How Kinship Systems Change: On the Dialectics of Practice and Classification 9781800731677

Table of contents :
Contents
Illustrations
Introduction
Part I Terminological Change
1 Kinship as Classification: Towards a Paradigm of Change
2 Terminology and Alliance in India: Tribal Systems and the North-South Problem
3 From Tetradic Society to Dispersed Alliance
4 Why do Societies Abandon Cross-Cousin Marriage?
5 Dravidian and Iroquois in South Asia
6 Indo-European Kinship Terminologies in Europe: Trajectories of Change
Part II Crow-Omaha
7 On the Origin of Crow-Omaha Terminologies
8 Substitutability of Kin and the Crow-Omaha Problem
9 The Evolution of Kinship Terminologies: Non-prescriptive Forms of Asymmetric Alliance in Indonesia
Conclusion
Appendix: Publications on Kinship by Robert Parkin
Glossary
Index

Citation preview

HOW KINSHIP SYSTEMS CHANGE

HOW KINSHIP SYSTEMS CHANGE On the Dialectics of Practice and Classification

 Robert Parkin

berghahn NEW YORK • OXFORD www.berghahnbooks.com

First published in 2021 by Berghahn Books www.berghahnbooks.com © 2021 Robert Parkin

All rights reserved. Except for the quotation of short passages for the purposes of criticism and review, no part of this book may be reproduced in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system now known or to be invented, without written permission of the publisher.

Library of Congress Cataloging-in-Publication Data A C.I.P. cataloging record is available from the Library of Congress. Library of Congress Cataloging in Publication Control Number: 2021020413

British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library

ISBN 978-1-80073-166-0 hardback ISBN 978-1-80073-167-7 ebook

To N.J. ‘Nick’ Allen (1939–2020) In memoriam



Contents

List of Illustrations

ix

Introduction

1

Part I. Terminological Change Chapter 1 Kinship as Classification: Towards a Paradigm of Change

27

Chapter 2 Terminology and Alliance in India: Tribal Systems and the North-South Problem

61

Chapter 3 From Tetradic Society to Dispersed Alliance

81

Chapter 4 Why do Societies Abandon Cross-Cousin Marriage?

96

Chapter 5 Dravidian and Iroquois in South Asia

133

Chapter 6 Indo-European Kinship Terminologies in Europe: Trajectories of Change

159

Part II. Crow-Omaha Chapter 7 On the Origin of Crow-Omaha Terminologies

197

Chapter 8 Substitutability of Kin and the Crow-Omaha Problem

227

Chapter 9 The Evolution of Kinship Terminologies: Non-Prescriptive Forms of Asymmetric Alliance in Indonesia

260

viii

Contents

Conclusion

280

Glossary

287

Appendix Index

Publications on Kinship by Robert Parkin

291 295



Illustrations

Figures Provided by author 0.1 Symmetric or bilateral prescription 0.2 Asymmetric prescription (matrilateral ms, patrilateral ws) 0.3 Crow-type terminology 0.4 Omaha-type terminology 0.5 GEG marriage (symmetric) 0.6 GEG marriage (asymmetric) 2.1 GEG marriage (asymmetric) 2.2 GEG marriage (symmetric)

17 18 19 19 19 20 66 67

Tables Provided by author except where stated otherwise 5.1 Burushaski kinship terminology. Source: Ali (1983) 5.2 Hindi kinship terminology as example of north Indian pattern. Source: Vatuk (1969) 5.3 Kumaon kinship terminology. Source: Klingel (1989: 196–97) 5.4 Malpahariya kin terms (+1 consanguines). Source: Sarkar (1937) 5.5 Trajectories of change in South Asian kinship terminologies

136 142 149 151 152



Introduction

Overview of the Collection This volume brings together a number of articles of mine on the topic of kinship, all but one of which has been published previously. Although they are divided into two groups (Parts I and II), with one exception (Ch. 8), the chapters in both parts are all concerned with changes to kinship systems. In Part I, this applies especially to transitions between different recognized terminological types, though changes to marriage practices are also taken into account, especially in Ch. 4. More particularly, the chapters in Part I are mainly concerned to track possible trajectories of change between prescriptive terminologies (those connected with cross-cousin marriage, often called ‘Dravidian’ [symmetric] or Kachin [asymmetric], depending on type) and those terminologies called by Robert Lowie ‘lineal’ and by Rodney Needham ‘cognatic’ (sometimes somewhat misleadingly called ‘Eskimo’, a category to which the English terminology is often said to belong).1 The chapters in Part II focus on what I have elsewhere called CrowOmaha ‘pseudo-systems’ (Parkin 1997a: 109–17) in respect of both the possible derivation of such terminologies from prescriptive ones (Ch. 7) and their possible sociological correlates, especially in terms of relations between spouse-exchange groups over time (Ch. 8; this is therefore the exception I mentioned earlier, in that it is not specifically about change).2 Chapter 9 applies the insights of Chapter 8 to marriage practices in Southeast Asia that resemble Crow-Omaha pseudo-systems in this respect but lack the associated terminologies. The Appendix provides a complete bibliography of my published works on kinship at the time of writing (autumn 2020).

2

How Kinship Systems Change

Most of the articles are reproduced here in approximately their original form, the main exception being Chapter 3, which is a partial reproduction alone. Most of the articles therefore retain their original integrity, but they have been placed in an order that tells a story, so to speak, or rather two stories, one for each part; obviously, this is also the reason for their selection. Hopefully, therefore, the volume will strike the reader not as a conventional collection of randomly or ill-connected past articles, but more as an organic whole, or two such wholes, both which say something about the way kinship systems change and their possible reasons for doing so. Naturally, the close relationship of many of these chapters to one another means that there will inevitably be some repetition between them. While I have acted to reduce some of that repetition, in other cases I have chosen to leave it, not least because that may help reinforce the arguments I am trying to impress on the reader. While I have also tried to explain sometimes difficult points clearly, the reader will certainly benefit from having some prior technical knowledge and competence in matters relating to kinship. Consulting introductory books on kinship – especially, in this context, my own (Parkin 1997a) – may prove helpful in this regard. Diagrams of ‘typical’ kinship systems are also provided towards the end of this chapter. The age of many of the sources I use should also be acknowledged. This is partly explained by the fact that up-to-date work on the more technical side of kinship studies – though now flourishing again, particularly through the Kinship Circle run at the University of California by Dwight Read and Fadwa el Guindi – fell out of favour for several decades following the criticisms of it by David Schneider and the shift away from structuralism. Thus, in very many cases, older texts are all we have in the absence of more recent research. However, my use of older sources can also be justified in the present context because my interest is precisely in how kinship systems, and especially their terminologies, change over time. Comparing sources from different periods may itself be a source of evidence for change.

Evolution, Change and History One potential criticism of such work relates precisely to its use, now and then, of ‘evolution’ as both a term and a concept. In both senses, evolutionary perspectives have come to be associated closely

Introduction

with nineteenth-century theories using the notion of evolution to make now unacceptable discriminations between supposedly civilized cosmopolitans in white male-dominated Europe and politically subjected ‘native’ populations elsewhere in the world, as well as justifying colonial rule by the former over the latter.3 However, the functionalist doctrines of Malinowski and Radcliffe-Brown, for whom evolution was simply speculative history and of no use to their more synchronic, presentist, supposedly scientific ways of doing anthropology, meant that evolution as an idea suffered an eclipse in social and cultural anthropology after the turn of the twentieth century. This revolution was not complete, as in America evolutionism survived the attacks of the Boasians on Morgan in the work of neo-evolutionists like Leslie White, Julian Steward, Robert Carneiro, Elman Service and the early Marshall Sahlins. In Oxford, the late evolutionist Robert Marett was responsible for teaching social anthropology until the arrival of the functionalist Radcliffe-Brown as the University’s first professor of anthropology from 1936 to 1946, long after evolution had given way to the latter’s and Malinowski’s respective versions of functionalism. In the post-WWII period, moreover, evolutionary ideas were invoked in some of the work on kinship of Rodney Needham and especially N.J. Allen, originator of tetradic theory, mentioned in a number of the chapters in this collection. I myself have been profoundly influenced by the work of the latter two scholars, having been trained at Oxford at the time they were both active there.4 Evolutionism has also survived, of course, and without a break of the sort described above, in biological or physical anthropology. Indeed, it has quite explicitly reinvented itself in some circles, including at Oxford, as evolutionary anthropology. Whatever label one chooses, the facts and ideas of evolution are absolutely central to physical or biological anthropology, unlike social or cultural anthropology, with its more present-oriented emphasis, for which a consideration of history is not always required. However, for evolutionary anthropology and its biologically oriented predecessors, as well as for much archaeology, evolution has always been primarily a matter of transitions between different hominid species, especially after their splitting from the great apes, culminating in homo sapiens, the sole surviving non-ape hominid species at the present day. Even the long period of the existence of homo sapiens on the planet is far more than the evolutionary range my own work covers or needs to consider,



3

4

How Kinship Systems Change

mainly because the study of kinship in the social sense of the word ultimately requires living informants who can be asked questions, not material evidence from millennia ago. I therefore find it more appropriate to ignore this earlier history and prehistory in order to think more in terms of ‘cognitively modern humans’ and to stress that whatever changes in kinship systems social anthropologists talk about they are restricted to the latter, contemporary phase of human history in which such humans have existed.5 In any case, neither earlier hominid species nor early homo sapiens are around any longer to be asked about their kinship systems, though presumably the latter, at least, had them in something like the modern sense. While later hominid populations, like homo sapiens, almost certainly had some form of language, in the absence of writing they left no written records (ignoring cave paintings and the like), whether on their ideas and practices of kinship or on anything else. In fact, actual records useful to the history of kinship only go back perhaps two to two and a half thousand years at most in Europe, to Greek and Roman antiquity, and in the New World barely to the seventeenth century in Spanish dictionaries and certain other texts dating from that time. The developments discussed in this collection are therefore firmly rooted in what is very recent history in terms of geological time and not in prehistory as conventionally defined at all. They also assume that humanity is a single species with a single shared origin in Africa, not a species with multiple origins, which classic social evolutionism and more explicit right-wing political agendas have frequently argued gave rise to different races. Such arguments as the latter are another demonstration of the continuing ability of evolutionary ideas to be diverted to political ends and of the continuing need of academia to interrogate and, where necessary, combat them. Whereas the firm rejection of evolutionism by social and cultural anthropologists like Malinowski and RadcliffeBrown was largely intellectual, for Franz Boas in America it also involved a moral rejection of evolutionism’s discriminatory tendencies. Among other things, this established firmly the principle of cultural relativism, namely that any society should be examined in its own terms and not with reference to a hegemonic, standard, cosmopolitan view of humanity and human society, nor to discriminatory, often politicized agendas like those associated with classic (i.e. nineteenthcentury) social evolutionism. I hereby affirm that I adhere to this principle as well, despite my occasional use of this problematic word.

