or with a skin or mask (sudi) which the people shed as they leave the water. However, it is on the relation between river and land, rather than the mechanical means of the transformation, that we should concentrate. This relationship is given additional content by an alternative view that 'anaconda' is just a way of describing ancestors which has no reference to the characteristics of present-day anacondas. The ancestors were processions of people who walked up the dry river beds leaving the imprints of their feet and sacred paraphernalia in the stillsoft rock where we can see them today. It is only from the point of view of our land-based, present-day society that these ancestral paths are rivers and so the ancestors are 'anacondas'. We thus have two views: one, that ancestral anacondas in the water are equivalent to ancestral people on land and the other, that water and land is a latterday distinction which repeats the ancestral distinction between journey and stopping point. This ancestral distinction is perhaps better described in more general terms as one between transformation or creation, and results or finished product. In this way, the relationship between water and land on one hand and that between creation and finished product on the other becomes analogous. At the same time another relationship, between procession in single file towards a destination and dancing in a circle, is associated with the previous two. Linear procession and dancing are the physical modes typical of the water-creative and land-resultant phases. (west) 35

land results circular dance form

water creation linear journey

(east)

Social structure The significance of these analogous relationships will be discussed below (pp. 61, 67). From the set of relationships shown in fig. 2, certain critical features of anacondas emerge. They are at home in water and also on the river bank and they are long and relatively undifferentiated in form. These two features make them suitable for representing mediation between river and land and a hierarchical order respectively. However, in spite of the firm idea that Exogamous Groups have anaconda ancestors and are aquatic in origin, the names of the anacondas themselves suggest a higher-level classification of Exogamous Groups according to 'cosmic habitats' — water, land and sky. It is impossible to arrange Exogamous Groups into three neat and mutually exclusive categories on this basis, just as it is impossible to arrange Exogamous Groups into a pan-Vaupes phratric pattern. Nevertheless, Indians refer to these categories in direct conversation as well as in myth as if they did reflect descent categories. They claim, for instance, that Water Anaconda People (Makuna) and Fish Anaconda People (Bara) are all one people who should not intermarry because they are all from inside the water. In the same vein, they sometimes claim that it is really only these true water people who arrived in anaconda form originally, whereas other group ancestors are merely named like anacondas. In keeping with this, the charter for intermarriage between Fish Anaconda People and Yeba People (Bara and Barasana) is a myth in which Fish Anaconda's daughter, a true anaconda in form, comes on land and marries Yeba, who is initially a jaguar. Yeba means 'earth' in many Tukanoan languages (although in Barasana it is restricted in meaning to this ancestor) and the Barasana claim that their dead souls can be seen on earth as jaguars, while those of the Bara go to underwater houses. Now Yeba, besides being a jaguar, is son of Meni Hino, Meni Anaconda, who travelled from the Water Door. The Sky Anaconda People (Tatuyo) are the other major affinal group with respect to the Barasana and, as their name suggests, they are associated with sky beings, including eagles. The dual anaconda and non-water-people status of some groups is resolved in Indian thought by the same kind of relativism that we have seen to operate above in the case of anacondas and people. Indians say that anacondas can emerge from the water onto land or straight into the sky, swapping their anaconda masks for jaguar or eagle masks and thus, although each of the three major cosmic habitats has its large ancestral predatory creature, these creatures are simply transformations of one another. 36

The model applied I do not mean to press the ethnographic aspect of this rather complicated subject any further because the more minute details of group origins are embedded in numerous myths and we have seen enough no longer to expect systematic ideologies of origin. A more detailed analysis of the Tatuyo case is presented by Bidou (1976). This work serves to underline the problem, for two very different sets of data on ideology of origin were collected from adjacent Tatuyo sibs. This means that even the notion of an agreed Tatuyo ideology is impossible. However, if there are few generally agreed myths of origin, there are at least general principles underlying the variety of myths. The Exogamous Groups are all equivalent in their possession of anaconda ancestors and these ancestors establish their true territories as well as providing a model creative process for each. At the same time, rules of exogamy create similarities and differences between these groups. The analysis of the kinship and marriage system will show that a minimal model of a Vaupes marriage system requires three equivalent exogamous groups (be they phratries or of any other order). I suggest that this fact may well be associated with the tripartite classification of anaconda ancestors into land, air and water. The underlying unity of anaconda, jaguar and eagle fits the requirement that the exogamous groups be equivalent. The relations between the habitats are triangular, in the sense that one can move from any one to any other just as any one of the groups can marry any other. Real jaguars, eagles and anacondas are natural mediators between these habitats and the inclusion of Fish-Eagle Anaconda as an Exogamous Group ancestor (Tukano) covers the least obvious connection, that between CHIEFS DANCER/ CHANTERS WARRIORS SHAMANS SERVANTS BARASANIA ;

BARA

EARTH e.g. JAGUAR

A EXCHANGE OF WOMEN BETWEEN EXOGAMOUS GROUPS

B ANCESTRAL CREATURES AS MEDIATORS

C EXOGAMOUS GROUPe.g. TATUVO NO INTERNAL MARRIAGE

with C

Fig. 3 Tripartite classification of Exogamous Group ancestors compared with marriage relations between Exogamous Groups 37

Social structure water and air. The triangular pattern of relations between the three habitats and intermarrying groups contrasts with the hierarchical relations between sibs and siblings. I shall argue later that these, too, are reflected in the cosmic structure of the universe, in which the same elements — water, earth and sky - are related in a different way. These different relations in no way contradict the first set for, if my understanding of Indian ideology is right, the incorporation of elements into different structures or processes where they receive different meanings is not only possible, but actually an essential feature of such ideology. Origin of sibs At this point we must retreat further into the ancestral past, beyond even the Anacondas, to explain that all Exogamous Groups are children of the Primal Sun and each regards the Sun as 'father' of its own particular group ancestors. In fact, the Sun, the ancestral anaconda of a given group, and often a string of other named anthropomorphic beings, are fused together in a way which is best described as 'predescent'. These beings are sometimes arranged in generations of father-and-son relationships, sometimes called elder and younger brothers and sometimes said to be separate manifestations of the same entity; conversely, the same name often applies to father and son alike. Repeated questions from the anthropologist about who is son of whom and which beings are 'the same' leads to worse confusion. The only conclusion can be that the relationship between name and individual, between father and son and between elder brother and younger brother, are not three separate types of relationship in 'pre-descent' terms as they are today. Indians stress that the original ancestral beings, collectively called He masa, did not know sexual procreation and just came into existence, one from another. Since it is the periodicity established by marriage, sexual reproduction and growth of families that creates discrete generations and sibling groups and also, as I show later, serves to distribute names, it is understandable that these distinctions should not have the same significance before sexual reproduction. The origin of sibs marks the end of this 'pre-descent' era. While these are the body of the ancestral anacondas, other versions of their creation make them products of sexual reproduction. It is clear in the myth of Yeba's marriage to Fish Anaconda's daughter, Yawira, that 38

The model applied the origin of Barasana sibs occurs simultaneously with the development of sexual reproduction and alliance between exogamous groups - Indians specifically say that the myth describes the origin of all these things. In fact, normal sexual reproduction is developed during the course of the origin of sibs, for the ancestors of the eldest sibs are born after Yeba made love to Yawira with his fingers, and those of the younger sibs are the product of normal sexual intercourse (see appendix 1). The sib ancestors married women of opposed exogamous groups and the sib populations increased over the generations. Stories of these first generations of sib ancestors are typically concerned with fights, migrations, marriages and the stealing of sacred paraphernalia. Although they include 'impossible' episodes, they are very obviously different from the myths to which I refer in this book: the beings in the stories are members of existing or extinct groups instead of anthropomorphic creatures, celestial bodies or strange spirits. The stories in which they figure presuppose a world populated by descent groups composed of ordinary people (no longer/ife masa), in fact, a world in which the social and cosmic order is already established. Without entering into the difficult subject of the relation between myth and history, I can best explain the situation by saying that there may be a parallel between the way in which Indians regard the 'descent era' and the way in which the western layman regards history. The generation in which sib ancestors are established, and the following ones in which the present-day descent lines are established, can roughly be described in genealogical terms, and sometimes the group membership of the wives is included. There is little consistency between accounts for any given group except in the set of names involved and the positions of one or two key ancestors. The patrilineal genealogies are told with emphasis on the male sibling groups and on the correspondence between the seniority order of these and the seniority order of their respective groups of descendants today. The teller also invariably points out that present-day people possess the same names. In any genealogy extending beyond two generations the names of one generation are repeated again in the next but one and teller and listener alike get hopelessly muddled. This alternation of a limited stock of names, which is maintained in present-day naming practice, inhibits the accumulation of genealogical knowledge. It is also consistent with the marked stress on agnatic sibling ties at the expense of father—son ties. 39

Social structure Between the genealogies of the first ordinary people and the genealogies extending back from the living there is a gap. Occasional characters associated with raids, particular shamanic acts and so on may be picked out from these dark ages in group history, and it is clear that Indians do not believe that the one or two recently dead generations they can outline tack directly onto the founding generations. However, instead of needing consistent genealogical knowledge as a measure of time and identity, they are content with the conviction that there are genealogies. They frequently refer to a patrilineal genealogical principle, explaining that over the course of time sets of brothers give rise to descent groups ordered by seniority, but they can rarely fill in more than one or two of the key names in such a process. The minimal emphasis on genealogy is perhaps surprising in a society where patrilineal group membership depends firmly on descent, but I hope to show that this is quite consistent with other aspects of the relationship between the present and the ancestral past. Finally, the credentials of a genuine sib lie in its relationship with the ancestral past. The sib must have arrived from the Water Door as part of an anaconda ancestor and it must have its appropriate wakingup house. It should have a set of sacred He instruments (Yurupary: see initiation, pp. 142ff) and various other sacred items derived from the body of the anaconda (or one of his other manifestations) which act as a guarantee of this ancestral link. People prove that a sib — invariably a senior one — is claiming false status in precisely these terms. They say that a relatively recent forbear of a genuine sib made the members of the false sib out of wild creatures (mice and worms in two particular versions) in order to increase his workforce. Then, when he tried in turn to address the leader of his workers as 'affine' {tenyu), 'mother's child' {hako maku) and 'younger brother' (bedi), the leader refused to answer. In the end he was obliged to call the upstart 'elder brother' (gagu). The stories go on to show that the sacred paraphernalia possessed by such groups is nothing but a collection of commonplace objects. The longhouse and its inhabitants Composition of the longhouse group The longhouse group contains a local descent group as a core and the 40

The longhouse and its inhabitants rest of the population is made up of wives of descent-group members and individuals whom I shall call 'extra residents'. Longhouse communities are in a continual state of flux. Also, Vaupes Indians are constantly visiting other longhouse communities so that it is often difficult to know whether non-local descent-group members are regarded as temporary residents or long-term visitors. In any case, they are never fully integrated into the ritual life of the community and are only rarely allocated permanent screened compartments within the house. These extra residents may be from the same sib, the same Exogamous Group or an affinal Exogamous Group in relation to the local descent group. If they are from the same sib their status as outsiders to the local descent group derives from the fact that they have closer agnates elsewhere. However, the married males among extra residents are nearly always from an affinal Exogamous Group and are related to the local descent group by a direct affinal tie through wife or mother; they are usually waiting to establish their own houses and cultivations. There is evidently a strong resistance to the confusion or merging of those descent groups which lie within the same Exogamous Group and yet are separated by their relative positions in the seniority order, for these are very rarely represented in a single longhouse group. The local descent-group members are united by such close agnatic ties that the founding group of agnatic siblings is almost always represented among the oldest living members or the previous, dead generation. Thus the senior generation are either agnatic siblings or patrilineal parallel cousins. Smaller communities may contain a single nuclear family; however, these are invariably living in houses too small for communal rituals and thus are dependent upon a nearby, larger community for ritual life. In the past, when it seems that longhouse communities were very much larger, local descent groups must have been much deeper in structure and extra residents were probably much more numerous. In order to give an idea of the size and composition of present-day communities in the Pira-parana area, I include a table (table 2) based on thirteen particular communities. These are the community of our fieldwork location and twelve surrounding ones, which together occupy a continuous but irregular-shaped area centred on the upper reaches of Cano Colorado. Each of the twelve communities interacts with the home community during rituals but the regularity of interaction drops towards the periphery of the area. The table shows clearly that the incorporation of extra residents is 41

Social structure Table 2 Composition oflonghouse groups (in order of population size)

Headman Felisiano Santiago Cristo Rufino Bosco* (home community) Atuni Manueli Americo Domingo (1) Domingo (2) Benjamin Pedru Pau* fl

Local House size: Local descent-group dance/ descent-group members' Total Extra residents population non-dance members wives 20 14 15 18

6 4 6 6

3 8 2 0

29 26 23 24

D D D D

7 10 7 5 6 4 3 4 3

3 1 2 3 1 1 1 0 1

5 0 0 0 0 0 0 0 0

15 11 9 8 7 5 4 4 4

D D D D _ —

Pau's household is also part of Bosco's, since they retain their compartment in Bosco's longhouse and live there some of the time. The numbers given for Bosco's do not include Pau's.

restricted to the larger communities which invariably live in 'dance houses' — those with a central space large enough for communal dancing on ritual occasions. One of the factors determining incorporation of extra residents is the ambition of a longhouse headman to increase his followers so that he will have the necessary numbers and economic potential to hold frequent ritual gatherings. Extra residents have a special obligation to pull their economic weight in the community since they depend on the goodwill of the headman much more than local descent-group members do. The steady economic surplus produced by a large and successful community also plays an important part in making incorporation of extra residents possible. Extra residents who stay for a long time and fell cultivations are rarely organised into balanced productive units — nuclear families (complete or incomplete) containing at least one adult of each sex. Such a balanced productive unit would be much more likely to set up 42

The longhouse and its inhabitants a small house of its own than to tolerate the authority of a headman from a different descent group. For our purposes here, we may ignore the existence of extra residents and treat the life and internal organisation of the longhouse community as referring to the community of local descent-group members and their wives.5 Indians regard this as the ideal community and the strong desire to see children grow up among their close agnates, speaking their patrilineal language and taking part in descentgroup marriage arrangements and rituals, prevents a high rate of deviance. The ideal that a widow should marry the dead husband's brother is very often followed and this, too, prevents small children from leaving their local descent group. The longhouse setting In the Pira-parana area there is no recognition of fixed territorial boundary marks associated with either longhouse communities or more inclusive units. However, virtually every landmark in the forest or along the river has some significance in the myths of origin of one group or another. Longhouse communities should be close to the emergence site of the appropriate sib but there are many cases where they are not. In these cases, the community members stress the distinction between ancestral and present-day territory. Longhouses are built close to rivers or streams: there is a port on the river front where canoes are landed and people bathe and fetch water (fig. 4). A path leads from the river to the longhouse, to the men's door, which is supposed to face east, although it does not always do so in practice. The longhouse is set on a sandy plaza which is extended like an apron outside the main men's door and fringed by a kitchen garden where both magical and edible plants grow. Amongst these the peach palms (Gulielma gasipaes) stand high above the roof of the longhouse and mark the site long after the house has gone. Beyond the plaza is a large area mainly taken up by cultivation sites, both newly cleared and in different states of regression to forest. These gardens are scattered according to the suitability of land and 5 I know of only one community, in Cano Tatii, a western affluent of the Pira-parana, where this situation is not the ideal. Here, two affinal descent groups occupy the same longhouse. Now, one group is represented by a single adult male, but people say that traditionally his sib are 'side-of-the-house exchange partners' (wi karoka tenyua) and always reside with their affines, either on the opposite side or at the opposite end of the house from them.

43

Note = Hot drawn to scale

Fig. 4 The longhouse setting 44

The longhouse and its inhabitants some may even be cleared on the far side of the river, but conceptually, at any rate, they make up a continuous zone which surrounds the house except where it is cut off by the river front. There is another path from the opposite, female end of the house which leads to the gardens, often crossing a small tributary of the main river which serves as an alternative water source and washing place used by women for their daily chores. This frees the main port for male use. Beyond the gardens is the forest (makaroka), familiar at first and becoming more alien further away from the house. The house and its gardens make up a differentiated, controlled 'social' area which is opposed to the undifferentiated (from this point of view only) forest. The river which bounds the 'social' area on one side is seen as a spiritual power supply connecting the house to other communities and to the source of ancestral power in the east. The longhouse Although Vaupes Indians practise shifting cultivation of bitter manioc, it is very difficult to determine the relation between this agricultural method and settlement patterns over time. I do not possess the historical or ecological data even to attempt this. However, good agricultural land is certainly an important consideration in choosing a house site and exhaustion of land is one of the many reasons given by Indians for a change in house site. On the other hand, it seems that some house sites have continuously supported sizeable populations for decades. Therefore, in some places, there is evidently sufficient good land for Indians to establish continuous cycles of cultivation and forest re growth. Other reasons for changing house site are the following: to move away after an important community member has died; to move away after quarrels or hostilities with co-residents or neighbours; to move nearer to communities with whom recent marriage alliances have been made; to exploit superior fishing and hunting resources; to recolonise ancestral territory; to avoid whites or, more recently, to move towards them to obtain merchandise and education. I estimate that the maximum natural life of a longhouse structure is about ten years. It is possible to replace the roof in the meantime, but finally the main posts and beams become unsound. However, a new house is very often built before the old one has decayed. This may be because the community wishes to move for any of the reasons 45

Social structure given above, or because they wish to change to a larger house on the same site with all the opportunities for increased prestige this brings. A change of house, particularly if it is also a change of site, accentuates the personal relations within the old community, because the members must either co-operate in helping the man who initiates the move or else split off from the community. The building of a large longhouse requires a long period of communal labour and, when a distant new site has been chosen, there is a difficult interim period while new cultivations are being established. The man who successfully initiates, organises and recruits the labour for such a move achieves considerable prestige and, usually, the leadership of the new community. However, since he does a disproportionate amount of the work himself, he cannot be an old man. If an important member of the senior generation remains in the community, this man may retain ritual leadership while the young man, his son or brother's son, gradually assumes political and economic responsibility. There is frequently a period in which the community is ambiguously referred to as belonging to the old man and being of the young man's house, for house-ownership is always ascribed to the builder. Although leadership is ideally determined by descent and falls to the senior sibling of the oldest generation, there are many examples of a younger brother assuming the role of longhouse headman. The building of a house is the most important step in the process of gaining power, but an ambitious man also builds up his reputation as a dancer, chanter or shaman. Ideally, these ritual specialist roles are separate from leadership (according to the model of five siblings with interdependent specialist roles) but, in practice, personal initiative tends to be directed towards more than one specialist role (see p. 28). The longhouse interior I describe the house with reference to fig. 5. All houses conform to this basic plan: not all have space for an adequate dance path and not all have a rounded female end, but this does not affect the layout of the interior. The longhouse mainly consists of an enormous roof which almost meets the ground along the side walls. This roof is supported by massive vertical posts which structure the floor space within. The roof is oriented on a male—female axis; it is gabled at both ends on a 46

WOMEN'S DOOR

- C p..* ... X

I DANCESTANDING V X SPACE ^

MEN'S CHAWT

COMING-IN-STANDING SPACE

MEM'S DOOR Fig. 5 Ground-plan of longhouse interior 47

Mote' household objects hob to scale

Social structure square-ended house and only at the men's end on a round-ended house. The wall at the male end is covered on the outside by paintings on bark. Apart from this special painting surface, the roof and walls are all made of the leaves of various species of palm. The roofing leaves are woven onto palmwood slats and the remainder of the timber is hardwood; joints are bound with vine. There is a women's door opposite the men's door and family compartments are strung along the side walls towards the female end of the house. Within the longhouse community, individuals are grouped into nuclear families. Each married man of the local descent group has a compartment which houses himself, his wife and his children. This compartment contains a hearth around which the family hammocks are slung. Family stores of food are preserved in the smoke above it and the women do the cooking over the fire. All personal possessions, including many items of household equipment, are kept in this private family area. Single descent-group men do not have compartments of their own since these are the domain of conjugal life and child rearing. After initiation, a boy may continue to use his parents' compartment in the daytime, but he must sleep outside it in the open part of the house. When he brings a new wife into the community, she is attached to the household of his mother at first, and may even sleep in the parental compartment. The new couple gradually establish their economic independence and some time before the birth of their first child they are allotted a compartment of their own. The stated rule, which is usually but not invariably observed, is for the headman to occupy the compartment furthest towards the women's door on the right-hand side of the house (when looking out of the men's door). Apart from this there is no strict pattern other than a tendency for senior family heads to be nearest to the women's door. This follows from the way in which new compartments are screened off from the adjoining, existing ones. Visiting families occupy hearth spaces on the periphery nearest to the men's door but any initiated males travelling without wives sleep apart from their families in the middle of the house. In the open part of the house, communal meals are held; the formal men's group meets nightly and coca and manioc are processed. In general, both sexes use the central part while men monopolise the space in the centre front and women the space towards the female door. Both communal activities and private family activities are part of the life of every individual but the parts are differently balanced 48

The longhouse and its inhabitants according to sex. For men, life in the longhouse is mainly communal — much productive property and most ritual goods are communally used, while for women it is relatively private — most productive property is privately used and women are excluded from the nightly male gatherings and many aspects of communal ritual. In everyday activity, the sexes do not keep strictly to their appropriate ends of the house. However, the more formal the occasion, the more carefully the spatial differentiation is observed. Formality is appropriate when there are visitors, especially large numbers of visitors from far off, and it reaches its maximum during communal ritual gatherings when several complete longhouse communities gather together. With the presence of visitors, not only does the male-female axis become more binding, it also takes on the added significance of visitor/resident. On arrival, the visitors remain just inside the main door: the visiting men remain in the centre front of the house throughout their stay (apart from eating and dancing in the middle and sleeping on the periphery), but the women and children become absorbed into the female activities of the residents after a while. Social and economic organisation of the longhouse community The basic organisation of the longhouse population is thus into conjugal families headed by local descent-group members. The position of these round the periphery of the house towards the women's door contrasts with the sleeping position of initiated, unmarried men in the middle of the house. From the point of view of the development of the local descent group, these youths are in an intermediate position between childhood and fatherhood. In terms of longhouse space, then, the male life-cycle consists of two halves — a journey from periphery to centre, followed by a reverse journey back to the periphery. The female life-cycle is lived out entirely in peripheral compartments but it, too, is divided into two halves — the first living with agnates and the second living with affines in a separate community. The role of the conjugal family as a reproductive unit goes hand in hand with its role as an economic unit. The separate family units are almost autonomous in food production. Each possesses its own gardens cultivated by the wife and daughters. These may be felled communally by the men and then partitioned off, or each husband-and49

Social structure son team may fell for the women of the family. Sometimes an initiated youth approaching marriageable age sets up a felling-andcultivating unit with the unmarried sister next to him in seniority. The pair add their produce to that of their parents, the sister filling her mother's manioc basket and the brother adding his meat and fish to the family supply. In this way, the brother—sister pair are only partly differentiated from the rest of the family. The economic unit they form is intermediate between their role as children, still learning under the authority of their parents, and the truly responsible and independent unit formed after marriage. Pira-parana Indians observe the rule of sister-exchange, which means that the marriage of sons is accompanied by the loss of daughters. Thus, with the marriage of a sibling pair, the family economy receives a severe shock. The shock may be gradual if the daughter-in-law assists her mother-in-law at first and the couple continue to contribute to the family production, but, by the time the first child is born to the new couple, the original family has lost both a male and female producer. This loss certainly affects the parents of the original family, but relations between father and son remain close. Relations between mother and daughter-in-law are more variable and, directly or indirectly, the co-operation of the mother-in-law and unmarried sisters-in-law have a great influence on the success of new marriages. However, any tension that new marriages cause within the local descent group tends to be between the new husband and his siblings. The marriage of the oldest brother—sister pair begins a process of gradual separation of the economic interests of male siblings. As each marries and the number of small children he supports increases, so his economic effort becomes more oriented towards his family of procreation. His life-style becomes increasingly different from that of the older initiated youths who co-operate closely and are able to spend their spare time and energy in visiting and attending rituals. In general, these youths are relatively free from economic stress for they are members of family units with high proportions of productive members. Their married brothers head family units at the other end of the scale with a high proportion of unproductive members (small children). Although marriage is essential for the perpetuation of local descent groups, the division of economic interests that inevitably follows it means that it is also a threat to descent-group unity. The same kind of separation between descent-group members can be seen spatially, 50

The longhouse and its inhabitants because the unmarried, initiated men, who all sleep in the open centre of the house, are later housed in separate compartments. While all kinds of different biological, social and psychological factors are doubtless involved in the dispersal of sibling groups over time, there is good reason to describe the process in terms of food production, because this is such a sensitive matter in the internal politics of longhouse communities. If each productive unit consumed its own produce there would simply be different levels of consumption at different stages of the family development cycle. But, instead of this, family production is ideally followed by general consumption, each family offering its produce at communal meals in the centre of the house (see p. 172). In practice, there is a delicate balance between an acceptable level of private family consumption and an unacceptable level, which is interpreted as anti-social withholding of food from the community. On the one hand, to the extent that communal consumption is practised, it does indeed relieve the burden on relatively unproductive units. On the other, if people think that these unproductive units are choosing to eat privately rather than contribute their share to the community, resentment builds up. In a situation where each family is aware that every other family is continually exercising choice over how much food to release, there is ample room for speculation and ill-feeling. When these feelings become intolerable the community splits. Of course, there are many other reasons for community splits and I do not wish to deny their importance. People say that others are bad-tempered, adulterous, bewitching people, too bossy, too violent and so on. However, besides being the real basis for many disputes, food production and distribution is the perfect medium for expression of bad feelings. These processes play a very special role in longhouse society: food production is intimately tied to the reproduction of local descent groups and food distribution is a concrete manifestation of the tension between family and community interests that arises from the reproductive process itself. So far, we have seen that the local descent group is founded by a male sibling group. This is ordered by seniority and, ideally, it is a replica of the system of five specialist roles. Thus, it should give rise to an ordered and well-integrated longhouse community — the more so because the founding siblings are usually the senior living generation. However, according to the marriage rules, the group must exchange sisters for wives of diverse outside origin in order to perpetuate itself. Thus, new families are set up: the generations become 51

Social structure separated and the unity of sibling groups is destroyed. This divisive process is a natural feature of the reproduction of any group, but it is particularly potent in a patrilineal community with patrilineal residence and virilocal marriage. Besides, in Vaupes society this process is underlined by the economic system which stresses the simultaneous unity and separation of the various family units making up a community. It is precisely the child-bearing and child-rearing role of wives from outside that is responsible for the separation of generations and dissolution of unitary sibling groups; therefore marriage plays the same role in ongoing society as it does in the overall development of Indian society. According to Indian ideology, in this grand development, marriage and sexual reproduction bring the original ancestral 'pre-descent' era to an end, and initiate a new 'descent' era in which patrilineal groups receive separate identity and become dependent on women's reproductive powers (see p. 38). In the 'descent' era, which continues at present, the same process can be seen repeated on a small scale for each generation; marriage and sexual reproduction bring the unity of the sibling group to an end and replace it with several family units in which men depend upon wives instead of sisters. This process sets up an opposition between two situations which may be expressed or manifested in various forms. The unified community is opposed to the family organisation; men in their capacity as descent-group members are opposed to men in their capacity as family members; women as descent-group sisters are opposed to women as wives in affinal communities; men who provide the structure of the local descent group are opposed to women who co-operate in reproducing affinal groups. Finally, these oppositions are related to that between the unified, hierarchical internal structure of descent groups and the external relations of equal reciprocity formed with affinal groups, which were discussed above. These are not just fanciful patterns - they are the result of processes which order the reproduction of the longhouse community. In the remainder of this book I hope to show that they order many other aspects of social and economic life. In summary, we can isolate two aspects of male life and two aspects of female life. Each sex has a descent-group aspect and a procreative aspect. Men are brothers and husbands while women are sisters and wives. In the case of each sex we shall see that the two aspects represent opposite phases of a repeated cycle; we already have an example of this for men, in the way in which they move between separate 52

The longhouse and its inhabitants family compartments and the communal centre of the house during their life-cycle.

53

The set of specialist roles

Erosion of the system The set of five specialist roles and the practical functioning, or lack of it, of the system have been briefly described already. This chapter is about the internal structure of the role-set and its analogy with other structures or processes. First of all, the ideal behaviour associated with the roles must be presented in more detail. The difficulty with this, as suggested before, is that sib roles barely function at all at a practical level and therefore I am obliged to present a system, which above all applies to specialist groups, in terms of individual specialist roles. Of the specialist roles performed by individuals, there are only three that are well defined at present. These three are shaman, chanter and dancer, all of which belong to the metaphysical domain and all of which are recognised in a title added to the expert's personal name (thus X-kumu, Y-yoamu and Z-baya are shaman, chanter and dancer respectively). These three roles only cover two of the five roles in the complete set, since dancer and chanter belong to a single category. The others - chief, warrior and servant - must be described from indirect sources, including statements about the past. The content of the warrior role is the most problematic, for the ideal behaviour of chiefs and servants can be seen in attenuated form today as, for example, in the relationship between an important longhouse leader and a youthful 'extra resident'. The role of servant is also clearly regarded as a less extreme version of the role of the Makii, who still inhabit some parts of the Vaupes today. A further complication lies in the fact that the present-day absence of three of the five roles has probably led to modification of the remaining two. The missing roles of chief, warrior and servant indi54

Erosion of the system cate the importance of (a) hierarchy and (b) use of physical force in social organisation. While both factors remain as ideals, they are tempered in practice by the opposing ideals of equality (among all Tukanoans) and peace. Of course, tension between the paired ideals must have always existed because rigid social hierarchy and indiscriminate killing are incompatible with the perpetuation of smallscale, forest-dwelling societies. However, confrontation with whites and the consequent acculturation have shifted the equilibrium point further away from hierarchy and violence and further towards equality and peace. Even so, people associate both hierarchy and violence with all the advantages of the past - the larger scale of ritual life, the independence of Indians from white outsiders and the physical and cultural strength of past generations. They regret that no one knows how to give orders or how to take proper revenge for wrongdoing today, but, even so, those who do so are resented. Instead of being regarded as 'chiefly' or 'fierce', they are 'bossy' or 'evil-tempered'. It may well be that it is this ambiguity of public opinion which influences individuals with leadership ambitions to become specialists in the metaphysical domain, where they may gain legitimate respect. Of the metaphysical experts — dancers, chanters and shamans — shamans make by far the widest reputations. Although shamans are respected by those who use their services, people from other communities often fear and dislike them for their dangerous powers. Clearly, the shaman, who can kill at a distance, has the ultimate power when direct killing is prohibited by public opinion. As the roles of chief, warrior and servant become increasingly irrelevant to the ongoing social process, the shaman becomes the best candidate for the position of primus inter pares or community leader. If we can assume that formerly there actually were specialist warriors, chiefs and servants, we can guess that the shaman was effectively controlled by his dependence on others. His metaphysical killing powers were matched by the concrete powers of the warrior, and people actually say that dangerous shamans were often killed for abuse of power. Both warrior and shaman were under the authority of the chief and economically dependent on the productive system based on chiefservant relations. This would make the shaman more of an equal to the dancer/chanter (with whom he co-operated in the metaphysical care of the community) than he is today, now that his powers have been let loose. 55

The set of specialist roles The expansion of the shaman's role into the politico-economic domain has made present-day shaman—leaders relinquish certain aspects of the ideal shaman's role which are incompatible with leadership. Shamans should withdraw from the procreative aspect of women while leaders must be married, preferably polygamously, in order to maintain their position (see below). There are shamans who are not leaders: these are sometimes unmarried and often living a relatively isolated social existence. The shaman-leaders, by contrast, are invariably married. People comment on this situation, saying that leaders cannot be good shamans. They also recognise inconsistencies in the seniority position of shaman-leaders within their own sibling groups. A shaman should be a younger brother and a leader should be a first born: therefore, a shaman-leader is bound to be wrong on one count or the other. The specialist roles I mentioned that the organisation of the set of specialist roles may be conceived of in two separate ways - first, as a simple hierarchy derived from the association of the set of roles with sibling order from first born to last born; and second, as a bilaterally symmetrical organisation based on the interrelations of three domains (p. 20). If we concentrate on the three domains, these may be arranged concentrically with the politico-economic domain on the outside, the externally oriented domain of warriors in the centre and the metaphysical domain mediating between the two. In this way, a model which

CHIEF DANCER/CHANTER, WARRIOR. SHAMAKI SERVANT

EXTERNALLY OMENTED DOMAIN

Fig. 6 Relation between concentric and hierarchical arrangements of specialist roles 56

The specialist roles simultaneously represents both hierarchical and concentric organisation can be constructed. In the following description, the specialist roles are ordered according to the three domains. At the outset, I should stress once more that I am discussing the ideal conceptions of these roles rather than the sociological functions of contemporary institutions. My account of the content of the shaman and dancer/chanter roles is largely based on present-day behaviour of individual specialists; there are no true chiefs or servants, although there are powerful longhouse leaders and dependent 'extra residents': there are no warriors, although there are individuals who are described as particularly 'fierce'. Politico-economic domain: chiefs—servants Chiefs

Servants

uhard (chiefs) rihoana (head-people) rotird (people who order) umato moari masa (instigators of work)

hosaa (servants) gahariana (end-people) muno yori masa (cigar lighters) moari masa (workers)

a

also the name for the true Makii.

As 'first born', the chief comes first on all occasions where an order of precedence is observed — greeting, receiving of food, coca etc., sitting for formal conversation and travelling by foot or canoe; the servant as an 'end-person' comes either last or not at all. The relationship of chief and servant is the extreme expression of the principle of hierarchy. Both are outside the ongoing social process: the one from above and the other from below. The chief represents his group in formal relations with outside groups. Actual communication with outside groups by means of shamanism, warfare and ritual gatherings should be ordered and/or initiated by the chief and then carried through by the appropriate specialists. However, there is a remaining form of communication with outside groups which is closely bound up with the others and which is reflected in the extreme positions of chief and servant in the hierarchy of specialist roles. This is marriage. Polygamy is the prerogative of the chief, while the servants are described as 'marrying each other like animals' instead of making exogamous marriages like members of other sibs. Today, there are several cases of longhouse headmen with two wives in the Pira-parana, and one has three wives. Polygamy is directly associated with politico-economic power: it is an 57

The set of specialist roles expression of superior ability to make affinal relations. On a practical level, the surplus of manioc products it brings to the leader's household allows him to hold rituals and thus gain prestige outside his immediate agnatic group. According to informants, both the frequency of polygamy and the number of wives per leader have decreased over the last two generations. The endogamy of servants is important ideologically, since it indicates the relatively sub-human status of servants and their lack of engagement in external social relations. However, in practice, endogamous marriages are extremely rare and are invariably described as wrong and incestuous. In spite of this, there was a striking example of marriage between close agnatic kin among the servant sib that corresponded to our host sib (of warriors). People said that such marriages were commonplace amongst this sib in the past. This case was one of the few examples of presentday practice of specialist roles at sib level that was pointed out to us.1 Among the activities ordered by the chief and carried out by servants, people mention building houses, felling manioc gardens, cultivating and processing crops, bringing food and materials from the forest, cleaning the house, lighting cigars, keeping the resin or taper light burning in the house during the evening and putting away ritual possessions. The roles of both servant and chief are sometimes seen as divided by sex, so that the male tasks are performed by male servants directly for the chief, while the female tasks are performed by female servants, immediately under the direction of the chief's wife, but ultimately also for the benefit of the chief. If we concentrate on the male aspect of the servants' role then this too appears sexually ambiguous. The male servant may be called on to fetch water and clean up - tasks which are allocated to women in the normal way. During ritual occasions, the servants must take charge of fire by lighting cigars and lighting the house. Domestic fire is usually controlled by women and kept separate from male ritual life, and thus the servants' ritual role is to protect other men by taking on the 'anti-ritual', female element. Their sexual ambiguity is consistent with their tendency to endogamous marriages. 1 Data provided by Silverwood-Cope (1972) and Reid (personal communication) show that the Makii conceive of their own society as divided into patrilineal exogamous groups internally divided into sibs or 'clans', which are ordered according to seniority. This ideal model corresponds closely to the Tukanoan model, but both the above authors found that in practice the Makii are much less strict about observing exogamy than the Tukanoans. Makii residential groupings and also the constituent hearth-groupings frequently contain affinally related males. By marrying each other, members of the Tukanoan servant sibs are behaving like the true Makii.

58

The specialist roles Although the servants share the designation of the true Makii as hosa, they are clearly differentiated from these.2 Their descent status as members of Tukanoan Exogamous Groups gives them the right to live and work amongst their 'elder brothers', while the true Makii live away from the Tukanoan masters and far from main rivers. The relationship between Tukanoans and Makii, besides being one of master and servant, is one of exchange between cultivated products from Tukanoans and forest products and labour from the Makii. The forest products are curare poison and game, as well as the animal skins, teeth and bones, and fruits, all of which are used to make ritual ornaments. Although the Tukanoan servants are regarded as civilised beings in comparison with Makii, there is a sense in which they share the Makii's proximity to the natural order. They are marginal to the social life of the community and provide a transition between the natural and social. Their labour converts forest to cultivation site and raw materials to processed ones: these are the productive processes that precede communal consumption on ordinary social or ritual occasions. With respect to the order of events on ritual occasions, the chief greets the guests, so initiating the ritual sequence, and the servants put away the ritual possessions at the end. Besides actively ending the ritual period, they take the 'anti-ritual' fire-lighting role during it. Finally, the case of ritual sequence may be used to underline the dual nature of the chief-servant relationship. On the one hand, they are opposed as superior and inferior, positive and negative extremes of the social hierarchy, but on the other, they are alike in belonging to the politico-economic domain. While from the point of view of ritual sequence the chief comes first and the servant comes last, from the point of view of the entire sequence including preparation, the productive unit formed of chief and servant creates the transition between secular and ritual time by transforming raw materials into the manioc beer, coca and tobacco required during rituals. Thus, the politico-economic domain is prior to the metaphysical domain in the sense that its products are the prerequisites of ritual life. This priority has a positive aspect manifest in the ordering and initiating role of chiefs and a negative aspect manifest in the natural and menial role of servants. 2 Hosa also means 'difficult' and therefore it is possible that the name shared by Tukanoan servant groups and Maku alike refers to their recalcitrance (from the point of view of dominant groups).