Introduction

If I use ‘evolution’ and its derivatives at all, it is in the context of describing and trying to account for social change alone, not to predict social or cultural improvement by supposedly lower levels of humanity, much less alleged races, as in classic nineteenth-century social evolutionism. Another objection has been raised recently by Bradley Ensor (2019;6 2017: 4–6), namely that studies such as mine tend to treat human populations, quite unrealistically, as bounded societies, each with its own language and culture. Furthermore, for Ensor, while those perpetuating such studies may well lack the discriminatory attitudes of nineteenth-century evolutionists, their perspectives are still what he calls ‘anachronistic’ and damaging to political efforts to produce a more equal and inclusive sense of humanity by sustaining a world view that tends to essentialize human populations and often treats them accordingly in discriminatory ways. In acknowledging this susceptibility, however unintended, I would agree with Ensor that social reality is more complicated than essentializing descriptions suggest, that many populations or local communities are historically of mixed origin, and that they may accordingly be culturally varied and linguistically polyglot. Indeed, this has been acknowledged for decades in anthropology, though admittedly not always problematized, at least not until the arrival of figures like Fredrik Barth and his work on the Pathan (e.g. 1969; cf. Evans-Pritchard’s bland observations [1940] that many Nuer are descended from captured Dinka slaves). However, kinship terminologies are rooted in language, and almost always in one specific language, without that making it impossible for them to borrow terms from one another. I therefore suggest that the term ‘speech community’ might be more appropriate as a way to avoid this difficulty, though as this has admittedly been a quite recent epiphany on my part, my use of it in this volume is not at all consistent. Nonetheless, whenever the word ‘society’ crops up in this volume, the alternative phrase ‘speech community’ may actually be more appropriate. One practice I am sticking to, however, is use of the equally contested term ‘tribe’ when writing about India, where it is an official legal and administrative designation for certain population groups (distinct from ‘caste’), and is thus hard to avoid, as well as approximating to many indigenous ‘tribal’ identities. This is also a suitable place to point out that, within speech communities or other relevant population groups, spouse-exchange groups are not necessarily



5

6

How Kinship Systems Change

identifiable as something else, like a lineage or clan, though they may be territorial, like a village.7 Frequently, however, they can only be defined in terms of themselves. The fact that kinship terminologies are linguistic is significant in another sense. Whatever objections might be raised to the treatment of kinship in evolutionary or developmental terms, there can be no doubt that languages change, including in how they form the basis of semantic classifications, and therefore that kinship terminologies do so as well. We also know this from other semantic domains, such as colour terms (Berlin and Kay 1969) and life-form terms (Brown 1984).

Technical Stuff Throughout the volume, I distinguish emic or indigenous kin categories and the terms that identify them from etic or analytical kin types: for example, English has the term uncle as a native category with its own term, but this can be divided analytically into the kin types father’s brother, mother’s brother, and often aunt’s husband (father’s sister’s husband, mother’s sister’s husband) as well. In many other speech communities, such distinctions may be emically much more crucial, such as the practice of distinguishing mother’s brothers from father’s brothers using different kin terms. For example, this may happen with cross-cousin marriage, where a man is expected to marry the daughter of the former, but must strictly avoid the latter, relations with whom will probably be considered incestuous; this differentiates male ego’s relations with these two uncles accordingly. Thus kin types are useful if not essential in comparing different terminologies and the category-terms that inform them, but they are to be treated as analytical rather than indigenous categories.8 The following abbreviations are used throughout for both kin types (etic) and categories (emic), combined as necessary: F = father, M = mother, B = brother, Z = sister, S = son, D = daughter, H = husband, W = wife, ms = man-speaking, ws = woman-speaking, ss = same-sex, os = opposite-sex, e = elder, y = younger. Where any of the last six formulae occur at the end of the abbreviation, it denotes that they apply to the whole of that abbreviation and not just one of the individual kin types found within it, as would otherwise be the case. A frequently found example of these abbreviations in combination is MBD, ‘mother’s brother’s daughter’; another example is

Introduction

FB, father’s brother. As is conventional, ‘ego’ is used to identify the person from whose perspective a kinship system or an aspect of it is being described. ‘Alter’ is sometimes used to specify the individual with whom ego is in a relationship for whatever purpose – that is, ego’s opposite number in the relationship. Generations are also identified here with reference to ego’s generation (sometimes called ‘the level of reference’, e.g. by Needham), while generations on either side of ego’s are identified by ‘+’ and a number if ascending and ‘−’ and a number if descending: thus +2 refers to the grandparents’ generation, +1 parents’, −1 children’s and −2 grandchildren’s.9 Sometimes the term ‘genealogical level’ or just ‘level’ may appear. While ‘generation’ implies similarity of age, ‘genealogical level’ suggests actual genealogical relations, regardless of age. This distinction takes account of the fact that, for example, ego’s uncles and aunts might be younger than ego (if they were born after ego), or that a large family of siblings may encompass a larger age range than the typical generation lasting a notional 25 years or so. The term ‘classificatory’ should also be noted, as it is routinely used in discussions of kinship to denote that a kin term translated with reference to the closest genealogical position to ego it covers also refers to remoter relatives who are seen as equivalent to that position by the speech community concerned. An example is when the term for FB also covers FFBS, FFFBSS etc.; another is terms for cousins like MBD (above) applying beyond the genealogically minimal first-cousin range. Genealogically speaking, invoking the classificatory idea means extending the range of genealogical connections to include remoter ancestral generations; thus, while FB is a father’s brother traced back to the +1 generation, FFBS is a paternal grandfather’s nephew traced right back through the +2 generation, though both FB and FFBS are +1 relatives to ego. For diagrams, see the last section of this chapter, just before the references.

Prescription In light of past controversies regarding the definition of prescriptive systems, mostly associated with forms of cross-cousin marriage, I should make my own position clear regarding what I mean by ‘prescription’, ‘prescriptive’ and ‘prescriptive alliance.’10 These terms all



7

8

How Kinship Systems Change

literally suggest a situation in which all egos are expected to marry specific relatives, usually interpreted as types of cousin in the first instance. It is therefore tempting to call such marriages ‘prescribed’, and indeed that was the early practice of some authors dating back at least to Alfred Kroeber and Robert Lowie, both writing in 1917 (Kroeber 1917: 384; Lowie 1917: 172; cf. Needham 1973). However, given arguments that, even where such rules exist, not all egos actually follow them, the objection repeatedly surfaced that ‘preference’ would be a better way of describing what was at issue than ‘prescription’. In any case, it was argued further the rule could not be followed one hundred per cent for practical reasons: not all male egos have MBDs, for example (an argument that depended on taking ‘MBD/FZS’ marriage literally as involving the minimal genealogically defined referents alone). One other aspect of this was that ‘preference’ and ‘prescription’ were opposed to one another: some societies had the former, others the latter. One authority who took this line was Maybury-Lewis, in a much cited article (1965: 225–27). Ultimately, Needham (1973), who was at the centre of a good many of these controversies, suggested that prescription is best seen as a property not of affinal alliance or marriage practices but of the kinship terminology – that is, of those terminological patterns that have equations and distinctions expressing the operation of these marriage rules. Therefore, one should speak of prescriptive kinship terminologies (or ‘relationship terminologies’ in Needham’s language), not prescriptive alliance, although the cumbersome phrase ‘relationship terminology of asymmetric prescriptive alliance’ was apparently admissible (it appears often in the writings of this school, and Needham himself was certainly not consistent in his usage over time). Fundamentally, however, in the case of asymmetric prescription (i.e. between classificatory MBD and FZS), one would expect to see terminological equations and distinctions like MBD = W, MB = WF, MBS = WB, ZS = DH, MB ≠ FZH, FZ ≠ MBW, MBD ≠ FZD, WB ≠ ZH etc., underpinning a rule or norm – indeed, an ideology – that the proper form of marriage is that between classificatory MBD and FZS, not between classificatory FZD and MBS, nor marriages involving direct exchange between any two affinal alliance groups or lines (in which case one has symmetric prescription). However, another property of prescriptive terminologies identified by Needham (1973) was their tendency to redefine kin who marry against the rules or with marriage partners with no known existing

Introduction

kin ties, as if the marriage rule had been followed. That is, a prescriptive terminology imposes its own interpretation on the marriages (or affinal alliances) that do take place. In the words of Maybury-Lewis (1965: 219), ‘it is characteristic of a prescriptive system that all marriages are treated as if they fall into the correct category.’ Followed consistently – which, of course, it may well not be – this would produce a fit between terminology and alliances, redefined or not, regardless of actual genealogical connections, and also regardless of the known fact that marriages may well be dispersed between affinal alliance groups. This provides an answer to the demographic argument outlined in the previous paragraph: for example, it may well be the case that few egos marry their genealogical MBD or FZS, that most marry a classificatory equivalent, and that, even if they do not, their spouses will be treated as such by being redefined as such. It is also an answer to those, like Harold Scheffler and Floyd Lounsbury (1971: 220, 223), who dismissed the notion of prescriptive systems of alliance between social groups altogether on the grounds that too many egos break the supposed rules (though they did not question the existence of such rules as such when applied to individuals), a position that assumes that only the genealogically closest referents are married. Prescriptive systems are distinctive but primarily because of their kinship terminologies, which are, of course, classifications. Unlike a scientific classification, however, they impose an ultimately ideological construction (‘this is how one should marry’) on the facts; that is, they do not necessarily faithfully reflect the facts (here, the prior genealogical connections of marriage partners, if known) but may indeed distort them. By virtue of this redefining property, the figure for marriages defined as proper could well approach 100%.11 An alternative expression, associated especially with Louis Dumont, is ‘positive marriage rules’ – that is, rules specifying the approved kin category from which ego should take a spouse (as opposed to negative rules stating whom ego should avoid in marriage). This is certainly simpler, but it leaves tacit the question of the nature of the terminology. Nonetheless, the phrase has its uses, and I will occasionally use it in what follows. Rather more frequently, however, I use the phrases ‘(a)symmetric (prescriptive) system’ or ‘(a)symmetric prescription’ as the least objectionable ways of describing cases where there is both an (a)symmetric prescriptive terminology and a marriage rule that can be associated with it logically. This is not meant to imply that only the genealogically minimal prescribed ref-



9

10

How Kinship Systems Change

erents will be married in such cases, nor that the alliances of any one line or alliance group cannot be dispersed, nor that, in the case of asymmetric alliance, ego only has one wife-giving and one wifetaking line available for alliance purposes.

The Chapters: Summaries and Acknowledgements Hopefully, I can mostly leave the various chapters to speak for themselves, but the brief summaries below should also be useful. Chapter 1 was originally published in the Journal of the Anthropological Society of Oxford Online 4(2) (2012): 183–211. It describes the main recognized forms or types of kinship terminology and discusses how they might be related, especially in developmental or evolutionary terms. It is especially interested in the phenomenon of zero-equation terminologies and what comes next. In this connection, it also postulates that cognatic terminologies are in a sense the end point of global processes of change in terminologies, as there is no obvious way in which they can evolve further without changing into an already recognized pattern, all of which are hypothetically prior in world history terms. Changes in the version reproduced here are limited to minor revisions of wording and some additional references. Chapter 2 was originally published in Contributions to Indian Sociology 24(1) (1990): 61–76. It represents a first attempt to deal with the question of differences and possible changes in kinship terminology in India involving the kinship and marriage systems of tribal populations in the centre as transitional between south and north India. South India has been known since Morgan as an archetypal example of bilateral cross-cousin marriage with ‘Dravidian’ or what Needham would later call ‘symmetric prescriptive’ terminologies. North India, by contrast, lacks both features and frowns on marriage with any cousin, cross as well as parallel (see also Ch. 5). In marrying preferably (but probably not prescriptively) into GEG categories (i.e. with siblings’ spouses’ siblings), marriage groups in tribal populations like the Juang exchange spouses, often quite intensively, within a generation, but forbid realliances between the same groups for the next one to three generations.12 In respect of linking ego with alter, GEG categories are not traced through previous generations in the way that cross cousins are, although as with cross-cousin marriage

Introduction

they can be seen formally as the basis of exchanges of siblings in marriage. Also, cousin specifications are not included in the GEG category as they are with prescriptive alliance but are likely to be merged with siblings and parallel cousins, as in north India. They thus have features of both south and north India, which in evolutionary terms they may link. Thirty years after this chapter was first published, I would now take the view that while tribal societies like the Juang were fully prescriptive at one stage in history, as in south India, for status reasons they have been influenced by north Indian norms to give up cousin marriage and develop terminological features that reflect this change. This is supported by the fact that in Koraput other tribal populations, both Munda- and Dravidian-speaking, have systems of bilateral cross-cousin marriage with fairly standard symmetric prescriptive terminologies to match. In addition, as Section VI of the original article showed, we can see potential evidence of change in two tribal terminologies of Bihar.13 More speculatively, in the course of a history now closed off to us, some of these tribal populations might have become castes and turned entirely to north Indian norms in kinship and marriage, as the Bhuiya of Odisha appear to have done or to be doing (see Ch. 5). Conversely, I would not now claim that the Juang and allied terminologies are prescriptive today in the rather decisive way I was prepared to do in the original article. I now accept that prescriptive terminologies should be defined in part by their cognate-affine equations, which the Juang etc. lack. However, the cognatic specifications in such equations might very well survive after the affinal ones have split off: for example, the equations MB = FZH and FZ = MBW might continue to exist even after the specifications EF and EM respectively have ceased to belong to them. This seems to be the case for the Juang and similar terminologies in central India, where the reduced, purely cognatic equations exist, and this was what induced me to exaggerate their prescriptive nature. There is still a lot of uncertainty about these tribal societies, especially whether they do in fact expect their members to marry in certain ways with reference to kin categories and, if so, what the kin terms are for such preferred or possibly prescribed categories. One other caveat is that, in a brief visit to the Juang area in 1998, I and my assistants were told that cross-cousin marriage did occur and that it was more frequent than GEG marriage, though the latter also took place. This tends to conflict with McDougal’s data, used in Chapter 2,



11

12

How Kinship Systems Change

which even in 1998 were some 35 years old. However, it does not contradict them entirely, and in any case other tribes in the area are described as having the same features that McDougal described for the Juang, though in less detail.14 Chapter 3 was originally published in the Journal of the Anthropological Society of Oxford-online 5(2) (2013): 194–206. For this volume, it has been cut, rearranged, somewhat reworded and provided with a new introduction, and the section numbers have been removed. As a result, this version omits most of the discussion of Crow-Omaha terminologies and their possible derivation from Allen’s tetradic model to concentrate on the further implications of treating the intermarriage of sets of siblings within a single generation as marriage to GEG categories, as in Chapter 2 and my description of it above. As Chapter 2 focuses on India, Chapter 3 also describes possible examples of this marriage practice elsewhere in the world. In this chapter, I also stress the frequent tendency in this form of affinal alliance to intensify alliances between such sibling groups as much as possible within the same generation. This circumstance, though generally underemphasized in the literature, may itself be crucial evidence of a previous practice of cross-cousin marriage in which everyone in the society was expected to marry in the same fashion, so that all egos, at least formally speaking, repeat the marriages of the parental generation. However, the ban on renewing alliances in the immediately following generation(s) with GEG marriage ensures that this will not be possible, while at the same time dispersing the alliances of any one spouse-exchange group to other such groups over time and not just confining them to one other such group. Taken literally, the latter situation would involve a moiety system of just two affinal alliance groups in the society perpetually exchanging spouses. However, as cross-cousin marriage does not need moieties to function, assuming this situation had existed, it may have come to be seen as irksome at some point in history and to be broken down into a larger number of smaller spouse-exchange groups, allowing alliances to be dispersed. Preventing alliances from being renewed in the immediately following generation or generations would have increased this dispersal still more. In other words, this hypothetical society might have 1) started with just two moieties exchanging spouses through cross-cousin marriage, 2) abandoned moieties but retained cross-cousin marriage, and then 3) abandoned cross-cousin marriage but retained the practice of intensifying marriages between groups of

Introduction

siblings within the same generation. Steps 2 and 3 would have had the effect, but not necessarily the intention, of dispersing alliances among spouse-exchange groups. Chapter 4 was first published in Warren Shapiro (ed.), Focality and Extension in Kinship: Essays in Memory of Harold W. Scheffler, Acton: Australian University Press (2018). That is, it originated as a contribution to a Festschrift for Prof. Scheffler, which quickly turned into a memorial volume, as the honorand died while it was still in production. As Scheffler was insistent in denying the link between prescriptive terminologies and cross-cousin marriage, he deserves some attention here, but as its title indicates the chapter also discusses the possible reasons for societies abandoning cross-cousin marriage and the circumstances in which they might do so. Although kinship terminology is discussed, the focus is more on changes to norms and patterns of affinal alliance in this chapter. It therefore constitutes a useful continuation of the themes discussed in Chapters 2 and 3. Changes to the original version of Chapter 4 are restricted to minor alterations in wording. Chapter 5 was first published in Maurice Godelier, Thomas Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship, Washington and London: Smithsonian Institution Press (1998). It returns the focus to South Asia and to the terminologies, first identifying one Iroquois-type terminology in the far north-west of the region, that of the linguistic isolate (i.e. unrelated language) Burushaski, and then investigating its potential as an intermediate form between the terminologies that are typical of south India (symmetric or bilateral prescriptive) and north India (a form of zero equation) respectively. It may therefore seem like an alternative or contradiction of the arguments set out in Chapter 2, but in fact Chapter 5 ends by suggesting a developmental sequence that incorporates all these terminological forms. Changes are therefore restricted to minor alterations in wording, and the first footnote in the original has been replaced. Also, the present version of this chapter ends with two paragraphs written especially for it, the abbreviations used for kin types throughout have been changed to be in line with those used in the other chapters in the present collection, and a map in the original has been deleted. Chapter 6 was first published in the Journal of the Anthropological Society of Oxford-online 7(2) (2015): 205–33. It looks at IndoEuropean kinship terminologies in Europe primarily from a diachronic



13

14

How Kinship Systems Change

perspective, though it also establishes the details of a number of these terminologies synchronically. Generally, it seems that there has been a shift in patterning among these terminologies across Europe, a trend which to some extent can be seen as geographical. Thus most Slavonic and Baltic terminologies in the east have zero-equation features,15 in some cases rather residual, while those in the west, principally Romance, Greek, Celtic and some Germanic terminologies, like English, Dutch and German, as well as Czech, are or tend to be cognatic in nature. However, there is some evidence, notably in the Baltic, German and Scandinavian terminologies, that these were also zero-equation terminologies well into historic times, having become cognatic (German) or being on the way to becoming cognatic (Scandinavian, Lithuanian) at the present day. The Latin terminology can be interpreted similarly, having zero-equation features itself, while the terminologies of its present-day daughter languages can only be described as cognatic. The basic hypothesis is that, in order for a prescriptive or post-prescriptive terminology to evolve into a cognatic one, its equations need to be broken down and re-sorted on a different basis. It is the stage of breakdown, naturally enough, that has been given the label ‘zero equation’. Once this stage has been reached, the very different cognatic pattern can be arrived at, a pattern that most closely accords with the biological realities of two parents of equal status with children, the fact that grandchildren, (grand-)parents and +1 and −1 collaterals are not distinguished terminologically as to side of family, and an amorphous category of cousins similarly not distinguished either.16 The fact that some of these terminologies, especially in eastern Europe, are zero equation in type and therefore resemble typologically Indo-European terminologies in north India should be emphasized here, as it forms one of the key arguments of this collection. Given the shared background of membership in the same language family, one can postulate a situation in which either the two branches both evolved into a zero-equation pattern before branching off from one another or underwent a similar evolution in parallel after their separation. There is also the fact that, while prescription is not found anywhere in the Indo-European language family in terms of its own long-term genealogical heritage, some lexically IndoEuropean terminologies in South Asia have become prescriptive historically through a process of lexical replacement (referred to in Ch. 5). Taking the data and arguments in the two chapters together,

Introduction

therefore, one might also posit an evolutionary development within this language family from symmetric prescriptive (Dravidian) to cognatic (e.g. English, German, Romance languages) via GEG marriage (north Munda), Iroquois (Burushaski) and zero equation (e.g. north India, Baltic). In this hypothesis, the zero-equation stage would be significant in enabling terminological categories to be re-sorted into a cognatic pattern. This would be easier than positing a direct change from prescriptive to cognatic, though something very similar has been suggested for the Arctic (Ives 1998). The original version of Chapter 6 has not been altered substantially: the original note 3, on kin term abbreviations, has been deleted as having already been given elsewhere in the present volume; the last footnote, 34 (previously 35), has been added too, and a new section on Celtic has been added immediately before the conclusion. Chapter 7 has not been published before. It starts the second part of the book, which, as already noted, is devoted to issues surrounding Crow-Omaha systems and some marriage practices in eastern Indonesia that resemble them sociologically but lack the associated terminologies. In Chapter 7, the theme of a transition from prescriptive to non-prescriptive terminologies is retained, but in the context of the possible origin of Crow-Omaha terminologies from prescriptive ones. Two theories are discussed: the derivation of Crow-Omaha terminologies from systems of asymmetric prescriptive alliance (‘Kachin type’), and their derivation from systems of symmetric prescriptive alliance (‘Dravidian’ type), possibly in some cases via Iroquois-type terminologies, which are symmetric like Dravidian but not prescriptive. The former hypothesis has so far proved far more popular in the literature, with specific theories being devoted to it. As well as discussing these theories, therefore, the chapter also points out the potential for the latter hypothesis as an alternative. Chapter 8 was first published in Structure and Dynamics: eJournal of Anthropological and Related Sciences 11(1) (2019): 73–95. It puts forward a theory of the possible sociological correlates of CrowOmaha terminologies in terms of the substitutability of kin in different generations in intergroup relations: i.e. it is hypothesized that those kin types that can substitute for one another in different generations are linked by vertical terminological equations of the types known as Crow-Omaha. Its contribution to the evolutionary theme, therefore, is that it suggests possible reasons for the development of Crow-Omaha terminologies out of prescriptive ones.