59

The set of specialist roles Metaphysical domain: dancerI chanters-shamans Dancer/chanters

Shamans

keti masa (possessors of myth) bayaroa (dancers)

kumua (shamans) werea koa baseri masa (shamanisers

of the beeswax gourd) muno baseri masa (shamanisers of tobacco) The dancer/chanter category as a whole covers the performers of ritual, the guardians of keti. While keti refers to any story about recent or mythical events, in this context it means stories of 'old people' or ancestors (bukurd keti); a body of oral literature roughly corresponding to our 'myth'. The dancer/chanters recreate ancestral times by alternate dancing and chanting on ritual occasions. The dancing is led by an expert dancer, or pair of dancers, who stands in the middle of the dance line, with the other dancer/chanters graded towards the two ends according to expertise. All are painted and dressed in the full complement of ritual ornaments. In the standard form of dance the line moves round the dance path in alternate anticlockwise and clockwise directions, singing and often using maracas or other instruments. Women join in at a given point after the beginning of the dance and leave before the end; they often dance with babies on their hips. In between dance sessions, the community of male dancer/chanters sit on long benches, the guests opposite to their hosts, just inside the men's door. An expert chanter, sitting in the middle of one of the benches, leads the chant and the others repeat his phrases in unison. My understanding of these chants is scant, since their linguistic form is very different from ordinary speech and many substitutes based on metaphor, metonymy and mythical allusion are used instead of everyday vocabulary. They deal with mythical events that are arranged within the broad framework of an ancestral anaconda journey upriver from the east, and particularly with the original receipt of the plants, ritual materials and elements of behaviour used in the ritual performance as a whole. Thus, they are closely related to narrative myth, but are especially concerned with the link between the present and ancestral past. The stages of anaconda journeys, for instance, are virtually untellable in narrative form - if asked about them, informants invariably lapse into chant. The term for chanter, yoamu, literally means 'one who travels far' and the expertise of the chanter is judged by the yoamard (chanters)

60

The specialist roles detail with which he follows the ancestral journey: the best chanters go up and down all the tiny sidestreams incorporating the lesserknown mythical sites into their chant. The appropriate chant sessions for any given ritual are performed in a set order and synchronised with the events to which they relate. If we isolate the dancing and chanting elements of ritual, these two activities are analogous to the two opposed phases in the anaconda journey. Travelling upstream in anaconda form is analogous to chanting and stopping on land to dance is analogous to dancing: chanting represents ongoing creative activity and dancing represents the results of creation (see p. 35). Since the final emergence onto land coincides with the transition from the 'pre-descent' era to the 'descent' era, it is appropriate that women and babies are incorporated into the dance sessions but are precluded from the chant. The specialisation into dancers and chanters within the comprehensive category of keti masa thus corresponds to a basic opposition between the two phases of creation. As a unitary category, the keti masa are like the ancestors when they were in the process of becoming like present-day people. Today, only the officiating shaman remains apart from the dancing and chanting and does not wear the feathers and ornaments of the dancer/chanters. Whatever the situation in the past, the rest of the initiated population all participate, and only the leaders of each activity need a specialist skill. If the dancer/chanters re-enact ancestral activity as a group, it is the shaman as an individual, or sometimes two paired shamans, who makes this possible. He treats the beer, drugs, paint, ornaments and other ritual substances which change the bodily state of the dancer/ chanters from secular to ritual, and from mortal to ancestral. The shamanic activity both protects against the dangers of ancestral contact and promotes the beneficial effects. The shaman's transformational power is received from the ancestors whom he contacts by travel to the sky. The shaman's role is not performed entirely during ritual, for he treats categories of food for individuals undergoing life changes (p. 122), and cures illness as the need arises. In the past, shamans were also responsible for releasing game animals from their mystical houses in the forest. The houses appear to ordinary people as outcrops of rock, each one associated with a particular species. The shaman travels in spirit to these houses and opens the door, releasing a controlled number of animals (with a jaguar to prey on them accompanying 61

The set of specialist roles each batch) for which he must pay with children's souls (p. 200). Shamanism is also intimately associated with the weather, particularly with thunder, as one would expect from the vertical shamanic journeys into the sky. In conversation, a distinction is made between shamans who only 'see' the ancestors and those who really travel to the sky; some say that only the great shamans of the past could travel. There are also grades of shamanic activity, the least demanding being shamanism of foods after minor life changes, and the most demanding being shamanism of the sacred beeswax and tobacco gourds during initiation (see above for the expressions denoting 'shamans). The act of shamanising, base-, is mythically associated with the adventures of Manioc-stick Anaconda (pp. 88, 261) in which it is stressed that shamanic powers derive from the Sun's tobacco snuff and involve crossing the vertical layers of the Universe. Birth shamanism is acquired by Manioc-stick Anaconda in the form of a tick, also base, sucking his host's blood. Ideally, shamans should remain apart from the procreative aspect of women, for 'women and children are like illness (nyase) to them'; they should not marry, have sexual intercourse or eat hot foods (which are associated with female sexuality), although they may have women working for them. Shamanism is acquired by a protracted learning period involving a special initiation rite in which the pupil is secluded and receives drugs. The pupil eventually replaces (wasoa-) his master, so that we may conceive of shamanism as a perpetual link with the ancestral state which is maintained by a succession of mortals. The positive, life-giving and protective powers of shamans are accompanied by destructive powers. Ideally, the former are used on behalf of agnates while the latter are directed towards members of outside groups. If we compare shamans to dancer/chanters, we see that shamans are of this world and yet able to mediate with the other, ancestral world, while dancer/chanters are actually part of the ancestral world when performing their role. Both are mediators with the ancestral world, but the shaman's transforming power is more concerned with the ever-present bearing of the ancestral world on the natural or physiological processes of this world, while the dancer/chanters are concerned with the developmental creation of society and cultural elements. Thus the shaman deals with food, drugs, illness, growth, sexuality, supply of game and so on, while the dancer/chanters reenact the development of Tukanoan culture using the elaborated 62

The specialist roles plant products, paints, ornaments and instruments which are its expression. The complementary opposition of the two roles is also expressed in their different spatial emphases, for the shaman operates in vertical space, crossing cosmological layers, and the dancer/chanters in horizontal space, travelling upstream from the east. Domain of competitive intergroup relations: warriors Warriors guamard (fierce ones) guariyaia (fierce predators - literally 'jaguars') masa siari masa (people-killing people) As far as can be ascertained, the causes of raiding and killing before pacification were revenge for evil shamanism, revenge for direct killing, disputes over women and desire for ritual goods and, possibly, for land. It is impossible to know whether people actually decided to kill for ritual goods and land or whether these were simply redistributed as a result of hostilities begun for other reasons. Indians deny that they killed strangers simply to get wives. The warrior intending to kill covered himself with red paint. It seems that there was special shamanism associated with killing and that the killer observed stringent prohibitions afterwards. Various elements of male ritual, particularly the 'imitating spearing' during fruit rites and male initiation (p. 143) are designed to make men into fierce warriors who 'see their enemies like game animals'. Bathing in the river before dawn and the general avoidance of laziness and sleep are also conducive to the male energy manifest in fierceness. It is clear that, in the past, warfare, communal rituals, shamanism and marriage were interrelated. These were the principal modes of intergroup communication and one mode would easily give way to another. People say that both the ritual giving of food appropriate between affinal groups and ritual greeting in general are like fighting'. There is certainly a strong element of aggressive competition in ritual gatherings as both hosts and guests try to outdo each other in their ability to 'take' their yage and to drink enormous quantities of beer. Today, the joking which fills pauses in the chanting and dancing routine readily becomes insulting and stories are told of dances at which treacherous killing or evil shamanism were performed. Such treachery is directly linked to ritual food-giving in a myth where the Wood Ibises (Mycteria americana) bring a ritual gift of ants to their 63

The set of specialist roles mother-in-law (Romi Kumu) and are attacked and killed by Pouncing Jaguar (Tdho yai) as they enter her house. It seems that hitting and clubbing duels were also held among young men. Killing must therefore be regarded as an extreme act belonging to a set of types of combat ranging from friendly, rule-bound competition to be unrestrained use of lethal weapons. This past situation, with graded aggression, treacherous feasts and shifting alliances between groups, is reminiscent of that described for the Y^nomano (Chagnon 1968). A further example of graded hostilities occurs in the range of wifegetting procedures. Marriages which are part of a pre-arranged exchange take place peaceably, while the removal of the potential wife becomes more and more violent as social distance between groups of the potential spouses increases. Although potential wives are often taken by raiding, there is a distinct difference between raiding to kill and raiding for a wife, as the Tukano text on p. 99 shows. Of course, this does not preclude the truly warlike raid from resulting in the abduction of women, nor the wife-getting raid from ending up in killing; it merely means that wife-raiding is classified as a less hostile act than killing. Put in another way, destruction of another group's patrilineal procreative power (killing a man) is a more extreme form of violence than augmenting the procreative power of one's own descent group at another's expense (by stealing a women and thus avoiding reciprocation of a sister). All these examples of graded hostilities suggest that the ideal warrior role which is specifically related to killing ('people-killing people') exists as an extreme negative form related to other types of concrete interaction between individuals (graded types of physical competition) and between groups (graded marriage procedures, ritual gatherings etc.). I take up this point below (pp. 69ff) and then again in the discussion of marriage practice (pp. 93ff). Analysis The above description of the five specialist roles should leave no doubt as to their interdependence, for each one is necessary to the maintenance of the life and wellbeing of any corporate group within a traditional setting. The fact that the Exogamous Group is not, and may never have been, an effective corporate group does not prevent this model from providing a cogent expression of its unity. The analysis that follows treats the set of roles as an ideological system and 64

Analysis explores the relations between the five with the help of the data already provided. Analogy with life stages We call the structure hierarchical because the set of specialist roles are associated with the 'hierarchical' birth order of the occupant 'siblings' (groups or individuals). The birth order of these siblings may be understood as a political order, with the older sibling always having the superior power but, when we look at the specialist roles in isolation from sibling relations, it seems that this political order is only reflected in the outer rbles of chief and servant. The position of these roles gives the entire set the appearance of a hierarchical order but, in fact, there is nothing obvious in the intrinsic nature of the middle three roles (dancer/chanter, warrior and shaman) to suggest why they should occur in this particular order. Quite the contrary for, in terms of political power, there are reasons for considering shamans 'superior' to dancer/chanters, but this 'superiority' is contradicted in the role order (pp. 55f). Since a hierarchy must have a single organising principle, we must look for alternative explanations which account for the qualitative differences between the roles. Further reflection on the composition of the sibling group suggests an opening, for, at any one time,, the birth order is manifest in the different ages of the individual siblings - that is, in their attainment of different stages in the life-cycle. The ordering of the roles by analogy with the life-cycle does provide an adequate explanation of the upwards sequence of shaman, warrior, dancer/chanter. Let me begin by presenting the evidence for this, starting at the bottom with servants. First of all, servants are outside or beyond the mainstream social process, separating their superiors from the elementary productive and menial tasks that tie society to its natural environment. In the same way, childhood comes between the dead, unborn state and adult social competence. Servants are also sexually ambiguous: there are both sexes of servant, but even the males belong to the female domain in relation to the rest of male society. The position of servants is like that of children who may be of either sex but are not properly sexually differentiated until the males reach initiation and leave the female domain. Like children, servants are under the authority of all other sections of society. Just as children are pre-social in the sense that they will be integrated into social life at initiation or menarche, ser65

The set of specialist roles vants are pre-social in the sense that the products of their labour are later consumed or used in the social life of their superiors. Servants are denied communication with outside Exogamous Groups because they are excluded from active participation in ritual. They are further isolated because they marry each other. In the same way, children are kept secluded from visitors and do not greet them; they are sexually inactive and therefore detached from the general preoccupation with marriage exchanges. The shaman's role contrasts markedly with that of the servant. Instead of being an undifferentiated member of the female domain, the shaman regards women and children as threatening. Instead of marrying his sister he marries no one. In this, he is like a boy initiate who is abruptly severed from the domain of women and protected from their procreative aspect (ch. 5). His role as supreme mediator, between present-day and ancestral times and between secular and ritual time within the contemporary cycle, is parallel to the position of the initiate, who is midway between childhood and the society of unmarried men who represent descent-group unity (p. 52). We have seen that the opposition between family units and descent-group unity is analogous to that between the present-day and the ancestral beginnings: it is this same analogy that applies to the already initiated youth who has been united with other unmarried descent-group males through contact with ancestors. It is precisely the shaman who achieves the transformation of children into unmarried youths by releasing his 'returned ones' into heterosexual society after initiation rites (p. 145). From the point of view of life-cycle development, the initiated youths have reached the stage when they will obtain wives prior to setting up family compartments and rearing children. The direct contact which warriors make with outside Exogamous Groups is therefore analogous to the direct contact made by youths in procuring wives. Just as the warrior role is the one directed outwards in the specialist-role set, so the marrying stage is the one directed outwards in the life-cycle process. It is followed by the creation of the family unit and the production of children, who are members of the descent group and are thus 'insiders'. It is important to digress here in order to underline the fact that this analogy between marriage and the warrior role only applies in the context of the wider analogy between the life-cycle process and the specialist-role system. While the warrior role is the only member of the set to be exclusively outwardly directed and not complemented 66

Analysis by a paired Internal' role, warfare is but one of a number of modes of interaction with outside groups. Shamanism and ritual gatherings are other modes which also depend upon activation of specialist roles. Besides this, if warfare and marriage are considered as alternative modes of interaction they are clearly differentiated as more hostile and less hostile respectively. These facts have already been mentioned above (pp. 63f), and while they do not undermine the argument here they do indicate the need for further analysis of the significance of the set of roles for external relations of the Exogamous Group. These are dealt with below (pp. 69ff) and again in relation to marriage rules and practices (pp. 93ff). Dancer/chanters are parallel to married men in the phase during which these build up a family of procreation. In their dancing and chanting, the specialists reproduce the creation of hierarchically ordered descent groups, and in their family roles married men create ordered sibling groups of children. The two phases of ritual, chanting (upriver) and dancing (round on land), can be seen as analogous to the relationship between penis and womb - the opposed creative elements in the process of sexual reproduction - which actually results in sibling groups of children. This completes a series of analogies, some of which have already been established (pp. 35, 61): land water results creation dancing (with women) chanting (without women) descent era pre-descent era family units descent-group unity womb penis The particular analogy between the phases of ritual and elements of sexual reproduction is supported by the exclusion of women and children from chant sessions and their incorporation into dance sessions. In fact, there is a special, intermediary female role of 'answering' the dance, (basa kudi-). The woman who performs it stands in the centre of the house surrounded by dancers. She holds a small child on her hip and sings on a high-pitched, continuous note which contrasts with the deep, rhythmical sounds of the dancing. It is tempting to regard this woman as representing the product of the opposed sexual elements - as the foetus itself. I suspect that this woman-and-child unit stands for the pregnant Meneriyo in the house of the dancing Jaguars, although I have no direct evidence to this effect. Warimi, Meneriyo's child, suffered a violent and premature 67

The set of specialist roles birth when his mother was killed by the dancing Jaguars, and he was known for his high, unintelligible speech and crying (S. Hugh Jones 1979 : M.4.D). The position of chief is parallel to that of the married man whose children are growing. Like the chief who has complete authority over the servants, the father has complete authority over his children until they are initiated and come under partial control of the shaman and other ritual kin (p. 145). An analysis of the relationship between the specialist-role system, the categories of participants in initiation rituals and the different types of flutes and trumpets used during these rituals is contained in Stephen Hugh-Jones's analysis of initiation rites (1979). His analysis produces very much the same conclusion about the homology between life-cycle development and the set of specialist roles as my own. However, by independent analysis of a more comprehensive set of data mainly drawn from outside the context of He wi rituals, I have been able to present this analogy in more detail. In general, the data on initiation, particularly those concerning the classification of flutes and trumpets according to specialist roles, provide strong support for the analogies I have described. The five-stage process described in relation to the five specialist roles clearly does not cover a complete life span from birth to death, since it begins and ends with a parent—child relationship. It is a cycle of social reproduction which spans a single generation and contains the conversion of a male descent-group child into a father. In the basic analogy drawn above between the set of specialist roles and the five stages of reproduction of a descent-group generation, there is an obvious correspondence with the organisation of longhouse space. In the discussion of development in the longhouse group, we saw that childhood is spent in the family domain represented spatially by the peripheral family compartments. Later, initiation transfers growing boys into a unified society of descent-group youths who sleep in the centre of the house. Sometime after these youths have married, they start families of procreation, which moves them back into peripheral compartments in the status of fathers. This process is one of cyclical creation and dispersal of single unified generations of local descent-group males, the creative movement operates from periphery to centre and the dispersing movement from centre to periphery. In the parallel specialist-role system, the intermediate transformative stages of initiation and sexual reproduction are represented 68

Analysis by shamans and dancer/chanters respectively. Appropriately, both of these operate in the metaphysical domain concerned with mediation between the ancestral past and the present. We have already seen that the unified group of youths contrasts with the parent-and-children units within the family compartments in the same way that the 'predescent' era of the mythical beginnings contrasts with the contemporary 'descent' era, in which reproduction depends upon women. Thus the shamans and dancer/chanters occupy positions which are metaphorically in-between the past and the present in accordance with their specialist abilities. In support of this, we may note the position of the dance-path, which is between the central space bounded by the four main house posts and the peripheral compartments, and the position of the chanting group between the main door and the centre of the house (see fig. 5). The analysis so far is summarised in fig. 7. Analogy with external relations I have emphasised the aspect of progression through the five life stages which is internal to the descent group, but the reproduction of the descent group is dependent on the external relations with those groups which provide the focal community with wives. The separate substance of opposed Exogamous Groups, manifest in their marriageable daughters, must be brought under control through marriage and transformed into descent-group children by sexual reproduction. The passage from childhood to parenthood is marked by a change from the absence of affinal relations to the formation and utilisation of these relations in the post-initiation phase of life. If we turn back to the set of specialist roles, the extreme positions of chief and servant are appropriately marked by exaggerated exogamy (polygamy) and endogamy respectively. This is enough to suggest that the series may be read as a progression from negative relations with outside groups towards positive relations with outside groups. The steps of the progression are not directly apparent in the characterisation of the middle three roles, for although shamans should remain unmarried (therefore they are neither endogamous nor exogamous), there is no mention of the correct marriage practice for warriors or for dancer/chanters. As with the political hierarchy, the chief—servant relationship sets the extreme positions and the inner three roles mark the transformation between the two in a less direct 69

The set of specialist roles manner. Following the earlier observation that each of these three roles is relevant in external affairs, it is now possible to show that they may be graded from negative to positive according to the character of the external relations they create. Servants, at the bottom of the hierarchy, have no external contact. Shamans communicate with destructive metaphysical powers, 'seeing' or 'travelling to' the enemy, but remaining at home in physical form. The warrior actually sets out to the enemy longhouse using the same physical energy to travel and to fight: the combat is concrete and face to face. Dancer/chanters perform opposite the guest dancer/chanters

PARENTHOOD -

-CHIEF

PROCREATION

-DANCER/CHANTER

YOUTH

WARRIOR

INITIATION

SHAMAN

CHILDHOOD-

-—SERVANT

Fig. 7 Concentric organisation of specialist roles and longhouse structure 70

Analysis in a situation of semi-friendly contact: their interaction with outsiders is at the same time co-operative and competitive. Chiefs, at the top of the hierarchy, organise and initiate all these different modes of contact. Leaving aside the extreme roles of chief and servant, the outwardly directed aspects of the three middle roles may be arranged thus: shamans

warriors

metaphysical aggression

direct physical aggression

dancer/chanters direct physical co-operation mingled with controlled aggression

Using the evidence contained in the description of the specialist roles, it seems that each of the five roles is basically dual, having an internal and an external aspect. The servants support the material existence of the descent group and have no contact with outside groups; the shamans promote life and growth within and use the same powers to promote illness and death of outsiders; the warriors defend and protect the descent group and attack and kill outsiders; the dancer/chanters provide a link with descent-group origins within und perform in the context of communal ritual gatherings to which outside groups are invited; the chiefs order internal affairs and represent the entire descent group in conducting external relations. Set out in this way, the internal functioning of the differentiated descent group appears to be intimately related to its external functioning. The series can be read from the bottom upwards as a transformation of self-sufficiency, which implies negative relations with outside groups, into positive interdependence between groups. The extreme of endogamy is connected by increasingly secure and controlled external relations to the opposite extreme of polygamous exogamy. Warfare, for instance, although it is a risky form of negative communication, does actually require physical contact and there is always the possibility that the grievance will be resolved by formal discussion instead of escalating violence. Warfare is therefore more controlled and potentially constructive than aggressive shamanism, which can only result in metaphysical or warlike reprisals. Similarly, exchange of ritual requiring the activities of dancer/chanters represents more friendly relations than does warfare. However, in the event of the ritual gathering, the result of interaction between hosts and guests may be either consolidation or deterioration of the friendship (p. 63). Metaphysical aggression, warfare and exchange of ritual invitations are therefore alternative modes of communication between 71

The set of specialist roles groups, and one may give way to another. They are associated with the hierarchically organised roles of shaman, warrior and dancer/ chanter in such a way that the set of roles represents the process of befriending outside groups when read upwards, and of estrangement from outside groups when read downwards. If this interpretation is accepted, then it would seem that exogamous marriage, which is emphasised in the polygamy of chiefs, represents the extreme form of 'friendship' or interdependence between groups. While this is true of certain marriage links, it is not true of others, for the degree of friendliness and mutual trust between groups linked by marriage is highly variable. For the moment, we may anticipate the discussion of marriage rules and practices in order to state that the closest, most friendly and mutually beneficial relations between groups who are not agnates are between groups related by ties of perpetual marriage alliance. In chapter 4,1 shall present evidence to show that relations of perpetual alliance between groups are regarded as more secure than relations involving exchange of ritual invitations. Provisionally, I assume that the position of chief in the specialistrole system is analogous to that of perpetual alliance in the hierarchical order of modes of communication between Exogamous Groups. The results of the analysis so far are set out in fig. 8. The above interpretation of the specialist-role hierarchy portrays it as a progression from isolation and self-sufficiency towards mutual social interdependence between groups achieved through the exchange of women. This is entirely consistent with the previous finding that the hierarchy is analogous to the passage of growth from childhood to fatherhood. The uninitiated child has no significant interaction with outsiders; initiation and future development represent the beginning of such interaction and its consolidation in marriage and procreation. It also follows from my interpretation that the stages - either modes of communication with outside groups or specialist roles — also represent relations with outsiders across decreasing social distance reading from bottom (too far for any contact) to top (close affinal neighbours). The sequence may either be read as a temporal progression in befriending a single previously unknown group or as a synchronic structure in which all outside groups may be graded according to social distance from the group of reference. In as much as social and geographical distance correspond to one another, we would expect accusations of shamanic killing to be levelled at the most dis72

Analysis tant known groups, war waged against less distant groups, ritual invitations exchanged with moderately close groups and perpetual marriage alliance relations to be established with the closest groups. It is difficult to establish the existence of such a pattern in the absence of warfare, but if warfare is merged with aggressive shamanism the resulting model of intergroup communication does fit the present-day situation moderately well (fig. 8). There are two points to make about the application of this model to the present-day situation. The first and more specific one concerns the exchange of invitations to ritual. This is common between neighbouring communities linked by repeated marriages as well as between slightly more distant groups. However, ritual gatherings vary considerably in type: the opposition between hosts and guests is most emphasised, and latent hostility is most apparent, during food-exchange rituals. Ritual food exchange is usually organised between moderately distant communities, and during the ritual itself it is common to see the principal donors or recipients incorporate their closest affines into their party. The ideal relationship between partners in ritual food exchange is discussed later (p. 90) and the variety of ritual gatherings is mentioned during the analysis of male initiation rites (pp. 142ff). The second point is that the local descent group, rather than the Exogamous Group as a whole, is the meaningful unit in most types of interaction between communities. I have implied so far that 'outsiders' are all potential affines but, from the point of view of any given longhouse group, this is not true, because many of the surrounding communities will belong to the same Exogamous Group. Indians say that aggressive shamanism, warfare and ritual food exchange are improper modes of communication between agnatically related communities, but accusations of shamanism, stories of past battles and actual patterns of food exchange do not always bear this out. However, the Exogamous Groups always retain their function of exogamy so that, in spite of indulging in certain 'improper' exchanges, agnatically related communities do not intermarry. To sum up so far: the five specialist roles are arranged in hierarchical order, according to the birth order of the founders of the sibs which possess these roles. They may also be arranged in a concentric order according to the three domains among which they are distributed. While the opposition between the extreme roles of chief and servant is self-evident from the nature of the roles, the ordering of the intermediate roles - shaman, warrior and dancer/chanter — 73

FUNCTIONING OF EX 0 0 A MO US GROUP

INTERNAL ASPECT CONTROL Of INTERNAL FUNCTIONING

RITUAL CONTACT WITH ANCESTRAL WORLD

SPECIALIST

ROLES CHIEFS

DANCER/ CHANTERS

MODE OF COMMUNICATION WITH OUTSIDE EXOGAMOUS GROUPS

EXTERNAL ASPECT

EXCHANGE OF WOMEN

CHANOE IN INTEREXOGAMOUS GROUP DELATIONS CL-OS;E AFF IN

BS

TBhiUOVS CO-OPERATION

1 PHYSICAL PROTECTION FROM OUTSIDERS

WARRIORS

PHYSICAL DESTRUCTION • now J suppressed

METAPHYSICAL CARE

SHAMANS

8

?

*

i

y METAPHYSICAL DESTRUCTION >

PRODUCTIVE LABOUR

SERVANTS

NO CONTACT

Fig. 8 Functioning of internal and external aspects of specialist roles 74

I

Dl 5T>KNT UNkNO IVKI GROUPS

Analysis requires explanation. Two related hypotheses have been presented to account for the ordering of the five roles. The first treats the set of roles as analogous to five stages in the passage of an individual male through a single descent-group generation, and the second treats the set of roles as a model for relating the internal functioning of Exogamous Groups to the range of modes of communication with outside groups. We have seen that each role has an Internal' and an 'external' aspect, and that the external aspects represent increasing closeness and security in relations with outside groups as we move from servants to chiefs. While this argument appears rather speculative as it stands, it is reinforced by the analysis of O-generation kin categories and marriage which follows. On the one hand, since marriage rules and preferences are phrased in terms of kin categories which require considerable explanation, it would not have been practicable to incorporate the discussion of marriage into the present chapter. On the other, postponing it has had certain disadvantages because, as I have frequently indicated, marriage is intimately related to other types of intergroup communication. For this reason, I reserve my comments upon the general value of this second hypothesis until the end of the following chapter.

75

Kinship and marriage

Introduction The Barasana kinship terminology is a variation on the basic Dravidian type (Dumont 1953).1 In this basic type, members of ego's generation are either 'classificatory siblings' or 'classificatory cross-cousins'. Classificatory siblings of the opposite sex to ego are prohibited in marriage, and classificatory cross-cousins of the opposite sex to ego are potentially marriageable. In the Barasana terminology, there are further distinctions within these O-generation categories. Although the fundamental opposition between marriageable and unmanageable members of the opposite sex exists, there are degrees of marriageability expressed in terms of the sub-categories. It would not be relevant to embark on a detailed discussion of either the entire terminological system or even the O-generation categories here. Instead, a brief review of O-generation classification, the application of the Ogeneration terms and the marriage rules and preferences expressed in terms of these will be enough. A great deal of data on Vaupes marriage practice and on Bara O-generation kinship terminology (collected in Cano Inambu) has been analysed by J. Jackson in a number of works to which the reader should refer for additional information and alternative analytical perspectives (1972, 1973a, 1976, 1977). The terminological system and set of marriage rules and preferences described by Jackson is very similar to the material presented here. However, the terms for undifferentiated bilateral cross-cousins 1 I also collected terminologies in Bard, Tatuyo and Makuna, together with data on the application of terms by speakers of these languages. The terminologies are so close to one another that terms may be directly translated from one language to another without any confusion. There are just a few exceptions of a minor order; for instance, one language may contain a separate address form of a certain term while another does not.

76

Introduction (tenyua pi., tenyu m., tenyo f. in both Barasana and Bara) which are of great importance both to the Barasana and to their immediate Bara neighbours in Cano Colorado, are apparently not used among the Bara of Inambu. The emphasis will be on O-generation categories, because marriage rules and preferences are phrased in terms of these. These categories may be arranged in a series according to the marriageability of women in each, and there is evidence to suggest an analogy between this series and the specialist-role hierarchy as discussed in the previous chapter. Demonstration of this point is followed by a discussion of marriage practice, which shows that wife-getting procedures may be related both to the series of O-generation categories and to the series of modes of communication between groups which was derived from the specialist-role hierarchy in the previous chapter. The O-generation categories The O-generation terms specially referred to in this section are classificatory ones: they are presented in fig. 9A, with the English glosses which I shall use as substitutes. Hereafter, when the glosses appear in the text in inverted commas it is to indicate that the terms to which they refer cover both primary and classificatory kin. A simplified model of the terms used by a male ego is set out in fig. 10. This does not include all the terms listed in fig. 9A, nor the individualising terms. It may be compared with the comprehensive list of terms used by males and females in appendix 2. Presented in this way, the terminological system is consistent with a 'dual organisation' marriage system in which sisters are exchanged between two opposed groups. However, in practice there is an open-ended number of Exogamous Groups and members from any one may take marriage partners from among a variety of other Exogamous Groups. The subcategories of 'agnatic siblings' and 'uterine siblings' on the one hand and 'patrilineal cross-cousins' and 'matrilineal cross-cousins' on the other can only be understood in the context of this situation. In ego's generation, agnatic siblings, members of ego's Exogamous Group, are distinguished from those who are uterine, that is, those having the same mother as ego or a mother from the same Exogamous Group as ego's mother. If a 'uterine sibling' is also an 'agnatic sibling', then the agnatic ties take precedence. Henceforth, I shall adopt the Barasana patrilineal bias and refer to agnatic siblings as 'siblings' and 77

Kinship and marriage

A

0-GENERATION TERMS & ENGLISH 6L0SSES

ngava

older sibJmg*

bedera

younger sibling

gag*

older brother

bedi

younger brother

gage

older s\ster

bedeo

younger sister

hako na

mother's children

tenyua

bilateral cross-cousins

hako ynaku

mother's son

tenyt*

bilateral male cross-cousin

hako wako

mother's

kevyo

bilateral female? cross-cousin

wekaho na

father's sisters

ynekaho wiaktt

father's

sisters

son

Mekaho mako

father's

sister's

daughter

na

daughter children

mother's brother's children

wiaku mother's brother's son hakoarumtt mako mother's brother's daughter

B

DIFFERENTIATION

BETWEEN OLDER' ^VOUNOER' SIBLINGS

I I A 6

A 6 older siblings

iUU D EGO

1

LI

A

o

Note : numbers indicate birth order witUin sibling groups Key • 0

male or female

Fig. 9 O-generation kinship terms 78

A o

younger siblings

A

6

The O-generation categories AFFIMES

A6KJATES cT

FF/MF

A7/>tt/

+2 FM/MM /77/CO

9 0*

FB

bt/amt?

MB

hakoarHtnv

9

FZ

mekaho

MZ

btfama

cf

Be M65/FZS

tenyts

MBO/FZD

tenyo

+1

By bedi

0

Ze

gago

9

bedeo

cf

S

makv

zs

haroagtt

9

D

mako

ZD

haroago

-1

cf

55/Z55

hananii

-2

9

SD/ZSD

hanenyc

Note the individual\s\nq terms for F,M, WF, W M , W are not included, nor is thtf full r^n^e of O-generation terms: see appendix 2 for other

Fig. 10 Simplified kinship terminology for male ego (in 'Dravidian Box' form) 79

Kinship and marriage uterine siblings as 'mother's children' — a direct translation of the Barasana term, hako ria. While a sibling relationship is covered by the expression singu ria or singu hanerd batia, 'one man's children' or 'one man's scattered descendants', there is no ego-oriented term embracing all siblings . Every sibling is either 'older' or 'younger' than ego. Actual age is only relevant in a true agnatic sibling group; the relative seniority of classificatory siblings depends upon the birth order within sibling groups belonging to past generations. It is the birth order of the father's sibling group that determines the relative seniority of true patrilateral parallel cousins and so on. People do not need to memorise these genealogical facts because, as mentioned earlier, the seniority system gives rise to a consistent internal ordering of descent groups in which each individual has a unique position. The position of ego's sibling group is determined by the position of the father vis-a-vis other members of his generation, and ego's position within his own sibling group is determined by birth order (fig. 9B). The fact that true relative age is ignored outside the sibling group produces a situation in which 'older siblings', 'father's siblings' and even 'grandfather's siblings' can be younger than ego in terms of real age. There is no older/younger distinction among 'mother's children': the terms are used reciprocally by ego and alter (with the appropriate male or female ending). Other uses of the terms, apart from those in which the common group identity of ego's and alter's mother are recognised, are between co-wives and wives of 'siblings', between cohusbands (who to my knowledge only exist in myth) and husbands of 'siblings' and between those who are 'affines of affines' to each other, for example, between the spouse of a 'cross-cousin' and 'crosscousin' of a spouse. It is impossible to appreciate how this last usage operates without an understanding of the way in which sub-categories of 'cross-cousins' are differentiated, for only some of ego's tenyua become 'mother's children' if they marry others of ego's tenyua; this depends on the relative strength of the cross-cousin ties which will be discussed below. In general, all the uses of the 'mother's children'terms fit the general description 'affines of affines', for in each case three Exogamous Groups are involved: ego's, a second from which ego's mother or spouse comes and a third which also receives women from the second (see p. 93 for further discussion of this point). The relationship 'mother's son'/'mother's son' is thus one of competition between members of separate Exogamous Groups for women of a third such group. 80

The O-generation categories The category of 'cross-cousins' embraces all members of ego's generation who are neither siblings nor mother's children. According to the logic of the terminological system set out in fig. 10, classiflcatory 'cross-cousins' are children of both mekaho, 'father's sister' (= MBW) and hakoarumu, 'mother's brother' (= FZH). While the Barasana certainly volunteer this as a general proposition, in some contexts they choose to differentiate between 'patrilateral cross-cousins', mekaho ria, and 'matrilateral cross-cousins', hakoarumu ria. In this case, 'patrilateral cross-cousins' are children of women who are of the same Exogamous Group as ego and of ego's father's generation, and 'matrilateral cross-cousins' are children of men who are of the same Exogamous Group and generation as ego's mother. It is thus possible to distinguish between cross-cousins according to the precise Exogamous Group membership of alter's parents in relation to ego's parents. The unilateral cross-cousin categories complete a set based on the common Exogamous Group membership of ego's and alter's parents. If alter's mother is ego's father's Exogamous Group 'sister', but alter's father is also ego's mother's Exogamous Group 'brother', then in choosing a unilateral term, the patrilateral link takes precedence just as it does in the case of bilateral parallel cousins (fig. 11). These sub-categories leave a residual category of cross-cousins among tenyua, whose parents do not have Exogamous Group membership in common with either of ego's parents, but who nevertheless belong to outside Exogamous Groups. They are people reckoned to be members of ego's generation who neither merit 'sibling' terms nor 'mother's children' terms. From remarks Indians make about their choice of terms, it is clear that these 'unrelated cross-cousins', as I shall call them, barely merit the 'cross-cousin' terms. These terms are used for want of any others when it is necessary to address unrelated O-generation members; however, a female ego will usually avoid using any term when greeting these people. It is these 'unrelated crosscousins' who become 'mother's children' once they marry a 'patrilateral cross-cousin' or a 'matrilateral cross-cousin'. This discussion has been somewhat simplified in order to provide a clear background for the discussion of choice of marriage partners. Additional O-generation terms for in-laws are described in appendix 2. I should also add that tenyu and tenyo may be glossed as 'male and female exchange partners'. This is illustrated in the use of He tenyu, the non-kinship term for a partner with whom sacred regalia are ritually exchanged. Once a marriage occurs, whatever the pre-existing 81

Key. . J_J

jA

A A=6

A

6

E

A

E A

. indicates either true or [_J classificatory siblings he. of the same Exo^amous Orowp. Eao Alter

E

=6

OjA

oA

A

6

g&go/bedec

E

A

mother's daughter hako tnako

A=O E

A

res'tdval cross-covsin tenyo or no term

T E

A

mother's brother's datgHtei ntako

Fig. 11 O-generation kinship relations (according to relations between parents of ego and alter) 82

A

o

father's sitters daughter mekaho mako

A=o

older sister or younger sister

6j

E

older sister or younger sister gago/bedeo

O7A

=A

o=A

6=A

E

A

i

father's sister* daughter mekaho mako

Marriage rules and preferences kinship ties, the husband and wife's brother become tenyua par excellence, so that if one asks who a man's tenyu is, the immediate answer is his wife's true brother or his true sister's husband. This man may be neither a mekaho maku nor a hakoarumu maku. This means that my discussion above applies to the determination of O-generation classification by marriages in the ascending (+1) generation. As I shall show, it is this set of categories that are relevant for ego's marriage, but, once he and his siblings begin to marry, the new relationships are taken into account. Although the main significance of the unilateral 'cross-cousin' subcategories is for marriage, these categories are not merely descriptive reference terms. They are used in address, but principally between individuals of opposite sex or between two women. Men tend to use tenyu in address to their male 'cross-cousins' even if the sub-categories are applicable. Except between two men, the use of tenyu and tenyo in address is considered technically correct but often too embarrassing in practice, because it implies potential marriage. The fact that relations between male 'cross-cousins' are not subject to this embarrassment is presumably related to their status as the active participants in the arrangement of marriage. The unilateral cross-cousin terms also imply potential marriage, but instead of emphasising the freedom of the present generation to choose marriage partners they emphasise the dependence of the present generation's marriage patterns on the marriages of the past generation. In other words, the general terms tenyu and tenyo imply 'we could make a relationship through marriage', whereas the unilateral terms imply 'we are related through a past marriage - which might constrain us to become related through a present one'. Marriage rules and preferences

Having established that the Exogamous Group membership of the four parents of ego and alter is critical in the distribution of Ogeneration terms, we may turn to the marriage rules and preferences which are expressed in terms of these categories. Negative rules First of all, there is no specific Barasana word for incest. Marrying into a forbidden category is covered by the verb yoke-, which is also 83

Kinship and marriage applied to the actions of insane people, the appearance of evil omens and manifestations of evil spirits. People who make prohibited marriages are likened to forest animals, particularly jaguars (who are seen as the closest forest counterparts of man in many other ways, too). Such people are also said to have lost their way' or 'muddled their path' {ma wisi-). In principle, the only permitted marriage is with a tenyo for a man and a tenyu for a woman. It is forbidden to marry either a 'sibling' or a 'mother's child', but a 'sibling' marriage is considered worse than a 'mother's child' marriage. In practice, the former are very rare and very conspicuous because of the common Exogamous Group membership of the partners: the latter are more frequent and less conspicuous because of the separate group membership of the partners. In both cases, a closer genealogical link (between the 'siblings' or, in the case of 'mother's children', between their mothers) is considered a greater bar to marriage than a more distant one. Marriage across generations is also forbidden, although it is reasonably frequent. Often, the generational status of alter is ambiguous depending on which way the relationship is traced, but if there is a close genealogical link between ego and alter, such intergenerational marriages lead to conspicuous inconsistencies in the use of kinship terms. People criticise the partners to cross-generational marriages and joke a great deal about how to address them, but nevertheless a marriage into the wrong generation of an outside Exogamous Group is usually considered only mildly wrong. With regard to all marriages outside the prescribed category of tenyo, and yet between partners from separate Exogamous Groups, it is significant that the original wife-receivers in a single marriage tend to lay stress on the correctness of the marriage and thus on the legitimacy of their claims. However, the wife-givers, when they demand reciprocation, are more concerned that the exchange should be completed than they are with the precise relationship between the potential spouses.2 Positive rules It is always stated that a man must marry his tenyo and that the cor2 I know of cases of marriage with the true sister's daughter and also with the sister's husband's daughter by a previous marriage. These are justified by the need to complete an exchange when age and sibling-group structure prevent a sister-exchange. Although they are considered a 'good' solution from the point of view of fair exchange, people do not

84

Marriage rules and preferences rect form of marriage is sister-exchange.3 The ideal marriage occurs when two men exchange true sisters within a single generation. According to the ideal family structure, boys and girls are born alternately, so that there are sibling pairs formed of an older brother and a next-born sister bound together in an exchange unit. However, reciprocal exchange is still recognised when the sibling links are classificatory but within the same local descent group, or when the reciprocal marriage occurs in the following generation (FZD marriage). Sister-exchange is consistent with the recognition of a unitary category of bilateral cross-cousins, so thus far the operation of the marriage rule presents no problems. However, among the women classified as tenyo, the ideal partners are the true or close father's sister's daughters classified in the sub-category of mekaho mako. Debate among Indians about marriage claims invariably centres on the rights of a man to his mekaho mako. Failing a close mekaho mako, it is preferable to marry a true or close hakoarumu mako and, failing either of these, a rather distant mekaho mako or hakoarumu mako or an 'unrelated cross-cousin' is chosen. The least appropriate choice is an 'unrelated cross-cousin' — a woman who has neither a mother from ego's father's widest Exogamous Group nor a father from ego's mother's widest Exogamous Group. Although I write as if there were a fairly straightforward hierarchy of choices, in practice many different factors influence the choice of which family or community should be approached for a wife: some of these are discussed below (p. 95). For the present discussion, the marriage preferences may be summed up as follows: the ideal partner is a genealogically close 'patrilateral cross-cousin' (FZD), the next best is a genealogically close 'matrilateral cross-cousin' (MBD) and both these genealogically close types are preferred to the other classificatory cross-cousins. Of course, genealogical distance is a matter of both degree and context; however, having noted the general priority of closer sibling links between alter's and ego's parents over more distant ones, I shall simplify the discussion by presenting the preferences as a simple series: mekaho mako; hakoarumu mako; 'unrelated' tenyo. The Barasana themselves will often describe other individuals as being 'true tenyua' (tenyua goro), 'very much tenyucC or 'a little bit overlook the muddle caused by marrying a haroago - this may be compared with the Trio case where such oblique marriages are 'primary and fundamental' (Riviere 1969 : 273). 3 There is some evidence that it is thought desirable to marry a woman from an outside sib of equivalent ranking to one's own: for example, 'chiefs' should only marry daughters of 'chiefs'. However, this ideal is not reflected in existing marriage patterns.