15

16

How Kinship Systems Change

Chapter 9 was originally published in the Journal of Anthropological Research 74(2) (2018): 232–51. Building on Chapter 8, it puts forward a similar theory of substitutability in relation to certain speech communities in eastern Indonesia. However, their terminologies are not Crow-Omaha. Instead, this region is one of frequent though not uniform asymmetric prescriptive alliance, the form of alliance that is clearly the origin of these examples. This shows that substitutability of kin does not need Crow-Omaha terminologies any more than dispersed alliance does. This chapter has been changed significantly from the original published version, mainly in that the three paragraphs setting out my understanding of the terms ‘prescription’ and ‘prescriptive’ have been transferred to the present Introduction. There are also some minor changes in wording, and the abstract, key words, acknowledgements and maps in the original have been deleted.17 This means that endnotes 1, 2, 3 and 5, as well as the unnumbered note at the start, have also been deleted, though the original endnote 4, now note 1, has been retained. A new note 2 has also been provided.

Diagrams The diagrams or figures below are designed for consultation when reading the chapters that follow. The first two figures (0.1, 0.2) show the two main prescriptive systems: symmetric or bilateral, the prescribed cousin being a cross cousin relatable as both MBC and FZC; and asymmetric or matrilateral, the prescribed cross cousin being for male ego MBD and for female ego FZS (a patrilateral relative from her point of view, NB). The next pair of figures (0.3, 0.4) show the key Crow-Omaha features and equations. The remaining pair (0.5, 0.6) show the operation of marriages between sets of siblings within the same generation (cf. especially Chs. 2 and 3). In line with the practice of Needham and his followers, I have preferred these so-called ‘matrix diagrams’ to genealogical ones, as this emphasizes the status of kin terms as category words rather than genealogical positions.



17

Introduction

Figure 0.1. Symmetric or bilateral prescription Lineal and parallel kin

Cross kin

male

female

male

female

FF

MM

MF

FM

FFB

MMZ

MFB

FMZ

MMB

FFZ

FMB

MFZ

F

M

MB

FZ

FB

MZ

FZH

MBW

MZH

FBW

EF

EM

ego

ego

B

Z

PosGS (cross cousin)

PosGD (cross cousin)

PssGS (parallel cousin)

PssGD (parallel cousin)

H

W

ZH

BW

EZH

EBW

EB

EZ

Sms

Dms

Sws

Dws

BS

BD

ZS

ZD

DHws

SWws

DHms

SWms

EZS

EZD

EBS

EBD

PosGDS

PosGDD

PosGSS

PosGSD

SSms

SDms

SSws

SDws

DSws

DDws

DSms

DDms

BSS

BSD

BDS

BDD

ZDS

ZDD

ZSS

ZSD

18

How Kinship Systems Change

Figure 0.2. Asymmetric prescription (matrilateral ms, patrilateral ws) D female

male

B female

male

FFZH

FZHZ

FZH

FFZD

FFZS HF

HZH

FZD

A female

FFZ

FZ HM

C

male

female

male

female

FF

FM

MF

MM

FFB

FMZ

MFB

MMZ

MFZ

FMB

F

M

FB

MZ

MB

MBW

MZH

FBW

WF

WM

MBS

WBW

FZS** ego ( f ) ego (m) MBD* ZH

Z

B

W

H

PssGD

PssGS

HZ

HB

HBW

WZH

BW WZ

ZD

ZS

Dms

Sms

SWms

Sws

BD

BS

MBSD

MBSS

WZS

WBD

WBS

DHws Dws

E male

MMB

WB

DHms SWws HZD

HZS

FZDD

FZDS

DDws

DSws

ZDD

ZDS

HBD

HBS

WZD

MBDD MBDS DDms

DSms

SDws

SSws

ZSD

ZSS

BDD

BDS

SDms

SSms

BSD

BSS

Note: ego’s line (male and female) is A; B is the line of ego’s wife-takers; C is the line of ego’s wife-givers. Lines D and E are those of ego’s wife-takers’ wife-takers and wife-givers’ wife-givers respectively. However, in the minimal three-line form of this system, line D would coincide with line C and line E with line B. This would logically produce equations like MMB = FFZH, MBW = FFZD, WBW = FZD and HZH = MBS. The transfer of wives is from right to left. * = prescribed spouse for male ego; ** = prescribed spouse for female ego.



19

Introduction

Figure 0.3. Crow-type terminology FZ

FB

F and M

MZ

MB

FZD, FZS

FBC

male ego and Z

MZC

MBC

D and S (C)

FZDD, FZDS

Key equations: FZ, FZD, FZDD (bold); F, FZS, FZDS (italics); C, MBC (underlined). Other possible equations (non-diagnostic): M = MZ, F = FB and FBC = MZC (parallel cousins).

Figure 0.4. Omaha-type terminology FZ

FB

F and M

MZ

MB

FZC

FBC

male ego and Z

MZC

MBD, MBS

ZC

MBSD, MBSS

Key equations: MB, MBS, MBSS (bold); M, MBD, MBSD (italics); and FZC, ZC (underlined). Other possible equations (non-diagnostic): F = FB, M = MZ and FBC = MZC (parallel cousins).

Figure 0.5. GEG marriage (symmetric) opposite-sex sibling pair A / opposite-sex sibling pair B sister and male ego

ZH and ZHZ

(sister marries ZH, ego marries ZHZ)

(respective spouses: ego’s sister and ego)

Direction of spouse transfer: symmetric

20

How Kinship Systems Change

Figure 0.6. GEG marriage (asymmetric) same-sex sibling pair A / same-sex sibling pair B brother and male ego

BW and BWZ

(brother marries BW, ego marries BWZ)

(respective spouses: ego’s brother and ego)

Direction of spouse transfer: asymmetric

Dedication Lastly, I dedicate this volume to the memory of the late Dr N.J. ‘Nick’ Allen (1939–2020), formerly Reader in the Social Anthropology of South Asia at the University of Oxford and my own doctoral supervisor, who undertook much similar work on the theme of change in kinship terminologies that is also the focus of this collection and who inspired me, as his student, to do the same.

Notes 1. Gertrude Dole showed long ago (1972) that ‘Eskimo’ terminologies are var-

2.

3.

4.

5.

ied and that they are not ‘lineal’ (or ‘cognatic’) in all genealogical levels, unlike the English terminology. Of the latter two descriptors, I prefer ‘cognatic’ to ‘lineal’ as more logical. Chapter 8 therefore also incorporates time as an intrinsic factor, but here it is a matter of the perpetually repeated course of a social process, not of permanent, irreversible change to an entire kinship system. Some parts of the world, such as the Americas, Australasia and Russian Siberia, could be said to have both categories of population, as they combine the well-established descendants of European-derived settler societies with increasingly marginalized native populations with much longer histories of settlement. Allen, to whose memory this collection is dedicated, was my doctoral supervisor, but we kept in close touch between my subsequent graduation in 1984 and his death in early 2020. Needham’s impact on me mainly came from his publications on kinship, though I had personal contact with him from time to time, mainly through the research seminars he held at All Souls College. He died in 2006. This sentence requires some qualification. By ‘cognitively modern humans’, I certainly mean homo sapiens. By ‘contemporary phase of history’, I assume that this starts with the later Pleistocene and extends to what is rapidly becoming known, more informally, as the Anthropocene, the modern period of extensive human impact on the planet, destructive or otherwise. I am

Introduction

6. 7.

8. 9.

10.

11.

12.

13.

14.

15.

admittedly writing here with the benefit of only the most partial knowledge of these matters, but that is enough for my purposes. For fuller discussions of early human (including hominid and hominin) kinship, see chapters in Allen et al. (2008); for a brief overview of human evolutionary history, see Gowlett and Dunbar (2008) in the same volume. Unpublished article cited with author’s permission. In neither of the articles cited here does the author mention me personally. An example of a village being the operational unit in affinal alliance is Kédang, on the island variously known as Lomblen or Lembata, one of the Solor Islands in eastern Indonesia (Barnes 1974). Analytical and indigenous categories may, of course, sometimes be the same, as with the English term father. One leading authority on kinship, Louis Dumont, reversed this for some reason so that he marked ascending generations with a minus sign and descending ones with a plus sign. This section is taken from Parkin 2018: 233-4, the rest of which is reproduced here as Chapter 9. The present section has some minor changes in wording, the original note 3 has been deleted, and the original note 5 is now note 11 of this introduction (the next footnote). One example of this is Anthony Good’s study of a south Indian subcaste with symmetric prescription, for which he conducted a statistical survey of marriage choices (1981). He found that only 25% of marriages were with a first cross cousin but that when other qualifying alters were taken into account (second and remoter cross cousins, and others in the prescribed category for marriage) the figure rose to 95%. Good explicitly states that the Maravar terminology has the redefining property mentioned above. I know of no comparable account of a society with asymmetric prescription. In the case of a delay of one generation, this could be interpreted as repeating the marriage of an FF, e.g. marriage to FMBSD, a specification that often occurs in this context worldwide. This is rare in the case of these tribal groups as far as our information goes, with only Juang na having this specification alongside PM and MFZ (McDougal 1964), though it is found quite frequently in other societies around the world. See also Ch. 3. The references for these two examples were inadvertently omitted from the original: for the Malpahariya, see Sarkar (1937), and for the Malto or Maler, Vidyarthi (1963). One other update should be mentioned: in the original article references to Parkin n.d., The sons of man, it should now read Parkin (1992), The Munda of Central India. That is, a situation in which practically every close and medium-close kin type has its own term, or alternatively a composite, descriptive phrase (e.g. morbro for mother’s brother in Swedish etc.). This links zero-equation terminologies in Europe typologically to the north Indian terminologies discussed in Chapter 5. We owe the term ‘zero equation’ to Allen (1989), though the principle is also known under Murdock’s earlier term ‘Sudanese’ (1949: 224, 238–39).



21

22

How Kinship Systems Change

16. As already noted above, ‘cognatic’ use of a terminology is Needham’s term,

equivalent in meaning to Robert Lowie’s better known and more influential ‘lineal’. For Needham, a lineal terminology was one in which the terms were sorted into descent lines, as is the case with prescriptive, Iroquois and Crow-Omaha terminologies. I am not saying that the categories of relatives mentioned above cannot be further distinguished using, for example, genealogical reckoning, only that a cognatic terminology will not ordinarily distinguish them itself. 17. These were inserted into the original text at the request of the publisher; they can be dispensed with here.