85

Kinship and marriage tenyua\ Their own estimation of the strength of cross-cousin ties tallies with the marriage preferences and demonstrates the priority of the patrilateral link over the matrilateral, the priority of either type over no link and the general priority of closer genealogical links over more distant ones. Here we may note again that the 'unrelated crosscousins' are the weakest form of cross-cousin: this is consistent with the fact that they are readily converted into 'mother's children'. Obviously, to the extent that sister-exchange is practised, the preference for marrying a mekaho mako cannot result in a statistical preponderance of marriages with classificatory patrilateral cross-cousins over those with classificatory matrilateral cross-cousins. As we have seen, the term mekaho mako is used in cases of a bilateral link as well as in cases of a patrilateral link alone, and therefore it is possible for two male exchange partners related as 'bilateral cross-cousins' to both receive a preferred mekaho mako (fig. 11). But, where the exchange partners are 'unilateral cross-cousins', if one marries his mekaho mako the other will receive a hakoarumu mako in return, so that the preference of one partner only is realised. We must therefore look beyond the existing marriage patterns for an understanding of this preference.4 First, it is significant that marriage with the mekaho mako constitutes an exchange between two groups that is completed over two consecutive generations, while marriage with the hakoarumu mako constitutes a replication of the marriage of ego's mother in the previous generation. Thus, the two forms may be contrasted as delayed exchange and one-way flow of women. Secondly, the mekaho mako is not merely the ideal partner, she is also the rightful partner - the woman over whom ego has the strongest claim. Ego shares his closest mekaho mako with his true siblings, while he shares his closest hakoarumu mako with his closest 'mother's children' who also regard her as their own hakoarumu mako. While relations between siblings are hierarchically ordered so that competition is minimised, the senior unmarried siblings having the prior rights, relations between mother's children are equal rather than structurally ordered. Besides, the male 'mother's children' rarely interact, since they are members of separate 4 If an analysis of existing marriage patterns in such a small population were to show to what extent the preferences are translated into action, there would have to be data accurate enough to assess each male individual's field of choice of a wife. Since the relevant genealogical data would concern people neither I nor my informants had ever seen, and since each person changes residence so frequently, it is impossible to know whether age, place of residence, marital status etc. had rendered any given woman a weak, strong or impossible candidate for wife at any given time.

86

Marriage rules and preferences and often spatially distant Exogamous Groups. Therefore, we can compare the prior claim to a woman (and her sisters), which is allocated to the brothers of a single local descent group who all regard her as mekaho mako, with the secondary claims to such a woman, which are distributed among members of a number of local descent groups. These regard each other as 'mother's children' and the woman in question as hakoarumu mako. Mekaho mako marriage thus represents exchange over the generations and a neat, unambiguous allocation of rights, while hakoarumu mako marriage represents a one-way flow of women and competition. The mekaho mako is a relatively secure choice which implies continuing exchange of women between the two groups concerned, while the hakoarumu mako is a risky choice which increases the imbalance between the groups. Both may be set against the third possibility of attempting to marry an 'unrelated cross-cousin', who is not specifically forbidden in marriage but over whom ego has not the slightest claim. Analysis of marriage rules and preferences in relation to specialist roles We can now set out the marriage rules as a series of categories of women, which ranges from the absolutely forbidden to the ideal partners, as follows: 'younger' or 'mother's 'older sister' : daughter' :

'unrelated cross-cousin'

'mother's brother's : daughter'

'father's sister's daughter'

bedeo/gago

hako mako

tenyo

strongly prohibited

prohibited

not prohibited but not approved

hakoarumu mako approved

mekaho mako strongly approved

In the previous chapter, it was suggested that the outwardly directed aspects of the specialist roles could be read as a series representing increasingly positive, secure and controlled relationships with outside groups, and here I suggest that the series of O-generation categories has the same overall character. If the female categories above are replaced by those embracing both sexes, we have 'siblings' at one extreme representing the internal structure of patrilineal groups (corresponding with negative external relations), and 'patrilateral cross-cousins' at the other extreme representing enduring marriage 87

Kinship and marriage exchange with an outside descent group (positive external relations). In the middle are people who are not genealogically related by an affinal link in the parental generation at all. In fact, I have already mentioned that these 'cross-cousins' are really so called because there is no other way to address them - in reference they are frequently described as 'other people' to whom kinship terminology does not really apply. The intermediary positions are filled with 'mother's children', distant competitors whose females are unmarriageable, and 'mother's brother's children', whose females are marriageable. The series as a whole represents the transformation of sibling ties into ties of perpetual alliance between opposed descent groups. I have claimed that the series of O-generation terms as they are applied to individuals bears some resemblance to the series formed of the external relations implicit in the specialist roles. It also bears a resemblance to the specialist roles seen, by analogy, as life stages, because one is born a sibling and one gains an established affinal link with the consolidation of marriage on the birth of children. Both the external relations series and the life-stages series are derived from the specialist-role series itself and, in myth, a direct reference is made to the congruence of the O-generation terms with this basic series of specialist roles. The myth of Manioc-stick Anaconda is explicitly followed by the events concerning the first exogamous marriage between Manioc-stick Anaconda's son, Yeba, and Fish Anaconda's daughter, Yawira. Other parts of this extended myth appear in this book, but the reader is referred to Stephen Hugh-Jones (1979 : M6 and M7) for more complete versions. The version here is much reduced. Manioc-stick Anaconda and the origin of exogamous marriage Manioc-stick Anaconda created a younger brother, Macaw, and a wife all by himself. He had children by the wife and then Macaw stole her, initiating a set of aggressive acts in which each brother tried to get rid of the other. Macaw who, being the younger, had the shamanic knowledge, tricked Manioc-stick Anaconda into dropping into the Underworld where he met the Sun. Sun is clearly a transformation of Macaw (S. Hugh-Jones 1979 : 232). Manioc-stick Anaconda claimed Sun as his mother's son but Sun put this claim to the test. He did this by blowing fire — his shamanic tobacco snuff — at Manioc-stick Anaconda, saying, 'If you are not my mother's son I will burn you up: if you are my mother's son you will not disappear'. Manioc-stick 88

Marriage rules and preferences Anaconda not only managed to resist the heat, but also to steal some shamanic snuff and burn the Sun himself. When the Sun's macawfeather crown, the repository of his heat (and one of the factors that relates him to Macaw), caught fire, he acknowledged that Maniocstick Anaconda was indeed his mother's son and was in possession of shamanic power. Eventually, Manioc-stick Anaconda returned to this earth in the state of an initiate to find that, while he had been underground, Macaw had initiated his children. After more aggressive trickery, Manioc-stick Anaconda succeeded in burning up Macaw, the stolen wife and his initiated children with the Sun's snuff. Macaw and the wife became Yurupary instruments; the children became birds. Manioc-stick Anaconda was all alone and mourned. He went off to marry a Jaguar Woman who bore him a half-jaguar son, Yeba. Yeba made a one-sided marriage with Fish Anaconda's daughter, Yawira, which is acknowledged as the origin of alliance between Exogamous Groups. Yawira discovered biti fruits (Hevea sp.) which dehisced into her lap, and she took some to her father's underwater land; she returned with cultivated food plants. Fish Anaconda threatened to eat Yeba, but the marriage was consolidated by Yeba's ritual gift of biti to Fish Anaconda. Yeba learnt to dance and take yage at the ritual held in Fish Anaconda's house for the presentation of biti: this occasion is the charter for all dance rituals, and for food-giving rituals in particular. The small fish, Wania (Cichlidae sp.), Yawira's 'crosscousins' (tenyua), wanted to kill Yeba during the dance. Later, Yeba's gift was reciprocated when he held a dance at which Fish Anaconda presented him with umari fruit (wamu: Poraqueiba sericea); the umari fruit were women. The important feature of the mythical cycle in this context is that it shows a clear correspondence between (a) the mother's child relationship and shamanism and (b) unilateral marriage and ritual dance gatherings, particularly those held for ritual exchange of food. These correspondences strengthen the analogy between the set of O-generation categories arranged according to marriageability of women and the specialist-role hierarchy, (a) Manioc-stick Anaconda's wife is taken over by his younger brother so the two become cohusbands: their affinal ties with a common third party (wife) already make them less like siblings, whose behaviour should be constrained by birth order, and more like 'mother's children'. 'Mother's son' is the appropriate address term between co-husbands who are not classificatory siblings (p. 80). It is fitting that the end result of Macaw ignor89

Kinship and marriage ing his birth position and taking his elder brother's wife is the cancellation of the sibling tie by means of Macaw's transformation into 'dead' Yurupary instruments. This leaves Manioc-stick Anaconda 'all alone'. In the Underworld the direct connection between the mother's child relationship and shamanism is made, for it is shamanic equality that proves the mother's son tie between Manioc-stick Anaconda (MSA) and Sun, thus: / MSA older brother non-shaman \ \ / MSA \ Macaw younger brother shaman / \ Sun inequality

mother's son mother's son ^ equality

shaman ^ shaman /

(b) Yeba's marriage is the original alliance between groups and therefore we may assume that the Jaguar Woman represents, not an Exogamous Group, but the natural animal world. Yeba's marriage is unilateral: although the subsequent exchange of fruits with his father-in-law might seem like a metaphorical exchange of women, closer analysis shows that in fact the exchange re-enforces the oneway character of the transaction. First, biti and umari are related as male and female. Biti appears as a sexually active male in another myth, and in the present myth Yawira eats biti fruits that have dehisced into her lap. Comparison with another female fruit-eating episode suggests that this represents insemination (S. Hugh-Jones 1979 : M2A). The umari, on the other hand, are directly referred to as women. Yeba receives a wife and umari-women while Fish Anaconda receives a gift of 'male sexuality' which, so far as the communication of sexual elements is concerned, constitutes a duplication of his own gifts.5 This exchange can be identified with the repeated oneway movement of women, and thus with mother's brother's daughter marriage. Appropriately enough, the gift from son-in-law to father-in-law is described as the origin of dance rituals, and of ritual food exchange in particular. Therefore, dancing originates with the gain of an affinal 5 Yeba also gives meat to Fish Anaconda (S. Hugh-Jones 1979 : M.7.H). The meat is the flesh of his own people, the birds and animals, whom Yeba has killed with his blowpipe: other animals and birds accompany Yeba as co-guests. In view of the analysis of meat production (pp. 192-200), it might be argued that this gift of meat is a gift of female substance, which thus contradicts my theory that female gifts flow in one direction only. There are various reasons for rejecting this argument. Most important is that meat is a particularly male substance in other contexts (see analysis of meat meal, p. 223, and reference to the ritual gift of meat entering the house as a penis enters the womb, p. 208). It is not surprising that meat is sexually ambiguous since it is both opposed to the men who hunt it and identified with them as their catch. Because of these ambiguities I have only included the fruits, whose sexuality is not ambiguous.

90

Marriage rules and preferences woman and is opposed to shamanism, which originates with the loss of an incestuous wife. Furthermore, the nature of biti as a male, inseminating fruit gives the 'origin of dancing', caused by the entrance of biti into Yawira's natal home, the additional character of metaphorical insemination. This accords with my earlier interpretation of dancing (p. 67). Overall, there is a correspondence between the specialist-role series and the O-generation classificatory series, which may be set out as follows: siblings 'mother's 'unrelated 'mother's 'father's children' cross-cousins' brother's sister's children' children' MSA/Macaw MSA/Sun MSA/Jaguar Yebal (present-day Yawira ideal marriage) (warriors) origin of (chiefs) (servants) origin of shamanism dancing It is significant that there is no mention of exchange of women in marriage in this or any other myth, in spite of the practical preoccupation with sister-exchange in the real world. Thus, the mythical progression ends with the unilateral exchange of women, which is expressed in present-day terms as 'mother's brother's daughter' marriage. This suggests that the series of mythical events may also be seen as a development of the present-day ideal of sister-exchange between Exogamous Groups from the original pre-descent beginnings. To return to the terminological categories: it does not make sense to ask whether the series of categories regarded as categories of individuals corresponds to any ongoing process in the real world. For instance, although a sister may be exchanged for a father's sister's daughter in marriage, she herself remains a sister and cannot be transformed into a member of any other category from ego's point of view. However, if we consider the terms as they are applied to whole groups by members of other groups, then it is undoubtedly true that over time the term applicable to a descent line or a single named group may change. We have seen that other sibs or Exogamous Groups may be classified as 'brothers' (elder or younger) or 'mother's children' (ch. 2). They may also be classified as 'cross-cousins' (tenyua) — either strong ones whom 'we' marry a lot or weak ones whom 'we' marry a little — or ngahera, 'other people', whom 'we' do not marry. Of course, it is a matter of descent dogma whether or not a group are 'siblings' or 'people waking up mother's children' (pp. 2If), while it is 91

Kinship and marriage an estimation of practical existing marriage patterns that leads to their classification as 'de facto mother's children', 'cross-cousins' or 'other people'. Nevertheless, descent dogma can ultimately adjust to shed distant members and to incorporate foreigners, so that the relationship between any two descent groups is open to change over time. If we set out the series of intergroup relationships which corresponds to (and virtually duplicates) the kinship categories applied to individuals, it is clear that, reading upwards, it represents the dissolution of descent ties and the subsequent creation of strong marriage ties. GROUPS

'brothers'

'mother's children'

'other people'

'people we marry a bit'

'people we marry a lot'

dissolution of descent ties creation of affinal ties creation of descent ties

v

dissolution of affinal ties

This model process could actually occur. We may imagine sibling descent lines becoming geographically separated by their mutual affines to the point where interaction between them becomes infrequent and they redefine their relationship in terms of their shared link with the affines — they become each other's mother's children. If one group stops creating marriage links with the affines then the one-time siblings become 'other people' to each other. In the future, marriage ties between the unrelated groups may be initiated and built up to the point where they regard each other as very close crosscousins. Similarly, we may imagine the reverse transition from affines to siblings. In fact, there is contemporary evidence that shifts between the adjacent stages of this process (in either direction) do actually occur, for small descent lines are reclassified by others as a result of migrations and changes in marriage patterns. More abrupt changes also occur: for instance, a case of marriage between very distant sibling groups belonging to the separate Simple Exogamous Groups of a single Compound Exogamous Group was described to me. The informant said that, in time, the immediate descent groups of the marriage partners would repeat the marriage and the two groups would become established cross-cousins to one another. However, the model represents the shift from sibling to affinal relations as a more gradual process, incorporating all the alternative relations between groups into a single system. The five-part series presented here may now be compared to the 92

Marriage practice models implied by certain features of Pira-parana social classification, notably the existence of the 'mother's children' terms as distinct from 'sibling' terms (p. 77) and the evidence for a pan-Vaupes, tripartite classification of Exogamous Group ancestors (pp. 36ff). These tripartite models are closely related to J. Jackson's tripartite model of O-generation terms based on the distinctions between the 'sibling', 'mother's children' and 'father's sister's children' terms in Bara (1977 : 93). Fundamental to her model is the proposition that 'everyone who is not a Makii or white is a kinsman' (1977 : 87): this proposition is clearly related to the notion that there is a pan-Vaupes structure in which everyone is classified into one of three categories, according to the habitat of the anaconda ancestor (p. 36). These are models of systems in which everyone has a place and no one or no group is unrelated or unknown. My five-part model, on the other hand, applies to an 'open-ended' system in which there are unknown and unrelated people and groups - the 'unrelated cross-cousins' and 'other people', as well as the yet-more-distant people discussed below (p. 101). Normally, friendly or semi-friendly interaction is inappropriate between these people, but if people 'beyond kinship' do arrive during travel or to attend a ritual they are given kinship terms to draw them into a closer relationship. The point about these kinship terms is that they are not derived from acknowledged connections through descent or marriage, but are invented for the purposes of amicable interaction. There is actually an expression, yard kuti- (meaning literally, taking possession of kinsmen) for this inventing behaviour. It is my belief that a comprehensive 'closed' model and an 'openended' model of Vaupes social structure can, and do, co-exist in Indian ideology. Marriage practice Obtaining wives Amongst Vaupes Indians, marriage determines the legitimacy of potential children, gives the husband exclusive rights in the wife's sexuality and sets up an economic partnership between husband and wife in which each has extensive rights and obligations with respect to the other. There is no word in Barasana for 'marriage': 'getting married' is simply 'taking a wife/husband' (manaho/manahu kuti); after the birth of children the spouses refer to each other as 'the 93

Kinship and marriage mother/father of my children'. In Vaupes marriage practice the food restrictions, seclusion, shamanic rites and communal ritual, some or all of which accompany birth, death, menstruation and initiation (Yurupary ceremonies), are conspicuous by their absence. The process of acquiring a wife is essentially a secular one, involving the physical transference of the woman to her husband's group. If the marriage is to persist, she must be successfully retained so that she forms a sexual and economic partnership with a particular man and, eventually, bears children who inherit his status as local descent-group members. There is no precise point at which the marriage is established, but the likelihood of it lasting depends upon two main factors: the acceptance of the situation by the girl's local descent group, and the birth of children. Of course, if the husband's community finds the girl to be an unsatisfactory wife, she is likely to be sent home, but if she herself is unhappy, her fate depends largely on the attitude of her own local descent group. If the marriage is in their interests, and particularly if it is part of an otherwise successful exchange marriage, they may not welcome her back. Usually, if a woman is living away from her agnates amongst a local descent group of another Exogamous Group, where she is known to have formed a sexual and economic partnership with a man, the two are generally acknowledged to be husband and wife. However, husband/wife relations are also recognised where the union is incestuous, provided that the couple openly co-operate in production and child rearing (when there are children), and thus the status of 'husband' and 'wife' is dependent on a pattern of interaction between the two, rather than on any formal alliance between opposed groups. In most cases, it is perfectly clear whether or not two individuals count as 'husband and wife', but there are certain cases in which women may live among potential affines without their wifely status being certain. Newly acquired potential wives may not have settled down with a particular partner, or widows may stay on with their husband's local descent group without fully establishing a marital role with the dead husband's younger brother (the expected behaviour for widows). In these cases, it is the details of domestic interaction between the two that people count as evidence of their marital status - whether the man fells a cultivation site for the woman, whether she cooks his game and prepares manioc products for him, whether they travel together and so on. It is these domestic arrangements that separate a love-affair or casual sexual relationship from marriage. In nearly all 94

Marriage practice cases, a child born outside such a stable domestic arrangement is killed at birth, and therefore the question of legitimacy of the children of unmarried women rarely arises. Although the recognition of the status of husband and wife does not depend upon there being a correct kinship link between the two, the kin relations between them are of the utmost importance in the formation and breaking up of the partnership. The rare incestuous unions that exist would not have been tolerated if either party was a member of a well-established local descent group. We have seen that the ideal partner is a true, close mekaho mako, and this is the category of women which is considered first. Structurally speaking, a marriage with a mekaho mako constitutes an exchange for the girl's mother but, although this makes the marriage an exchange in itself, the practical claims and debts may not be perceived in this way by the two parties. For instance, the girl's mother herself may have been owed in exchange for the potential husband's father's brother's wife: there are many ways in which the failure to exchange, or the exchange of classificatory rather than true sisters, distorts the patterns of debt implicit in the ideal models. There is a number of factors that determines the willingness of the girl's parents and close agnates to let her go, and of utmost importance amongst these is their need to find wives for their sons. The existing set of debts and claims arising from previous marriages and the age- and sex-structure of the girl's sibling group are considered. The possible advantages of using the girl to make a new alliance with another group must be weighed against the disadvantages of offending the rightful claimants, who may be close neighbours. The general character of the relations between the two groups, as well as the character of the potential husband, are also taken into account. I never saw a girl removed by her potential husband or in-laws, but there is every indication that when the marriage has been arranged and agreed in advance, she simply accompanies the claimants back to their longhouse. In such cases her parents should advise her to go peacefully and her mother should present her with a few necessary female possessions.6 The Pira-parana Indians themselves contrast this 6 Other sources mention the institution of 'mock-capture' of wives in the Vaupe's (Briizzi 1962 : 414, Goldman 1963 : 142), in which the girl's folk feign various degrees of aggression. I found no evidence for this among Pira-parana Indians, for capture and the associated use of force was 'real' but nevertheless controlled. While the term 'mock-capture' suggests a ritualised and obligatory play-act, my argument will be that the degree of force is purposely chosen and is effective in bringing about a certain change in relations. It is

95

Kinship and marriage peaceful manner of obtaining women, which they say is appropriate for marriage with a mekaho mako or for any pre-arranged exchange marriage between friendly affinal communities, with seizing women. It is appropriate to seize a woman if a rightful claim has been turned down by the girl's agnates, and also if there is no prior claim to the girl either as mekaho mako or as the promised exchange for a sister already given. It is also appropriate to seize women from moderately distant or very distant communities, but use of force is incompatible with close neighbourliness. A party is made up of the father and other close agnates of the prospective husband; it usually includes the prospective husband himself and sometimes close affines of the local descent group concerned. These men set out for the longhouse of the woman they have chosen. They may intend merely to make a formal request for a woman, in which case they must decide upon a concrete offer of a sister in exchange. Otherwise, they may intend to seize the woman at night or when she is working alone in the manioc garden or fetching water. There are intermediate tactics, such as seizing the woman under circumstances in which they are not very likely to get away undetected; but, in any case, however determined the intention, the appearance of a potentially aggressive, united, all-male wife-getting party separates the occasion from the less formal and more interpersonal negotiations among close affines. When the raiding party come to put their plan into action the subsequent events are influenced by the reaction of the girl's community. The raiders may get clean away with their woman either before or after some fighting, and in this case the girl's community may decide to pursue them immediately or to make a counter-raid at a later date. If their verbal demands are turned down, or if they are caught in the act of seizing the girl, there is a range of possible responses from verbal negotiation, either polite or aggressive, to physical violence. The violence may range from mild tussling to the use of weapons. There are threats of killing even today, although it is impossible to say how frequently killing actually occurred on wife-getting raids in the past. It is my guess that the extreme cases of pertinent, however, that Goldman contrasts friendly inter-Cubeo marriages with more 'hostile' marriages between Cubeo and outside groups and the 'mock-capture' appears to be associated with the former. He describes an external marriage in which a Tatuyo man, part of an all-male wife-getting party, carries off a Cubeo girl (not the girl intended for him) by surprise and without any 'mock-capture' episode. The marriage repeats a marriage of the previous generation, in which a Cubeo woman was also given to that particular Tatuyo group, and thus the forceful style of the Tatuyo elopement fits my analysis below perfectly.

96

Marriage practice outright denial of a rightful woman (a mekaho mako or exchange for a sister already given), or of a determined raid on a distant and previously unknown group, could have escalated to killing. As mentioned before, although warfare and raiding for wives are different in that war raids are far more aggressive acts, it may well have been possible for either one to result in the other. Wife-getting methods as a function of social distance At the same time as they are partially controlled by the existing relations between the two groups concerned, the wife-getting events set the scene for future interaction between groups. Ultimately, there are three possible outcomes: the completion of an exchange through a sequence of encounters, which may be more or less violent, the establishment of a one-sided marriage or no marriage at all. Each of these results may be accompanied by a certain amount of bad feeling between the groups, and geographical distance is bound to influence the likelihood of reconciliation between them. Nevertheless, in general the three results can be arranged in an already familiar series. Exchange is an amicable solution of mutual advantage; a one-sided marriage gives rise to an uneasy but positive relationship between the groups; and no marriage at all results in no future interaction. Now, although the wife-getting raid may result in exchange through a counter-raid, the immediate aim of a wife-getting raid is to achieve a unilateral marriage. Besides this, raiding is incompatible with the friendly relations between close, af finally related groups which should be perpetuated by the amicable exchange of women. Raiding is compatible with interaction between semi-distant communities. Non-aggressive, informal ways of obtaining wives may be identified with the exchange of women and interaction between neighbouring communities. Further support for these identifications is presented below; for the moment the series may be set out thus:

97

no positive interaction distant group

friendly/hostile interaction semi-distant group

friendly interaction close neighbouring group

no marriage

one-way marriage

exchange marriage

wife-raiding

amicable wife-getting

Kinship and marriage To this may be added part of the analogous series of Ogeneration categories for individuals, and that for groups: INDIVIDUALS GROUPS

'unrelated crosscousins' 'other people'

'mother's brother's children' 'people we marry a bit'

'father's sister's children' 'people we marry a lot'

We may now return to the analysis of the ritual food exchange following Yeba's marriage. In addition to the connections already noted between unilateral marriage, ritual gift-giving of food and the origin of ritual dance, there was stress on the aggressive behaviour of the hosts towards Yeba. One myth-teller explained that Fish Anaconda was so aggressive because he had no woman in return. First, his father-in-law wanted to kill him and then the small fish, Wania the cross-cousins of Yawira — wanted to kill him. Similar aggression occurs in the only other mythical mention of a ritual gift of food, the Wood Ibises' gift of ants to their mother-in-law Romi Kumu (pp. 63f)This is also a gift from wife-receivers to wife-giver (Romi Kumu procreates by virgin birth and is thus a wife-giver), and the givers are also attacked in the host's house (by Pouncing Jaguar). Therefore, myth also emphasises the danger of interacting with unilateral wifegivers - a danger that echoes the practice of the wife-getting raid. I am not claiming that marriage with a 'mother's brother's daughter' is always achieved through a raid, while marriage with a 'father's sister's daughter' is always achieved peaceably. Nor am I claiming that ritual food exchange is always between partners to a unilateral marriage, or that unilateral marriages always occur between semi-distant groups. In short, I am not claiming that there is any necessary practical connection between the analogous terms of the separate series in any given situation. Instead, each series deals with a separate variable in the make-up of the 'total' soqial relationship between two groups, and the analogous terms in their separate series fit together in an ideal sense. Thus, on the one hand, it is in myth and in statements about ideal behaviour that the connections between terms are made in a clear-cut way. On the other hand, in practice, the relationships between individuals or between communities are complex and open to many different modes of change. They may also be perceived in different ways by either party. Besides, the entire range of traditional intercommunity relations has become modified by the pacification of Vaupes Indians. Very violent interaction is now virtually non98

Marriage practice existent, and consequently interaction between distant communities is much less dangerous than it used to be. Although I could give a string of marriage case-histories, both peaceable and semi-violent, to support the particular argument in this section about the relations between O-generation kin categories and wife-getting practices, these would be subject to the qualifications just mentioned. Since I am basically concerned with ideal models, it is more appropriate to draw on a remarkably clear model of marriage relations between Exogamous Groups contained in a mythical text from the Tukano Exogamous Group (Fulop 1954 : 132). Of course, a similar model relating to the Pira-parana area would have been preferable, but I was never given one. It is possible that the fragmentation of groups and numerous migrations in the recent history of this marginal area has made it more difficult for Pira-parana people to envisage whole Exogamous Groups as standing in a relatively fixed pattern of marriage relations. According to the text, a mythical law-giving hero instructs an ancestor of the Tukano on the correct way of obtaining wives. First of all, Tukano must marry outside their own group. The hero names three other Exogamous Groups from which they may obtain wives, saying that Tukano must not fight for these women, instead a brother and sister of one group must make an exchange with a brother and sister from another. He goes on to name three more Exogamous Groups, saying that if the Tukano want women from these they must fight for them, but the fighting is only permitted to extract the women — they must not kill. He adds that these six groups will be like brothers' to the Tukano. A footnote is added to the text glossing como hermanos (like brothers) as amigos intimos (intimate friends), which suggests that the ethnographer is puzzled by his Spanish-speaking informant's use of 'brother' in the context of affinal relations. The text clearly distinguishes three modes of behaviour: peaceful exchange of women, raiding of women and total absence of marriage — the first two of these are appropriate for specific sets of outside groups and the third for all others. It is implied that there is no exchange between groups who raid women and that killing is only appropriate between groups who do not intermarry. The three modes are precisely those mentioned in my analysis and, moreover, the connection between absence of marriage and killing supports the analogy between the specialist-roles series and marriage patterns between groups, for warriors occupy the middle specialist role, just as the absence of marriage links occupies the middle position in the marriage99

Kinship and marriage pattern series (p. 101). The association of the series with more general modes of interaction is underlined by the statement that the six marriageable groups will be like brothers'. General considerations Continuous and symmetrical organisation of the models Before leaving the question of marriage I must comment on the equation of affines with 'brothers'. In the field I was puzzled and disappointed to find that when I asked Indians how people in affinal kinship categories should behave, I was told like younger brothers' or 'like older brothers' in the case of cross-cousins, and 'like father's brothers' in the case of mother's brothers. I had assumed there was a fundamental conceptual divide between 'us' and 'them' - 'agnates' and 'affines' - which would be reflected in statements of ideal behaviour. Frankly, I thought these replies were senseless. However, it now seems to me that the element of security and interdependence in relations with outside groups is of enormous importance and that there is indeed a sense in which brothers are like close sisterexchanging affines. While brothers co-operate in the organisation of descent-group affairs and the life of the community, close affines co-operate in the provision of wives for each other. In practice, local descent groups are on far more intimate terms with close neighbours linked to them in perpetual alliance than they are with more distant 'brothers'. So close is the relationship between neighbouring affines that an individual may even participate in a raid on another segment of his own wider group in order to provide a wife for one of his affines. Thus, with the focus on the opposition between the hierarchy within descent groups and the reciprocal exchange between affines, relations between brothers seem diametrically opposed to relations between partners to sister-exchange. But, with the focus on the general qualities of security and interdependence, the two types of relationship are alike and together they may be opposed to the relationship with 'other people' who cannot be trusted at all. In this way, the series of kinship categories (whether applied to individuals or groups) has an element of symmetry about the central position which parallels the organisation of specialist roles into three domains. In the previous chapter, I suggested that the specialist-role series from servants up to chiefs could be read as a series of increasingly 100

General considerations positive and secure relations with outside groups. The servants have no relations with outside Exogamous Groups, and this is reflected in the statement that they are endogamous. The chiefs control external relations and this is reflected in their polygamy. However, it was claimed above that relations between siblings and relations between sister-exchanging affines are alike in being characterised by interdependence and security. These contrast with the middle position of 'unrelated cross-cousins'/'other people', with whom relations are least secure and interaction is minimal, whereas, in the series formed of the external aspects of the specialist roles, the contrast is between the two ends of the series. Since it was claimed that there is homology between the specialist-role series and the system of O-generation categories, this apparent inconsistency requires comment. On one hand, the outward-directed functions of the specialist roles make a series which, read from left to right, represents a build up of affinal relations: A affinal

affinal ~7

'

relations

'

7

~T~

relations

On the other hand, the series of O-generation categories may be read from left to right as the breakdown of descent relations, followed by the building up of affinal relations. If the arrows are made to indicate positive, building-up processes, we arrive at the following: B 'siblings' 'other people' 'people we + descent no descent or marry a lot' relations affinal + affinal relations relations Thus, there is correspondence between the right-hand parts of the two series (A) and (B), but not between the left. However, the left-hand part of the O-generation category series may be rewritten, extrapolating from the central position of 'other people' towards the extreme left, so that the series as a whole represents the building up of affinal relations with groups originally even more distant than the 'other people'. This rewritten series (C), corresponds to (A) above: C groups beyond 'other people'

'other people'

'people we marry a lot'

In practice, there are no kinship categories applicable to these groups 'beyond other people' because the very distant groups are beyond interaction - they are simply very distant 'other people'. 101

Kinship and marriage If the rewritten series, (C), is compared to the original series, (B), and the two are set under the specialist role series, the role of shaman appears to be ambiguously associated with a stage in the breakdown of descent relations and a preliminary stage in the construction of affinal relations, as follows: servant

shaman

warrior

B 'siblings' ^ C most distant 'other people'

dancer/chanter

chief

'other people'

^ 'people we marry a lot'

> 'other people'

^ 'people we marry a lot'

We already know that aggressive shamanism and warfare are held to be inappropriate between classificatory siblings but that nevertheless they do (or did) occur (p. 73). As we saw, the myth of the Maniocstick Anaconda also stresses the role of shamanism in the breakdown of sibling relations. However, in practice, shamanic killing is also frequently blamed on renowned individuals from distant groups with whom there is no interaction. Therefore it seems that the attribution of aggressive shamanism to different social categories fits with the ambiguous nature of the next-to-bottom position in the five-stage structural processes just described. In essence, this discussion brings me back to the point made about the alternative types of organisation of the specialist roles, for the rewritten model of O-generation relationships that I have just presented corresponds to the hierarchical organisation of roles, and the original model (with sibling and sister-exchange relations as extremes) corresponds to the symmetrical organisation into three domains (fig. 12). The rewritten and original models are really just different ways of looking at the same structure, for sibling relationships are the antithesis of sister-exchanging relationships and thus it is consistent that they should co-exist, at one extreme of the series, with total absence of relations with outsiders. General value of the models In summary of this and the previous chapter on specialist roles, I suggest that the value of the models I have set out is, first, that they accommodate change and, secondly, that they contain simultaneous expression of different types of relationship in the manner just described. The emphasis on change means that interaction between 102

O-CENERATION KINSHIP TERM5 CONTINUOUS fir SYMMETRICAL ORCANISATION OF INTER6ROUP RELATIONS

I

UNRELATED CROSS-COUSINS

BEYOND UNRELATED CROSS-COUSINS

1 >

MOTHER'S FATHER'S BROTHER'S CHILDREN SISTERS CHILDREN

MOTHER'S SISTER'S CHILDREN T

UNRELATED CROSS-COUSINS

/DISTANT, \ / DAM6EROU5 \ I INTERQROUP I . \ RELATIONS/

SIBLINGS

, \ I SECURE \ I INTER6 ROUP I V RELATIONS I

/ FRIENDLY-HOSTILE I INTER0ROUP >RELATIONS

B

SPECIALIST ROLES HIERARCHICAL CONCENTRIC 0R6ANI5ATIOW OF SPECIALIST ROLES

SHAMAN

SERVANT

I

)

.

WARRIOR

1

1 i

fVARRIOR EXTERNAL \ DOMAIN/

OAWCER/CHAWTER

1

) SHAMAN METAPHYSICAL /DOMAIN

Fig. 12 Symmetrical and continuous organisation of specialist roles and O-generation kinship categories 103

CHIEF

I SERVANT POLITICOECONOMIC DOMAIN

Kinship and marriage individuals and groups need not appear as a mere 'expression' or 'reflection' of some ideal relationship between the parties. Instead, it appears as the result of courses of action chosen from a set of alternatives by each of the parties. People may decide to maintain the status quo or to change it in favour of closer or more distant relations. They may do this by raiding wives, exchanging sisters, waging war, inviting people to rituals, adopting particular kinship terms and so on. They pick an appropriate strategy, but then their interaction with others during the attempt to put the plan into action may result in unforeseen change. For instance, an attempt to achieve a unilateral marriage through wife-raiding may result in a counter-raid; then the consequent sister-exchange may lead to close relations between the parties, and eventually to a change in residence to bring them geographically nearer to one another. The models are arranged so that the differences in kind may be understood as differences in quantity along the natural scales of distance and time (age). Although the models are constructed so that minimal changes will be between adjacent stages, in practice, of course, there may be greater changes leaving out intermediary stages. The dynamic aspect of these models and their property of expressing different types of relationship between the elements or stages will be discussed again in the chapter on classification of space and time (ch. 7). Finally, I should make it clear that I consider that these models are recognised by Indians, either consciously or at a deeper level, both in mythology and in statements about ideal behaviour and the nature of their social system. In present-day circumstances, it is inevitable that parts of this analysis concerning traditional life are tentative. However, I think that there is enough evidence to suggest that it is, at least, along the right lines. When I speak of the value of these models, I mean primarily their value to Indian groups who subscribe to them. They are well suited to an open-ended society (p. 12) composed of small, isolated groups bound together in a constantly shifting pattern of relations. However, I think they are also valuable to anthropologists, for they are models describing how processes of growth and change can occur while wider institutions and ideological structures remain the same. With the help of these models, it is possible to shed light upon a problem that has intrigued and puzzled those familiar with Vaupes ethnography. This is the co-existence of the opposed ideals of hierarchy and equality in a single social system. On one hand, we have 104

General considerations seen that the hierarchical internal organisation of descent groups is recognised in the use of kinship terms, and that the ordering of sibs and their sub-units is clearly a subject people care about. Differences of opinion over positions in the hierarchy give rise to strong emotions and are frequently discussed. On the other hand, this pervasive hierarchical system seems to conflict with the egalitarian life-style of Vaupes Indians and, indeed, the hierarchy has little practical content. The sib-roles are barely performed at all by their owners; individual specialist roles are achieved rather than ascribed according to birth order; and individuals who are formally differentiated by seniority meet on more or less equal terms. Traditionally, it seems that the ongoing social process bore a close resemblance to that in many societies from the New Guinea highlands, with their regulation by means of warfare, exchange and 'big man' leadership. Of course, one way of describing the contrast between hierarchy and equality is as a disparity between an 'ideal indigenous model' and 'the facts on the ground'. However, this type of description is not very useful because it leaves the relationship between the two to be explained. Besides, such a description does not do justice to the situation, because the 'ideal' and 'actual' cannot be so readily separated out. The combination of hierarchical and egalitarian principles clearly constitutes a problem for Goldman, who describes the Cubeo as having 'the skeleton of an aristocratic system that is fleshed out with an egalitarian ethos' (1963 : 92). The apparent inconsistency between these two aspects of Cubeo culture leads him to play down the hierarchical features in his general observations and to imply that they must be remnants of a higher culture. In spite of this, it is a measure of his remarkable ethnographic record that the Cubeo concern with descent seniority is well represented in the text. We are left with the impression that the hierarchical order is important for its own sake, and certainly relevant during ritual, but that it has little to do with secular life. Bidou takes the problem further, for he does succeed in relating the principles of hierarchy and equality; he claims that control and neutralisation of power is a function of the hierarchical system of specialised roles. However, Indians themselves like to stress the differential access to power, the right of older brothers to order their younger brothers around and the obligation of the latter to obey. The alternative interpretation given here is that the model of specialist roles actually serves as an indigenous explanation of the relationship between fixed hierarchy and the ever-changing patterns 105

Kinship and marriage of interaction based on the principle of equality. These two phenomena are treated as complementary aspects of the same structure — the set of five specialist roles — and thus they appear as transformations of one another. The internal differentiation can thus be turned outwards and deployed in the field of external relations to differentiate a field of 'equal' outside groups from a given point of reference. However, the specialist role system is an 'ideal' system in another sense. The complementary aspects are specifically divided between internal and external relations in such a way that the hierarchical order ought to be maintained within and yet change ought to occur without. This has to be so, to enable each Exogamous Group to maximise its position with regard to outsiders, in order to ensure its supply of women and its military and economic survival in a shifting social environment. The ideal is of a unit whose capacity for survival in a competitive world is a reflection of the suppression of competition within.

106

5 The life-cycle

Introduction Basically, the events and ideology associated with life-cycle change among Pira-parana Indians show a double theme which I suppose to be universal. They are concerned with (a) the difference between life and death and (b) the difference between the sexes. In this chapter, I regard the life-cycle as a dual process concerned, on the one hand, with the physiological and spiritual reproduction of the individual and, on the other, with the reproduction of the social structure. Of course, the two processes are related by metonymy, since the social structure is built up by the biological reproduction of individuals. I have already discussed many aspects of the reproduction of the social structure by means of exogamous marriage; here I turn to the ideology and ritual which accompany those changes that are concerned with the creation and development of the body and soul of the individual. In this chapter, the principal changes in this individual life-cycle process are related to the social structure, showing that ideology and ritual bring the biological facts of life, such as birth, growth, menstruation, sexual reproduction and death, into line with the existence of localised exogamous patrilineal descent groups.1 The end of life Life and death Life and death are alternate phases of a grand cycle. The dead are 1 We never witnessed the practices surrounding birth, death or first menstruation and, although a He wi, a male Yurupary ceremony (He House: see S. Hugh-Jones 1979), was held during our fieldwork this did not involve true initiation. These gaps in our joint

107

The life-cycle always calling to the living to join them and the living are always calling on the dead to succour them on ritual occasions. The living also install the souls of the dead in new-born children. In spatial terms, this alternation of life and death is represented by a vast circular river flowing over and under the earth. The upper half flows from west to east across the earth's surface; its lower reaches are called Ohekoa Riaga, Milk River. The lower half is the Bohori Riaga (boho-; to be lower), the Underworld River, where the dead go. The dead phase of the life-cycle is basically unknown and unknowable. On a practical level, Indians concentrate their efforts on shutting it out altogether. The body is buried below the earth's surface where the living never venture and various ritual acts 'cut off death' (bohori ta-) from the living. Those aspects of death concerned with the grave, the decay of the body and the disappearance of a once-known individual must not be allowed to overlap into the world of the living. One cannot mention the recently dead without adding 'the one who finished being' to the name. Signs of death, such as a gourd on someone's face or white worm-like creatures that live underground (see below) are treated with fear and disgust. A similar fear accounts for the live burials of very old or very ill people which have been frequently reported from the Vaupes area and definitely occur from time to time on the Pira-parana. This emphasis on the separation between life and death has two results. On the one hand, the dead phase of the life-cycle is the antithesis of life — a blank; a state of nonexistence. One Indian expressed this in response to questions about an afterlife when he said, T haven't died yet; have you? How should I know what happens?' On the other hand, it is necessary for the opposition between life and death to have some positive content. This is so both ideologically and practically, for, after all, people must do something with their dead, and they must be cutting off and avoiding something. Therefore, death is also conceived of as a reflection of life in which all elements and processes of life are reversed in a sinister way. The simplest demonstration of this is the reversal of river flow in the Underworld. There are several accounts of what happens after death, and these cannot be combined into a single, consistent theory, for they deal with different aspects of death rather than consecutive events. Yet all of these have a common antithetical relationexperience are simply due to chance and they illustrate the problems of working in extremely small and isolated communities. The accounts I give of life-cycle rituals which neither I nor my husband saw are drawn from informants' descriptions.

108

The end of life ship to life. They also leave vast areas of nothingness, for they are mainly known through the limited information provided by the few mythical episodes which refer to them. Here, I shall be concentrating on the separation of elements of the self at death but, before I begin, I must draw a distinction between the aspect I have been describing, which is bad, fearful and represents dissolution and decay, and the other aspect of death, connected with the 'dead' ancestors, which represents the origins of life. Contact with the ancestors, which takes place either through shamanism or directly at male ritual gatherings, is supremely dangerous but also an essential condition of continued life. T o live' and 'to recover from illness' are covered by one root, kati-. Similarly, 'to die' and 'to get ill' are covered by a single root, riha-. The meaning of these words reflects the Indian notion of life and death, because health and illness are inverse states of which life and death (distinguished by addition of a suffix denoting finished action) are extreme forms. Events after death The practice. Death provokes angry speeches from men who enlarge upon the cause of death, promising revenge if it was caused by human agents, and who remember the deeds of the deceased and lament the continual reduction of their descent groups. Women make sobbing speeches in the same vein but without the element of revenge. The body is put in a foetal position with a gourd over the face; it is bound with hammock rope and wrapped in the hammock of the deceased. Then it is placed with burial goods in a coffin made from a canoe cut in half and doubled over. Men perform the burial while women weep. A man is buried with ritual dance ornaments — feather head-dresses, monkey-fur tassels and so on - which are also referred to in other contexts as Bohori gaheuni, Underworld goods. A woman is buried with her little basket containing paint, a mirror and other personal things. Possessions not buried with the corpse should be destroyed and, if the person was important, the house should be abandoned. The grave is made in the centre of the house for a man and by the entrance to the family compartment for a woman; it is a deep hole with a side tunnel for the coffin. Once the grave is filled the shaman cuts off death (bohori ta-), cleansing the house with 109

The life-cycle tobacco smoke, burning beeswax and giving the occupants shamanised snuff. Then normal life resumes. There is sometimes a divination process in which cassava is left over from the grave while the occupants of the house retire. The marks in the loose earth indicate the sorcerer's identity, and the direction from which he came. One means of revenge upon a human sorcerer is to boil up a mixture of bodily substances removed from the corpse (hair, nail-clippings, a tooth) with chilli peppers. The sorcerer's spirit comes in the form of a bumble-bee and drops in the pot; at this point the sorcerer dies. People are terrified of the spirit, wdti, of the dead person until the shamanism is over. The shaman drives the wdti off to the forest. According to one informant, it is chased by a swarm of bees, werea, released by the burning of the beeswax. Until about twenty years ago a funeral dance with masked charades, many of them similar to those described for the Cubeo (Goldman 1963 : 2190, used to be held at an interval after the death of an important man. The body in the grave. The corpse remains in the grave where the rotten soft tissues of the body drain through a hole specially drilled in the canoe coffin. The entire mass of flesh, blood, fat and skin is called body liquid, ruhu oko, and thus is differentiated from hard bones and teeth. It is important that the earth should not touch the body until it has completely rotted — first flesh and then bones. This is the point at which the canoe itself begins to rot and the soul, usu, escapes, coming up to this earth. The soul takes the form of a bumblebee whose buzzing says, 'It is bad, you are living and I am disappearing'. When the living give the name of the dead person to a baby, the deceased is no longer angry, but happy because he or she lives again. The body in the Underworld. The Underworld is known through the myth of Live Woman's visit there. Since this myth is also important in my later analysis of concepts of space and time, I present it in more detail than many of the others. Live Woman in the Underworld As he was dying, a man told his wife to wait for him on the maniocgarden path two days after his burial. She took her small son and met her husband's spirit, which she did not recognise as such, in the way planned. He told her to cover her eyes to avoid fire ants and, so doing, 110

The end of life she arrived in the Underworld at the House of the Dead. Here she had several sinister adventures involving her aggressive spirit mother-in-law who loathed her for being 'alive'. This Spirit Woman sent her to fetch water in a pot which, from Live Woman's point of view, was an openweave basket. Later, in order to make a fire, she tried to catch a light from Spirit Woman's anus which glowed because of the fire in her body. When Live Woman's son was hungry, Spirit Woman offered him sweet potatoes which she roasted under her armpits and behind her knees, but from Live Woman's point of view these potatoes were white worms. At dusk the spirit husband called her onto the plaza to search for lice in his hair — a common conjugal act. She found his hair full of grave-earth and saw his eyes were fixed. She realised he was dead and wept. A tear fell on his knee making him furious. Although she protested that she was just sweating, he hid away in the manioc-fibre store, deserting her for good. Then she went down to the port on the Underworld River and, although Spirit Woman had expressly forbidden her to do this, she looked downstream. She saw that the river was choked with dance ornaments, the burial goods of the dead. Beyond was the port of Underworld Agouti house. She swam among the burial goods, arriving at the house a season later, to find Agouti Woman grating manioc. Again, Live Woman did not perceive things in the same way as her hosts: their pepper pot was shrimps and their discarded palm-fruit stones were edible palm-beetle larvae. Agouti Woman intended to present Live Woman's mother with a leaf-packet of fish and so agreed to take Live Woman back home. The 'packet of fish' was really chewed manioc root 'wrapped' in agouti faeces — the tell-tale sign of agoutis' manioc-raiding. Agouti Woman led the way, running, along a forest path only fit for animals, through hollow logs and prickly undergrowth, until finally they arrived at the manioc garden where the agoutis habitually stole manioc. There, Live Woman's mother was mourning her. Live Woman and her son were dead like spirits, but Live Woman was successfully shamanised back to life whereas her son died. This myth shows why women recover from severe illness while men die, and also why women outlive men. The oppositions between Live Woman's point of view and that of the Spirit Woman demonstrate that the Underworld is a reversal of life on earth: a continuous surface is full of holes (the spirit's pot), food for living people is worms that consume dead bodies (the spirit's sweet 111

The life-cycle potatoes) and this food is roasted in a manner associated with the burrowing of worms into parts of the body which are 'semi-orifices' (the spirit's armpits etc.). The opposition between the points of view of Live Woman and Agouti Woman is both weaker and more favourable to Live Woman, and this is consistent with the position of the Agouti house, midway between the world of the dead and the living. Indians say that the water of the Underworld River is composed of rotten corpses. The ritual goods are tied onto a stick when they are buried with a corpse and they are said to be the soul, usu, which helps the body to float in the river. This indicates a separation between the body, which is destined to rot and become one with the river, and the soul, which resists this destiny. The soul in the ancestral house. People also say that the soul goes to the ancestral house of the deceased's particular sib. They sometimes say the married woman's soul accompanies her husband's and sometimes that it returns to her own ancestral origins. Such houses, called people waking-up houses (masa yuhiri wi sing.), are geographical sites on the journeys of the ancestral anacondas and are located in mountains, rock formations and rapids (p. 33). They are beautifully painted longhouses in which the ancestors perpetually dance in fine feathers, drink beer and take coca. Shamans can travel in thought and see these places, but ordinary people can see them only on the point of death, when their souls have deserted their bodies. One woman told me that, when delirious from a snakebite, she stood on the plaza of such a house while her father and other dead kinsmen invited her in and the women begged her to help them make beer. If she had gone in, she would have died and joined her ancestral community. The elements of the individual separated at death First of all, it is important to clarify the concepts of usu and wdti which I have glossed as 'soul' and 'spirit'. Anatomically, the usu is located in the heart and the lungs. It is also breath (usu koa-, 'soultake', to breathe) and is connected with the animation of the body (usu hea-, 'soul-cross-over', to rest). A shaman's soul can travel away from his body at any time and experience the supernatural world. Beliefs about the nature of dreams and hallucinations are varied and often vague, but there is a general consensus that they do not depend upon the soul leaving the body; for the ordinary person this separation 112

The end of life happens only at death. The soul is thus portrayed as both concrete and substanceless, visible to anyone as heart and lungs, and yet also only visible to the shaman. It is not surprising that breath, which passes between the interior of the body and the external world, and is so clearly essential to life, is the medium of shamanism. Shamans blow spells, suck out foreign bodies and make extensive use of tobacco and snuff and cigars which are both inhaled and blown out. While each individual has a supernatural aspect manifest in the soul, from the point of view of the shaman, the aspects of the individual body and soul — are also aspects of the entire world. It is through his ability to separate the aspects of his own person that he can release his soul and travel between the two worlds. A wdti is a representative of the other, supernatural world which appears directly to the living. In order to appear to ordinary people it must have a concrete form or manifestation, but this is always the reverse of proper human appearance or behaviour. Most watia are terrifying forest beings, entirely evil, but there are also He watia, ancestral spirits, who appear to the living during male Yurupary rites (see p. 143), and the watia of the dead. The 'other world' is not a unitary concept in Indian thought: it embraces the distant natural world of the remote forest, the ancestral world and the sinister world of the recently dead.2 To return to the world of the dead, the different accounts yield a variety of conclusions. First, the two aspects of the individual are clear. In every account the body has one fate and the soul, or spiritual aspect, another. The body ends by rotting away: it becomes Underworld River water, or, in the grave, it liquefies and unites with the earth. In each case it goes down, down into the grave, downstream in the Underworld River and down into the earth. The soul resists ending, or disappearing (yayi-), as Indians put it. It is manifest in the feathers that float instead of being submerged, in the spirits of the Underworld house who remain while the bodies go downstream, in the bumble-bee who does not want to disappear, in the soul installed in a new child and in the perpetual ritual activity of the ancestral house. Compared with the bodily aspect it is always in a higher 2 There is a third category, wuho, semantically closely related to, and partly overlapping with usu and wdti. Wuho means 'imprint' or 'residue' - something made by the original being or object, but which is not actually that being or object itself. If a mythical character transformed himself into a bird, examples of that species of bird which exist today are his wuho. Wuho is evidently closely related to wuha, sleep, and is the word most commonly used for 'shadow', although wdti can be used as well.

113

The life-cycle position. The separation of body and soul is nevertheless incomplete; it is relative and contextual because the one cannot do entirely without the other. Thus, the rotting body has its soul ornaments and the beings in both the Underworld and the ancestral houses have bodies, although they represent souls. Besides this two-fold aspect reflecting the interdependence of body and soul, overall there is a two-fold fate of the soul itself, along the same lines. On the one hand, there is the fearful Underworld house, a reversal of ordinary life connected with the fate of the body in the grave and, on the other, there are the beautiful ancestral houses which are connected with the fate of the soul when it leaves the body before burial. To press the point home, we might note the diet of these different soul-beings. The Underworld spirits eat worms, which are way beyond the category of 'food' and are utterly disgusting, while the ancestors eat ritual substances, which are removed from ordinary food in the opposite direction as they are so pure. The separation of elements of the body initiated at death is summarised in the columns below. The analysis above has clarified why items can appear in both columns but, in spite of this, the overall contrast is one of 'bad disappearance' and 'good endurance'. Aspect of Individual Body Soul death

body buried with goods

burial

body rots away in earth

Underworld

body rots in river

body in river soul-beings

body becomes water spirits in Underworld

process evaluated

bad disappearance

soul escapes to ancestral house soul inhabits ne^ spirits remain in house ornaments float ancestors in origin house good endurance

The beginning of life

Instead of being an unpredictable and sudden occurrence, like death, the creation of life is gradual and more amenable to social control. For instance, the circumstances of conception (such as kin relations between the parents and the time and the place appropriate for sexual encounters), the integration of the newly born into the community 114

The beginning of life and the subsequent social development of the child are processes which can, to some extent, be socially moulded to bear out a particular ideological system. Besides, although the natural or physiological aspects of conception, foetal growth and the baby's postnatal development happen anyway, they are still open to various interpretations and different forms of purposeful guidance. The way in which a society promotes the growth process must surely reflect a corresponding theory of growth. Let us start with the theories of sexual creativity and the theory and practice of growth in the foetus and child. The events of birth and subsequent naming are best left until afterwards. Just as we found there were different accounts of the unknowable events after death there are different accounts of the unknowable manner of creation of the foetus but, in this case, too, there are consistent patterns. However, the general nature of growth processes is rather different from that of the death process, because the combination of sexual elements is so obviously crucial. Analysis of the beginning of life gives a sexual dimension to the body/soul dichotomy. Besides being set in motion by a combination of male and female reproductive elements, early growth has as its goal the repetition of the same reproductive process and this requires sexual differentiation of mature males and females. When we consider menstrual and initiation rites, it will become apparent that the dichotomy between the sexes replicates the dichotomy within each individual (of either sex) which is established in the early phases. Ante-natal development The ideal places for sexual intercourse are the port and the manioc garden, although hammocks in the longhouse are an alternative. The pregnant woman and the expectant father are called guda hako and guda haku, 'mother and father of the (belly) filling', names which already emphasise their joint responsibility for the foetus and correspond to the later teknonymy whereby each parent refers to the other as 'father/mother of my daughter/son'. The mother is not considered pregnant until she is 'filled up' by repeated intercourse. This is the point at which the soul-stuff derived from the father's semen has accumulated sufficiently to form a living' foetus which begins to grow. People say that female children are made of their mother's blood and male children of their father's semen, and that girls resemble their 115

The life-cycle mothers and boys their fathers. They also say that bones come from the father's semen and that the bone cavities are filled with semen itself {ahea badi). Alternatively, people say that the father makes everything - bones, skin, flesh etc. - with his semen. Again, people often say they do not know, because one cannot see it happening. No one actually said that the mother's blood creates the flesh of the child, although according to the logic of the system set up by the other statements this would appear to make sense. We may sum up the theories as follows: 1 father's semen : mother's blood :: body : girl 2 father's semen : mother's blood :: bones : ? 3 father's semen : mother's womb :: child : receptacle Here, I shall concentrate on theories 2 and 3, first showing why mother's blood should be considered responsible for the body liquid of the foetus. It is reasonable to assume that the relevant division is between bone and body liquid because this is regarded as the basic division of body substances in many different contexts. Menstrual blood is called women's body liquid, ruhuoko: the use of the general term 'body liquid' unites it with other soft tissues and liquid components of the body. Having children is thought to deprive the mother of life-force (katise) and fat, leaving her thin and dryskinned. Since it is also said that the life-force of women is repeatedly renewed by growth of menstrual blood, the loss of life-force after childbirth implies the loss of menstrual blood. There is also the physical existence of the umbilical cord and placenta. The cord is ritually linked with the river system of the earth and with the stems of cultivated plants: there is no doubt from the contexts of the link that it is regarded as a source of nourishment from mother's body to foetus. Again, contact with the red pigment used by men during 7/e wi rituals would make women bleed excessively from the womb, because it 'is' menstrual blood. Men apply the paint particularly to their navels, the very point at which the maternal link was originally severed, in order to replace their body liquid (p. 150). If all these factors support the association between female blood and the body liquid of the child, then why should no one have mentioned it directly? It does not really matter whether we consider the female blood a 'contribution' (theory 2) or a vehicle of nourishment of a foetus which is actually made of semen (theory 3), for the fact remains that the creative role of blood was only made explicit in the case of girl-babies (theory 1). It may have been chance that I never 116

The beginning of life heard this, but the omission does fit with other elements of Piraparana belief. First, there is a tendency to avoid distinguishing between flesh and bone in the context of death - we may note here that Pira-parana Indians do not (and say they never did) disinter their dead for the purpose of endocannibalism of burnt bones, as did certain other Vaupes groups (see, for instance, Goldman 1963 : 249-50). At death, the strong opposition between enduring soul and mortal body seems to eclipse the weaker division of the body into slightly more-enduring bone and less-enduring body liquid. This has the effect of emphasising the integrity of the living body and supporting the third theory of conception. Secondly, we shall see many instances of the lack of stress on the creative powers of ordinary mortal women which, I shall argue, follows from the patrilineal structure of descent groups. We saw that bone and body liquid alike decay in the grave but that bones, being harder, decay last. Not surprisingly, physical hardness and endurance are related together and opposed to physical softness (liquidity) and lack of endurance. Therefore, physical endurance of body substance is male like bone, and physical decay of body substances is female like blood or body liquid. Pira-parana Indians, in common with many other Amazonian groups, do actually regard men as 'hard' and women as 'soft'. They believe that men possess superior physical strength, are better able to keep ritual restrictions, resist sleep, go without food and so on. The sexual associations of endurance and decay immediately suggest a sexual dimension to the other items summarised under the body—soul dichotomy on p. 114. In particular, they suggest that the body is 'female' and that the soul is 'male'. This sexual dichotomy of the aspects of the individual is borne out at many other points in this book. Post-natal development If we look at the path of growth, it consists of foetal growth initiated by 'filling up' with semen and then this is followed by growth of the new-born child: these two processes are separated by the birth itself, to which we will return later. Whichever theory of conception is held, there is no doubt that both mother and father are essential. However, the soul of the foetus, created by the father's repeated insemination, is no discrete entity: we shall see that there are other contributions of soul as well. In fact, in 117

The life-cycle the context of life, soul (usu) sometimes has a very general meaning, more or less co-terminous with life itself. Anything necessary to life can be said to add soul, but there is also a particular occasion, naming, when a discrete soul is transferred to the child: I return to naming later. During the gradual growth of the new-born child, body and soul are drawn together. On the one hand, it is food that both adds soul and makes the body grow, thus milk, crops etc. are all said to give soul to the child. On the other, people say that it is shamanism that gives food these powers, and that without it all food would be lethal. Therefore, in the relationship between shamanism and food, there is a replication of the relationship between semen and mother's blood. In both cases, there is an initial life-giving element followed by a nourishing or transporting one. The first is male, since shamans are males, and the second is female, since women are responsible both for milk and then for the preparation of solid foods. Together, these elements give life and soul but, considered separately, they are related as soul (male) and nourishment (female). Like the act of insemination, which is subject to conscious control in a way that the female role in growing the foetus is not, shamanism represents the exercise of male control over female food-producing powers. Everything the child eats or otherwise uses for the first time should be shamanised. Some shamanism is designed to bring about specific changes in the individual and this is usually performed during a special ritual period, and other shamanism is designed to make things that the individual will often be doing safe and beneficial. Thus, kana fruits (Sabicea amazonensis) and mother's milk are the first things to be shamanised for a baby. The kana shamanism is a particular lifegiving act (see pp. 122, 125f), but the milk shamanism makes the sucking of milk safe thereafter. Neutralising danger and activating the beneficial effects of substances are often combined in single shamanic treatments. Before I describe the growth process, I should explain a little more about food shamanism. My information on this is far from complete and, in any case, I can only give a very general outline here. A great deal more information on food restrictions and the associated ideology is contained in Langdon (1975). I discuss the series of foods in more detail in chapter 6. Most of the potential dangers from foods derive ultimately from the mythical events in which various animals, plants and items of productive equipment figure. The danger comes not only from the nature of the raw materials but also from the productive processes of catch118

The beginning of life ing, killing, cutting, tying, burning, cooking, sieving etc. It may also come from the sex and the life-cycle state of the producers, and likewise the danger may apply differentially according to the sex and lifecycle state of the consumer. When an individual eats something or uses an object, he or she is not simply consuming the finished article, but is also absorbing powers associated with an entire process and all the social and physical relations involved in it. This view leads to a series of analogies between processes of production, patterns of consumption and processes of digestion, some of which will be set out in later chapters; it is also associated, not surprisingly, with a theory of the body which places extreme value on the regulation of exits and entrances. All I am able to do here is to summarise this view of the body and give a few illustrative examples. Dangers to the body are best explained by imagining the body as a vessel with special orifices through which foods, excreta, smells, sounds, breath and visual images should all go in or out in a regular and controlled manner. The outer covering of this vessel has differential permeability depending on context, so that some creatures and shamans can see through to the soul, and sometimes the covering opens up in sores and so on. Permeability of the skin and opening of the orifices is good in as much as good things go in and bad things go out, but it is bad in as much as bad things go in and good things are lost. Of course, the chances are that whether the orifices are open or closed the bad effect will come along with the good; this is prevented (a) by the regular alternation of opening and closing, a kind of dynamic moderation, and (Z?), by shamanism. I do not intend to present the evidence in support of this theory of the body here because it would mean duplication of much material that is presented elsewhere in this book. Illness gets into the body from the opening of the skin and orifices, and, correspondingly, it tends to show itself in malfunctioning of the skin and orifices. Stephen Hugh-Jones has shown, for instance, that when initiates are ritually opened up they are prone to a type of illness in which the body literally drains away through the excessively open anus (1979 : 198ff)- There are other ways in which illness is a repetition of events which precede it: some foods can introduce items of material equipment used in their production into the body and these implements attack the internal organs. Thus, if boiled maniocstarch drink is drunk on the wrong occasion (when the individual is 'too open'), the stick that was used to stir it will stir up the usu of the 119

The life-cycle drinker. Again, there is a significant category of foods described as 'grease-filling'. These are fatty foods, but more important seems to be the fact that they exude grease while cooking. When the individual eats these foods he or she is 'filled with grease' and then jaguars and snakes magically see the grease and perceive the eater as 'edible game'. Thus, the food exudes grease while cooking and then the consumer exudes grease after eating. These examples illustrate the general principle that relations and processes outside the body can be transferred to the body itself. It is just such a transfer that is made through shamanism of food. The shaman blows spells into a small sample of ready-prepared food and sends away the illness-causing agents associated with all the foods in the category his sample represents. Thus, if the category is 'large fish' he has to mention and deal with every sort of large fish. The client eats the food and is protected, so that he may eat any member of the category in future (until his ritual state undergoes change). The illness-causing agents are not sent away altogether; they are just sent away from those who eat the sample which has been shamanised. Thus, the effects of the shamanic process are transferred to the body of the consumer (see fig. 13). The new-born baby is utterly dependent on its mother for milk and general care; it has orifices which are small, undeveloped and uncontrolled. It sucks, sleeps, cries, defecates and urinates at irregular and short intervals. As it grows, it learns to control its orifices: crying gives way to speech - but here another form of control must be shown for the child must speak its father's language and not its mother's - frequent sucking gives way to more regular patterns of eating and fasting and, at the same time, more 'advanced' things are eaten; frequent excretion gives way to control of bowel and bladder and at the same time the child must go outside to excrete; sleeping is gradually regulated until it is co-ordinated with the day-and-night rhythm of longhouse life. In these ways, the child is gaining selfcontrol and independence from the mother. As a result of these changes, it is able to form social relations with other members of the community. It does this through the simultaneous regulation and enlarging of its orifices. In a sense, the two processes are naturally bound together, for larger orifices can let more in and out at once and thus can be shut for longer. Although the developing child comes to abide by the periodicity of longhouse life, as long as it is only a consumer, filling its waking intervals with play instead of productive activity, 120

The beginning of life a ACT OF SHAMANISM

b RESULTS OF SHAMANISM

Fig. 13 Transference of shamanic activity to body of client 121

The life-cycle the integration is only partial. As in most societies, girls change faster and earlier than boys in this respect. The latter go through an abrupt change at initiation. These bodily changes are ones of natural ability, but the new powers are socially guided so that the right language is spoken, and the right times and places are learnt for each activity. But they are socially guided in another way by shamanism. The new-born child is taken through a cumulative series of foods of increasing danger until it reaches a mixed and adult diet. This is the series that is repeated in the same or more condensed form after each ritual rebirth but, in the case of the new-born baby, the time-span between members of the series is longer than on any subsequent occasion. A condensed form consists of fewer categories, because one particularly important category may be taken to represent a number of consecutive ones. The nature of the myths, the productive processes, variety of raw materials and types of natural growth which are relevant to the construction of this series lead to a vastly complicated and cross-cutting system, but nevertheless there is a gross pattern to the system which contains the following relations: less dangerous small vegetable particles below ground . . . . above ground roots fruits, stems, leaves

water small fish

more dangerous large animals land . . large mammals

These relations have also been pointed out by Stephen Hugh-Jones (1979 : 93) and will be the subject of more detailed analysis in chapter 6. At the very beginning of the child's life, milk and kana are shamanised. The function of milk shamanism needs little explanation. Kana fruits, which are like small, soft, sweet-tasting pink gooseberries, are extremely important ritually and are certainly polysemic as life-giving symbols. Here, I would stress first, their ritual name as 'Sun's eyes' and secondly, the fact that they are supposed to cause white-eyed babies if eaten in pregnancy. The Sun is the source of both light and life and we may guess that much-desired white eyes are related to the 'Sun's eyes' and thus to lightness or clarity of vision. Since sucking and seeing are among the most positive abilities of the new-born baby, it is reasonable to suppose that shamanised kana opens up visual experience in the same way that shamanised milk opens up alimentary experience. In support of this, I might add that the next time young boys take shamanised kana is when they see Yurupary instru122

The beginning of life ments for the first time at initiation. When the weaning process begins, the child is opened up ritually by the progression from smallparticles vegetable foods to the meat of large animals, whilst simultaneously going through a two-phase growth process. In the first phase, from underground roots to above-ground plant parts, which starts and ends with products of manioc plants (p. 216), the child completes a process of upward growth analogous to that of a manioc plant; in the second phase, from small fish to large animals, it completes a process of emergence from water to land. Land animals (mammals and game birds) are called 'old or mature fish', wai bukurd, which confirms that the water-land change is indeed conceived of as a development. We might also have deduced this fact from the progression of the ancestral anacondas from Water Door to house sites. According to the sexual division of labour, we may describe the first phase as female and the second as male although, added together, the two represent a gradual weakening dependence on females alone and concurrent strengthening of dependence on a heterosexual productive system. Therefore, at the same time as achieving an alimentary opening up of the child, the shamanism of the food series achieves a gradual integration of the child, as consumer, into the socio-economic life of the community. We shall see that the structured growth process revealed by this very limited analysis of the food series is also contained in the sequence of events at birth. It is clear that the positive benefit of the shamanism, as opposed to the dispersal of danger, lies not so much in the individual acts, as in the relations between the items in the whole series of acts which promote a growth process. I believe that this is why people tend to emphasise the dangers avoided rather than the positive benefits bestowed when they mention isolated acts of food shamanism. The development of the child is represented in fig. 14. Let us now return to the events at birth which divide foetal development from that of the living child. Birth Birth should take place in the manioc garden, the mother being accompanied by an experienced woman but no men. After the birth, the cord is cut and the placenta buried. The mother and child return to the house, entering through a side door. Prior to their entry, beeswax is burned and all household goods - ritual items, pots, coca equip123

FOOD SHAMANISM

Q BODY

(JSOUL

[ MJLK ] [

UNDERGROUND • 5MALL

FOOD SEKIES

•OVERGROUND

JL

WATER.•

LAND LAROE

•cf-

FEMALE & MALE PRODUCTS MATURITV

TOWARDS

CHILD CROWS BIGGER 6- MORE OPEN

OPEN/APP£TITE SATISFIED

CHILD LEARWS SELF-CONTROL

CLOSED/ RESTRAINT

Fig. 14 Aspects of child development 124

The beginning of life ment and so on - are moved onto the plaza to prevent contact with the blood of birth which would injure those who use them later. Mother, father and child are secluded in the compartment for several days and both parents are restricted to a diet of ants and termites (of specific varieties which live underground), cassava (manioc bread), starch and water. At the end of seclusion, red paint is smeared on wife, child and husband and, after the river has been made safe by water shamanism, the three bathe at the port. The goods are removed again while the family re-enter the house, which has been cleansed with beeswax smoke. After their return, mother and father eat a meal of small fish boiled with pepper. The seclusion foods and those eaten afterwards must all be shamanised. The parents should not have sexual intercourse until the child is weaned, and the spacing of children suggests that the rule is generally kept. Close spacing is regarded as 'bad' as well as extremely impractical. Three years or more is the approved interval.3 Returning to the beginning, the baby is born in a female domain, the manioc garden. The birthplace and the path leading from it to the house are rihi rukuro and rlhi ma (tukuro, land; ma, path): rihi- means to whimper for maternal affection (n, blood or flesh ;hi-, to call) and is related to the words for children, ria (blood- or flesh-ones) and river, rlaga (-ga: suffix for certain containing objects). One alternative name for the birth path is 'manioc-stick path', a reference to the close association between women's procreative and manioc-cultivating powers. Yet another is watzr-kana path, kana being the fruit already mentioned in the context of birth shamanism. The significance of kana is complex: besides being connected with sight, the little pink fruits of kana are described as usua, souls or hearts, and kana evidently has a more general meaning in which birth is conceived of as a river-crossing. Birth shamanism was first gained through a rivercrossing, for it was originally obtained by Manioc-stick Anaconda when, transformed into a tick, he clung on to Tapir shaman as he crossed the river on the cosmological layer below the earth (see p. 261). The term 'to cross a river', kana-mo, perhaps relates this act to the plant whose fruit gives life to the new-born baby.4 3 Breaking this rule brings harm to the existing child rather than to the parents. The harm is in the form of mystical illness, and also in the direct damage caused by deprivation of the mother's milk and attention. 4 It is pertinent here that the kana plant itself has a stem that branches almost at right angles thus: Np . Small portions of the stem are tied together to make three-pronged painting sticks with which black paint is applied to the skin in basketry-like patterns (i.e.