References Allen, Nicholas J. 1989. ‘Assimilation of Alternate Generations’, Journal of the Anthropological Society of Oxford 20: 45–55. Allen, Nicholas J., Hilary Callan, Robin Dunbar and Wendy James (eds). 2008. Early Human Kinship: From Sex to Social Reproduction. Malden, MA and Oxford: Blackwell Publishing, for the Royal Anthropological Institute. Barnes, Robert Harrison. 1974. Kédang: A Study of the Collective Thought of an Eastern Indonesian People. Oxford: Clarendon Press. Barth, Fredrik (ed.). 1969. Ethnic Groups and Boundaries: The Social Organization of Cultural Difference. London: George Allen and Unwin. Berlin, Brent, and Paul Kay. 1969. Basic Color Terms: Their Universality and Evolution. Berkeley: University of California Press. Bowden, Ross. 1983. ‘Kwoma Terminology and Marriage Alliance: The “Omaha” Problem Revisited’, Man 18: 745–65. Brown, Cecil. 1984. Language and Living Things: Uniformities in Folk Classification and Naming. New Brunswick: Rutgers University Press. Ensor, Bradley. 2017. ‘Testing Ethnological Theories on Prehistoric Kinship’, CrossCultural Research 51: 1–29. ———. 2019. ‘A Critical Examination of Evolutionary Theory in Prehistoric Kinship Research’. Paper given to panel on ‘Issues of Kinship Theory: Solidarity or Logic?’, at the IUAES conference on ‘World Solidarities’, Adam Mickiewicz University, Poznan, Poland, August 2019. Evans-Pritchard, Edward Evan. 1940. The Nuer. Oxford: Clarendon Press. Good, Anthony. 1981. ‘Prescription, Preference and Practice: Marriage Patterns among the Kondaiyankottai Maravar of South India’, Man 16: 108–29. Gowlett, John A.J., and Robin Dunbar. 2008. ‘A Brief Overview of Human Evolution’, in Nicholas J. Allen, Hilary Callan, Robin Dunbar and Wendy James (eds), Early Human Kinship: From Sex to Social Reproduction. Malden, MA and Oxford: Blackwell Publishing, for the Royal Anthropological Institute, pp. 21–24. Ives, John W. 1998. ‘Developmental Processes in the Pre-contact History of Athapaskan, Algonquian and Numic Kin Systems’, in Maurice Godelier et al. (eds), Transformations of Kinship. Washington DC and London: Smithsonian Institution Press, pp. 94–139.

Introduction

Kroeber, Alfred. 1917. ‘Californian Kinship Systems’, University of California Publications in Archaeology and Ethnology 12(9): 339–96. Lévi-Strauss, Claude. 1991. ‘Cyclical Orders and the Problem of Hierarchy’, Man 26(2): 348–49 (with reply). Lowie, Robert. 1917. Culture and Ethnology. New York: Boni and Liveright. Maybury-Lewis, David. 1965. ‘Prescriptive Marriage Systems’, Southwestern Journal of Anthropology 21: 207–30. McDougal, Charles. 1964. ‘Juang Categories and Joking Relationships’, Southwestern Journal of Anthropology 20(4): 319–45. Murdock, George. 1949. Social Structure. New York: Macmillan. Needham, Rodney. 1973. ‘Prescription’, Oceania 43: 166–81. Parkin, Robert. 1992. The Munda of Central India: An Account of their Social Organization. Delhi: Oxford University Press. ———. 1997a. Kinship: An Introduction to Basic Concepts. Oxford and Malden MA: Blackwell. ———. 1997b. ‘Tree Marriage in India’, in Klaus Seeland (ed.), Nature is Culture: Indigenous Knowledge and Socio-cultural Aspects of Trees and Forests in Non-European Culture. London: Intermediate Technology Publications, pp. 51–56. Sarkar, S. 1937. ‘The Social Institutions of the Malpahariyas’, Journal of the Asiatic Society of Bengal, Letters 3: 25–32. Scheffler, Harold, and Floyd G. Lounsbury. 1971. A Study of Structural Semantics: The Siriono Kinship System. Englewood Cliffs, NJ: Prentice-Hall. Vidyarthi, L.P. 1963. The Maler: A Study in Nature-Man-Spirit Complex of a Hill Tribe in Bihar. Calcutta: Bookland.



23

Part

I 

Terminological Change

1 

Kinship as Classification Towards a Paradigm of Change

I The aim of this chapter is to consolidate accrued knowledge and understanding of the ways in which patterns of kinship terminologies may change over time.* Study of this topic, as of terminological patterns generally, goes right back to Lewis Henry Morgan’s pioneering work in the middle of the nineteenth century, and we now have sufficient data and understanding of the issues involved to attempt grander paradigms than have been possible in the past. I wish to broach this issue here as a first step in this direction, while also acknowledging some of the problems involved. The discussion will inevitably be quite technical and specialized and will assume that the reader already has a degree of knowledge and understanding about these issues. First, however, some preliminaries are in order. The study of transformations of kinship terminologies clearly invites an evolutionary approach, and Morgan himself counts as one of the great evolutionists of the nineteenth century. Typical of his time, he sought to link such changes to changes in social organization more generally, as well as putative stages in history and prehistory. Although he was less ethnocentric about the superior qualities of the Victorian period than some of his contemporaries, his writings in this respect do consider the question of the emergence of ‘civilization’ from ‘savagery’ via ‘barbarism’ in a manner that would be quite

28

How Kinship Systems Change

unacceptable in contemporary anthropology. Nevertheless, if more modestly, quite a number of anthropologists since his time have been prepared to identify changes in terminological pattern and to link them, where ethnographically justified, to changes in marriage patterns, given especially the association of various types of crosscousin marriage with specific terminological patterns – that is, if one is known to have changed, it is reasonable to investigate change in the other. An important example that is still relatively recent is the collection Transformations of Kinship (Godelier et al. 1998, based on a 1993 Paris conference), contributors to which pursue such issues without falling into the ethnocentrism embedded in terms like ‘civilization’. Although the Boasians were firmly opposed to Morgan’s evolutionism in particular on quasi-moral grounds, this collection and other more recent work does prove that it is possible to consider such changes in a non-ethnocentric manner. Nonetheless, to play it safe, there is now an increasing tendency to talk about ‘transformation’ or simply ‘change’ rather than ‘evolution’, as exemplified by the title of the volume just cited. For Allen, adopting a phrase often used by Marcel Mauss, the phrase ‘world history’ is another suitable and legitimate alternative (e.g. 1998: 318). Coming between Morgan and Godelier et al., a number of other anthropologists have made their mark on such work (see Parkin 1997: Ch. 14 for an overview). Early figures include Irving Hallowell, Fred Eggan and Alexander Spoehr on Native American terminologies (see Parkin 1997: 162–63), while both W.R.R. Rivers, a diffusionist, and Leslie White, a neo-evolutionist, defended Morgan’s general approach against his Boasian critics during the first half of the twentieth century without contributing significantly to such work themselves. Later figures include Elman Service (e.g. 1971: Ch. 7), Gertrude Dole, whose thesis on terminological change (1957) has never been published, though some articles have been (e.g. 1960, 1969, 1972), Elmendorf (1961), and Dyen and Aberle (1974). At Oxford, many of Rodney Needham’s basically structuralist studies of variations in patterns of kinship terminology used the idea of change as a mode of explanation for such variations (e.g. 1980), an approach that inspired both my own work and that of my doctoral supervisor, N.J. Allen, who had himself been taught by Needham initially. Allen’s notion of tetradic society (1986; also 1989a, 1989b, 1998), an extrapolation back in evolutionary terms from four-section systems with bilateral cross-cousin marriage like the Kariera of Australia, has caught the at-

Kinship as Classification

tention of other scholars and has been the partial inspiration behind two recent volumes, Transformations of Kinship (Godelier et al. 1998, already cited), and Early Human Kinship (Allen et al. 2008), the latter being explicitly evolutionary in the contemporary sense (the volume includes contributions from social anthropology, archaeology, evolutionary psychology, primatology and linguistics). Not all those associated with Needham’s influence have been equally taken by his ideas in this respect, one notable sceptic being R.H. Barnes, who has criticized Robert Blust’s work (1980; Barnes 1980) reconstructing early Austronesian kinship and more recently suggested bluntly that ‘the history of evolutionary speculation in anthropology has never produced anything like certainty in our understanding of how and why such patterns change’ (2012: 203). However, some of Barnes’s own early work (e.g. 1977: 151–52), together with that of Gregory Forth (1985, 1988, 1990), David Hicks (especially 1981, 1983, 1985, 1986) and C.H. Wheeler (1982) (the first two were students of Needham’s, the last briefly of Barnes), did point in the same direction, though in Barnes’s case already with evidence of the scepticism he was to express more strongly later (e.g. 1977: 152–53). One figure of note in Allen et al. 2008 and other much more recent publications is Christopher Ehret, who has worked on developmental problems in Bantu kinship in Africa (e.g. Ehret 2008). The work of the late Per Hage is also significant, especially in his often collaborative attempts to construct paradigms relating to the whole of human history under the umbrella of terms such as ‘Nostratic’ (see his memorial volume, Jones and Milicic 2010). In addition to Needham and Allen, an important inspiration for my own work in this area has been Stanley Witowski, who applied the methods Berlin and Kay (1969) developed for the study of colour terms to kin terms (Witowski’s thesis is unpublished but see Witowski 1972). Berlin and Kay compared colour terms and terminologies from a wide range of societies around the world and postulated their emergence in an evolutionary sequence that was, at least in part, predictable. Thus terms for colours like yellow, green and blue typically emerged after that for red, meaning that if a terminology has a term for one of these colours it will have one for red but not necessarily vice versa; similarly, if blue has a separate term, yellow and green will also have one each but not necessarily vice versa, as blue typically emerges later as a separate category in the evolutionary sequence (often being merged previously with green). Witowski



29

30

How Kinship Systems Change

showed that the same basic principle of predictability applied to kinship terminologies; for example, the presence of cognate-affine equations based on the practice of bilateral cross-cousin marriage implies the presence of equations between cognates of the type FB = MZH and MB = FZH but not necessarily vice versa. However, this is not because cognate-affine equations emerge later but because they break down earlier in evolutionary terms, possibly but not necessarily because cross-cousin marriage is no longer being followed. Nonetheless, this is no objection to Witowski’s adoption of Berlin and Kay’s basic method, which is implicit in the core analysis of the present chapter, as it indicates that changes to kinship terminology are more likely to take place in one direction than its opposite. Similarly, for example, it is evident that the existence of bifurcate merging or bifurcate collateral equations in generations 0 and/or −1 normally implies the same in +1 but not necessarily vice versa, reflecting a general trend towards the Hawaiianization of these genealogical levels (typically 0 before −1, then +1, if these adjacent levels change at all). The main point is not that changes in terminologies are predictable in a deterministic way but that the parameters within which such changes take place are limited and can be known, occasional ethnographic exceptions notwithstanding. The method is a predictor of trends, not of predetermined stages or changes. This approach to the predictability of change can also be invoked in respect of other types of equation and the order in which they typically disappear from kinship terminologies, namely: first, alternate generation equations (e.g. eB = FF, yB = SS); secondly, prescriptive equations expressing cross-cousin marriage (i.e. equating cross kin and affines); and thirdly, classificatory equations of the type F = FB = FFBS = FFFBSS etc., which rely on the principle of same-sex sibling equivalence. Alternate generation equations are central to tetradic theory (mentioned above and explained in more detail in sect. II of this chapter), though they also appear frequently in actual terminologies of bilateral cross-cousin marriage (e.g. the Kariera of Australia) from which tetradic society has been extrapolated backwards. The dissolution of prescriptive equations can normally be associated with the decline or abolition of cross-cousin marriage, which such dissolution nonetheless tends to lag behind. The fact that the equations between cross kin that are characteristic of terminological prescription frequently survive the breach in cognate-affine equations can be related back to Witowski’s observation that the existence of the latter