125

The life-cycle The womb is referred to as masa yuhiri wi and the exit as masa yuhiri sohe, and these names of 'peoples waking-up house' and 'door' relate the female physiological birth process to the origins of social groups from the east. The analogy between the journey from the womb and the anaconda journey from the east is repeated in the description of both the umbilical cord and the river system of the earth as kana ma, life-giving paths. The river system of the earth is actually described as a branching umbilical cord connecting Indian longhouses, which are souls/hearts and also the fruits of the kana plant, to the ancestral place of origin in the east. Using these metaphorical links, we can now set out four analogous processes as numbered in fig. 15. (1) The birth of the child from the womb to the outside world, (2) the passage of the child from the manioc garden to the social world within the longhouse, (3) the creation of human descent groups by means of the ancestral anaconda journeys from the east and (4) the growth of the kana plant from underground roots to above-ground fruits. Each process has a starting point and an end point connected by a path, and at either end of the path is some mark of discontinuity such as a boundary or door. However, so far we have not considered the physical tie between mother and child — the umbilical cord — which as the source of foetal nourishment and a kana ma is obviously of prime importance. If we compare the physical entity composed of placenta, cord and foetus/ child with the processes already set out, we see from fig. 15 that the placenta is clearly the starting point, since it begins by being attached inside the womb and is the source of nourishment. In fact, the name of the placenta - ri saniro, 'round enclosure of blood' — likens its shape'to the round womb and the round manioc garden. The cord is the 'path' and the foetus/child is the end point (5). However, immediately after birth, the cord is severed and the placenta buried (6). The continuous transformation represented by the path of nourishment is abruptly severed by human intervention and, at the same time, the position of the placenta with respect to the child is reversed for, instead of being above, it is below. Now, in general, human growth is upwards and thus the burial of the placenta artificially fills in the ante-natal stage of human life, making foetal growth and childhood development seem joined in a continuous process of upward growth away from the earth and away from maternal origins. We shall see weaving patterns). This suggests once more that kana is associated with right-angle crossings, which are the essence of weaving.

126

PROCESS

STARTING

\ BIRTH cW\\d

WOMB

PATH

BOUNDARY

END

CERVIX

VAGINA

ENTRANCE TO VAGINA

OUTSIDE WORLD (new-born child)

GARDEN BOUNDARY

GARDEN PATH

BOUNDARY

of

BRINGING-IN 2 OF CHILD

MANIOC GARDEN

child)

movement of

child

CREATION OF I HOUSE BEYOND', 2 DESCENT GROUPS) WATER DOOR WATER DOOR movement of anacondas/ ancestor* GROWTH OF KANA PLANT

UNDERGROUND! EARTH ROOT SURFACE

^ NOURISHMENT! 5 OF CHILD

PLACENTA IN WOMB

METAPHORICAL 6 OROWTH OFCHIlDi FROM EARTH

PLACENTA IN GROUND

L

Fig. 15 Metaphors of birth 127

JOIN WITH CORD

--'"EARTH SURFACE

MILK RIVER

UMBILICAL CORD

MOVEMENT \JP—•DOWHJ

FAMILY SIDE DOOR OF WEST — • EAST ^PARTMI COMPARTMENT! COMPARTMENT) isoaa) child)

PORT

(river

STEM

T

ORIGINAL iHOUSE SITES

STEM/FRUIT

FRUIT

JOIN WITH CHILD

CHILD

BELLY BUTTON

k ^

SEVERED CHILD

EAST — • WEST

DOWN—• UP

jDOWN—*-UP

The life-cycle that this conception of human life as a growth away from maternal origins occurs in other forms also. At the same time, the burial of the placenta serves to contrast the sudden and completed creative event of birth, a downward movement, with the gradual, unfinished upwards growth of new life. Seen in this way, childbirth occupies a position of 'sudden change' between foetal and childhood growth just as the severance of the cord does. In support of the metonymical relationship between cord cutting and the entire childbirth process, we may recall that the cord is a metaphorical nourishing river which is cut across the flow, while birth shamanism has its mythical origins in ariver-crossing,too. Childhood growth which follows the severance of the cord takes place in the family compartment and is shamanically initiated by the administering of kana fruit, so that all the end terms of the processes shown in fig. 15 are brought together in practice. There is another property of kana which is highly significant here and suggests a further theme of analysis. This is the cyclical nature of plant growth, for while kana fruit is the end of one growth process, the detached fruit is the beginning of another. In this way, the kana fruit is analogous to both the sexual reproductive elements of semen and female blood, which are at the same time products of one human growth process and initiators of another. We have already seen that the placenta is 'planted' in the ground and there is a sense in which semen is also 'planted' in women, since there is evidently a metaphorical link between the womb as the source of human life and the earth as the source of plant life. The nature of childbirth and the total exclusion of men, even in the essentially ritual act of burying the placenta, together with the location of the event in the manioc garden — a female domain — make this a female occasion par excellence. This 'natural', female birth of the child contrasts with the later seclusion and bathing which, I shall argue, refer to the child's family status. A child that is unwelcome or still-born is buried immediately without making the journey to the house. Infanticide occurs in cases of illegitimate children, when a child follows a run of children of the same sex (especially if they are girls) or when relations between the mother and father are strained. I believe such children are usually buried alive, although it is possible that they are sometimes killed first. In any case, the fate of the placenta is already reminiscent of the fate of the dead, and the burial of unwanted babies in the manioc garden merely brings home the parallel. At death, the body turns into 128

The beginning of life rotten liquid and amalgamates with the earth; at birth, the placenta is buried as it is about to rot so that the child is made to grow metaphorically from rotten body liquid in the earth - the state to which it will return at death. If we are to include insemination in this upwards and downwards cycle, then it is clearly a downwards movement which transfers semen into the womb/earth from whence the foetus grows upwards, emerging from the womb/earth at birth. In this respect, insemination is like death - a fact recorded in the myth about the origin of cultivation of coca, when the hero dies on the point of ejaculation and then grows upwards as coca plants (p. 212). Besides, we shall see that, in as much as insemination renews the body of the father in the child, it is a kind of death of the father because it marks the end of a generationspan (p. 164). The burials in the manioc garden by women, and in the house by men, correspond to the natural quality of the new-born baby and the social quality of the person who dies later in life. Nevertheless, analysis of part of the Live Woman myth (p. I l l , analysed on p. 187) will show that, after a death, it is actually the weeping of women that makes the body rot. This means that, in one sense, we may say that women control the natural growth and decay of the body — a conclusion which fits the male-female association of the soul—body dichotomy. However, although the entire life-cycle has a female aspect associated with the physical body and a male aspect associated with the soul, from another perspective it is the underground phase, during which there are natural decay and natural origins of life, which is female, and the socially controlled above-ground phase which is male. The growth from a female, underground domain to a male, above-ground one is confirmed by the shared root in umua-, to be tall or on high (e.g. Umuari Masa, Sky People) and umu, man. The same low/female : high/male opposition which provides the poles of individual growth also differentiates between grown members of the opposite sexes, for women sleep below men, sit closer to the ground, exploit low natural resources and so on. The life-cycle of the body with its male-female polarity is represented in fig. 16. While it is typical of Indian thought that there is a male-female polarisation within the life-cycle of the individual of either sex, this notion does present a problem. After the discussion of initiation, it will become obvious that boys do grow away from their mothers into a male adult world, but that girls remain in the female domain of 129

The life-cycle

ABOVE GROUND etc.

UNDERGROUND

Fig. 16 Life-cycle of the body 130

The beginning of life another longhouse on marriage. Girls move away from their mothers by developing female reproductive and productive powers equivalent to those of their mothers. However, the change from female child to female adult does bring with it the ability to make direct sexual and economic partnerships with men, and thus the girls' growth towards a 'male' maturity may be understood as a growth towards a complementary relationship with adult men. This complementarity is perfectly well represented by the gradual introduction of male-produced foods into the diet. At childbirth, it is notable that mother, father and child are secluded together and bathe together afterwards. Back in the house, the mother and father share a meal of shamanised pepper and fish, but the baby remains on its milk diet. Although the mother has been released from the seclusion restrictions, she still retains a dietetic link with the baby, for she moves through the series of dangerous foods that must be shamanised much slower than the father, for fear of passing these foods to the child in her milk. The same pattern occurs with physical contact for, after the joint seclusion in the family compartment, the father resumes normal male social life which takes place in the centre front of the house, while the baby accompanies the mother in her female existence towards the periphery of the house. These are strong indications that the seclusion is concerned with physical identification of the three and is contrasted with the later physiological tie between mother and child which excludes the father. If the father breaks the restrictions of the seclusion period, he is much more likely to harm the child than himself, although some particular dangers apply to him too. The fact that the harm to the child stems directly from his actions and does not require intention on his part is further evidence of the physical identification of the two. 5 This direct suffering of mystical consequences because of a breach of restrictions (a manifestation of lack of self-control) normally reflects on the person who commits the ritual offence, and therefore it implies a unified self interacting with an external world. If father and child are considered to be linked in a manner analogous to the linkage between 5 This may be compared with evidence from the Ge tribes of Central Brazil, which shows that members of nuclear families are believed to transmit illness to one another automatically. Both T. Turner for the Kayapd (n.d.) and C. Crocker for the Bororo (1977) consider that this demonstrates the physiological unity of the family group. Among Pira-parand peoples, the analogous physiological unity is fleeting because it gives way to relations in which illness is transferred by non-automatic means such as shamanism, contamination with dangerous foods and so on.

131

The life-cycle the actions and their consequences of a single being, we may take this as evidence that the father (representing action) is in the act of transferring his substance to the child (who represents consequences). The natural process which actually achieved this transference was the 'filling up' by repeated insemination and so, amongst other things, seclusion is a reference to the father's physical contribution to the child. The unity of father, mother and child represents the physical conjunction of the three elements — penis, womb and incipient child (semen). I give further evidence later from an analysis of the seclusion diet, that seclusion constitutes a kind of reenactment of pregnancy (or, more generally, the state intermediate between death and birth), in which particular aspects of female sexuality are eliminated. In this way, the development of the child before birth is brought under conscious, shamanic control and removed from the hidden, involuntary control of women's natural procreative powers. However, the replication of the father's body in the child is not the only reason for the seclusion. Indians say that the mother should be secluded anyway on account of the loss of contaminating blood, and we may guess that the child should be secluded because it has passed from death to life and from contact with female blood in the womb to a relative independent existence outside. In fact, the association of seclusion with the actual birth process is reflected in the threat of Taking-in People', which affects mother and child more than the father (p. 267). The seclusion is therefore concerned with a number of concurrent transformations: I return to that aspect of it which concerns female renewal and the loss of blood when I discuss menstruation. The seclusion after childbirth, like other ritual seclusions, ends with a ritual bathe followed by a resumption of normal life. The removal of household goods after the ritual bathe makes it an obvious repetition of the original journey of mother and child into the house after the birth. This 'second birth' from the river recalls the emergence of the ancestral groups from the river at the first house sites. We have already seen that the identification of the womb (people's waking-up house) with group origins implies an analogy between natural childbirth and the origin of descent groups. The progress from manioc garden to house and the progress from river to house represent the separate terms of this analogy. If we now summarise the whole process from manioc garden to incorporation into the normal communal life of the longhouse, we see that a natural mother—child 132

The beginning of life tie (birth in manioc garden) is followed by the addition of the paternal tie which decides the child's social identity (seclusion of family unit) and then by the rebirth of the child and parents (entrance after bathe). There are two important features of this overall process to notice: the first is that the child is born from the ground and then reborn from the river so that the birth process has exactly the same structure as the shamanic food series which controls subsequent growth, and the second is that the family relations between father, mother and child established during seclusion stand midway between the natural mother—child tie and the future incorporation of the child into the local descent group through naming. Naming So far, I have concentrated on contributions to the living body rather than the soul, although we have seen that the separation of body and soul is carried over into the relation between the sexual contributions to the body: female blood and semen. Naming is the occasion when the child receives soul in a pure and non-concrete form. All my evidence suggests that naming is distinct from the seclusion and I do not believe it ever happens during that period. Nevertheless, Indians say a baby should be named within a few days of birth, for otherwise it would not live. The ethnography suggests that the interval between birth and naming varies considerably over the Vaupes as a whole.6 A baby should be named after a dead patrilineal relative of the second ascending generation and appropriate sex, so that a boy should be called after his paternal grandfather (or FFB) and a girl after her father's father's sister. This name is a 'shamanic-name' (baseri wame) because it is transferred shamanically, with red paint and milk acting as the vehicle.7 The naming also 'changes over the soul', usu-wasoa-, of the dead ancestor into the child so that this soul is prevented from 6 Koch-Griinberg on Tuyuka (1909/10 : 313): after bathing rite. McGovern on Taiwano (1927 : 252): third day after bathing. Briizzi da Silva on Brazilian Vaupes region (1962 : 429), Reichel-Dolmatoff on Desana (1971 : 140), Stradelli on Vaupes in general (1928/9 : 537): after three years. Goldman on Cubeo (1963 : 171): after one year. I never witnessed the whole sequence of events after birth but the isolated events I did see while travelling suggested that the seclusion is not observed as strictly as it might be, at least not after the birth of a third or subsequent child. 7 Men usually have a joke-name (ahari wame) as well. Both men and women also have Spanish names; these are called gawa wame, white-people's names. Joke-names, and some of the common Spanish names, are linked to shamanic-names, so that separate individuals may have all three names in common.

133

The life-cycle 'disappearing'. The important thing about the soul is that it belonged to the recently dead — 'three years' was the interval after death given by one informant. It is quite clear that naming anchors the child to the local descent group by establishing its patrilineal kinship to other members with reference to a recently dead member. Naming therefore completes the tripartite series of mother-child tie, family ties, local descent-group ties. It is the latter that articulate the community as a whole. This is the point at which to summarise the more definite manifestations of soul. First, there is the division between the body liquid and the bone. The bone, because it is filled with semen, which is a manifestation of father's soul, is linked with soul, but, nevertheless, it is also firmly attached to the flesh, and the pair are opposed to the soul in the grave. Next, there are heart and lungs, both 'soul' (p. 112), both part of the fleshy 'liquid' body substance and yet differentiated as receptacles of blood and air respectively. They must both be in working order at birth, for Indians recognise that a live baby breathes air and has a beating heart. Together, these are the most female manifestations of soul, but, taken separately, they are opposed as blood to breath or body to soul. After death they rot with the liquid parts before the bones. 8 The name is substanceless but endures better than soft organs and even bone, since it comes back to the world of the living. We also have good evidence to suppose that the 'liquid' is female and associated with the mother's contribution to the foetus, while the bone is male and associated with the father's contribution; the name is patrilineal and comes from the second ascending generation but belongs to a set distinguished by sex. The soul of the individual thus has a one-generational and a two-generational cycle. The 'one-generation' soul rots, which does not matter for it has been passed on to a child during its foetal development, but the 'twogeneration' soul is kept for the patrilineal descent line (see fig. 17). Menstruation The practice At the onset of her first menstruation, the girl is secluded in a 8 Although I have no direct evidence that the lungs last longer than the heart, it is tempting to regard the floating ritual goods, the soul of the rotting corpse in the Underworld river, as the lungs. Animals are normally gutted into the river and the lungs float downstream attached

134

Menstruation

HEART

LUNGS

(BLOOD)

(AIR)

SOFT ORGANS

BONE

NAME

(SEMEN) ROTS F I R S T

ELEMENT FROM MOTHER

ROTS SECOND

Cf ELEMENT FROM FATHER

1ST ASCENDING GENERATION QIFTS OF BODY SUBSTANCE

DOES NOT ROT

C?&Q ELEMENTS FROM FF&FFZ 2ND ASCENDING GENERATION OlPT OF SUBSTANCELESS SOUL

Fig. 17 Elements of the soul

screened-off portion of her family compartment designed to protect her from fire. She observes the starch cassava, ant and termite diet typical of seclusion periods. If they are available she eats umari fruits. She keeps many other restrictions, such as avoiding her hammock, mirrors and other objects besides. Her seclusion ends in a ritual bathe accompanied by vomiting and, like a woman after childbirth, she enters the house by a side door after the household goods have been removed and beeswax has been burned. Ideally, the period of seclusion lasts five or more days. During the seclusion, the girl has her hair cut close to her head by an older woman. Thereafter it is allowed to grow much longer than before, but only married women or those well past the menarche have fully grown hair. Various shamanic acts are performed for the girl: the most important are the treatment of seclusion foods and then the series of crops, fish and meat which is also treated for the new child. Pepper shamanto the submerged remainder. Besides, the lungs, in contact with the air, suggest continued breathing and thus an affinity with life rather than death. The theme of floating lungs as mediators between water and air and also between death and life also appears in the key myth of Mythologiques I (L6vi-Strauss 1964).

135

The life-cycle ism, which occurs at the end of seclusion, is the most significant point in the series.9 For subsequent menstrual periods a weaker version of the same procedure is followed. The girl is secluded in the family compartment and eats the minimal diet (although ordinary cassava may be substituted for starch) for one to three days. She returns to the normal diet after bathing, vomiting and eating shamanised foods. The series is reduced to its essential elements — in an extreme case fish cooked with pepper stands for the entire cumulative series. Menstrual restrictions are less marked as life progresses but, whatever her age, a menstruating woman does not cook for others or do any manioc work for at least a day. The nature of the menstruating woman The terms I heard for menstruating women were as follows: guabeko, (the one who is) not going bathing bedigo, keeping ritual restrictions ngamo huyago, sitting as a female who has reached the menarche kerea soego, singeing sloth kerea soe-bago, singeing and eating sloth nyama sesogo, smoking deer hetaga ma keo-soego, clearing and burning the path to the port ngunanya roago, boiling up leaves of cariauru (Bignonia chica) for red paint The references to avoiding bathing, keeping restrictions and sitting describe her state literally. Bedi- is very probably related to bedi, bedeo, younger brother, younger sister, who are therefore 'forbidden ones' (see p. 219). Nagmo means specifically 'one who has reached the menarche', but more generally it means 'free female partner'. A boy initiate is ngamu, 'free male partner', and the root ngam- means 'to reciprocate' or 'to do in both directions', for instance, 'to fight each other' is ngameri sia. The other references suggest a variety of associations between menstrual blood and fire and another between menstrual blood and red paint. The burning of the port path links the path from river to house with the vagina, a metaphor which is 9 I do not know whether pepper shamanism is as crucial in the childhood series as it is after menstruation and initiation, when it marks the end of seclusion. Since (a) no special childhood pepper rite was mentioned and (b) the child is not leaving seclusion at the time it first eats pepper, I assume that pepper is more significant as an agent of transition in later life.

136

Menstruation consistent with the previous metaphorical link at birth between womb and vagina and ancestral origins from the river. Menstruation in myth Various myths about Romu Kumu, the female creatress, are essential to an understanding of menstruation. Here I give extremely condensed excerpts, because the mythology relating to menstruation and male initiation has been dealt with in detail by Stephen Hugh-Jones (1979). (1) Romi Kumu lives up in the sky and is the first grandmother of us all; she is immortal because she has the sacred beeswax (werea) gourd with her. She grows old during the day, bathes at dawn and becomes young and white again. She alsb renews her red face paint, urucii (musa;Bixa orellana, used exclusively by women), and takes off a layer of skin with the old paint. This paint is her menstrual blood. Her name means 'Woman Shaman' but she is like a man. (2) She tried to initiate the ancestral people with shamanised substance in a gourd. They refused the substance, saying it was bitter and stank of her vagina. She hid this powerful gourd behind her in anger, and snakes, spiders and white people stole it, hence their skin-changing powers (and the clothes of whites). The ancestral people pursued her and got a second-best gourd: that is why Indians are mortal. (3) She was Poison Anaconda's daughter. She stole the Yurupary instruments from the port and played them at dawn when her brothers were too lazy to get up and bathe. While she had them, men were like women - they menstruated and worked manioc. The men pursued her, retrieved the Yurupary, and punished her by making her menstruate - that is the origin of female menstruation. (4) She had fire in her vagina. In order to obtain it from her jealous guard, her grandchildren asked her to singe a game animal for them. She squatted over urucii sticks to light them and, as they caught fire, the grandchildren stole them away. That is the origin of domestic fire. (5) She was a monstrous, sexually voracious woman with fish-poison for pubic hair. She raped Warimi, who was in the form of a bumblebee, and took him to her father, Poison Anaconda. Warimi got inside Poison Anaconda's body and stole his gall-bladder to make curare. While he was boiling it up snakes, spiders, etc. stole it: that is the origin of their poison. 137

The life-cycle The myths link menstruation to the shedding of outer skin, bathing and the powers of the beeswax gourd. They also link menstrual blood, red paint and fire. They give menstrual blood, which comes from women's wombs, a male counterpart in the poison which comes from a man's liver. They also show that Yurupary and menstruation serve to differentiate the sexes. With reference to fig. 18, the progression (2), (3), (5) is particularly concerned with the loss of shamanic power to men. In (2), Romi Kumu has shamanism in her body and the ancestral people depend upon this. The shamanism, in its form outside her body, is the beeswax gourd for which she is successfully pursued by the ancestors. In (3), the ancestors have become Yurupary instruments (which bear the names of the ancestors in (2) today). The instruments are now the object of pursuit and theft, while menstruation — a transformation of Romi Kumu's immortality - is a punishment. Romi Kumu is an underhand thief who steals when no one is looking. She is like the snakes and spiders in (2). In (5), she is worse still; her shamanism which became menstruation has now become a dangerous sexual appetite. It is Warimi who has shamanic power; he is a creature of 'pure soul' (hence his transformation into a bumble-bee — see p. 110) and can enter the bodies of Romi Kumu and Poison Anaconda.10 Thus, in the series, shamanism is removed from women and attached to Yurupary instruments which are opposed to menstruation. Menstruation is like poison. However, the true poisons belong to Poison Anaconda and Romi Kumu as parts of their bodies: curare, used to kill game animals, is removed from the body of Poison Anaconda by a person of pure soul and it therefore attains an independent existence (like a soul). Fish-poison remains part of Romi Kumu's body. From these myths we may draw a set of creative relations which are already familiar: body : soul :: women : men. To these may be added another pair — fish : game animals — on account of the final ownership of poisons. The female-male relationship between these two has already been demonstrated (a) in the relationship of Yawira (fish) to 10 There are many reasons for calling Warimi 'pure soul' besides his transformation into a bumble-bee. One of his names is Ruhu Mangtt, One without Body. He was the product of incest, and instead of his mother releasing him from her womb, jaguars released him by eating her up. He had a second gestation in the river and thus the normal opposition between male and female elements were not present in either his conception or birth. The female element was 'too male'. Besides this, he had no substance or fixed bodily form — he could escape through the narrowest basket-weave and he could change into all kinds of creature. For the story of WarirnVs life, see Stephen Hugh-Jones (1979 : M.4.).

138

Menstruation Yeba (land animal) and (b) in their relative positions in the shamanicfood series. Menstruation in the life-cycle Menstrual blood is retained in the womb as if by a dam and then MYTHICAL EPISODE

LOSS OF GOURD TO MEN

9

cf

SKIN-CMAhJ—^

progre&siort of single individual

>

progression of

%

application of black p

O

application of red pai

Fig. 21 Body paint in He wi cycle 152

MALE SECULAR WORLD ADULT MAW EXCHANGES £ COMMUNICATES WITH ADULT WOMEN

ETC."

Male initiation (He wi) matic form. First, He wi is shown as an extreme form of ritual in fig. 20 (a point which has been made in a different way on p. 146) and then He wi is shown in its relation to the male and female secular worlds in fig. 21. Fig. 21 shows that all the significant categories of people in his life experience paint the individual at He wi: the mother he leaves, the 'dead' men he joins in the ancestral world, the live' women he joins in the adult secular world, then the repetitive cycles set in and he is painted by the live men in the secular world, then internally repainted by the dead ancestors in the ancestral world. Finally, it should not escape us that the overall change from solid black from the mother to solid red from the henyerio is an effective skin-change since the red is applied when the black has worn off. The black (we, Rubiaceae sp.) is a dye rather than a pigment and so, by the time it has gone, a layer of skin has been lost.13 The skin-change makes He wi a metaphorical menstruation (see p. 140) for which women are responsible. Besides this, the identification of black paint with 'dirt' makes bathing and black—red painting sequences analogous forms of skin-changing. In fact, this could have been guessed before from the daily renewal of Romi Kumu in the river. I do not intend to go deeper into the role of bathing in the He wi ritual except to point out that it serves the double function of changing the bather and giving him contact with the ancestral origins (see discussion of birth). The sacred instruments and female rites The account of He wi allows us to develop the relationship already derived from the myths about Romi Kumu — women : men :: menstruation : He instruments. There is a parallel between the opening up of male penises by the playing of He instruments and the opening of the female womb. This fits with an informant's statement that women see their hair as it falls to the front of their faces and this is their He which makes them menstruate. Tom Langdon reports that wom£n dream of He immediately before their first blood appears (1975). In diverse ways, women's natural menstrual cycle is linked to the concept of He and, conversely, He wi is linked to women's menstrual cycle. We were told directly that He wi is like women's menstruation, but that women really do menstruate while He wi is bahi kesoase, imitation. This statement implies that both are forms of 13 This black dye is not genipa and it is very possible that when other Vaup6s specialists refer to use of 'genipa' they are mistaken.

153

The life-cycle renewal repeated throughout a person's adult life but that basically menstruation occurs naturally, in spite of all the shamanism and ritual that surrounds it, whereas He wi has to be socially contrived: unfortunately the ancestors do not come out of the river of their own accord. This situation is related to the fact that the onset of menstruation is a sudden and obvious mark of female physiological maturity, while male maturation has no equivalent focus. Just as myth shows that male shamanic power came from Romi Kumu's beeswax gourd, so the actual use of the gourd during He wi appears as the source of male renewal (or initiation). The beeswax gourd is even more sacred than the instruments themselves and is kept secluded with the shamans. The climax of the He wi rites is the coincidence of the burning beeswax from this gourd with the initiates' first attempt on the Yurupary instruments. In other words, the sexual potency of the initiates is made to coincide with the release of the contents of the female womb. After the appearance of the He wdtia, which confirms that the living are now in maximum contact with the ancestral world, the participants identify themselves with female sexuality by actually eating from the gourd (they eat coca impregnated with beeswax aroma). These actions bring male and female sexual processes together so that the female opening (release of beeswax) assumes joint responsibility with male opening, the seeing at first and then the blowing of the instruments, for the change in state of the initiates. Blowing the instruments up and down the beeswax-filled house and eating from the beeswax gourd can be regarded as acts of ritual fertilisation. Pira-parana Indians say that the house represents a womb, and they also say 'penis is eating' as a joking metaphor for sexual intercourse, so that this would be entirely consistent with their own ideological forms. Sexual intercourse is a time when both men's penises and women's vaginas are opened for the womb to receive the seminal soul-stuff which will develop into a foetus. In He wi, as in the joke, the fertilisation is reversed and the man is a mouth who eats the contents of the womb instead of losing his own soul-stuff through the penis. Therefore, during 77e wi, female sexual material is passed from an ancestress, Romi Kumu, to the men, who have become identified with ancestors for the purpose. Furthermore, the previous analysis of the nature of shamanised samples showed that eating these samples transfers processes outside the body to within it, and thus it could 154

Male initiation (He wi) be said that the men are made to experience an opening which repeats that of Romu Kumu (pp. 119f). Although the ultimate creative power is allocated to female sexuality in the rite, real women are excluded. Even seeing them, let alone touching them or having intercourse, is prohibited. The prohibition protects men and women alike, for ancestral sexuality, whether male or female, is dangerous to ordinary women, just as the sexuality of these ordinary women is dangerous to the 'ancestral' men. By possessing the beeswax gourd (amongst other things, including ancestral red paint) the men appropriate the ultimate female powers of sexual reproduction for themselves and so maintain their control over women. This reminds us of one of the fundamental differences between the menarcheal girl and the He wi community — while the girl must voluntarily withdraw to the edge of the house, the He wi community of men do not withdraw from the women, they send the women away to the periphery. The natural and social We have seen the social character of He wi both in its 'contrived' occurrence and in the numbers of individuals involved. We have also seen that the men are hierarchically divided both by broad age-based categories and by ritual occupations such as shaman and chanter. By differentiating between male social categories and integrating them into a single occasion, He wi stresses the interdependence and cooperative character of male roles and groups. Menstrual seclusion, on the other hand, sets individual women apart in an order which is purely random from the social point of view. The only distinction is between those who have reached the menarche and those who have not (who are never secluded). Although He wi is a social occasion in comparison to menstruation, both are basically concerned with the physiological changes in the individual and changes in the soul. The difference is in the fact that male physiological change and the accompanying soul-change is related to a change in social status, whereas the female one is not. The menarcheal girl does not change sleeping place; she does not learn particular skills; she does not leave the female secular domain and she does not enter a social hierarchy. The thoroughly natural character of menstruation and the semi155

The life-cycle social character of male physiological maturity are demonstrated by their concurrence with other cycles. He wi, together with fruit rites, makes up an annual cycle controlled by the stars, the maturation of the fruits themselves and the rhythm of wet and dry seasons. Menstruation obeys a monthly cycle and is thus related to changes in the moon. The Cubeo make this explicit by saying that the moon copulates with a girl, bringing on first and subsequent menstrual periods (Goldman 1963 : 180-1). Pira-parana mythology says the moon copulates with menstruating women and that during an eclipse of the moon, called the 'dying moon', the moon becomes a small red ball of menstrual blood which comes to earth and fills the house and its objects. The annual cycle is of great economic significance and He wi occurs at the point where the annual cycle renews itself. This is the change from the long dry season to the long wet season. Rain is conceived of as both a skin of the universe and also the menstrual blood, or sometimes urine, of Romi Kumu, and thus He wi occurs at a time of cosmic skin-change. This is consistent with the idea that He wi succeeds in renewing the natural processes of the world. While He wi renews the natural world as a whole and also the souls of its male inhabitants, the fruit ceremonies renew particular species of fruit. Thus, the annual ritual cycle in general is concerned with the maturation or renewal of the natural products upon which man depends, and at the same time with the maturation or renewal of men. The renewal of women, being associated with the moon, is also associated with the opposition between day and night. There is a myth in which Sun and Moon quarrel about which shall 'dry up the vaginas of women' and which shall be the night (S. Hugh-Jones 1979 : M.3). The benefits of the night are described as 'water for cooking and the rainy season'. We may conclude that the opposition between day and night is linked to that between dry and wet. Therefore the menstrual cycle with its alternate phases of loss and retention corresponds to both the daily cycle and the annual cycle. Now, we have already seen that the annual cycle is the menstruation of Romi Kumu. The daily cycle is also associated with the skin-change of Romi Kumu because she bathes at dawn (p. 137). However, the daily cycle has Underworld connotations too, because sleep is regarded as bad and akin to death. Besides, Romi Kumu's bathing at dawn recalls the point in mythical progression at which she had already lost some of her power, for she no longer had the beeswax gourd when she played 156

Male initiation (He wi) the men's He at the port at dawn. Remembering that ancestral female sexuality is opposed to sexuality of ordinary women, we may construct the following analogous set of relations: waning moon waxing moon menstrual cycle ordinary women Romi Kumu powerless (loses He, punished with menstruation)

night - day secular producing and consuming cycle ordinary women and ordinary men Romi Kumu semipowerful (loses gourd but plays He at dawn at port)

wet season dry season He wi and fruit ritual cycle Romi Kumu and male ancestors Romi Kumu allpowerful (possesses beeswax gourd)

Now we can concentrate on the Moon, who is closest to ordinary women. (1) He copulates with his younger sister without revealing his identity and makes her pregnant with Warimi who is 'pure soul'. (2) He comes down to earth and eats the bones of recently buried men who copulated with menstruating women during their lifetime. (3) He eats cooked agouti and swells and then goes hungry. This is his waxing and waning. He is associated with the Underworld. He swells by eating agouti, an Underworld animal, or alternatively he satisfies hunger by eating Underworld bones of the recently dead. This sinister activity is mythically associated with copulating with menstruating women. He also has incestuous sexual relations, and the lack of proper social differentiation between mother and father results in a child who is so entirely 'male' that he lacks a proper body. It therefore seems that the actions of the ancestral Moon have the common characteristic of being 'too natural'; they ignore the proper distinctions made by social rules, bringing things together that should be separate: brother and sister, penis and released menstrual blood, the living and Underworld bones. The rules are that one must not commit incest, copulate with menstruating women or contact the Underworld. There is no rule prohibiting agouti as food (in spite of the Underworld associations) but this is the weakest and 'latest' action of the Moon, since it is what he does today instead of in ancestral times. Even so, meat is last in the food series and thus meat-eating indicates a position fur157

The life-cycle thest in time from He wi. If we take the action of the Moon to characterise the menstrual cycle as a whole, we may say it is 'too natural', confusing women with the natural world and counteracting the principle that human blood should stay inside human bodies. If this is excessive conjunction, then He wi is excessive separation, for no women, no copulation and no food are allowed at all. In ordinary life there is a balance: one can copulate with women from other groups, one can copulate with women who are not menstruating and one can eat a mixed diet of vegetables, fish and meat. If menstruation is a move away from this happy norm in one direction, then He wi is a move in the other. In this respect, they are equal but opposite: equal in their opposition to the norm but opposite in the form that this opposition takes. However, it has often appeared that menstruation and all things female are regarded as bad, while He wi and all things male are good: thus, Underworld bones are bad and ancestral bones are good; decay of the body is bad and life of the soul is good and so on. This amounts to an assertion that social distinctions are good, but negative, because they create a world in which nothing can happen; everything is too far apart. A lack of social distinctions is bad but it is positive in that anything can happen because everything is too close together. A balanced world is one in which there are social categories which are just the right distance apart for ordered exchange between their members. I set out some of the relations between menstruation, He wi and ordinary life below. The first set emphasises the simple opposition of both male and female events to ordinary life, while the second emphasises the opposition of male and female events with ordinary life as an intermediate point. These are alternative ways of looking at the same things.