Kinship as Classification

implies the existence of the former but not necessarily vice versa. Classificatory equations can also survive this abandonment of crosscousin marriage but cannot survive further changes in the direction of generational, zero-equation or cognatic terminologies (stages 6, 8 and 9 in the list below). Classificatory equations too are implied by the existence of either or both prescriptive and alternate-generation equations but do not necessarily imply either of the latter where they (the classificatory equations) do occur. The question of the evidence for such changes should also be addressed, as this raises problems in its own right. Terminologies can be found that are internally logically consistent in their patterning throughout; for example, the Tamil terminologies linked to bilateral cross-cousin marriage recorded by Beck (1972, Appendix F) or Good (1991: 56), or the highly symmetrical English terminology. That is, one can use logical consistency to establish logical terminological types that may correspond to ethnographic realities but do so rather infrequently. However, a great number of recorded terminologies lack this consistency, which itself has often suggested a terminology in flux and therefore a situation of change (e.g. Allen 1975, 1976; also Parkin 1992: Ch. 7) – that is, logical inconsistency becomes a proxy in examining terminological transformation. Because of these considerations, Needham called for the abandonment of types of whole terminologies and advocated instead a concentration on the features, very often mixed, whereby terminologies might individually be described and characterized (especially Needham 1971). However, while I sympathize with this view as regards analyses of actual terminologies, here I am more concerned with general patterns in transformations. For this reason, the necessity is to concentrate on uniform and logically consistent types, not on individual and possibly internally consistent actual cases, nor internally inconsistent cases. Nonetheless, one must always remember that it is one thing to set out logical transformations step by step, quite another to prove the actual changes they imply historically. Another possible objection is that what appears to be inconsistent actually reflects kinship practice in the society concerned – that is, it has meaning in synchronic terms. This latter tension is exemplified by an interesting contrast between the approaches of Gertrude Dole, who put forward an evolutionary explanation for inconsistencies in the pattern of the Xingu Carib terminology (1969), and Ellen Basso, who rejected this in favour of an explanation couched in terms of actual kinship prac-



31

32

How Kinship Systems Change

tice in the present (1970). In essence, though, one can argue that, if cross-cousin marriage has ceased to exist in a society that very probably had it historically, it has also ceased to have present-day meaning. In other words, there is no necessary incompatibility between the two positions: changes in kinship systems take place because circumstances change along with the meanings attached to them. The fact that the Xingu Carib have a meaningful kinship system and terminology in the present does not mean that these features have not changed in the past. Kryukov discusses possible sources of evidence for change in terminologies (1998: 298–99), namely historical records (e.g. for Chinese; Fêng 1937), old ethnographic and similar reports compared to more recent ones, myths etc. as presenting older forms, emerging distinctions between the address and reference terminologies in which the former may point to future developments, and using older informants. Other methods identified by Forth (1990) include the comparison of different terms in cognate languages, and of whole terminologies in cognate languages (cf. Parkin 1992: Ch. 7 on Munda), and redundancy – that is, where a kin type is covered by more than one term, which itself may be evidence of change, being an instance of the lack of consistency mentioned above (e.g. the examples in Parkin 1998: 264; cf. Ch. 5 this volume). Kryukov (ibid.) also discusses the ways in which terminological change may be effected: in particular, derivations of other terms, periphrastic phrases involving other terms, or loans from foreign languages may all be used to introduce fresh distinctions (1998: 300–1). Forth offers a similar list of possible modifications in linguistic terms (1990: 376 n.7): borrowings, new coinings, compounding, and vertical or horizontal extensions of particular terms (i.e. vertically to other generations and, e.g., horizontally of terms for lineal kin to cover also cross-cousins). A related issue is the direction of change and evidence for it. The dominant assumption in the literature is that terminologies expressing bilateral cross-cousin marriage have an evolutionary priority among attested types – that is, that they are the starting point for any paradigm of transformation or evolution and that all changes away from it are irreversible. While there is some historical evidence for this assumption, it is not accepted by everybody. For example, in relation to parts of eastern Indonesia, where there is ample evidence of change in both terminology and affinal alliance, both Fox (1984, 1987) and Guermonprez (1998) have suggested non-prescriptive

Kinship as Classification

cognatic terminologies as the starting point, while Blust (1980) has claimed that the Proto-Austronesian terminology was asymmetric prescriptive in type. Fundamentally, the theory of an original system of bilateral cross-cousin marriage is still a hypothesis profoundly influencing the direction of work in this area. Hypothesis-driven research is, of course, perfectly respectable, but we now need to bring together the evidence that supports the hypotheses informing research on terminological change. This is one aim of my longer-term research, though I can do little to take this forward in the present chapter. Nonetheless, there is a general though not universal assumption that a simple structure will become more complex rather than that complexity is simplified: whereas the former process may but need not involve the creation of terms, or alternatively their re-sorting, the latter would involve their deletion, this being less likely. In fact, deletions certainly occur, but I would argue that they do so at the end of the world-historical process of change and (in line with Allen, e.g. 2008) agree that they do not do so at its beginning. The view that deletions of terms are unlikely is also found in Elmendorf (1961: 370–71, 376). Fundamentally, therefore, I adopt the by now conventional view that symmetric prescriptive terminologies expressing the practice of bilateral cross-cousin marriage are basic among attested forms and that they have changed into other forms but not vice versa. In cases where this seems to be contradicted – that is, where a speech community appears to have adopted such a terminology from some other type – the explanation is usually that it is the language of the speech community that has changed, not the pattern of the terminology. Thus, in India, there are a number of Indo-European languages with symmetric prescriptive terminologies that are not otherwise known in this language family. One of the best known is Sinhalese, whose speech community has probably long been practising cross-cousin marriage, as does most of south India. The assumption here is that at some stage of history the Sinhalese adopted an Indo-European language in place of, probably, a Dravidian one but retained the old terminological pattern, and also the associated marriage practices, in doing so. Finally here, we must remember that not all research on kinship terminologies is evolutionary or transformational. This tension has already been mentioned in relation to the differences between Dole and Basso mentioned above. For one thing, there are countless



33

34

How Kinship Systems Change

synchronic analyses of particular terminologies in particular ethnographic situations, only a minority of which consider possible questions of change, even in passing. Secondly, there are comparative analytical approaches that, implicitly or explicitly, are synchronic in type. Historically, two in particular stand out with neither exclusively being used for kin terms, though most of the work in both cases has been in this area. One is componential analysis, pioneered by Ward Goodenough (e.g. 1956) and inspired by a combination of Bloomfieldian linguistics and Kroeber’s features (1909), like side of family, age and generation – and, indeed, more recently the term ‘feature analysis’ has been suggested for this approach (D’Andrade 1995: 21). The other approach, Chomskyan rather than Bloomfieldian in inspiration, is formal semantic analysis associated primarily with the psychological linguist Floyd Lounsbury and the anthropologist Harold Scheffler, whose work on Sirionó kinship (1971) is typical of the method, as well as having been devised as a foil to Needham’s structural analysis of the same system (1961). Neither componential analysis nor formal semantic analysis is an inspiration to me, largely because they both involve synchronic analyses that end with what is essentially the resetting of genealogical formulae in a different mode. In addition, Kroeber’s and Goodenough’s features are incidental to the sort of analyses I and my mentors have pursued, while the Lounsbury–Scheffler approach underplays the aspect of cross-cousin marriage that is the one area where terminology and social practice are clearly linked; in particular, Scheffler denies that Dravidian terminologies have anything essential to do with cross-cousin marriage (1971). In fact, cross-cousin marriage cannot be ignored in work on terminological change generally, even though it is not relevant to all actual cases, as a lot of such change seems to be associated with changes in the attitudes towards this form of marriage, causing its modification or abandonment (see Parkin 1997: 53–56 for an anonymized account of these two approaches). Dwight Read’s work (e.g. 2001) appears to continue in the same synchronic vein, except that, unlike Scheffler and Lounsbury, whose work he criticizes for its circularity of reasoning, he is concerned to find a way of representing kin ties that does not show them as genealogical by invoking what in effect appear to be categories. Indeed, far more contemporary scholars of kinship studies appear to approach them synchronically rather than diachronically or in evolutionary terms, my main reason for disregarding them here.



35

Kinship as Classification

II Abbreviations used for kin terms are those listed in the Introduction to this volume. To recap what I said there, I associate different kin terms with distinct categories, one term per category, and further distinguish the category-term, which is emic or ethnographically specific, from the notion of kin type, which is etic or analytical and comparative, though not necessarily irreducible. For example, the English term ‘uncle’ defines a category that consists of two kin types, FB and MB, and often two others, FZH and MZH. The Tamil term mama, by contrast, unites the kin type MB only with FZH of the above, and additionally covers WF and HF. In Tamil, the kin type FB has a separate term (or rather, two terms, split by age relative to ego), since to equate FB and MB would be a category mistake here with adverse moral implications. The equation MB = FZH = HF = WF is a diagnostic of a two-line prescriptive terminology associated with bilateral cross-cousin marriage; it is an equation absent from the English terminology. Finally, I generally use Needham’s paradigm of labels for terminological types or features, as set out in anonymized form in my introductory book (1997: 72 ff., starting with the last paragraph on p. 72). In particular, my use of ‘lineal’ follows Needham’s usage, not Lowie’s earlier usage, which is equivalent to Needham’s ‘cognatic’ (e.g. the English terminology); for Needham and myself, ‘lineal’ indicates the sorting of category–terms into descent lines, as generally in symmetric prescriptive or what Needham calls ‘two-line prescriptive’ terminologies (below often marked as ‘TLP’) and its immediate derivatives, three-line (asymmetric alliance) and four-line (symmetric prescriptive specifying marriage to second cross cousins). Here I am ignoring claims sometimes made in the past that, for example, six-, eight- or even ten-line terminologies occur. Usually, these can be reinterpreted as two- or four-line (cf. McKnight 1971 on Connell’s false analysis of the Wikmungkan terminology in Australia), which, together with three-line, are the only numbers of lines I am recognizing below. In this context, Needham’s notion of ‘line’ is a property of the terminology and in itself has an unavoidable unilineal bias (patrilineal or matrilineal): it says nothing necessarily about the mode of descent in the society concerned. For example, many societies in the Amazon, as well as the Sinhalese in Sri Lanka, have bilateral or cognatic (i.e. non-unilineal) descent but two-line prescriptive kinship terminolo-