Simple opposition of like extremes to normal life female renewal

ordinary life

male renewal

separation from opposite cooperation with opposite

separation from

sex

sex

opposite sex

no sexual relations

sexual relations allowed

no sexual relations

shamanised foods of minimal diet

ordinary diet

shamanised foods of minimal diet

158

Summary of the life-cycle Opposition of extremes menstruation

ordinary Me

He wi

woman alone

men and women in small social groups

men in large social group

seclusion at edge of house

sexes together in middle of house

seclusion in centre front of house

Moon's meat diet

mixed diet

ancestors' coca and tobacco diet

Underworld bones (rot)

people's living bones (endure for life-time)

ancestors' bones (endure for ever)

Moon's sex with sister

sex with affine

no sex or ancestral sex

sex with menstruating woman (in myth)

sex with nonmenstruating woman

no sex

Summary of the life-cycle We may now summarise the life-cycle. The new-born child is composed of a maternal contribution (female blood/body liquid) and a paternal contribution (semen/bone), and it soon receives a paternal grandparent's contribution (name/soul). Its growth to maturity is conceived of as a growth away from female origins towards male destiny, but nevertheless it takes place within the female domain. Growth is promoted by female food and male shamanism which (on the whole) influence body and soul respectively. It is both an opening up of the senses and orifices and a gaining of control over these. The menarche for a girl and initiation for a boy mark a turning point in this gradual growth, for they give way to repeated cycles of renewal thereafter. In girls this is achieved through an internal skinchange which is an involuntary fusion with the natural world. In boys it takes place through a ritually induced soul-change, partially modelled on the female creative cycle, which is achieved through contact with the ancestral world. The major male soul-change takes place at He wi, which is the climax of an annual cycle of fruit rituals. The physiological turning point of first menstruation or first He wi is the potential start of the process of conception, birth and growth of the next generation. From the mother's point of view, pregnancy, birth and breast-feeding form a two-phase cycle which conforms to 159

The life-cycle the same structure as the menstrual cycle although it is more drawn out in time and has a product — the child. Once the father's physiological relationship with the child has been acknowledged during the post-natal seclusion, he returns to his male domain from where he contributes shamanism (if he is competent) and, later on, maleproduced foods. At death the body rots in the Underworld while the soul is installed in a new child. The rotting of the body completes a bodily cycle begun by the growing of the child from the ground: the freeing of the soul completes a soul-cycle begun by naming. Although the male and female life-cycle have very much the same basic structure (continual growth followed by repeated renewals) when they are described in this way they are really very different, because male ritual experience has a cumulative aspect as well as a repetitive one. This means that men start growing or 'changing upwards' in a ritual sense in the second period of their lives which follows the first period of childhood growth. Correspondingly, adult male society is differentiated by age to a much greater extent than female society. So far, I have made no attempt to incorporate marriage into the life-cycle apart from drawing a parallel between the position of the warrior role in the set of five specialist roles and the position of marriage in the life-cycle. Apart from this, marriage was described as an event in another domain — that of kinship and intergroup relations. Although there is a sense in which marriage obviously is a life-cycle event, it is not ritualised like birth, menstruation, Yurupary rites and death. The physiological possibility of a new generation has already been ritually recognised in initiation and first menstrual rites performed for all individuals of all groups by the time marriage occurs. In this context, we must therefore conclude that marriage is solely a redistribution of female reproductive powers which makes the new generation socially possible. Conversely, the menstrual and initiation rites are full of physiological sexual references; the sanctions for their rules are inability to give birth, inability to father sons and so on. It is considered wrong to have sexual relations before these rites have been performed; it is my impression that boys do so less than girls, because they are usually initiated before the question arises. Boys approaching initiation are sometimes involved in homosexual teasing which takes place in hammocks in public: this play is most common between initiated but unmarried youths from separate 160

Perpetuation of Pird-parand society Exogamous Groups. After initiation or first menstruation, it is common for young people of both sexes to have illicit heterosexual affairs, many of which are classified as incestuous. These affairs invariably incur the disapproval of the girl's parents. If an unmarried girl bears a child it is almost always killed at birth. Although He wi and the onset of menstruation release the sexual powers of individuals, they in no way determine whom these individuals should marry. As we have seen, that is taken care of by rules phrased in terms of descent-group membership and kinship categories. The maturing of sexual powers and redistribution of rights over women are different kinds of precondition of the next generation, while birth of children to an exogamous marriage is the realisation of this new generation in both its physiological and social aspect. All this is to state no more than the obvious — that socially defined descent groups are made up of biological individuals and that, therefore, birth of descent-group members is also birth of biological individuals. The family is both a meeting of members of opposite Exogamous Groups in order to create a new generation of the father's group, and also a meeting of opposed sexual powers in order to make children whose sexes are biologically rather than socially determined. It is the latter physiological aspect of reproduction that I have dealt with so far in this section but now it is possible to bring the two aspects together. Perpetuation of Pira-parana society To consider the case first of all from the point of view of the social structure, we may remember from chapter 2 that the married women remain permanent outsiders within their husbands' longhouse communities. They are essential to the continuity of the patrilineal group and yet they are never incorporated into it. The growth of longhouse descent groups through marriage and birth of children has its corollary in the dissolution of united sets of brothers into separate family units and ultimately into separate longhouse groups. Thus, marriage breaks up the sibling groups of the established generation at the same time as it creates the new united ones of the next. Neither process is accomplished in an instant, but they are related over time, for as children grow up their fathers focus more on the family of procreation than on the family of origin. Seen in this way, women are both creators and destroyers of the 161

The life-cycle generations making up the path of patrilineal continuity over time, yet since they come and go in each generation, they do not appear as stages on this path. Women shuttle back and forth while generations of men accumulate: it is the interrelation of these two processes that creates the social persons and groups of Pira-parana society. In one sense, the female shuttle and the male accumulation process coincide every generation when the men reach adulthood and swap sisters for wives (or otherwise lose sisters and gain wives in forms of marriage other than sister-exchange). Yet, in another sense, in as much as the ideal marriage with the mekaho mako is achieved, the processes diverge in one generation, as sisters are lost, and coincide in the next when the daughters of these sisters are returned. In the first generation a descent group receives a woman who is physiologically equivalent, but socially opposed by different Exogamous Group membership, to the sister who was lost. In the next generation it receives a woman who, in addition to these qualities, has the advantage of being directly linked as child to the sister who was lost, as shown in fig. 22(2). Thus, it is in the second generation that the true fruits of the original gift are recovered and the continuing alliance is established. Now we may turn back to the earliest phases of the life-cycle in which we have seen that a one-generational cycle is associated with male bone and semen and with female flesh and blood, while a twogenerational cycle is associated with the inheritance of patrilineal names. While bone endures longer than flesh, together they make up the physical body which rots after death and is opposed to the soulstuff inherent in the name which is handed on in perpetuity. If we simplify the life-cycle and impose a midpoint at which marriage and procreation occur, we may juxtapose the cycles of social reproduction of patrilineal groups, the life-cycle of the individual of each sex and the inheritance of names (fig. 22). From this exercise it is apparent that ideally the re-creation of society depends upon the interplay of no more than two living generations - parents and children, for as the parents die the children's children are born and take on their grandparents' names. In the case of female children, the names which are overtly a function of patrilineal status may also be derived from the father's father's sister through the female line, for ideally this woman is also the mother's mother. The inheritance of names therefore coincides with the recombination of a socially stressed male line and a socially unrecognised female line. The principal reason for thinking that this 'unrecognised' female line is significant is that separation 162

Perpetuation of Pird-parand society and recombination of sexual elements is typical of creative process in Pira-parana ideology. Also there is actually a sense in which blood and semen as reproductive potentialities are passed on in same-sex lines: this is expressed in the theory of conception which states that girls are made from blood and boys from semen. Returning to the contributions of semen, blood and names, at the level of social structure we see that, on the one hand, semen and blood DESCENT LINES 1 ONEGENERATION MODEL

2 TWOGENERATION MODEL

SOUL TRANSMISSION

LIFE-CYCLE

SEMEN NAMES

KEY I patriUneal ties

I ^ women from outside Exogawous Groups

j female (mother-daughter) ti« '

X 6- p men's y 6 C[ wovuen's names

Fig. 22 Aspects of alternation of generations 163

The life-cycle represent a man incorporated into his local descent group and an incoming woman. On the other hand, the naming system implies a dynamic relationship between the two opposed groups such that the female products of one are transformed into the female products of the other in the next generation. The endurance of names, as opposed to the ephemerality of flesh/blood and bone/semen, corresponds to the dynamic character of the two-generational system, as opposed to the simple static opposition of groups in the one-generational system. This conclusion raises an important problem, for it suggests that the transmission of substance from father to son - which actually creates patrilineal structure and continuity — is ideologically less powerful than the naming system as a model for perpetuating patrilineal groups. We have already seen that, in spite of the firm patrilineal structure, there is little emphasis on genealogy in Pira-parana society, and that the repetition of names over the generations assists in the destruction of genealogical memory (p. 39). We have also seen that the seniority structure within patrilineal groups allows new-born members to derive their exact position in the hierarchy from their fathers' positions without any reference to more distant generations. The further finding that Pira-parana social reproduction hinges on the relationship between only two living generations (parents and children) means that father—son transmission is only relevant to the world of the living, for ideally one father-son tie is severed by the death of the father just before the son himself is transformed into a father. Father-son transmission is thus a 'here-and-now' phenomenon which places a new child at the end of a path, the previous stages of which have already been obliterated by time. Naming, by contrast, serves to keep the stock of patrilineal names which existed in the beginning in circulation, so that, ideally speaking, each alternate generation consists of the very same names and the very same souls. Thus the bonds between father and son, having a firmer physiological base, are eroded by time while names which are consciously and ritually bestowed transcend time. The naming system is superior as a guarantee of continuity in another way because it overcomes the fluctuations and failures in the reproduction of the group through seminal (father—son) ties. This derives from the very nature of names which bestow separate identities on members of a set in a way which is never wholly a function of kinship or other relationships between these members. For instance, in the case of Pira-parana society, names are ideally - and sometimes 164

Perpetuation of Pira-parana society in practice - passed from FF to SS and FFZ to BSD; but this ideal could never be kept up in all cases (even if grandparents did always die before their grandchildren were born) because the same sex siblings would have the same name. The name would therefore no longer serve the distinguishing purpose of a personal name. The fact that names can be chosen in other ways than the ideal one, although still from a limited set belonging to a wider patrilineal group, means that both personal identity and social continuity can be maintained in spite of group expansion and premature death of descent-group men. As well as providing an ideologically less hazardous means of social continuity than do seminal paths, the type of classification provided by names is prior to seminal transmission in time. Seminal ties are a means of transmission of group identity but they operate in a similar way within each group and in no way account for the separate identity of groups in the first place. This latter lies in the separate identity of the ancestral anacondas, who were alike in form and place of origin but had separate names. It is the original separation of Fish Anaconda, Stone Anaconda, Sky Anaconda and so on that founded the exogamous groups and made marriage alliance between them possible. Finally, the link with the ancestral past provided by the inheritance of names is recognised and contrasted to the fate of the physical body in a statement made by a Barasana Indian, 'If we did not take on names we would die out like a rotting corpse'. According to the one-generational and two-generational models discussed above, women's contribution to social continuity has appeared as complementary to that of men, the exchange of female reproductive powers being associated with the repetitive cycles within the cumulative development of patrilineal groups. Although these female movements are very much less susceptible to genealogical record than is patrilineal succession, there is a permanent testimony to them in the spatial separation of longhouses. After sexual maturity, women move physically to another longhouse in another exogamous group territory and are thereby transformed into wives. From the point of view of any one longhouse community, the resident women divide up in the same way into descent-group sisters, who are forbidden in marriage, and wives, who are from affinal exogamous groups. Over time, sisters are swapped for wives, who bear a new generation of sisters and so on. These two phases of the female life-cycle are reminiscent of the 165

The life-cycle two phases of the menstrual cycle which we saw was also divided into a creative phase (retention of blood) and a potentially creative phase (loss of blood). There is confirmation of the parallel in the myth of the Frog Wife, where a female state which may be identified with menstrual loss is associated with a visit home. Frog Wife A Frog Woman (of unidentified species: goha) arrived from far off and settled down with a man. She visited her own people each umari season: she went to the cold, white Mildew House far down the Milk River in the east and became a mildewed umari lying on the ground. Although she forbade her husband to follow her, he disobeyed, failed to recognise her and pinched her mildewed umari flesh. She was 'grease-filling', a Snake Woman, and so when he got home he was bitten by a snake and died. The decay of the Frog Wife's outer skin, her periodic journeys away from her husband and her dangerous properties during her secluded periods liken her visits home to menstrual seclusion. Besides this, umari is said to be an ideal food for menstruating women. If the assumption that Frog Wife is metaphorically menstruating is correct, then the link between menstrual loss and female residence among agnates is established. The married life of Frog Wife is divided into cycles of alternation between her husband's and her own natal community. Each cycle is a small-scale replica of her life as a whole which divides into a pre-marital and a post-marital phase. It might occur to the reader that the patterns of alternation of residence in a woman's life suggested by the myth of Frog Wife is seemingly different from the patterns of physiological growth and decay during the female life-cycle, which are set out in fig. 20. The myth makes the return to the natal group, which is metaphorical menstruation, a small-scale repetition of the childhood period spent among agnates. This makes good sense, since pre-pubescent women and menstruating women are both useless from the point of view of procreating affinal groups. But the physiological pattern of growth and decay equates pre-pubescent with non-menstruating women instead; this is because both are undergoing periods of growth. However, if we look at the kinds of growth, it is growth of the prepubescent woman herself in the childhood phase and growth of wombcontents which (if not lost) become incorporated into a child in the 166

Perpetuation of Pird-parand society post-menarcheal phase. From the point of view of a continuing female line, these two types of growth are integral parts of the same process, but from the point of view of patrilineal exogamous groups, they represent growth of members of opposed exogamous groups. All this is simply a rather laborious way of saying that the physical continuity between mother and child, and likewise that between the first and second half of the female life-cycle, are severed by the rules of exogamy, patriliny and patrilocality. The menarche occurs before marriage and so exogamy does not make an exact break between the pre-menarche phase and the postmenarche phase: for a while the girl's patrilineal group is in control of her mature reproductive powers but it is forbidden to make use of them except as an item of exchange. In this way, female physiological development and female social development may be seen as parallel processes: in real life the social process can only control the voluntary behaviour associated with the physiological process, but in myth the two processes are made to coincide through use of metaphor. We have seen that the male and female modes of continuity are respectively cumulative and repetitive, the male mode creating the patrilineal path from ancestral times to the present, while the female one is broken up into repeated alternate phases consistent with the exchange of women between groups. The male mode represents continuous extension over time, and the female one represents continuous renewal. Although the male mode is the dominant one from the point of view of the basic division of society into exogamous and longhouse groups, there is an inherent advantage in the female mode. The very construction of the patrilineal path removes the living progressively further from the source of power in the ancestral past, while the repetitive female cycles allow constant regeneration. In order to tap ancestral power, men periodically go back in time and become 'people of an underneath generation' by participating in the He wi rituals. The discussion and analysis of such rituals showed that they incorporate female metaphors of regeneration and are compared directly to menstruation. It therefore appears that female repetitive cycles are an essential element in the negation of cumulative patrilineal time necessary for 'soul-change' during the life-cycle. It is important to note here that integrating the collective reversal of patrilineal time in identification with ancestors achieves a personal reversal of life-cycle time for each participant, since each undergoes soul-change. If we turn back to the introduction of patrilineal soul-stuff in 167

The life-cycle naming, we see an analogous process taking place in the socialstructural domain, for the accumulation of patrilineal generations is counteracted by the reincarnation of timeless soul-stuff. This reincarnation of soul-stuff, it was argued above, is determined by the cyclical exchange of women's reproductive powers between affinal descent groups. This cyclical exchange of women represents the ability of the patriline to shed sisters and take on wives in a regular, dependable fashion. The female line is thus like a skin which the patriline puts on and takes off according to the social rules governing marriage. Here again, the voluntary control over clothing or cultural 'skin', which can be separated from the wearer without being lost, contrasts with the involuntary control over natural skin, which grows naturally and can only be separated by decay. Thus, when men put on ritual ornaments and 'change soul' they are imitating the powers of women, but at the same time doing what women cannot do. Many of the points made in this chapter will be incorporated into the analysis of concepts of space and time in the final chapter.

168

Production and consumption

Introduction It is the very same people who produce and consume on the one hand and who grow, experience ritual change, marry, reproduce and die on the other: Pira-parana Indians make the most of this fact. In this chapter, we shall see that the processes of producing and consuming various substances are analogous to the processes of reproducing both individuals and the social structure. The relations between the individuals who make up Pira-parana society are ordered through production and consumption in the daily routine and on ritual occasions. In promoting these ordered relations, socio-economic processes play a positive part in social reproduction. We saw that when individuals absorb substances, they also absorb the processes, whether mythical, shamanic or practical, which contribute to the production of these substances: here, the same argument is extended to the 'relations of production' and also to the relations between natural species (as conceived of in Indian ideology). I must warn the reader that the ordering of this chapter is rather complex. Because the same substances appear in different contexts, the analysis is built up on many different fronts from one part to the next. Throughout, the conclusions about production and consumption are used to further the analysis of the life-cycle begun in the previous chapter. For subsistence purposes, the longhouse group is a virtually selfsufficient unit. It is divided in two principal ways — into family units or 'family productive units', and by sex. Most of the objects exchanged between members of separate longhouse communities, pepper, tobacco, pottery, basketry, weapons, canoes, red paint etc., could equally well be produced at home. The most important goods entering from out169

Production and consumption side the Pira-parana area are poison and raw materials for ritual ornaments from the Makii, and manioc-grating boards from the Arawakspeaking Kuripako. The exchange of ritual paraphernalia between Tukanoan groups forms a separate system or 'sphere'. Such items are given at communal rituals as part of a perpetual exchange between geographically distant communities who refer to each other as He tenyua (p. 81). We have seen that smoked meat, fish and insects accompanied by tree-fruits may also be exchanged at communal rituals. In addition, the provision of fruit for He rika soria wi is sometimes conceived of as a similar type of intergroup exchange. The sexual division of labour The most significant facts about the sexual division of labour are summarised below. Although broad generalisations such as these obviously do not apply to every concrete productive act, they hold for nearly all productive activity. For the sake of brevity, I do not attempt to do justice to the variety of the Pira-parana diet or material culture. Men

Women

felling and burning manioc gardens

planting, harvesting, preparing manioc

hunting and

cooking all food

fishing

cultivated 'drugs'*

cultivated 'foods'

cultivated tree-fruits

cultivated vegetables and lowgrowing fruits small-scale gathering

large-scale gathering production of wild food for exchange

ritual

preparation of wild food for immediate consumption

use of high, above-ground resources

use of low, underground resources

manufacture of all wooden and basketry items (including house)

manufacture of all pottery items

manufacture of ritual goods and ornaments ritual

manufacture of paint and garters for use

J

The 'drugs' are tobacco, coca {Erythroxylon coca) and yage (Banisteriopsis sp.).1 show later that they are distinguished from 'foods' which nourish the body rather than the insubstantial aspects of the individual.

170

The sexual division of labour Some of these relationships bear out the analysis in the previous chapter. For instance, the sexual associations of low and high positions are consistent with the model of growth of children away from maternal origins towards 'male' maturity. Men's role in producing foods for ritual exchange, cultivating the drugs essential for ritual performance and manufacturing almost all ritual ornaments serves to reenforce male control of ritual life. On the other hand, the few ritual items produced by women - paints and garters (yuta gasero, 'wovenfabric skins') — are associated with skin-changing, which we have seen to be a natural female ability. In the opposition between manioc cultivation and basic protein provision (hunting and fishing), the most important relationship in the day-to-day feeding of the community, there is a clue to a more general factor in the sexual division of labour in food production. Most of women's time is spent in producing manioc, which is a continuous-cropping and extremely reliable source of food. Success in hunting and fishing, the most time-consuming of male contributions to food production, is far less certain. Nevertheless, fishing is more certain than hunting and, although meat carries more prestige as a food, fish makes up the bulk of the protein supply. Women never hunt, but occasionally they fish. The method they use most often is poisoning: this is very different from other fishing methods, because the water is made to yield up the fish so that they may be gathered from the surface. Poisoned fish are actually referred to as 'gathered fish' (huari wai). It is significant that when the sexes co-operate in poisoning, the men dam the river or stream and then introduce the poison into the water upstream. The women remain downstream, gathering up the fish as they suffocate. Thus, men do the trapping and killing and women the 'harvesting'. Again, women do collect seasonal plant and animal species from the forest, but they do so less often and in smaller quantities than men. These fruits, insects etc., are less reliable food sources than manioc on two counts: first, because they are seasonal and secondly, because they have to be searched for. However, they are more reliable than game or fish because, once located, they are there for the taking. Although seasonality is by no means the same as the unpredictability which makes wild foods difficult to find or catch, there is a similarity between these two qualities, for both are manifestations of the independence of the natural order from social manipulation. 171

Production and consumption Seasonal cultivated fruits are mainly controlled by men; they are also lumped together with wild seasonal fruits in the category He rika, and thus seasonal tree-fruits cross-cut the cultivated-wild distinction. Overall, the allocation of tasks in food production does suggest that women are given the most reliable ones, for which manioc cultivation provides the ideal model, while men are given the least reliable ones, for which hunting provides the ideal model. Success against the odds makes meat, followed by fish, the most highly esteemed food, while the absence of seasonal products for most of the year makes these more highly esteemed than the dependable daily fare of manioc products. Thus, prestige, risk of failure and male food production go hand in hand. This division of labour in food production is reflected in communal eating patterns. There are two basic forms of communal meal; as far as I know, they do not have names in Indian languages but they are readily distinguished in practice by a number of interrelated factors. Both consist of hot, boiled foods seasoned with chilli pepper and served in the centre of the house. Cassava is supplied by each one of the female family productive teams (mother and daughters) and this is dipped into the hot food and eaten. In general, the first type of meal consists of food supplied by a man, cooked by his wife or sister and served to all the men, women and children of the community. A single family may provide the meal or several may produce dishes simultaneously. The second type of meal consists of manioc-based dishes or other readily available food produced by women. Each female family team produces a pot, and men eat first while women and children eat afterwards or not at all. It is said that women and children have private access to these foods in any case. The second type of meal is held at dawn and only at other times if there have been no meals of meat or fish. Even then, the women only co-operate to provide a meal if the men have been engaged in useful activities, which have prevented them from hunting and fishing, or if there are visitors present. Thus, the second type of meal is a response to legitimate' hunger and is regarded as a poorer substitute for the first. In practice, while fish and meat are the most regular elements of the first type of meal, large quantities of insects and some cooked tree-fruits may be served to the entire community in the same way. Meat is always served on its own at a separate meal, whereas both fish and insects may be mixed with vegetable products in the same meal or even in the same dish. Also, while the second type of meal consists 172

The sexual division of labour mainly of manioc-based dishes, small quantities or leftovers of insects, fish or tree-fruits may be incorporated into these. The more ritually dangerous fish, mainly the largest species, are rarely served at meals of this second type. Therefore, meat and some kinds of fish are set apart because they are always served alone at separate meals of the first type; other fish, insects and tree-fruits may be incorporated into mixed dishes and served at either type of meal, the deciding factor being the quantity available. Manioc-based dishes, unless they contain a high proportion of fish, are only suitable for the second type of meal. The minimal form of the second type consists of pots of hiari, caramelised manioc juice highly seasoned with pepper (alternatively called 'pepper pot', bia sotu). This food is always available, for a pot of it sits in the crook of the stand supporting the family cassava supply. It is the standard food offered to visitors on arrival. In these meal patterns meat, the least reliable food, stands at the opposite extreme from hiari which is readily available at all times. Large, ritually dangerous fish are closest to meat, and then large quantities of naturally occurring or seasonal foodstuffs eaten alone are followed by these same foods eked out with manioc products; finally, pure manioc products form the least-esteemed meal and hiari is the 'poorest' extreme. We can see that the same related sets of contrasted characteristics which underlie the organisation of food production are also apparent in the patterns of communal consumption. There are good meals and substitute meals, just as there are prestige foods and the basic foods which are taken for granted: both are distinguished according to the contrasts between the following: male producer

female producer

wild food unreliable food large quantity of wild food

cultivated food reliable food small quantity of wild food

Although women's cultivating activities associated with maximum reliability are, in a sense, the strongest form of social control over natural processes, this control is routinised. The very reliability of the crop suggests that women harmonise with manioc rather than pitting their physical and mental powers against it. Just as in the case of menstruation and the very much more dangerous soul-change at He wi, dependable female powers are less valued than the male ability to succeed in risky encounters with the natural world, whether this be the natural world of ancestral forces or the opposed one of natural species. 173

Production and consumption Manioc Production The garden sites are chosen by men who clear the undergrowth, fell the trees and burn the dried wood. Main gardens are felled each year from October onwards and are burnt between November and March in the long dry season. The site is ritually cooled with beeswax by a man, and then the women take over. Manioc is ready to harvest a year, or slightly less, after the sticks are planted. With continuous replanting of the harvested areas, a main garden may be kept in manioc production for about three years. Of course, there is a sharp decline in yield after replanting, and in the end the small size of the roots and the encroaching weeds make the effort worthless. However, people continue to visit old gardens for other crops long after the manioc is finished and, in this way, each family productive team has a number of manioc gardens, which together represent the entire cycle formed by felling, cultivating and encroachment of the forest. Women sometimes receive help from men with carrying huge bundles of vigorous manioc sticks from a newly grown garden to a newly burnt one. Thereafter, women plant alone. The sticks are broken up and each section is poked into the earth, with about half a metre of space left between plants. The entire garden is covered. In practice, the fallen logs form natural boundaries between one day's work and the next. The half-grown crop must be weeded, and then the later harvesting and replanting falls within the women's regular daily routine. This daily routine can be divided into (1) harvesting, (2) separating the three basic raw materials: fibre, starch and juice and (3) baking cassava. This three-stage process is better understood with the aid of fig. 23. A complete schematic account of the processing and use of manioc plants is given in fig. 24; it is intended to demonstrate the enormous versatility of the crop and show the production of foods not included in fig. 23. (1) Harvesting. Women leave the house soon after dawn when breakfast is over. They spend the time until midday working in the gardens, digging up manioc, replanting the sticks, peeling the roots and loading them into baskets. These baskets of manioc are shaken violently in 174

SEPARATI HO

HARVESTING

DRYING & BAKING

Manioc

STARCH [

'

RIVPR » CARRIED

ROOTS stOFR«RC

TOHOVSe

MANIOC GARDEN KEYS

RIVFR

HOUSE

, MANIOC- HUSBAND'S CONTROL OVER WIFE'S REPRODUCTIVE POWER

PREGNANCY

MEN RAID PO&EI0N WOMAN

BIRTH

BROTHERS

SISTERS SOCIALISATION'

C GENERAL MODEL (FOR A &B ABOVE & fig. 28 ) CONJUNCTION^ ACQUISITION

LOSS

SEPARATION

Fig. 27 Comparison of meat production and reproduction 193

Production and consumption (5) boiling with pepper and salt (female; family compartment) (6) serving cooked meat (female; family compartment to centre of house) Stages (3) and (4) may be done by men in the case of a large creature. During butchery, the gut is allowed to float away; liver, heart and sometimes kidneys are kept. Animals are not skinned, nor are the feet and head discarded. At serving, the chunks are laid out on a banana leaf; they are eaten with cassava dipped in the peppery juice from the pot. First, the meat brought in by men is like newly obtained wives, particularly raided wives. One man made this clear when setting out on a wife-raid, for he announced his departure by telling different individuals he was going to hunt a different species of game (see also p. 223). To become edible, the creature must be divested of its hair - a sign of animal nature that differentiates birds and forest mammals from humans. In terms of the human cycle, the wife must be divested of her foreign nature before incorporation into her new community. Appropriately enough, the singeing of animal fur is related to menstruation: this occurs in the myth of the origin of fire, where fire to singe a game animal is released womb-contents (pp. 137f). Menstruation is also related directly to singeing animal fur in the expressions for menstruating women, who are 'singeing sloth or deer' or 'burning the path to the port' (the location of the singeing fire).2 If fire for singeing — stage (3) — is like released menstrual blood, then fire for cooking, stage (5) — which retains and transforms flesh in pots instead of burning directly, should be like retained menstrual blood. There is evidence that release of menstrual blood is opposed to retention in pots, because it is said that if menstruating women make pots these will crack. Conversely, there is no restriction on potmaking in pregnancy, which is the longest retentive phase in the female cycle. We may assume that of the two kinds of female retentive phase — monthly growth of blood and pregnancy — it is pregnancy that is most like boiling. Both pregnancy and boiling involve 2 Although the origin of fire myth confirms the link between menstruation and singeing, we have to explain why two of the expressions for menstruating women (p. 136) refer to alternative uses of fire. I do not think this damages the case because the inedibility of the species concerned plays the same role as the singeing. Inedibility is a metaphorical reference to menstruation since (a) one cannot 'eat' (have sexual intercourse with) menstruating women and (b) (prohibited) intercourse with menstruating women does not result in a child, just as singeing or otherwise cooking an inedible animal does not result in a meal.

194

Meat: analysis of production transformation of a retained substance to yield a valued product, whereas the monthly growth of blood is wasted when it is released. Besides, the salt and pepper added during boiling may be likened to semen and blood.3 Not only are their white and red colours appropriate for semen and blood, but also the Barasana terms for them relate them to the sexual roles in conception. Salt is moa, 'activating stuff (moa-, to move, to work) and pepper is bia, 'renewing stuff (bia-, to renew, to repeat). Furthermore, in myth, Tapir shaman's wife asks, 'Who has been stirring my pepper pot?' when she means, 'Who has been making love to me?' (S. Hugh-Jones 1979 : M.6.A). In this case, the 'stirring of pepper' is clearly like the activation of wombcontents during insemination. Although salt is added to boiled food whenever it is available, it is not nearly as important as pepper in Indian culture or ideology. Here, then, the boiling process is metaphorically associated with a particularly female aspect of reproduction in which salt is merely an activating agent. If we return to the intermediate stage of washing, gutting and butchering at the port, stage (4), this may be likened to the rebirth achieved by bathing at the end of menstrual seclusion. Not only is the carcass bathed, it loses the gut and reproductive organs which controlled the animal life process, and its physical form is destroyed. There is evidence to suggest that the end of the menstrual seclusion is not simply a practical incorporation of the wife into her husband's community; it is also a metaphorical incorporation, since myth relates menstrual seclusion to the married woman's contact with her descentgroup origins (Frog Wife's visit home: p. 166). Therefore, we may say that just as the meat loses its animal nature, the wife loses her foreign status to the extent necessary to render her a potential mother of her husband's descent-group members. Her integration after 'menstruation' marks the completion of the transfer of her productive powers to her husband's community. The butchering stage is followed by cooking/pregnancy and serving, or birth - stage (6). The community absorbs the food just as it absorbs the children born to wives from outside. Inside the bodies of community members, the eaten food is digested and turned into body substance and waste, and so we may suggest an analogy between socialisation of children and the internal processing of meat (see the theory of digestion p. 186). 3 Salt used to be made by burning an unidentified water-plant to ash: now it is obtained from whites.

195

Production and consumption The analogy between the digestive process, in which nourishment is separated from waste, and the growth of the longhouse community, through formation and dispersal of family groups, is implied in the Indian statement that the longhouse is a human body which lives when the house is occupied. If we follow this implication, the separation of nourishment from decaying waste has its parallel in the separation of family groups into their constituent units. This separation may be seen in two ways: (1) within the family as a whole, the new generation is separated from the decaying older generation or (2) within the new generation, the boys are separated off from the 'wasted' girls who cannot reproduce the community. We have reached a position exactly similar to that in the analysis of manioc processing, when we saw that the separation of elements was like both the separation of generations and the separation of male and female procreative powers (at first menstruation and initiation). However, in the meat-digestion process either the girls or the parental generation are the waste. The production of waste is essential in both cases, for it is the necessary complement to reproduction of life. Girls must be given away so that wives can be obtained, and parents must grow old and die so that new generations can be created and grow up. Similarly, waste must be eliminated by action of the gut or the body would become full of rotten food and we could not eat again. There is plenty of evidence to prove the rather obvious connection between faeces and decaying bodies (S. Hugh-Jones 1979 : M.7.J). There is also evidence that faeces are like lost daughters, for a man may joke, saying, 'I am going to deposit my daughter' when he means, 'I am going out to shit'. The meat-preparation and consumption cycle and the community reproductive cycle are not merely analogous, they are interrelated at various points. It is wives who enter the reproductive cycle and bear children and who also prepare the meat for communal consumption in the meat-production cycle. Besides, the hunting-and-cooking partnership is regarded as a 'sign' of marriage. Wives also provide the manioc that is necessary to all meals. The give and take of women is therefore closely connected to the obtaining and relinquishing of food supply and the catching and releasing of natural products (meat and faeces). An irritated headman told a shy member of his household, 'If you can't get yourself a wife, you'd better go and eat your own shit'. The human reproductive cycle, the community reproductive cycle and the meat-production and consumption cycle are 196

Meat: analysis of production set out in diagram form so that the analogies mentioned in the text are apparent. However, once again, I stress that the character of Piraparana Indian thought is such that the separate transformational process within a single cycle are metaphorically related, and therefore, my diagrams show one version of these analogies amongst many possible ones. The exchange of sisters for wives, the loss and regrowth of menstrual blood, the birth and regrowth of a new foetus and the use and re-creation of body energy in the production and consumption processes are all examples of cycles composed of two interdependent phases — the one leading on to the other. A myth about No-Anus Spirit provides abundant evidence that the cycles I have presented in the diagram are related, as well as illustrating what happens if the destructive phase of one cycle cannot be completed so that the renewal of the entire cycle is impossible. No-Anus Spirit is unable to complete the digestive cycle. No-Anus Spirit ('Wati Gude Mangu') Dragonfly's children refused to accompany their parents to the manioc garden. No-Anus Spirit came into their house with a basket of the pig-teeth belts which are part of the ritual dance dress. He dressed the children in these and led them in a mock dance. A weakling among the children told the parents, who instructed the children to cut the carrying strap (a tump-line) of the basket and so steal it from No-Anus Spirit. The children did as planned, and the parents hid the basket, but the children still refused to accompany them. Soon, the spirit arrived to take his revenge for the stealing of his belts. Pretending he was going to cook caterpillars, he got the children to light a fire by the lighting post and when the water boiled, he summoned them to watch and pushed them in. The weakling fell out, turned into Yago Mini, Mourning Bird (unidentified species) and warned the father, who was felling a garden. The father laid the cooked pieces of his children out on a banana leaf, arranged the bones so as to reconstruct their bodies, and coated them in tree cotton. They turned into gake hosa (Cebus albifrons) edible monkeys; the father whipped them and they went off into the forest. He did the same with the children's flesh, which turned into manioc pigs (kiyese, Tayassu tajacu). He told each group of animals that they would be 'edible game' from now on. No-Anus Spirit disappeared for a whole season until he met 197

Production and consumption Dragonfly scooping fish out of the water. Dragonfly farted and the spirit implored him to make him able to fart too. Dragonfly took the pole for stirring red pigment (ngunanya) while it cooked and pretended to bore out an anus for No^Anus Spirit. No-Anus Spirit asked if the hole would allow him to rid himself of the rotting food in his belly and was pleased to learn that it would. Dragonfly hammered the pole in until he killed No-Anus Spirit and then deposited the body in various river beds, where the liquid and soft parts became potters' clay of different grades and colours. The sequence of the myth is set out in fig. 28. It begins with children asserting their independence from their parents and a spirit unable to digest. The transformation of the children begins with dancing in animal teeth, dancing itself being an other-worldly activity. However, the use of this single piece of equipment heralds their transformation into animals rather than ancestors (the teeth are taken from pigs destined for the pot). The acquisition of teeth results in the children being cooked, except for the weakling younger brother who has not grown so far away from his parents. The fire is in the position of the night-time lighting fire and is thus the opposite of the daytime cooking fire. When they are served, instead of being eaten the children are put together and are thus reborn. They are given hair and sent off as game animals, divided into bony monkeys and fleshy, round pigs. They are sent off by whipping, which is also employed at initiation to make bodies grow. We can now outline a backwards cycle. The children assert their independence by obstinately ignoring the productive rhythm instead of taking a useful part in it. They and the spirit do not complete their respective digestive and socialisation processes. Instead, the children become identified with the sinister world of the spirit (dancing), are cooked, put together, given hair and sent to the forest. Human productive energy is changed into game animals by a backwards movement through the cycle which normally changes game animals into human productive energy. When 'a season has passed', the activity is transferred backwards into the end of the previous cycle, where it continues in reverse. Farting, literally 'anus-blowing', is a socially approved act that is made as noisy as possible and draws laughter and jokes. Possession of an anus would both make No-Anus Spirit socially acceptable and enable him to digest and live'. But in fact he achieves neither, because he is killed by the pole which goes 198

• CHILDREN RECONSTITUTED CHILDREN GIVEN COTTON 'HAIR' tWHIPPED

NO-ANUS SPIRIT COOKS CHILDREN

[TRANSFORMED INTO EDIBLE AWIMAL5]

MONKEYS £ PIG5A OFF TO FOREST WO-ANUS5PIRIT 6- CHILDREN DANCE TOGETHER

SEASON PAS5E5

SPIRIT PROVIDED WITH ANU5 6-klLLED

Fig. 28 Reverse cycle made in the myth of No-Anus Spirit 199

CHILDREN NO-AWUS SPIRIT WHO CANNOT DIGEST

Production and consumption in where the rotten food should come out and thus a complete cycle is made. It is appropriate enough that this sinister food-retaining spirit should become the clay from which pottery retainers are made. The cycle described in the myth is set out in diagram form so that it may be compared with the previous forwards cycles. Turning children into game animals in the myth of No-Anus Spirit has its counterpart in the shamanic practice of swapping new-born babies for game animals (p. 61). In fig. 28, the point of entry on the far left represents the 'other world', and the far right represents the interior of the community (its nourishment with either food or children) and so, once more, the shaman is performing his characteristic mediating role. Before leaving the meat-preparation cycle, we should briefly consider how this would apply to fish. Fish come straight out of the water and do not need singeing, and thus they enter the cycle at the washing and butchering stage. If we follow the analogy set out above, they are born from the water as potential food rather than reborn from water as potential food, or, alternatively, we could say that meat is reborn in the water as fish. This fits the conclusion above that fish are more susceptible to human control than game animals. As mentioned at the outset, this analysis suggests a preliminary distinction between manioc preparation and meat preparation. Manioc preparation is metaphorically concerned with physiological and ritual reproduction and differentiation of the sexes, while meat preparation is metaphorically concerned with marriage and the reproduction of the descent group through its constituent family units. In the following discussion of the structuring of time and social relations by production and consumption, we shall see that manioc and meat are also opposed in other ways related to this distinction, and the analysis will yield a more specific formulation of the relationship between them. Structuring time by production and consumption Secular production and consumption Production. The daily cycle is governed by the sun, but, within the framework set by the alternation of day and night, time is structured by the activities of the longhouse community. The most important of the regular, daily productive activities are manioc production and cooking meals by women, and hunting or fishing and coca production 200

Structuring time by production and consumption by men. Other activities are either seasonal or occasional, while these regular activities are part of every normal daily cycle. Since coca production is the only regular activity not described so far, I make a brief digression to outline the process. Coca production. Old coca bushes are stripped and broken into sticks which are planted in long straight rows several plants wide (see fig. 29). The long intersecting rows make up a huge geometrical form representing a spreadeagled human body (p. 212). The coca bushes planted simultaneously with the first manioc crop remain in production after the garden has been abandoned as a source of manioc. Men pick coca in a group, conversing, joking and passing cigars and coca as they go. The leaves are carried in a single basket and tipped out into a large shallow basket in the centre front of the house. They are toasted by stirring in a deep round-bottomed pot over a fire laid in the side front; meanwhile, someone is sent to fetch dried leaves, preferably of the jungle grape {Pourouma cecropiaefolia). The toasted leaves are pounded in an upright cylindrical wooden mortar with a heavy wooden pestle. The grape leaves are set on fire, also in the centre of the house, and the resulting ash is gathered and added to the coca poundings in a male gourd. The mixture is sieved through a bark-cloth bag tied onto the end of a long pole, which is rhythmically banged inside another, much longer, cylindrical tube tied up to the 'ritual post'. The particles that remain in the sieving bag are repounded and resieved, usually twice more. The sifted powder is transferred to the gourd, from which it is served to the community of men. This gourd sits on a wicker stand, similar to the one supporting the cassava basket, placed by the ritual post. The process is usually completed at dusk. Picking may be in the morning or afternoon depending on the quantity required. The processing of coca is followed by formal speeches recounting the preparation process and the stages are ordered, roughly rather than precisely, according to prestige. Toasting is a skilled task allocated to an experienced elder who often sings ritual dance-songs as he works. Pounding requires great energy and is allocated to young men, while sieving is skilled and is usually performed by the owner of the growing crop. Fetching leaves for ash is a menial task typically given to the youngest men, to extra residents or even to women. The differentiation of processing tasks therefore serves to bind members of the productive team in interdependent relationships. 201

Production and consumption We have seen that the harvesting and separating stages of the manioc process take up a day, dividing it neatly in half between the two and providing a culmination point when the manioc juice is boiled up and served to the men at dusk. Cassava making — stage (3) — and cooking foods for communal meals are fitted into this day as