36

How Kinship Systems Change

gies expressing direct exchange or bilateral cross-cousin marriage. Similarly Crow-Omaha terminologies are not always found with the mode of descent suggested by the respective vertical equations that characterize them (matrilineal and patrilineal). In the main, I also avoid ethnographically specific labels for types or features of wide incidence. One exception is ‘Iroquois’, as well as ‘Dravidian’ when I am comparing it with Iroquois in terms of how they respectively treat cross kin (explained under stage 4 below), otherwise I call ‘Dravidian’ symmetric or two-line prescriptive (below, ‘TLP’). The distinction between Dravidian and Iroquois corresponds to that between Type A (Dravidian) and Type B (Iroquois) crossness in Trautmann and Barnes (1998). Another exception is ‘Crow-Omaha’ (also ‘Crow’, ‘Omaha’, separately), for which no satisfactory ethnographically neutral term has been coined (stage 5 below). A third, occasionally, is Hawaiian for a type of equation (stage 6 below), for which ‘generational’ is to be preferred. I do not claim to be entirely consistent in these respects. As already noted, most though not all writers on this subject seem to accept that some form of two-line prescriptive terminology (TLP, associated with bilateral cross-cousin marriage) is the base point from which all other terminologies have derived historically. The basic reason is that there are no obvious signs that TLP is derivable from, or has been derived from, any other type of KT. Only logical reductions seem possible, such as Allen’s tetradic society, featuring a terminology with just four terms along two dimensions, one consisting of vertical moieties, the other of horizontal moieties consisting of sets of alternating generations (+2/0/−2 etc. opposed to +1/−1 etc.; see Allen 1986, reissued 2004, for the main statement reduced to its essentials here). One vertical moiety is marriageable for ego, the other is not; similarly, ego’s horizontal moiety is marriageable for ego, that of ego’s parents and children is not. Tetradic society can be considered either in terms of its basic four sociocentric categories or from the perspectives of individual egos. It is also sex-neutral, in the dual sense that it does not require the genders to be identified and can also be shown with either patri- or matri-moieties as the vertical dimension (or more strictly, helical dimension; see Allen 1989). The two dimensions it brings together are attested in the real world both individually and in combination but not combined with the intensity and purity the model sets out.

Kinship as Classification

What, therefore, might be the steps away from TLP? I list the main possibilities below and assume that the stage of tetradic society, which is in any case hypothetical and ethnographically unattested, has already been left behind. Stages 2 to 6, at least, should be considered alternatives, not a series. Stages 1 to 3 will normally have a prescribed spouse identified by kin term and a tendency to redefine categorically wrong marriages in accordance with the prescription. They will also not need separate terms for affines, or rather affines will be identified with some sort of cross cousin as marriageable relatives and the latter’s cognatic kin. (1) Two-line prescriptive (TLP) or symmetric prescriptive terminologies expressing bilateral cross-cousin marriage. Found with or without sociocentric groups; for example, moieties, marriage classes, sections. For Allen, sociocentric systems have an evolutionary priority, as an example of the whole of society being coordinate with the universe of kin. This accords with tetradic theory, but not all societies that have symmetric prescriptive marriage and terminology have such groups, nor are such groups necessarily connected with affinal alliance (especially moieties). Also known as ‘Dravidian’, a label that will be used on occasion below. (2) Expansion into a ‘Kachin’ three-line scheme expressing matrilateral cross-cousin marriage (although the actual Kachin have a five-line terminology; cf. Leach 1945). This involves splitting the category of cross kin plus affines into two classes of wife-takers and wife-givers while preserving both prescription (now unilateral) and cross-affine equations. Dravidian crossness (not Iroquois crossness) should therefore be consistent throughout (see stage 4 for the difference between them). Common locally in Southeast Asia and possibly Siberia, with isolated examples reported or claimed to exist in the Americas (Tsimshian, Sirionó, Txicao). Some of these terminologies have vertical equations in lines adjacent to ego’s, typically ascending (equating senior kin with wife-giving affines) or descending (equating junior kin with wife-taking affines) in accordance with the status differences between wife-givers and wife-takers that are typical of such systems of affinal al-



37

38

How Kinship Systems Change

liance. Some writers have sought to derive Crow-Omaha terminologies from this feature, though this has been contested (see further below; also Ch. 7, this volume). Only matrilateral (MBD/FZS) asymmetric prescriptive terminologies seem to occur; ‘pure’ patrilateral (FZD/MBS) asymmetric terminologies are unknown, despite the theoretical possibility and informants’ declared preferences for these kin types as spouses in some societies. The conventional labels ‘matrilateral’ and ‘patrilateral’ assume a male perspective; for female egos they should be reversed. (3) Expansion of stage 1 into an ‘Aranda’ four-line scheme (Australia); transition from stage 2 would be awkward, involving restoration of symmetric equations and mass deletions of terms, so choosing to move from stage 1 to either stage 2 or stage 3 probably represents a clear and irreversible choice. In fact, it may be easiest for stage 3 to emerge directly from an expanded variant of the unattested tetradic society (N.J. Allen, personal communication). Stage 3 bans first cross cousins in marriage but prescribes second ones, and the terminology consistently maintains Dravidian crossness, not Iroquois crossness, throughout. It also maintains cognate-affine equations. If there are sociocentric groups linked to the marriage system, they are sorted equally between unmarriageable and marriageable (see Korn’s account of the Aranda, 1973: 26–32). Apparently restricted to Australia, where they have nonetheless come under question recently (Dousset 2002). However, McConvell (1997) has suggested, on linguistic evidence, that the actual Aranda system developed out of a Kariera-type or stage 1 system no later than 1,500 years ago. Stages 1 to 3 are all undeniably prescriptive – that is, they express various types of cross-cousin marriage. (4) Iroquois: very common worldwide, including Australia, North and South America, and South Asia (see Parkin 1998 on Burushaski), though not Europe; much discussed in Godelier et al. 1998. Non-prescriptive, generally banning first cousins in marriage but allowing second, third etc. cousins without prescribing them (cf. Viveiros de Castro 1998: 364). Iroquois terminologies lack cognate-affine equations, thus apparently

Kinship as Classification

making it difficult to maintain Dravidian (or TLP) crossness consistently beyond first-cousin range in ego’s level, PG(E) range in +1 or (E)osGC range in −1 – that is, the minimal genealogical specifications within these ranges (cf. Hornborg 1998: 179). They therefore tend to rely on absolute-sex rather than relative-sex distinctions in determining crossness beyond these ranges, whereas Dravidian does the reverse, and does it consistently throughout the terminology (see the table in Tjon Sie Fat 1998: 69). Iroquois is therefore to be distinguished from Dravidian, as well as the genuinely prescriptive form of second cross-cousin marriage represented by the Aranda in stage 3 above (cf. Viveiros de Castro 1998: 356). Another difference is that the ‘Aranda’ system neatly divides marriageable from unmarriageable groups (there are equal numbers of each), whereas Iroquois tends to treat all groups as potentially or actually affinal, apart from ego’s (cf. ibid.: 360). This encourages a move away from sociocentric groups (if any) towards super-categories of consanguines and (potential and/or actual) affines. As we have just seen, first-cousin marriage will generally be banned, but the system may nonetheless allow (but not prescribe; cf. Viveiros de Castro 1998: 364) at least some second, third etc. cousins as spouses. This often seems to be associated with generational delays in the ability to renew alliances between groups or lines, in the fashion of the Munda (Parkin 1992: Ch. 8; also this volume, Chs 2–5), though decisive proof is still lacking regarding whether or not any Munda terminologies are Iroquois. In Viveiros de Castro’s words (1998: 355–56), this suggests ‘longer cycles of repetitions of the exchange’ than in Dravidian. However, this last feature has also occasionally been associated with north Indian (see Parkin 1998, on the basis of ethnography by Parry 1979, Krengel 1989) and Crow-Omaha (by Lévi-Strauss 1966, Héritier 1981 etc. [the Samo case]; see stage 5, below), opening up the question of transformational links between these terminological types and Iroquois. The property of allowing marriage to second etc. cousins but excluding first cousins may be a development of an earlier trend for a prescriptive system to disallow genealogical first cross cousins but to go on permitting or prescribing remoter ones. The change then becomes formalized terminologically, producing Iroquois and



39

40

How Kinship Systems Change

no longer Dravidian crossness and breaching affine-cognate equations. Affines may retain the old composite terms, leaving cross cousins to find substitute terms (thus preserving the main sense of difference between consanguines and affines). Apart from coining entirely new terms, which in principle is always possible (perhaps leading to zero-equation terminologies; see 8, below), there are in principle two other ways of coping with this shift, namely by extending other terms to cover cross cousins, either vertically or laterally (see respectively 5, Crow-Omaha, and 6, Hawaiian, below), and by borrowing vocabulary from another language (e.g. Munda loans from Indo-European, Parkin 1992: Ch. 7). The hypothesis of a shift from prescription to Crow-Omaha is less certain and needs testing, though Trautmann and Barnes (1998) have hypothesized a shift from Iroquois to Crow-Omaha in North America (see Ch. 7, this volume). Some terminologies combine ± 1 Iroquois with Hawaiianization in ego’s level, a situation called ‘Cheyenne’ in North America, ‘Tupian’ in South America, and ‘bifurcate generational’ by Dole (cf. Viveiros de Castro 1998: 384 n. 63; Dole 1969: 106). Several commentators have remarked on the difficulties in establishing the nature of the crossness of particular systems (Allen 1998: 327–28; Héritier 1981, cited by Viveiros de Castro 1998: 345). Viveiros de Castro’s answer (ibid.) is that this is so but that the easiest way of doing this is to examine the terminological equivalences of remoter relatives with primary kin (including first collaterals). The problem seems especially acute regarding Iroquois crossness, but it also affects how to diagram the +1 level with Dravidian first-generation collaterals and lineals (cf. Viveiros de Castro 1998: 337–38; Allen 1998: 317, 331 n. 4; Parkin 1997: 81). For Viveiros de Castro (1998: 369), Dravidian in South India treats M as parallel kin, FZ as cross, while Australian systems reverse this. This has been rejected for Australia (e.g. Dousset 2002). The problem is marriage, as (assuming patrilineal descent and group exogamy) mothers marry into ego’s group from their natal group, while father’s sisters marry out of ego’s group into their conjugal group. They therefore straddle the divide between marriageable and non-marriageable groups. No diagram can cope with this conflict, and a choice therefore has to be made regarding how to treat them.