1

replanted in new garden

COCA kahi I—

DRY GRAPE

LEAVES

\Pourouma cecropiae folia] udye hu

set alight

ASH oha

t

\

LEAVES

STICKS

kahi hu I

kahi rttktt

toasted by stirring \v\ pot over fire

t.

pounded in \Apnght mortar

i POWDER

returned to sieve

sieved through bark cloth in leaning tytoe coarse bits emptied from bark cloth

FINISHED COCA Fig. 29 Coca processing 202

Structuring time by production and consumption the need arises. However, the men's day is less rigidly structured, because coca production and hunting or fishing are alternative pursuits, neither of which need take up a whole day. For one thing, the communal sharing of both protein food and coca means that a man may temporarily rely on the production of others, but it is regarded as a very bad reflection on a woman if members of her family draw on the cassava supply of other women. For another, coca production, although elaborate, can be completed in less time than manioc harvesting and separation. If a large quantity is prepared, picking and processing may occupy the morning and afternoon respectively (in the manner of manioc), but if a small quantity is required, picking usually starts relatively late in the afternoon when the men have returned from hunting or fishing. Nevertheless, coca production must be planned to end at dusk, and it is evidently seen as an important factor in structuring time, for, in a myth describing the origin of night, it is explained that when there was no differentiation of day and night, people got muddled up with their coca processing. We may therefore sum up by saying that both coca and manioc production serve to structure the day and bring on the transition to night at their completion, but that manioc is more practically effective in this respect than coca. Men's protein production does not structure time, and is an alternative to coca production. While production is confined to the daytime (although cassava baking and coca processing may slip over into the dark a little), consumption is divided, so that food consumption is confined to the day while coca consumption is not. Meals are initiated by breakfast at dawn and then held whenever appropriate until dusk. After each meal, drinks are served. These are made of manioc juice, starch and fruits, but they do not contain pepper. They are supplied by each female family productive team in the manner of pots of food for the maniocbased meal and cassava. After the meal, everyone washes carefully and then the men hand round coca and cigars amongst themselves. These must never be taken in conjunction with food and, since they are the substances off which the spirits or souls of the ancestors live, they may be described as 'soul-foods', as opposed to the cassava and peppery dishes which nourish the body. Men who are smoking or chewing coca may accept drinks and therefore, in some respects, these may be said to mediate between the food and soul-food. The cigars are smoked a puff or two at a time, and the coca is taken from a round gourd on a bone or leaf scoop; it becomes a sodden 203

Production and consumption wadge in the side of the mouth and is gradually swallowed. One man, acting as host, passes first the cigar and then the coca to another, who uses them and then either returns them to the host or hands them on to the next man. This slightly formalised distribution of coca and tobacco is especially appropriate after meals before the community disperses for productive labour, but it is also typical of any time when several men are together and no food is being served. When alone outside the house, men draw on their individual pouches of tobacco and cigar stubs. However, the 'soul-foods' come into full force after dusk when the men's circle gathers. The men sit on stools in a circle towards the right-hand centre front of the house (see fig. 5) and converse, tell myths and joke together. Throughout, coca, cigars and frequently snuff as well, are handed from the headman along the line and back (for in spite of the circular seating pattern the line has a beginning and end). Like its production, the consumption of coca is accompanied by formal speeches and verbal acknowledgements, which may be lengthened and exaggerated to suit the formality of the occasion. Towards midnight, the headman disperses the circle by initiating an exchange of 'goodnights' and shutting the men's door. Women, who spend the time after dusk in small groups around the outside of the men's circle, usually leave for their hammocks before the men. They are less strict in observing the prohibition on eating after dusk. It is clear that the daily pattern reproduces the smaller-scale pattern of the meal. This is represented in fig. 30. While the latter contains the sequence: food, drink, coca and tobacco, dispersal, the former contains the sequence: food during the day, manioc juice at dusk, coca and tobacco in the men's circle, sleep. Although the day as a whole is opposed to the night according to the presence and absence of food consumption, the night is also internally divided into the men's circle and sleep - the transition ideally occurs at midnight. We should also note that the organisation of time by consumption is different for the opposite sexes, because women are virtually excluded from tobacco and coca consumption and retire much earlier than men. The patterns evident in the organisation of the daily socio-economic cycle can be identified more clearly by comparison with ritual occasions (fig. 30A, B and C). Production and consumption on ritual occasions Indians attend communal rituals every few weeks. These gatherings 204

food

A MEAL CYCLE

cowmumiry gathers

drinks

MIDDAY food

B DAILY CYCLE

DAWN

DUSK

C SECULARRITUAL CYCLE

Note Women do not norw\c\\\y participate in consuvnptiovx of coca, tobacco byage or observe seclusion after communal ritual. Fig. 30 Structuring time by food and drug consumption 205

Production and consumption break the normal routine of maintaining the material basis of existence of the community. The daily pattern is suspended, and both food and sleep are forbidden during the ritual (strictly for initiated males and less so for women and children). At the same time, longhouse communities lose their autonomy and become temporarily drawn into a wider social unit, which represents the interdependence of separate longhouse groups for marriage and political purposes. Since the presence of visitors increases the formality and ritualisation of daily life, in one sense the ritual occasion may be seen as a product of the gathering of longhouse communities. Members of the host community visit the guest communities some days before the dance to invite them 'to drink beer', adding the nature of the ritual and the names of the dances to be performed. The invitation speech sets the day by giving the calendar of the preparation of beer, which is made up as follows: 'bringing-in-manioc day', 'pairing-off bringing-in-manioc day' (in the case of a large gathering), 'beer-working day' and 'beer-drinking day' on which the ritual begins. The invitation also mentions the coca preparation throughout the days previous to the ritual. Visitors who arrive before the appointed day are drawn into the preparations. The beers are made by particular women on the orders of the male host; they may be based on different vegetable materials and are kept separately — the main one in the beer canoe and others in tall pots. The rest of the women assist the beer-makers. It is clear both from the phrasing of the invitations and the actual sequence of events that the beer preparation, and, to a lesser extent, coca preparation, structures the preparation period and actually achieves the transition into 'ritual time'. During the preparation, certain elements of the ordinary daily pattern of work are extracted and stretched out while others, hunting and fishing and cooking dishes for communal meals, simply continue in desultory form. Making beer involves the harvest and separation of roots and boiling up juice, which form part of the normal manioc process, but these are drawn out over two or three days instead of being confined to one. Cooling the juice prior to serving is replaced by pouring juice onto chewed solids followed by overnight fermentation. The fermentation is actually a cooling process too, but it is delayed in time by heaping hot ashes round the base of the fermentation vessel. The fermentation transforms daily foods and juice into a ritual drink which must be kept separate from the foods and juice from which it was made. This 206

Structuring time by production and consumption transformation takes place most appropriately in a canoe (idire kumua, which literally means 'beer canoe') which is the means of transport between communities, transport from the ancestral beings in the east to the present inhabitants of the Pira-parana and also transport from life to death, since corpses are buried in canoes. Pots are also transformative symbols, as we have seen from their use in manioc and coca processing and also in cooking. Finally, drinking beer when the ritual begins is like drinking manioc juice at the beginning of the night on ordinary days. While manioc-beer processing has its own momentum, culminating in the correct degree of fermentation, the coca leaves must be toasted fresh: the whole process is repeated daily and the product accumulated. Nevertheless, the few hours spent on coca in the normal day are extended to fill the whole day and sometimes a number of hours after dark, too. The obliteration of the division between day and night can also be seen.in women's pre-dawn sieving of fermented beer (p. 180). Therefore, overall, preparation is like a long day devoted to coca and manioc work and relatively devoid of protein production and family cooking. During the night of fermentation which achieves the transition into ritual time, the men make their own transition. They sleep very little and anticipate the next day by finishing their coca, preparing tobacco snuff (rather similar to coca preparation) and starting protective shamanism of sacred substances to ward off the potential dangers of contact with the ancestral world. They also start to play musical instruments and, if He instruments are to be used, they play them outside the house during this pre-ritual night. Although, in one sense, it is the women's beer-making that brings communities together, we should not forget that it is the men who issue the invitations and, during the ritual itself, wear the sacred ornaments, chant, lead the dance and consume the sacred substances. Of the latter, only beer is freely consumed by women and children and, even so, it is reserved for men if there is a shortage. The female-promoted gathering of groups and the essentially male interaction which follows during ritual may be compared with sex roles in the creation of the social structure. Women are the substance of affinal links between groups; they also differentiate and order the members of a sibling group through their child-bearing cycles and yet relations between groups, whether they are affines or 'brothers', are conducted by men, who wage war, exchange ritual invitations, take 207

Production and consumption women in marriage and perform either dangerous or beneficial shamanism. The role of women in providing the beer which creates the ritual occasion has yet further significance. It transposes one aspect of the everyday relation between the sexes — the flow of manioc products from women to men — to the relationship between the opposed groups of hosts and guests. The guests, as invited recipients of beer, are like men in relation to the 'female' hosts. This is confirmed in the use of house space, for the guests are confined to the men's end of the house (where they enter) and the hosts travel back and forth serving beer and maintaining a female position in relation to the guests. It is significant that the guests bring coca, but no manioc products, with them. Furthermore, there are two kinds of ritual at which food changes hands: one is the cassava distribution after reintegration of initiates, when the cassava passes from host to guest, and the other is the food exchange already discussed, in which the guests bring the gift of wild products. Kaj Arhem reports that an informant from Cano Komeyaka actually likened the guests entering with their ceremonial gift of fish or meat to a penis entering a womb (personal communication). This accords perfectly with the relative positions of guests and hosts, as well as with my own material on the symbolism of house space (ch. 7), and is a further indication that symbolic fertilisation may be important in the structure of all communal rituals. There is actually a point in communal ritual proceedings at which the chief guest and host retire to the extreme female end of the house to chant together, which may well represent the climax of the symbolic fertilisation process (but unfortunately I do not know the content of the chant). Returning to the parallels between the daily and ritual structuring of time, many of the essential features of the communal ritual appear as stronger versions of corresponding features of the men's circle. The loose arrangement of participants on a circle of stools below the box of feather ornaments (normally suspended from the roof) is replaced by two rows of men (roughly 'visitors' and 'hosts') who wear the ornaments and sit nearer the men's door. The cigars, coca and snuff consumed each night are present in greater quantities in the dance context and they are accompanied by yage. Yage is called idire kahi, 'drinking' coca, as opposed to coca which is bare kahi, 'eating' coca, or simply kahi. This suggests that yage may be regarded as a liquid extension of coca, in much the same way that beer is a liquid exten208

Structuring time by production and consumption sion of manioc. Beer and yage are actually served as a pair; yage, which is very bitter, cannot be kept down without a draught of beer immediately after drinking it. These same sacred substances that are an essential feature of all communal rituals may also be used to differentiate between moresacred and less-sacred rituals. The different varieties of yage, the vessels and implements for taking snuff and coca, and the special ceremonial cigars believed to derive from ancient times are graded according to their degree of sacred power and are chosen to suit the occasion. In this, they are like certain ornaments and musical instruments (for instance 'imitation' and 'true' He) and the factor of 'absence of women', all of which indicate the sacredness of the occasion. All these sacred elements make the dance participants enter an alternative world identified with the ancestral past. This alternative world is present in a weak form as an aspect of everyday life accessible to men, but hardly at all to women and children. Since yage is only the cultivated 'drug' used exclusively during ritual, it obviously plays a special part in distinguishing ritual life from everyday life. It provides what might be described as the 'ritual frame' — the other-worldly experience of the participants. Being both produced and consumed during the ritual itself, it is not associated with the cyclical alternation of ritual and secular time in the same way as manioc beer. Instead, it refers to a mode of time in which the beginning of creation in the mythical past is the point of reference, for it is said to show the men how to chant and dance — activities which reproduce the original creative process, as we have already seen. Besides creating the ritual in the sense of making the participants perform in the appropriate way, the yage also creates an alternative experience of time and space to the everyday one. Those who drink it see and hear things beyond everyday reality. Although their souls remain within their bodies it is, as one informant put it, 'as if we were so large that the universe is small by comparison and we can see everything'. Indians also say that the house becomes the universe during ritual so that the men's door becomes the eastern entrance to the earth we inhabit. Much of the chanting 'shown by yage' is about the mythical associations of the landmarks on the journey upriver from this eastern entrance made by the group of ancestors. The chanters actually identify themselves with these ancestors, saying, 'We, Fish Anaconda Son (or some other group ancestor) came up river 209

Production and consumption and . . . ', demonstrating that, although they do not move from their stools, they are travelling in time and space. The accounts of yage visions which are filled with anacondas, jaguars, ancestral houses and distant geographical settings fit with this 'ritual frame' achieved by yage. Indians say that the hallucinations are produced by yage in the form of an anaconda which, having entered the body, thrashes around inside and slides out as a stream of vomit. Thus, the anaconda form of the ancestors within the universe is united with the anaconda form of the yage within the bodies of the participants. We are now in a position to draw together the scales of organisation of time. The relationships which follow are represented in fig. 31. In the men's daily cycle the day is divided between protein provision and coca production, while the night is divided between the men's circle and sleep. The ritual is like an extended men's circle and the preparation period is marked by the extended production of coca. Similarly, for the women the daily cycle divides into manioc production and cooking meals, while the night is filled with a weak form of social interaction after dusk, and sleep. The ritual is like an extended form of this interaction and the preparation period is filled with the extended first two stages of manioc production. Ritual, together with the preceding preparation, may thus be seen as an extension of selected aspects of the daily cycle. The aspects that are not selected are production, cooking and consumption of protein (and other peppery dishes), stage (3) of manioc production, cassava consumption and sleep. These may be contrasted with the selected aspects: stages (1) and (2) of manioc production, coca production, consumption of the product of manioc; stages (1) and (2), beer, consumption of coca and formal social interaction. We have already seen that the provision of communal meals takes place within the family unit, involves both husband and wife (protein provision and cooking) and also refers metaphorically to the process of formation and dispersal of family units. For everyone except the unmarried, initiated youths, sleep takes place in the family compartment where husband, wife and children are enclosed together. It therefore seems that interaction between the sexes within family units is incompatible with ritual time, while separation of the sexes in their respective manioc and coca work and during the ritual proceedings themselves is appropriate during the preparatory phase and the ritual itself. Although the sexes dance together during the least 'sacred' rituals, the partners are not husband and wife. At the most 210

Structuring time by production and consumption extreme form of ritual, male initiation, the sexes are totally segregated. The chief problem with this conclusion is that although the manioc process has a product at the end of stage (2), beer, which is closely

DAILY CYCLE PROTEIN

MANIOC STACE 3

MANIOC STAGES 18-2 COCA

RITUAL CYCLE MANIOC STAGES

COCA KEY PRODUCTION

NO PRODUCTION OR NO CONSJMPTlON

CONSUMPTION

SEX OF PRODUCER / OR CONSUMER

Fig. 31 Production and consumption of protein, manioc and coca 211

Production and consumption parallel to coca, it has another at the end of stage (3), cassava, which, as a food, is opposed to coca. However, there is plenty of evidence that coca and manioc are paired both as crops and as final products (cassava and coca). First, coca and manioc are the most important male and female crops. Secondly, they are planted in a pattern which recalls the bone—flesh relationship, the coca resembling a skeletal human form and the manioc 'covering' it to make the circular maniocgarden shape. Thirdly, as they are originally planted in Yeba's garden — the 'first garden' — coca is the body of Yeba's brother and manioc is Yawira's sisters (p. 183). Origin of cultivation of coca When Yawira had brought the cultivated food plants to Yeba from her father, she ordered Yeba to fell her a manioc garden and not to care for his younger brother too much. She got Yeba's younger brother, Nyake, to help her carry the bundles of manioc sticks ready for planting. As he did so she seduced him and they made love on the path, at the edge of the garden and finally in the middle of the garden. Here he died on the point of ejaculation and she laid out his body to make the rows of coca. This was nyake-coca, the variety owned by Yeba's people ever afterwards; it was different from Fish Anaconda's coca. When Yeba went to pick his coca, the bushes bled with human blood. Later, Yawira bore a son, the ancestor of Nyake Hino Ria {Nyake Anaconda Children), also called Rasegana, a dancer/ chanter sib. Thus, Yeba obtained coca from his brother's body and Yawira obtained manioc from her sisters' bodies. The myth suggests that male descent-group relations and coca are associated (the more so since the creation of coca also leads to the birth of a descent line) and that these are contrasted to female descent-group relations. In addition, the male descent-group members need the creative powers of females from opposed descent groups to perpetuate their lines. In one sense, Yawira is obviously like the fertile earth on which she lies and Nyake's ejaculation is a 'planting' act, for both the coca and the child bear the same name. The identification of coca with a male descent line and of manioc with a woman of opposed descent-group origin will be discussed again later (pp. 228f). Fourthly, Indians consider that the vegetative reproduction of both manioc and coca create continuous 'paths' over time. This type of 212

Structuring time by production and consumption reproduction distinguishes the two from pepper and tobacco (which have fertile seeds) but it likens them to yage which, we saw, shares the name kahi with coca.4 I return to other aspects of these similarities and differences on pp. 228f. Finally, processing manioc into cassava and processing coca take place in the open female and male part of the house respectively. Both make use of dry toasting (gate-) rather than boiling (roa-), the cooking method appropriate to the peppery foods which accompany cassava. In both cases, some of the productive equipment is communally used (although it is owned by the maker or purchaser). At the end of processing, each is set up on a wicker stand. All these factors suggest that, although cassava is a food rather than a soul-food, it is a female counterpart of coca and is opposed in many ways to the foods into which it is dipped. Like other foods, it nourishes the bodies of consumers rather than their souls, but, as will be shown below, it is connected with a particular aspect of bodily growth opposed to that connected with the other foods. However, first we must consider how the various foods are related together in the re-establishment of secular time after ritual, pointing out some of the salient features of the order in which they are shamanised. Reintegration through food

consumption

The serial shamanism of foods which occurs at the beginning of life or after the onset of menstruation or ritual participation was briefly described on pp. 122f. Here, I use the post-initiation period, during which reintegration through foodstuffs is most carefully ordered, to bring out the basic principles of this process. The sequence is represented in fig. 32. There are slight differences between this food series and the one at the beginning of life, since, as the baby only takes milk, it does not undergo a seclusion diet. At first, the diet consists of thin starch cassava, raw and toasted ants and termites, some of which are caught by the participants, and drinks of toasted starch crumbs in water. All must be eaten cold. Basically, this diet lasts until the pepper shamanism, but certain forest fruits and then cultivated vegetable foods are added. At first, the only cooked foods allowed are toasted ones; even coca is reshamanised after the ritual, in order to make the toasting fire, which is used to 4 Pepper may be reproduced either by seed or by planting the sticks of an old plant.

213

._

END OF RITUAL

SECLUSION 'ANTS & (TERMITES

FRUITS 6- ROOTS

SALT COLD MANIOC JUICE

•STARCH

j .

i

PEPPER | SHAMANISM |

PEPPER & HOT MAWIOC JUICE

I SMALL FISH£ J MANIOC ] LEAVES ICOMPLETE 'CASSAVA i i

UNDERGROUNDTOASTING BOILING

HOT FOOD FOOD COOKED WITH

214

SMALL FISH SMALLS— LOW-LIVING GAME

LARGE GREASEFILLING FISH LARGE & HIGHLtVlNG GAME

ABOVE GROUND WATER

Fig. 32 Sequence of food shamanism following He wi

SECULAR LIFE

->LAND

Structuring time by production and consumption make the new batch, safe. With the vegetables, boiling is introduced, but food must still be eaten stone cold. Shortly before pepper shamanism, cold manioc juice (also boiled) and salt are added. Pepper shamanism introduces pepper and hot manioc juice and removes the danger of fire. It is followed immediately by shamanism of small fish (wai ria, which literally means 'fish-children') caught in funneltraps by the initiates and cooked with manioc leaves and pepper. These are followed over a period of months by other fish roughly in order of size, but also arranged into categories according to fishing methods, cooking methods and special natural characteristics referred to in myth. Long after the least harmful fish, but before the largest 'grease-filling' fish (p. 120), come ground-living game birds and small mammals. Later on, other categories of mammal and the most dangerous fish are introduced; the series culminates with tapir. The sequence in the post-pepper shamanism part of the series depends partly on availability of the rarer foods. Complete cassava (made with starch and fibre) is introduced immediately after pepper shamanism. The series has certain dominant themes: one of these is the gradual emergence from the cold world of He wi. During the emergence, various elements associated with domestic fire are introduced in an ordered fashion: toasting, boiling, hot food, pepper and direct contact with fire. Firewood is obtained by women from cultivation sites, and is thus a female-controlled by-product of men's forest-burning. The once-burnt wood which lies in the clearing is also the burnt bones of Manioc-stick Anaconda, a product of the burning process which we saw was paired with the burning that resulted in men's He (p. 184). The term for both fire and firewood, hea, also suggests a link between these and the sacred instruments and, besides this, it is said that mad people try to play pieces of lighted firewood as if they were He. Just as the manioc stick was shown to be a female equivalent of He, fire is a female inversion of He. Another theme is the gradual emergence from the 'underneath layer' represented by He wi. The ants and termites in the seclusion diet are classified as 'inside the earth' creatures, and in the full version of Manioc-stick Anaconda's competitive relations with Macaw, the emergence of initiates from their ritual experience is specifically linked to the seasonal emergence of the flying form of one of these permitted species of ground termite (p. 261; S. Hugh-Jones 1979 : M.6.A). In eating these, the participants are identifying themselves 215

Production and consumption with this emergence from underground. Also, while starch may be considered the lowest, underground part of the manioc plant, the first ordinary meal of boiled fish and pepper is mixed in with manioc leaves, the uppermost parts of the manioc plant, which are called kl gaha, upper end of manioc. Thus it seems that the participants are also identified with the vertical structure of the manioc plant which spans the space between underground and sky. This is consistent with the general analogy between birth and manioc cultivation set out before (p. 125), and with the prohibition on any foods that do not grow from the ground during seclusion. After pepper shamanism, the overall progression is from wai rla (small fish) to wai bukura (game animals), just as human growth is from ria masa, children, to bukura or bukura masa, elders. Thus, during seclusion, the participants emerge from the ground and afterwards they grow from water onto land, becoming harder and more mature. The contrast between the starch of the seclusion diet to the complete cassava eaten with the normal communal meals following pepper shamanism once more implies a change from origins to maturity (see analysis of manioc process p. 189). Yet another general change throughout the series is from foods which may be obtained by both sexes to ones that are exclusively got by men. Although the participants are supposed to obtain at least part of the non-female-cultivated portion of their diet themselves as an integral part of their growth process, all the foods included before pepper shamanism may be collected by women also; in fact, women did supply the participants with some of their foods during the seclusion period we witnessed. The foods introduced after pepper are in an order roughly corresponding to certainty of success, as discussed above (pp. 171—3); fish come before game animals and tiny fish caught in traps come first; poisoned fish follow these, preceding larger fish and game. The first of the game creatures are tinamou (ngaha), ground-creeping birds which were also traditionally caught in large numbers by trapping. We saw before that certainty of success or reliability in production of food correlates with allocation of production to women, and lack of reliability with allocation of production to men. Overall, the dietary progression is from female to male productivity and from reliable to risky food sources. By obtaining the foods to be newly introduced for themselves the participants are learning both a progressively more male economic role and to 216

Interpretation of foods overcome the independence of the natural world by building up fishing and hunting skills. An additional progression easily observable in the series is from small to large. The contrast between the extreme diets of thin starch cassava and ants or termites on the one hand, and thick cassava containing coarse fibre with tapir meat on the other, is sufficient to demonstrate this. A composite change which may be derived from the particular changes noted above is from a cold, underground diet eaten in seclusion to participation in communal meals. These meals incorporate the very elements — boiling, addition of salt, addition of pepper, and consumption while hot - which we saw to be metaphorically linked to the female gestation process and the family as the focus of sexual reproduction of community members. Meat, the ideal substance for such meals, was shown to represent the transformation of a woman obtained from outside into children born into the husband's community. It is appropriate enough that pepper should be the focal point of this change, since it stands for the essence of this female sexuality — the creative blood. The coincidence of pepper shamanism, the henyerio rite, and the reintegration of participants into the heterosexual community with its secular daily routine serves to underline the ability of the ritually reborn individual to partake in this world of creative conjunction between the sexes. Interpretation of foods Starch The secular world is characterised by hot sexual conjunction, family relations, a complete diet of foods and productive economic activity. It contrasts strongly with the world of He w/, which is cold, dead and characterised by consumption of soul-foods, separation of the sexes and emphasis on ancestral female sexuality. These hot and cold worlds are the worlds of natural conjunction and social restraint mentioned before (p. 158). The seclusion period is both intermediate in time and of an intermediary character: the sexes are still separated, but the participants depend on a constant flow of starch products from women: they eat food, but only when it is very cold, and there is no dipping cassava into other, peppery foods as in normal times. 217

Production and consumption There is some productive activity, but the ants and termites collected are not shared with women and the basketry artifacts made are not given to women until after pepper shamanism. In the analysis of birth, I suggested that the seclusion after birth is a re-enactment of foetal development. The general position of seclusion periods between metaphorical death and life makes them open to the same interpretation but, of course, the interpretation must take account of the different natures of the occasions for seclusion: birth, menstruation and He wi. The principal seclusion food is thin starch cassava, and, by looking closely at the ritual role of starch, we can relate the different types of seclusion. Although it is prohibited as food during i/e wi (pp. 144f), starch is rubbed into the incised patterns on some of the He instruments. During the initiates' seclusion, marked by a starch-based diet, there is the elders' pepper shamanism rite at which the initiates throw pieces of thin starch cassava on bark-cloth strings into the roof. The initiates retire to play He on the plaza, while the elders receive their shamanised pepper and hot manioc juice and afterwards dance with women — all inside the house. The initiates must not see this activity, but the elders periodically dance out onto the plaza, revealing themselves to the initiates and the He. On this occasion, and on the later occasion following pepper shamanism for the initiates when thick rounds of starch cassava are distributed, the house is said to be 'white all over' with starch. Indians say that the house is a womb and that manioc juice is milk. Remembering that starch is like semen in various ways, we may see the initiates during seclusion as conceived but unborn creatures. The thin starch cassava is the seminal stuff of which they are made and the house is the collective (male-constructed) womb to which the starch—foetuses are attached by umbilical cords (strings). The reintroduction of hot manioc juice, explicitly identified with milk, into the diet coincides with their 'birth' out of seclusion and thus represents the milk drunk after birth. While the initiates are still identified with the cold, dead world during the elders' pepper rite, the elders themselves receive pepper and dance with women inside the 'womb', as well as seeing He, without coming to any harm. This is evidence of their established periodicity, their mastering of the transition between the hot and cold worlds. The thick rounds of starch cassava (about four times thicker than the seclusion cassava) distributed at the initiates' reintegration can be identified with the fully grown foetus 218

Interpretation of foods which is released into the living world or, in terms of the initiates, into heterosexual society in a new adult status. While this interpretation rests on the identity of starch and semen, we should not forget that cassava is produced by women. In myth, both starch and fibre were represented as sexually desirable women, and we also saw that in certain contexts starch may be contrasted to fibre as female maturity contrasts to male maturity. An interpretation of the role of starch in terms of cold seminal origins opposed to the hot female sexuality embodied in pepper is not sufficient by itself, because it seems that starch is also associated with mature female sexuality. However, the cold female sexuality compatible with He instruments and likened to semen is obviously very different from hot sexuality. While cold starch cassava is provided by women who are segregated from the initiates and observing ritual restrictions, hot peppery food is cooked by women engaging in normal heterosexual relations with men and co-operating with them in production of protein meals. Starch is thus metaphorical sexual nourishment from women, who are actually sexually forbidden. Birth of new descentgroup members depends upon correct exogamous marriage and this, in turn, depends upon relinquishing sisters. Thus starch, in representing the sexuality of forbidden women, can also be said to represent the forbidden sisters in their positive aspect — it is these sisters who 'nourish' the descent group with wives. Put in another way, it is the descent-group aspect of women that is represented by starch. This is opposed to the natural sexuality represented by pepper, and so the relationship between starch and pepper is one of cold, social classification to hot, natural conjunction. Both the opposed descent-group membership of wives (exogamy) and the natural sexuality (the female reproductive cycle) are essential to the reproduction of the husband's group. I believe that there is linguistic support for the association of ritual restrictions, starch and prohibition of incest. It seems highly likely that the terms for 'younger brother' and 'younger sister', bedi and bedeo, are related to bedi-, 'to observe ritual restrictions'. If this is so, then the root bed- may be given the general gloss 'to be forbidden or closed off'.s In Makuna, the closest of all Tukanoan languages to Barasana, starch is bede, a term which apparently utilises the same root. While the natural female sexuality associated with pepper resides 5 The gloss 'closed off also covers the appearance of bed- in bedo, ring (a closed-off line).

219

Production and consumption in the menstrual cycle and is therefore adequately represented by menstrual blood, it is not yet clear what represents the cold, descentgroup aspect of mature women. Although women's descent-group membership, like men's, originates in the father's contribution of semen, we have seen that the interpretation of starch as semen does not account for the specific associations of starch with female sexual maturity. Here I would like to suggest tentatively that starch may have an additional reference to vaginal fluid, and that this substance is regarded as a product of sexually mature women which constitutes a transformation of the semen from which they originated. Although I have no direct evidence to support this suggestion, the term for vaginal fluid, ahea badi (ahea means both penis and vagina), is the same as that for semen, and the whiteness of the two also unites them. This identification also makes sense of the reference to starch in the mythical episode of Yeba's penis (p. 185) and that of Dragonfly's daughters (pp. 185f) as well as the use of starch on He instruments. In the episode of Yeba's penis, Yeba eats starch. Since eating is metaphorical sexual intercourse, we may say that it is sexual contact with Yawira's cold, vaginal fluid representing her separate, underwater descent which transforms Yeba's penis to the human position. This interpretation accords with my association of starch with the descent-group aspect of women and also with the Indian view that the Yeba— Yawira cycle is about the interdependence of civilisation and exogamy. In the episode of Dragonfly's daughters, the cold, white Starch Woman was hidden away from strangers by her father. However, the stranger who succeeded in making love to her became white and alive on his front. The fact that the starch-skin of this woman is transferred to the stranger during sexual intercourse fits my interpretation of starch as vaginal fluid. The starch is associated with natal origins, since Starch Woman is closely guarded by her father, and yet it is life-giving to outsiders: this is consistent with its association with the descent-group aspect of women. The starch rubbed into the He instruments is also open to interpretation as vaginal secretion, especially in view of the analysis of the He wi rite as a symbolic fertilisation, in which the phallic instruments 'eat' from the house filled with beeswax aroma — a cold, ancestral version of the blood-filled womb (pp. 1540Now the starch diet during menstrual and post-natal seclusion can also be explained. The seclusion of women following the onset of release of blood is a type of social control imposed upon natural 220

Interpretation of foods female periodicity. The women are set apart spatially and they are also sexually forbidden, so that their relation to the male community becomes like one of sister to brother. The first menstrual seclusion which is imposed by the girl's descent group may be seen as evidence of their control over her newly matured sexuality, while menstrual seclusions following her marriage may be seen as recognition of her opposed descent-group origin. Thus, while the menstrual cycle as a whole is a natural cycle, social rules divide it into a natural, hot phase which coincides with natural retention of blood, and a ritual, cold phase which coincides with the release of blood. In the myth of Frog Wife, there is evidence that the seclusion phase is associated with the descent-group status of a woman, while the intervening period is associated with wifely status. We have seen that Frog Wife's return to the cold, white Mildew House in the ancestral east where her own people live is actually a metaphorical menstrual seclusion (p. 166). In eating starch, the menstruating woman is recreating her own seminal, descent-group origins and at the same time she is reinforcing the cold aspect of her sexual maturity. When she takes shamanised pepper she moves into a phase of hot sexuality again. In the post-natal seclusion, father, mother and new-born baby withdraw together into the family compartment. By eating starch, the father identifies himself with semen and thus stresses his new parenthood, but he also performs metaphorical sexual intercourse with cold female sexuality. Meanwhile, the mother emphasises her own descent-group membership and cold female sexuality. Therefore, besides stressing the father's creativity (semen) at the expense of the mother's (blood), the starch consumption also refers to the social aspect of family structure, its foundation in exogamy. Of course, these two aspects of the seclusion are related. I have suggested that starch represents semen and also vaginal fluid, while pepper represents menstrual blood. Although vaginal fluid may be said to represent the descent-group status of sexually mature women, there is no evidence that it contributes to the foetus — the only female contribution to the foetus is blood. We saw that both semen and female blood are related to the physical substance of the body but that semen, unlike blood, establishes the descentgroup membership of the child. However, the male power to create the physical substance of the child with semen contrasts with the male power to install or renew the soul of the child at naming and at He wi. This explains the position of the starch diet midway between 221

Production and consumption ritual periods, characterised by a diet of soul-foods, and secular periods when all food is eaten with pepper. Starch represents the physical aspect of the descent group manifest in the bodies of its male and female members, rather than the spiritual aspect manifest in names or souls. Thus, although starch is the whitest and purest of foods, it is still food for the body rather than food for the soul. We now have a simple model relating consumption patterns to the elements in the reproduction of the individual, as follows: time sequence: food consumed:

ritual soul-food

associated element of individual:

soul

seclusion cold food (starch)

secular life hot food (pepper)

, semen

female blood, body Although the model above describes the reproduction of the individual, it omits the descent-group aspect of women which is crucial to the creation of an exogamous marriage, and which is also embodied in starch. Likewise, it omits an aspect of men associated with pepper. Pepper is not only eaten; the juice of raw peppers is sniffed by men to induce fierceness and it is also rubbed into the skin to draw out grease, which forms a base for red paint (ngunanya). Red paint is also associated with fierceness, for those intending to kill paint their faces and bodies solid red. Again, the red paint applied to men's navels during He wi induces fierceness. The chain of associations runs full circle, because the sacred red paint is a male, ancestral version of menstrual blood, and we already know this to be associated with pepper (p. 195). Besides, pepper, red paint, the blood of killing and menstrual blood are all linked by their red colour. The association of male aggression with pepper completes a paradigm which opens the way to an interpretation of the ideal meal: white (starch) red (pepper) male descent-group substance natural destruction female descent-group substance natural creativity Cooked meat The ideal meal eaten on secular occasions consists of thick cassava made of starch and fibre and meat boiled with pepper. The cassava represents mature males (fibre) and mature females (starch) in their 222

Interpretation of foods capacity as descent-group members (pp. 218ff). The pepper represents hot, female sexuality and the meat hot, male aggression, since it is the result of male killing. These last two are the paired natural aspects of sexuality concerned with the increase of the community and its success in competition with outside communities. The first two are the paired social aspects of sexuality concerned with the exchange of female reproductive powers between descent groups. Just as the cassava is dipped in the peppery meat juice, the cold descent-group substance of males and females must be imbued with the hot natural powers of female sexuality and male aggression before children can be born. Or, put the other way round, these hot powers must be cooled by exogamy for the purposes of social reproduction. This fits with the practical stress on the immorality of eating meat without cassava.6 The salt, which I have omitted so far, was earlier interpreted as semen in its minimal role as 'activator' of the female creative powers. I suggest that this 'female-activating' role of male sexuality contrasts with the descent-group role embodied in manioc products, and is conceived of as the least spiritual manifestation of the male reproductive capacity. The distinction between the cold and hot constituents of the meal may be represented as that between the separate descent origin of husband and wife and the natural conjunction of the two. As with the specialist-role series, the contrasted elements may be conceived of as differentiated by time — the transformation of brothers and sisters into husbands and wives - or, alternatively, as opposed aspects of each sex which are present all the time. Another way to conceive of the same relationship is as the one between the two basic modes of obtaining a wife, described in chapter 4: sister-exchange and the aggressive raid. According to the practice of sister-exchange, sisters actually represent future wives, and opposed groups are able to reproduce in a socially controlled manner. However, wife-raiding implies the reproduction of the most aggressive group at the expense of the weaker group. This is a 'natural' form of marriage compared with the first and, as we saw from the Tukano text, it is just one stage above warfare. My discussions of the analogy between killing game and raiding wives in the context of meat production provides further evidence that meat is related to wife-raiding. Also, Indians say 6 The word uko-, to complement or balance, is used for eating cassava together with peppery food or any animal product. The same root exists in the word for medicine, uko, and a great many medicines are designed to cool the dangerous heat of disease.

223

Production and consumption that the warrior 'sees his enemies as game animals', and thus killing people and killing animals are basically the same kind of activity, different only in degree. We may therefore conclude that wife-raiding and game-killing are alike in being one stage more social, or at least one stage less antisocial, than warfare. Pepper pot Having considered the ideal meal of meat with cassava, I turn now to the minimal meal in which boiled meat is replaced by pepper pot, hiari. I have already shown that hiari is paired with cassava in the wicker stand. It is also opposed to cassava, as follows: cassava hiari colour white red—brown consistency dry solid viscous liquid seasoning no pepper lots of pepper cooking toasting boiling position above on stand below on stand container basket pot container colour pale brown with black white leaves According to parts of the previous analysis in this section, these qualities should liken hiari in its pot to retained menstrual blood in the womb. The meaning of hiari as 'bitter/sour stuff supports this, as does the reference to Tapir shaman's wife's 'pepper pot'. In spite of these qualities, hiari is made from manioc juice which is explicitly likened to milk — a pure white substance. The production of hiari involves the transformation of thin, sweet, straw-coloured juice to a viscous, bitter, dark caramel by prolonged boiling, that is, by absorption of a great deal of heat or fire. The obvious inference to be drawn from this is that menstrual blood in the womb is a hot, concentrated product of milk in the breast. I heard no direct statement to this effect, although at least one Amazonian group holds this theory. 7 However, my discussion of birth and growth shows certain relations between these two manifestations of female body liquid: first, they are related as respective sources of ante-natal and post-natal nourishment of the child; secondly, from the mother's point of view, they are retained and released nourishment respectively; and thirdly, at 7 The Kayap6 believe that the female contribution to the foetus is milk which drips down from breasts to womb inside the body (T. Turner n.d.).

224

Interpretation of foods birth, the released placental blood becomes dangerous and is buried, while the milk takes over as nourishment and is retained in the growth of the child. We may therefore construct the cycle in fig. 33, according to which retained milk in a high position — the mature girl's breasts — is heated and transformed into retained menstrual blood, which in turn nourishes the foetus via the umbilical cord: the newborn child is a hot, dangerous product of this blood, but its link with the blood is made to decay in the ground and is replaced by cooling milk from above. If we return to the parallel domain of cooking, manioc juice is produced by boiling the liquid separated from manioc solids and it is MOTHER,

[FOETUS]

CHILD

HIGH MILK IN BREAST

PLACENTAL BLOOD IN WOMB

CHILD DECEIVES MILK& LIVES

FOETUS RECEIVES PLACENTAL BLOOD IN WOMB

birth

CHILD & PLACENTAL BLOOD BORN

BLOOD DECAYS

LOW Fig. 33 Transformation of female body liquids 225

IN GROUND : DANGEROUS HEAT LOST

Production and consumption turned into hiari by further boiling. Therefore, manioc juice may be said to mediate between cold cassava and hot hiari. It is thus a perfect transformative symbol which belongs between hot and cold worlds; this fits with serving partly cooled manioc juice at dusk — the transition between day and night within the daily cycle which is itself intermediary between the hot and cold worlds (see pp. 156f). By analogy, milk is both a transformative product of cold descent-group status manifest in semen and vaginal fluid and the material out of which hot natural female sexuality and male aggression are made. We shall see this analysis supported in the following discussion of relations between cultivated plant products. In the present context, that of the relationship between the minimal meal and the most valued meat meal, it confirms that the hiari represents an entirely female aspect of sexual reproduction. This is consistent with the fact that hiari is produced by women to feed men, who have withdrawn from food production (pp. 209f). Cultivated plants and social models It has already been shown that the meal consists of cassava dipped into food boiled with pepper and that these foods are forbidden at times when coca and tobacco, the soul-foods, are prescribed; that in addition, freshly boiled manioc juice mediates between food and soul-food during secular time, but for ritual it is fermented into beer — together, beer and yage differentiate ritual from the men's circle in secular periods; and finally, that starch cassava and the absence of both hot manioc juice and pepper distinguish the seclusion period from the re-established secular time when complete cassava, hot manioc juice and pepper are normal fare. Therefore, although other food-stuffs, body paint and a host of other substances are significant in the regulation of ritual and secular life, a small handful of cultigens are of particular importance. These are manioc and pepper, grown by women and consumed by everyone, and coca, tobacco and yage, grown and consumed by initiated men. Furthermore, the products of male cultivation are paired off with the products of female cultivation, and these pairs of sexually contrasting products are cultivated, consumed and otherwise used in such a way that their interrelations provide a model of the process of reproduction of social groups. However, they do not simply refer to this process of reproduction, they enter into it by uniting and separating the sexes, 226

Cultivated plants and social models families, longhouse communities and even the living and the dead at different times and in different ways. Much of the evidence for the existence of these male—female pairs has already been given. We have seen that cassava and coca are paired (pp. 211 — 13) and that their consumption is alternated. The same relationship exists between yage (or liquid coca') and freshly boiled manioc juice, for when fresh manioc juice is turned into beer instead of being drunk immediately, yage takes over. The pairing of yage and beer during the ritual and the drinking of hot manioc juice at the pepper-shamanism rite, when the effects of the He have disappeared, make up the following sequence: secular life : preparation : ritual : seclusion : secular for ritual life reestablished fresh manioc : manioc juice : beer : — : boiling hot juice fermented consumed manioc consumed juice shamanised no yage : yage pre- : no yagd pared and consumed Furthermore, liquid manioc juice together with starch-and-fibre cassava accounts for the whole root of manioc (ki), while yage and coca are also liquid and solid members of the same category (kahi) (pp. 208f), as follows: solid liquid kahi coca yagei ki starch + fibre manioc juice Since cassava and pepper, as complementary foods, contrast with coca and tobacco as complementary soul-foods, there is reason to assume a relation between pepper and tobacco which is similar to that between cassava and coca. Both pepper and tobacco are 'essences' rather than solids, because pepper infuses cooked food and tobacco is taken as snuff or smoked. So far we have the following relations: solid liquid essence male coca yagd tobacco female cassava manioc juice pepper There is evidence that pepper and tobacco are regarded as paired substances in their cultivation. Normally, pepper is grown dotted over the round, mounded beds and tobacco is grown in rows in 227

Production and consumption mounded oblong beds: both are frequently cultivated around the plaza border instead of in the manioc gardens. People state that the two crops should be grown in conjunction on sites where abandoned houses have been burnt, and that the pepper beds represent the compartments. The soil is worked into neat beds which are almost exclusively planted with these two; gourd vines are an optional extra. The only such site I saw was of a small, non-dance house which had not been planted by the owner and was therefore probably less constrained by formal rules than is usual. Nevertheless, the planting was consistent with the differential use of house space by the sexes, for the tobacco was mainly in oblong beds in the centre and towards the men's door, while the pepper was mainly in smaller round beds on the periphery and towards the female door (fig. 34). The shape of the beds is also consistent with the sexual pairing of the crops (see ch. 7). The association of these crops on old house sites is parallel to the association of coca and manioc in the manioc garden. Also, in contrast to manioc, coca and yage, both tobacco and pepper are used as trade items. The analysis of the manioc process, together with the discussion of starch and complete cassava in the reintegration of initiates, showed that cassava represents the cold descent-group substance manifest in the physical existence of the descent-group members. The myth of the origin of Yeba's coca suggests that coca, which is paired with manioc in many respects, is also identified with male descent lines. The ideal is that each Exogamous Group owns a different named variety of coca, just as Fish Anaconda and Yeba did, although in practice, communities often cultivate 'other people's' varieties. If coca represents an aspect of descent-group structure, then this must be an incorporeal aspect since coca is soul-food. However, yage is also related to descent-group structure, for the set of varieties owned by a descent group are said to originate from inside the set of Yurupary instruments owned by that group. The vine itself is described as 'men's life path' (kana ma), and as such is opposed to manioc which is 'women's life path'. Moreover, when fathers drink yage, their children are said to experience the effects: this suggests that the yage has the power to travel along human descent lines of its own accord. We may therefore hypothesise that coca and yage represent separate but related aspects of descent-group structure. The use of coca in everyday life, the united and yet internally differentiated structure of the male local descent group that picks and processes it, and a particular element of ritual proceedings in which hosts and guests offer each 228

Cultivated plants and social models

9

r^

for

)

! ©

1

f

lines of cental I house-posts. {

Key: P : T= G : S :

PEPPER TOBACCO GOURD VINE SUGAR CANE

Fig. 34 Old house site planted with pepper and tobacco 229

Production and consumption other their own coca, all suggest that coca represents the fixed structure of the descent group in its spiritual or incorporeal aspect. The more formalised or ritualised the production and consumption process is, and the more sacred the implements used, the deeper the descentgroup structure elicited by the coca and thus the wider the presentday descent group referred to. Yage, on the other hand, is not described as food for ancestral souls and this, I believe, is for good reason. The function of yage is to transport people into an ancestral state (pp. 209f); this, and its ability to flow along descent lines, suggest that it connects past and present and thus enables travel into the ancestral past. Thus, if coca represents descent-group structure built up over the generations, yage represents the ability to go back to the beginning and repeat the process today. Thus, the more sacred the ritual occasion, the stronger the yage required, the closer the contact with the original ancestors and the greater the potential danger to participants. To translate the coca—yage relationship into concrete terms, we might say that if coca is the structure of the path, yage is the vehicle, or, for a riverconscious people, if coca is the structure of the river beds, yage is the water or the canoe which travels over them. Indians do actually say that the river system of the earth is a yage vine connecting longhouse communities to the ancestral east in the same way as an umbilical cord: they say that when the vine is cut for use, the scar this leaves is a navel. Besides this, yage is identified with an anaconda — the supreme self-propelling water vehicle (pp. 209f). Its liquidity, as compared to the solidity of coca, is also consistent with its transporting power. So far, the analysis has shown that coca and cassava are differentiated as descent-group structure created by relations between men and descent substance created by women. It was argued above that yage represents the connection between present-day men and the foundations of descent-group structure in the ancestral past. If my model of the relations within the set of cultivated plant products is correct, manioc juice should be related to cassava in an analogous manner: it should represent the connection between present-day men and women and the origin of their descent-group substance. This origin is not in the ancestral past, for the physical substance of the descent group is formed at the birth of descent-group members. In the discussion of hiari it was shown that the relationship between manioc juice and cassava is analogous to the relationship between 230

Cultivated plants and social models milk and descent-group substance (pp. 224-6). This confirms that manioc juice is indeed a female 'milk' connecting the child to the cold male and female elements which contributed to its conception. Besides, yage is also described as milk (Romi Kumu's milk), and thus yage and manioc juice are paired directly as ancestral milk for the 'dead' and ordinary female milk for the living. The statement that yage is milk also fits its identification with the river system of the earth, for all rivers flow eastwards into the Milk River. The remaining pair is tobacco and pepper. Pepper represents hot, natural conjuction manifest in the menstrual cycle and cooking fire. The myth of Manioc-stick Anaconda and Macaw shows that tobacco is derived from an opposed kind of fire — the Sun's heat (pp. 88f). At the same time, it shows that tobacco is the foundation of shamanism which is, in effect, the power to cross between cosmic layers. The shaman's cigar is said to be his 'eye' with which he sees the mystical causes of illness: the rising smoke is associated with his travel to the Thunders, the Sun and other inhabitants of the upper cosmos. The special power of shamans is due to their ability to let the soul leave the body, and thus tobacco is associated both with the independent existence of the soul and the 'direct line' to the ancestral forces. Ordinary men are not as powerful as shamans in this respect, but they are all capable of minor shamanic acts and of soul-change during ritual. Thus, male use of tobacco is associated with the shamanic ability of men as opposed to women. The shaman is concerned with the hot natural activity of the ancestral world rather than its cold 'dead' aspect, which is contacted through ritual alone. He refers to the mythical events about killing, snakes, sexual intercourse, poison and other dangers which are an ever-present influence on the processes of human life. However, tobacco is associated with cooling, harnessing and controlling these forces, while pepper represents unrestrained natural activity. If tobacco is linked to the insubstantial soul, pepper is definitely linked to the substantial body, as its association with menstruation and growth of the foetus would suggest. We were warned not to eat pepper before going in aeroplanes for fear of being too heavy. Therefore, together, pepper and tobacco may be said to represent the dichotomy of female body controlled by natural forces and male soul which controls natural forces. The body/soul distinction embodied in the relationship between pepper and tobacco is an aspect of each individual, independent of descent-group membership, 231

Production and consumption and is thus a relatively 'natural' one. In keeping with this, both substances are items of trade. Since we already know that the house with its resident community represents a living individual, we can also see why pepper and tobacco should be planted after the house is destroyed. These impart a plant-equivalent of the body/soul relationship within the former house and thus mediate between the organised longhouse community and the later encroachment of unrestrained forest growth. These findings can now be brought together as follows:

male drugs

female foods

1 aspects of the descent group

2 link between present individuals and origins of 1

3 aspects of 'natural' individual

insubstantial structure

link with origins of insubstantial descent-group structure soul

(coca)

(yag^)

(tobacco)

substance of members

link with origins of descent-group substance

substantial body

(cassava)

(manioc juice)

(pepper)

Coca and cassava represent the insubstantial and substantial aspects of descent groups. Yage and manioc juice represent the link between present descent-group members and the origins of these insubstantial and substantial aspects of their descent-group status. Tobacco and pepper represent soul and body as aspects of the natural individual. The male products are thus associated with insubstantial phenomena and the female products with substantial ones: this fits their respective associations with change of the soul and change of the body. The sequence, as I have set it out, represents a progression from cold to hot. Within the set of female products the progression: cassava — manioc juice — pepper is closely analogous to the process of producing hiari, or 'pepper pot', in two separate boiling stages (pp. 225f). Alternatively, according to the interpretations above, the progression is from social to natural, from group to individual and from separation to conjunction, for at one extreme groups are separated according to rules of descent, and at the other, body and soul are united together in the natural individual. These are all oppositions which have previously been associated with the one between cold and hot. If we turn from the nature of the plant products to the context of their use, we find that the division into male and female products is 232

Cultivated plants and social models associated with a division into cold and hot, for the male products are used during 'cold' ritual occasions when fire is avoided, while the female products are consumed during 'hot' secular periods when heat is allowed, as shown: Cold (social) Hot (natural) Cold (social) coca yagd tobacco Hot (natural) cassava manioc juice pepper We thus find that the supposedly 'cold' world of He wi contains 'hot' tobacco while the supposedly 'hot' secular world contains 'cold' cassava. If the axes are rewritten as 'social' (cold) and 'natural' (hot) we find again that the supposedly 'social' world of He wi contains 'natural' tobacco while the supposedly 'natural' secular world contains 'social' cassava. In fact, these superficially contradictory phenomena are represented in the qualities ascribed to the products, for we have seen that tobacco is associated with the division of the natural unity of the individual, and that cassava is associated with the substantial or 'natural' aspect of descent groups. We have also seen that tobacco is a substance derived from the Sun's heat and yet used for cooling, and that cassava is a cold substance which gets heated by being dipped into peppery foods. Consideration of these points, which support the structure of the model above, suggests a reformulation of the previous description of He wi as 'social' and 'cold' and of secular life as 'natural' and 'hot'; instead, it would be more accurate to say that during He wi people exercise social control and cool things down, while during secular life they utilise natural forces including heat. This change from adjective to verb is not as trivial as it might seem, for it implies that the positive aspect of He wi, far from being 'cold and social', is actually the natural, hot activity of the ancestral world. Cooling and social control is only necessary in order to avoid the danger of excessively close contact. Similarly, I suggest that there is a sense in which the secular world of human society is really too cold and social; it requires purposeful use of natural forces to heat it up and sustain life. The dipping of sterile, boring cassava into nice, hot, tasty food is the perfect expression of this fact. In the different parts of this chapter we have seen various ways in which processes of production and consumption of various foods and drugs are both directly and metaphorically related to the processes of physical and social reproduction of social groups. Throughout the 233

Production and consumption discussion, it has been apparent that these creative processes are related to one another as wholes, so that the metaphorical relations between them do not simply exist between fixed elements; they exist between the different types of transformation with which the processes are concerned. To give an example, in certain contexts I have said starch represents semen, but I have shown that this symbolic identity is actually based on the analogy between the processes in which semen and starch are involved. Semen is separated from penises of mature men in the same way as starch is separated from fibre and, in turn, insemination belongs in the wider process of reproduction and growth of individuals which is analogous to the manioc process in which starch separation belongs.

234

Concepts of space-time

Introduction The source of the contemporary natural and social order is the ancestral past known through myth. However, it should be clear from the material so far discussed that the ancestral past is also an alternative aspect of the present which can be contacted through shamanism and ritual. In this chapter, the universe is treated as a conceptual construction which contains the activity and power associated With ancestral creation. In order to contact this alternative reality, people must transpose the system of the universe with its creative processes onto the concrete systems which they are able to control, or at least change, through practical action. To do this, they construct their houses to represent the universe. They also conceive of their bodies, their sexual reproductive systems, their natural environment known through direct experience and the structure of their patrilineal groups in such a way that these too correspond to the structure of the universe. Thus, the concrete world is derived from the Imaginary' ancestral world, but it also provides the way to it. Countless examples of the transposability of different systems have already been given; instead of repeating them all here, this final chapter is focused on the more fundamental concepts of space—time upon which such transposition is founded. I start out from a set of statements made by Indians: the house is the universe, with the male door in the position of the Water Door; the house is a man who lives when there are people inside; the Exogamous Group is derived from the body of an ancestral anaconda who has both human and anaconda form; the main river of the earth is the gut of the universe with the ancestral anacondas as intestinal worms; the universe is a womb. These statements suggest the anal235

Concepts of space—time ogous conceptual organisation of five spatial structures: the universe, the house, the anaconda body (with its division into specialist roles), the human body and the womb. To these can be added a sixth — the longhouse setting, which comprises the differentiated space around the longhouse within which the productive and social life of the community is lived. This extra structure is not a bounded entity like the others: it represents the concrete, known natural world for any single longhouse community. It is important to include it, because, whatever people are doing in metaphorical terms, in real terms they are acting within the confines of the longhouse setting and its focal point, the longhouse. Thus, when they identify themselves with ancestors, they are actually putting on feather head-dresses, drinking yage and so on. The products they use come from the various domains of the longhouse setting — the river, the forest and the gardens. None of them originates within the house itself, for this is merely the place of manufacture or preparation. Indians say that the house becomes the universe during ritual: the longhouse setting is crucial in this transformation because it mediates between house and universe, as it is both in contact with the inaccessible parts of the universe through movements of natural species, heavenly bodies, river flow, etc., and also, as a result of human activity, with the life inside the house. The longhouse setting represents an alternative structure to that of the anaconda body. Each is concerned with a different aspect of the relations between the longhouse community and the external world. The longhouse setting is the natural environment within which the community exists, and the anaconda body represents the social descent structure within which the local descent group or agnatic core of the community takes its place. This follows from the fact that the anaconda body is ideally made up of five sibs which are strung along a river in seniority order. It is now possible to arrange all six structures into a hierarchy of scale so that they fit together in chinese-box fashion: womb human body longhouse river territory of universe Exogamous Group or longhouse setting If we temporarily ignore the longhouse setting which is concerned with the non-social base of longhouse life, the series corresponds in a general way to the construction of Tukanoan society which may be 236

Introduction conceived of as a series of 'creative levels', as follows: construction of society

sexual reproduction

corresponding spatial structure

(womb)

individual

| (human body)

local descent group (longhouse)

I Exogamous I Tukanoan | Group | society

I

(river)

I I (universe) I i

The details of the analogies between my six structures are built from a variety of data concerning the movements and changes made within each by people, food, mythical heroes, heavenly bodies etc. Tftus, the structures themselves, conceived of as static physical forms in space, are actually the products of changes in space and time that occur within them. The spatial structures are like empty shells; as physical entities they are built or otherwise created by human, ancestral or natural forces, but as conceptual entities they are maintained and ordered by the processes that occur within them. Thus, the house is built with a male and female door, but the sexual identities of the doors are maintained by the movements of the longhouse residents. The creative character of ongoing activity is recognised in many Indian beliefs: for instance, the house only lives when people are inside it; the sites on the ancestral journey upstream did not exist until the ancestors stopped at them; the rivers did not flow until their beds had been trodden out by the ancestors and so on. Henceforth, I shall refer to the six structures I have isolated as 'spacetime systems' in order to convey the idea that the movements they contain are all-important. Each of these space—time systems has three spatial dimensions, but in order to simplify the analysis I shall deal first with the horizontal planes or ground-plans, and then with the vertical axes of certain systems, particularly concentrating on the universe. From the Indian point of view, this is working backwards, because the present order, which is represented on the horizontal plane of the earth, is derived from the vertical order of the universe as a whole. However, precisely because life is mainly lived out on the earth's surface, the horizontal order of the various systems I discuss is more clearly and consistently differentiated than the vertical order (see pp. 2570- this is why it is best described first. In can then be shown that the horizontal and vertical orders of certain systems are transformations of one another. For the purposes of this argument, the systems may be divided into 'movable' and 'immovable' as in fig. 35: in the movable 237

Concepts of space—time systems, the human body, the anaconda body and the womb — the principal horizontal axis (head to toe of the prostrate human body, head to tail of the anaconda and recess to opening of the womb) — may be rotated to a vertical position, just as the human being can lie down or stand up. This transformation of a horizontal axis to a vertical one provides the key to the relationship between the horizontal and vertical planes of the immovable systems. These immovable systems are the universe, the longhouse setting and the house, all of which have their central horizontal plane fixed, since it is provided by the earth's surface. These points will be easier to appreciate after the analysis of the horizontal planes. Horizontal space—time

Description of horizontal systems The earth. The earth is round and is structured by the river system MOVABLE SYSTEMS HUMAN BODY

ANACONDA BODY

WOMB

Ox

O

HEAD

SERVANT*

IMMOVABLE SYSTEMS HOUSE

LONGHOUSE SETTING

UNIVERSE SK Y

HIGH EST KIATUR AL LIFE

ROOF

\

/ / EARTH'S

|

(J

FOREST

RIVER

\ y'jEAST

WEST

SURFACE

/

. --^ G"R*

LOAI'EST NATURE«>L LIFE UNDERWORLD

Fig. 35 Comparison of movable and immovable systems of space—time 238

Horizontal space—time draining into the Milk River, as shown in fig. 36. The rivers of this system have their headwaters in the west and culminate in a single flow at the Water Door in the east or, in Indian terms, an upstream journey corresponds to the daytime path of the sun. Besides the Water Door in the east there is a Back Door {Hudo Sohe) in the west, and two side doors, or Rib Doors (WarubuSohe, sing.), to the north and south. These doors are important as the exit points for dangerous forces dispersed by shamanism. According to some accounts, there is also a river encircling the earth's perimeter. For the moment, let us imagine a single river bisecting this earth. This river provides a continuum between pure forest and river mouth, or land and water, the river itself becoming wider and more differentiated from forest as it runs downstream. The Tukanoan groups of the Vaupes are river people, bbth ideologically and by virtue of economic emphasis. Longhouses are built near rivers; communication is mainly by river; fish is the principal source of protein. Towards the headwaters the availability of fish gradually declines. The waterfalls and rapids act as natural barriers to many species of fish, particularly the larger ones, so that those who live in the headwaters are described as 'eaters of tiny fish'. The headwaters people are forced to rely more on forest products, not necessarily because these are more abundant, but because the fish are scarce. Furthermore, there is an ideological link between downstream location and cultivation, since crops were originally brought upstream from Fish Anaconda's underwater house. Political relations between longhouse communities, being largely determined by exchange of invitations to rituals, are dependent on a surplus of manioc and coca. Thus cultivation has a direct bearing on political dominance. Although this association between downstream location and superior cultivations linked with political dominance is neither proved nor disproved by Pira-parana patterns, it is clearly described for the Cubeo (Goldman 1963). Also, in the Vaupes area as a whole it is notable that the large groups, such as the Tukano and Tariana, occupy the downstream regions. In the past, these groups dominated the westerly, upstream groups, such as the Bara, Tuyuka and those on the Pira-parana. If these differences between upstream and downstream location serve to differentiate the Tukanoan populations which live in the middle of the earth, they are also associated with different types of supernatural power located beyond the people in the extreme east and west. The mouth of the Milk River is the source of origin of 239

N

WB DOOR

W

EVIL FOR.EST a deification of preuntordtr SPIRITS * . - j w v n w . ^^ soc*»>_

Fig. 36 The earth's surface

240

,

MILK RIVER development of Ti/I

^

Horizontal space—time human culture, from whence the ancestors and all aspects of culture came, while the headwater forests are the home of cannibalistic spirits in which normal bodily characteristics and processes are reversed. Both ancestors and forest spirits are dangerous, but the former are essential for the well-being of society, while the latter are unambiguously threatening. The ancestors represent the pre-social origins of society, and the forest spirits represent the realm outside present-day social control: while the ancestors are 'before' in time, the forest spirits are 'beyond' in space. Indians insist that they live in the middle of the world between these opposed supernatural powers. Their land is hot from the overhead sun, whereas that of the whites, which is on the periphery of the earth beyond a circle of mountains, is cold.1 Now that the associations of the east—west axis of the earth have been set out, it remains to adapt the model to take into account the so far neglected natural shape of a river system. The headwaters are spread out and the successive confluence brings them into a common flow. This means that the mouth has a single location in the east, while the headwaters form an arc on the western periphery, so that the qualities associated with the west are spread, as is shown in fig. 37A. According to this model the journey upstream, whatever its actual direction, has the characteristics of the dominant east—west axis described above. The fanned river system, it will be noticed, leaves the north east and south east quadrants of the circular earth empty of significance — this corresponds to the lack of any mention of them in Indian ideology.2 The anaconda body. The association between the seniority order among sibs and their position in the body of the Exogamous Group anaconda ancestor, as well as the ideal correspondence between downstream residence and seniority, have already been discussed (pp. 24, 29). I also mentioned that the servant sibs are modelled upon the true Makii who live in the extreme headwaters (interfluvial areas) and who are believed to have no rules of exogamy, no houses and a bare minimum of artifacts. Both in location and life-style, the 1 The model is adapted to local features, so that in any particular region the dangerous mythical spirits are said to live in specific watershed areas and the ancestors are said to have come from downstream. 2 Although I never obtained a drawing of the earth as a whole, my models may be compared with the Tatuyo drawing in Bidou (1972 : 80).

241

B CREATION OF DESCENT GROUPS

RIVER SYSTEM OF THE EARTH

FORE ST ^ ^fflggS.-,...-—.;- 112r .. -....„

^^tSJ^^S^^'^rs ^..g^^^^^^liWA^Tg

r

) |

SHAMAH5 SHAMAN

WARRIOR

[ DA MC 6ft/CHANTfftSJ C HI g *S

SPIRITS C

SINGLE-ANACONDA MODEL OF EARTH'S RIVER

D

DOUBLE-ANACONDA MODEL OF EARTH'S

|

ARRIVAL OF ANCESTRAL ANACONDAS

Fig. 37 Alternative models ofriversystem of earth 242

^^^^^^^^^MJKPBW

Horizontal space—time Makii mediate between the servant sibs of the Tukanoans and the forest animals themselves in their 'animal houses' (p. 61); this position is borne out in their role as suppliers of meat and forest products to Tukanoan groups. There is therefore a series of relationships associated with position along a river in which the river and anaconda body are treated as analogous systems. Superimposing this anaconda body model upon the main river of the earth, we arrive at fig. 37C. The statement that this river is the gut of the universe with anacondas for intestinal worms, and the association between gut and anaconda skin in the form of the tipiti (pp. 187f) support this model of the earth's east—west axis. Conversely, the Exogamous Groups originated from the east in the form of anacondas and swam towards the west with the last-born sibs situated in the tails. This order was maintained until the anacondas reached the centre of the earth, at which point the sibs were distributed in the reverse order, with the last born furthest upstream. The whole process may be represented by a double anaconda model of the earth's river (fig. 37D). The model expresses the paradoxical position of the first-born as both oldest, and therefore furthest from the position of origin, and also most powerful and therefore most closely associated with ancestral origins. If this double anaconda model is superimposed on the fanned river system of the earth in such a way that the separate Exogamous Groups are dispersed at the centre, each choosing its affluent (which is how Indians describe the origin process), we arrive at fig. 37B, a model of the earth which contains the totality of Tukanoan society. Comparison of the single and double models of the earth in terms of anaconda bodies shows that the main river may be described as either a single series of hierarchically arranged terms, or a double series of terms arranged symmetrically about the centre. I showed previously that there are aspects of bilateral symmetry in the ordering of specialist roles arising from their membership in three separate domains. The incorporation of both hierarchical and symmetrical (concentric) principles of organisation into the ordering of sibs lends itself to expression in these alternative models (see pp. 56f)« The longhouse setting. Here, it is important to distinguish between the actual situation of the various domains of the longhouse setting, which is extremely variable and subject to a host of natural and social factors, and the conceptual model presented below. In this model, the 243

Concepts of space—time differentiated domains are ordered in such a way as to bring concrete human activity into line with the Imaginary' system of the universe. In order to develop this conceptual model, reference is made to the description on pp. 43ff, summarised in fig. 4. The nature of the river on which the longhouse has its main port has been discussed at various points and may be summed up as follows: it is the 'ancestral water path'; it is an umbilical cord which connects the souls (the longhouse communities) to the mouth of the Milk River; it is the branching stem of the kana plant; it is the yage vine. All these images treat the river as a spiritual power supply connecting the longhouse community with the ancestral powers in the east. This means that, understood from the point of view of a single community, there is a sense in which the river stops at the longhouse port, for that is where the power reaches the community. If the model of a typical longhouse setting, fig. 38 A, is altered to accommodate the notion of a nourishing river, it appears as in fig. 38B. The latter model resolves the apparent inconsistency between the main axis of the house and the direction of flow of the river. Whilst in reality these two are often at right angles, Indians claim that the house faces east and that rivers flow east. This point brings out the special qualities of the metaphorical representations of the river: the twisting yage vine, the branched kana stem and the umbilical cord are all conceived of as 'paths' {ma). The special characteristic of a path is that although it may twist and turn, it leads to a given point — in this case to the source of life and growth. These path metaphors are thus able to relate positions that are only conceptually east to the true east where the sun rises. If the conceptual setting of the longhouse is compared to the river system flowing across the earth, the common structure of the two is clear, for the river, a single point, is opposed to the forest, an arc, in either case. In both systems, social life exists between the two. It is important to stress that this conceptual model is mine. Although it has heuristic value here, I believe that it is recjundant from an Indian point of view, because the flow of water and the movements of natural species along or above the ground connect the river and forest of their experience to the extremes of the earth in a direct manner. While, from my analytic point of view, the universe is an 'imaginary' extension of the known world, from the Indian point of view the known world is determined by the structure of the universe and is also supplied from its extremes. This direct relationship 244

A

LONGHOUSE SETTING

B CONCEPTUAL LONGHOUSE SETTING

Fig. 38 Comparison of typical longhouse setting with conceptual longhouse setting 245

Concepts of space—time between the outer domains of the longhouse setting and the extremes of the earth itself perhaps explains why the longhouse setting does not appear in the set of analogies drawn from Indian statements (pp. 235f). The longhouse interior. Here, I refer back to the description on pp. 46—9 and to fig. 5, which is reproduced in simplified form as fig. 39A. Inside the house, the social connotations of space are best understood as the outcome of two principles of organisation working in combination. There is a linear male—female axis running between the two doors, and also a concentric pattern in which the periphery represents private, family life and the centre represents public, communal life. These principles may be deduced from the fixed positions of male and female doors and family compartments, and from the daily and ritual use of house space (many examples of which are contained in previous sections). The principles are not independent of one another, for in concrete situations there are associations between male activity and communality on the one hand, and female activity and domestic privacy on the other. 3 While these associations are objective features of socio-economic and ritual organisation, since, on the whole, men do things together while a woman does things on her own or with her daughters, they are also features of Pira-parana social structure with its solidary local descent groups and isolated unmarried wives. In spatial terms, these associations mean that the centre of the house is identified with the male end while the periphery is identified with the female end. This is manifest in the position of the family compartments which are clustered around the female end, but it is made even more obvious during He wi and He rika soria wi. During these rituals, the men conduct their communal activities in the centre and front of the house, shut off from the women by a screen erected across the middle of the rear end (see fig. 39A). The women and children, confined to their family compartments and the small communal cooking area, may only communicate with one another via the plaza. 3 A third principle of organisation, which I do not explore further, differentiates the leftand right-hand sides of the house. The right contains the box of ritual ornaments, the headman's compartment and the men's coca, stools etc., while the left contains the beer canoe and the( post topped with tree-resin that is burnt for night-time light. The resin is metaphorically linked to hot female sexuality since 'fetching resin' (guhe umagu) is a joking reference to sexual intercourse or to fetching a new wife. The positions of the various objects support a general association of the left-right organisation with the previous two, thus: right : left :: male : female :: communal : private.

246

A HOUSE VISITORS

FAMILY COMPARTMENTS ; • •

•

•

• • • •

•

•

•

•

•

•

WOMEN'S WOMEN'S DOOR

AREA

C.RC?E

FAMIUY COMPARTMENTS:

B

HOUSE AS BODY

'S

DOOR

VISITORS

---VeRseo TO BODY

O

FAECES

VOMIT FOOD BREATH fi" BREATH

i leo

C

TO BODY

HOUSE AS WOMB

Fig. 39 Models of house as body and womb 247

^

Concepts of space-time The body. The house is a person, a mythical figure called Roofing Father (Muhi Haku), who is an enormous bird 'the size of a house' with a tapir's head. The different species of palm leaf used for roofing grow as feathers from the various parts of his body. His head is at the male end of the house, his anus at the female end, and the sides of the roof are his rib-cage. The association of the male-female axis of the house with the digestive tract is suggested by analogy in the statement that the main river is the gut of the universe, for we know that the house is the universe. It is also reflected in the use of space. First, the front of the house is kept painted and clean like the face. Speech, a function of the mouth related to breathing and patrilineal descent (since languages are descent-group property), is clearly associated with the male end of the house. The men's circle is towards the male end, the ritualised version of this — the chanting group - is placed just inside the men's door and visitors are greeted either on the plaza or just inside the men's door. Vomiting, induced by drinking yage during ritual, also takes place just by the men's door and, appropriately, the yage/vomit is an anaconda which is simultaneously swimming in the river of the universe, the gut of the drinker and the front part of the house (pp. 209f). On one hand, drinking yage and vomiting, listening to speech and answering and breathing in and out are all two-way processes involving entrance and exit through the orifices of the head, appropriately associated with the comings and goings through the men's door and ritual processes of patrilineal regeneration. Eating and digesting food, on the other hand, is an irreversible process producing both nourishment, which is absorbed from the gut, and rotten waste that is eliminated at the anus. In keeping with this, the plaza outside the women's door, narrow and littered with debris, is used for intimate family activities and women's crafts.4 The occupants of separate family compartments convene in the centre of the house to receive nourishment from communal meals and then disperse to resume productive activity. This dispersal of the nourished community members may be likened to the dispersal of nourishment within the body from the centre of the digestive system to the peripheral tissues (see the theory of digestion p. 186). Although the community normally produces its own food, thus receiving it from 'inside', when it receives food from outside communities in 4 In the Cubeo mourning ceremony the masked figures representing dung-beetles are stationed at the female door (Goldman 1963 : 233). 248

Horizontal space—time ritual exchange the gifts enter through the men's door or 'mouth'. In fig. 39B, the house is represented as a body with a twisting gut, drawn in the way that Indians draw the gut inside the outline of a person. The womb. The image of the house as womb equates the men's door with the vagina and therefore with the passage from womb to outside world. This fits the use of the men's door as entrance and exit for journeys between longhouse communities. The women's door, conversely, is an 'internal' passageway to the manioc gardens, the nourishing base of the home community (see fig. 39C). The spiritual power inherent in the river contrasts with the power of the gardens to nourish the bodies of the community members in the same way as semen and womb or male and female reproductive powers. Besides the 'pure' cold manioc which enters by the female door, there is a 'hot' lining to the womb (house) made up of the cooking fires in the family compartments, and thus both aspects of female sexuality contribute to the foetus (growing community). Overall, food from the western periphery combines with ritual life supplied from the east to create growth. The correspondence of the house and womb, as I have just described it, is shown in fig. 39C. The analogy between universe and womb may be described in the same terms, with the Milk River representing a pathway into the womb up which the inseminating anacondas swim. The conjunction of blood and semen or forest and river occurs in the middle of the womb/earth when the patrilineal descent groups finally emerge onto land as in fig. 40A. However, there is also reason to conceive of the upriver anaconda journey as a birth process — it is literally the birth ('waking up') of Exogamous Groups, the river is an umbilical cord and Romi Kumu, the female ancestral source of humanity, is located in the far east. These facts suggest that it would be appropriate to give the womb image an about-turn, so that the passage from the east to the centre of the earth represents the birth of society — the passage from the 'pre-descent' to the 'descent' era. Returning to the birth of the individual (pp. 123ff), which is necessarily located within the longhouse-setting system, we may recall that the natural, female birth in the manioc garden is contrasted to the later 'rebirth' from the river. The two births are made analogous by the identification of the womb with the ancestral waking-up houses, of the river with an umbilical cord and so on. In spite of this, they were opposite in 249

Concepts of space—time

A SINGLE MODEL

RIVER/WATER.

W

B DOUBLE MODEL

W

Fig. 40 Horizontal models of universe as womb 250

Horizontal space—time direction and concerned with the natural role of the mother and social role of the father respectively. Since the longhouse-setting system is analogous to the system of the universe, we may transfer the movements associated with birth to the ground plan of the earth, thus producing a double womb image of the earth (fig. 40B) similar to the double anaconda image in fig. 37D. Just as the double and single anaconda models were consistent with the alternative hierarchical and concentric principles of organisation of the series of specialist roles, the double and single womb models are consistent with the alternative ways of viewing the relations between the sexes. 'Male' and 'female' are equal but opposite natural powers or, alternatively, they are contrasted as 'good, social creativity' and 'dangerous, natural creativity' (pp. 158f). Similarly, the east—west axis of the earth has aspects of symmetry about the centre and aspects of negative progression from east to west. Synthesis of horizontal systems Five of the six space—time systems are set out in fig. 41. The anaconda body is missing, because it is linear in physical form rather than rounded: I return to it below. The models in fig. 41 are constructed according to the hypothesis that each is composed of a linear and concentric ordering of space-time working in combination. In fig. 41(1), the shapes generated by the combination of a linear and concentric order provide a basic model which adequately describes the concentric organisation of the 'western' half of each system and the linear order of the 'eastern' half. The existence of these two types of organisation is most obvious in the house system, but the similarity in shape and analogy between the positions and processes belonging to the separate systems suggests that they conform to the same basic model. It should be clear from fig. 41 that the adequacy of the basic model depends upon there being some correspondence between the 'western' half of the linear axis and the radius of the concentric order. The comparison of the values associated with C and C' and again with B and B' shows that such correspondence does exist. Sometimes there is perfect identity and sometimes, as in the house and the body, the west of the linear axis is differentiated from the periphery of the concentric order, in spite of being like it in some respects. In the house, for instance, the female door is replicated in the peripheral doors and there is a strong association between women and family life, although 251

SYSTEM [proc€<s

CONCENTRIC ORDER

LINEAR ORDER

COMBINED ORDERS

j

1 BASIC MODEL

-