Kinship as Classification

(5) Crow-Omaha or C-O: very common worldwide, possibly including in Europe, at least historically (e.g. the linking of MF, avus, with MB, avunculus, in Latin). In C-O, cross cousins are linked with kin in adjacent levels whom it would typically be incestuous to marry (see Parkin 1997: 110–11 for details). This does not explain the ‘matrilineal’ (Crow) and ‘patrilineal’ (Omaha) skewing of the two forms, however. Needham’s view (1971) was that they cannot be seen as a discrete class of terminologies, reminding us of the failure to associate them with any particular sociological feature (cf. McKinley 1971a; also Barnes 1976, 1984, who criticizes McKinley’s own ‘sociology of knowledge’ explanation, McKinley 1971b). Nonetheless, as equations they exist, and my own working hypothesis here is that they may be explicable as a way of effecting change in many other, more formal types of terminology (cf. Allen 1975 on Byansi in the Himalayas, 1976 on Sherpa). On that basis, other, non-C-O generational merging might be included here (i.e. without the lateral bias of C-O); for example, in ego’s own line (see Needham on Gurage, 1969). One hypothesis in terms of Crow-Omaha/Iroquois comparisons that needs further testing is the apparent tendency for Crow-Omaha patterns not to separate second etc. cousins from first cousins. For Trautmann and Barnes (1998), as we have seen, C-O features are likely to have emerged from Iroquois-type terminologies, at least in North America (opposing Kronenfeld’s argument [1989] that they derive from Dravidian crossness). Barnes has repeatedly pointed out (1984; also 2012) that Héritier’s definition of the Samo terminology of W. Africa as Omaha (1981) ignores important differences between it and the actual Omaha terminology. A new collection has recently come out revisiting the problem of interpreting C-O terminologies (Trautmann and Whiteley 2012; cf. Chs 7, 8, this volume). (6) Generational or Hawaiian (Murdock): very common worldwide, except in Europe. In particular, in developmental terms, PosGC come to be identified with G = PssGC terms, perhaps leaving their old terms to cover affines only where these had previously been applied to PosGC as well. This will normally signal the unmarriageability of PosGC, who are now equated



41

42

How Kinship Systems Change

with both PssGC and siblings. This process may eventually extend to −1, then +1 (typically in that order; see below; cf. also my remarks on Witowski, above). This stage could in principle emerge directly from any of the above schemes, apart, perhaps, from 5 (C-O). Being non-lineal in Needham’s terms, this scheme represents a definite breach with prescription and the ‘post-prescriptive’ types discussed above (1 to 5). It is also now less likely to be treated as a ‘type’ in its own right than it was by Morgan. For Morgan, it was the simplest terminological type (which he called Malay) and evolved into, not from, Iroquois (my stage 4, his Turanian). (7) Bifurcate merging changing into bifurcate collateral – that is, breach of lineal–same sex collateral equations, such that, for example, the pattern F = FB = MZH becomes F ≠ FB = MZH. Can come at any stage. Not very significant in itself diagnostically but may lead on to next stage if all collateral kin come to be distinguished through separate terms. (8) Zero equation (Allen; Murdock’s Sudanese). Each primary kin type, as well as each first collateral, has its own term (e.g. F, FB, MZH, MB, FZH), though, as in Hindi and other north Indian terminologies, there may be equivalences between each such kin type and remoter relatives (see Parkin 1998, also Ch. 5, this volume). In other words, it can be seen as a development of bifurcate collateral in which the remaining PG(E) equations, a relic of prescription, are broken up and replaced by isolating terms (e.g. F ≠ FB = MZH [stage 7 above] becomes F ≠ FB ≠ MZH, and MB = FZH becomes MB ≠ FZH). There are therefore likely to be a relatively large number of terms (as in Hindi). Also found in European history (Latin, German, Slavonic), and still today in Scandinavia (but cf. Swedish and Danish) and eastern Europe (Polish, Hungarian). Periphrastic phrases for some terms may occur (e.g. Arabic bint [D] amm [F, FB] for FBD), rather than primary terms (i.e. periphrastic terms do not constitute a separate type, despite Murdock’s Sudanese). Arabic should probably be placed here, associated with FBC marriage. Many African terminologies are also zeroequation ones, while others show Hawaiian features, but Ehret’s data (2008) also show bifurcate collateral and bifurcate

Kinship as Classification

merging features, with a tendency for +1 and −1 to be more conservative than ego’s genealogical level (i.e. bifurcate rather than Hawaiian, in line with Witowski’s predictions). (9) Cognatic (Needham; Lowie’s lineal). Historically may represent a re-sorting of terms after a terminology has reached stage 8, such that lineal kin are consistently distinguished from collaterals, which may well share a single term. For example, from a hypothetical pattern of five separate terms for F, FB, MZH, MB and FZH, a re-sorting may take place that ends up with a more economical two-term pattern of father ≠ uncle. This, of course, would involve the deletion of three terms (assuming the other two were applied to father and uncle respectively). The terminology therefore lacks the descent lines (Needham) of prescriptive and other post-prescriptive systems (stages 1–5). This deletion of terms may be happening in Polish, where stryj (FB) seems to be giving way to wuj as a term for ‘uncle’ rather than just MB (Parkin 1995), and it appears to have been a common transition in Europe that some terminologies have not effected or not yet. Pace Murdock, this terminological pattern is not exactly the same as ‘Eskimo’, since the latter tend to retain classificatory equations in +2 (cf. Dole 1960), though it is not only found in ‘complex’ and/or modern societies. There is some intergenerational but non-lineal merging, especially of terms for cousins, while in Italian and Romanian GC = CC. English has separate terms for GC and CC and only ‘cousin’ as an intergenerational term so is a pure example of the type. The end point of terminological development? It is certainly hard to conceive of a further logical development away from English etc. that is not already covered in previous stages, though only time will tell whether this assumption is correct. One practice, sometimes identified separately, namely oblique marriages of the type MB/ZD and BS/FZ that occur locally in the Americas, especially South America – also standard in South India according to Good (1980) – should really go with TLP, since such marriages are basically an alternative to bilateral cross-cousin marriage within the same system – that is, they tend not to occur on their own as an autonomous type (cf. Parkin 1997). As for FBC marriage, associated



43

44

How Kinship Systems Change

mainly with Islamic societies in the Middle East and Western and Central Asia, I have tentatively placed this under stage 8. Although clearly a marriage preference, this practice cannot be seen as a prescriptive system in the sense of stages 1 to 3. Repeated FBC marriage does not produce a closed formal system, as stages 1 to 3 do (see Parkin 1997: 119), nor is there any tendency to redefine marriages to other kin according to a terminological requirement (indeed, since it is only a preference, other kin types can be freely married without being redefined, except as ‘spouse’, by the terminology). Indeed, there is no terminological scheme associated with this marriage preference, which coexists with the Arabic terminology in the Middle East but with the quite different Urdu or Punjabi terminologies (both Indo-European) in Pakistan (more periphrastic phrases in Arabic, more primary terms in Urdu and Punjabi, though in all of them FBD can only be identified periphrastically, e.g. Arabic bint ‘amm). Moreover, the preference relates to genealogical referents, not a category, as usually with cross-cousin marriage and its immediate successors. However it arose, therefore, there is no systemic link between terminology and marriage practice, and it is best placed in stage 8, a non-prescriptive stage, on the basis of the zero-equation tendencies of the terminologies with which it mostly coexists ethnographically. Ehret suggests that FBD marriage was present in the proto-Semitic speech community as long ago as 6,000 years BP (2008). Stages 1 to 5 are linked by three things that stages 8 and 9 lack (6 and 7 being transitional in this view). First, they are either fully prescriptive, with cognate-affine equations and consistent Dravidian crossness (1, 2 and 3), or else derived from prescriptive systems, still expecting, preferring, allowing or tolerating (but not prescribing) marriage with remoter cousins and/or other ‘relatives’ but without cognate-affine equations or the terminological redefinition of wrong marriages (though Iroquois is uncertain here). The hypothesis of derivation from prescription is more speculative in the case of C-O (5) than Iroquois (4) and needs further testing. For Trautmann and Barnes (1998), as we have seen, it is more likely that C-O derives directly from Iroquois, at least in North America. Kryukov, however, claims that C-O should be derived from asymmetric prescriptive (3 above), not TLP (1998: 312). This revives similar earlier hypotheses by Lane and Lane (1959) and Eyde and Postal (1961), also more recently by Trautmann (2012), but rejected earlier by Coult (1965) and Barnes (1984). Since intergenerational equations may occur

Kinship as Classification

with both symmetric and asymmetric schemes, and since they are one way of effecting change away from prescription (e.g. Byansi and Sherpa; Allen 1975, 1976), the matter is not so clear and needs revisiting. However, as already noted above in part, whereas Iroquois is practically defined by its separate treatment of remoter cousins than first cousins, C-O terminologies do not seem to make such distinctions (another hypothesis that needs further testing). Secondly, stages 1 to 5 all have a bifurcate merging or bifurcate collateral pattern in at least +1, and possibly other levels too. Though not very useful in identifying different systems, precisely because many of the latter share these two patterns (cf. Kryukov 1998: 311), the threshold between their (partial) presence and complete absence does seem to mark the threshold between prescriptive and prescriptive-related (or post-prescriptive) systems on the one hand and those terminological schemes or types (8, 9) that have no systemic association with marriage rules and preferences on the other. The latter will, of course, distinguish affines from cognates post-marriage but not also marriageable from unmarriageable kin pre-marriage, as in a prescriptive system (incest rules generally ban unmarriageable kin, but not all societies specify marriageable kin terminologically). In other words, in these later stages affines exist after the marriage, not also before it, as is formally true of prescriptive systems. There is therefore no term for spouse before the marriage, as formally with prescriptive systems (or rather, it is the same as the term for the prescribed category), only afterwards. Hawaiian (stage 6) may eliminate all bifurcation too, thus crossing the above ‘threshold’, or may preserve it in +1 but eliminate it in ego’s level and possibly −1 only, thus remaining transitional. As already noted, although Morgan treated Hawaiian as a distinct type of terminology, there is less of a tendency to do so today. Thirdly, stages 1 to 5 represent terminologies that are sorted into descent lines (in Needham’s sense, i.e. including, but certainly not restricted to, actual vertical terminological equations – for example, the patriline formed by MF, MB, MBS, MBSS), whereas 6, 8 and 9 are not. In Lévi-Straussian terms, stages 1 to 3 are elementary systems (or structures, stage 5 semi-complex structures, stages 6 to 9 complex structures (Iroquois not being separately identified in his paradigm)). As already noted, bifurcate patterns often survive longest in +1. Indeed, Kryukov suggests that while ego’s level may be more subject



45

46

How Kinship Systems Change

to merger (Hawaiianization), +1 is more susceptible to differentiation (1998: 313), or at least to maintaining differentiation. This may be because, formally speaking at least, this level represents marriages that have already taken place. More particularly, it expresses past sibling exchanges in marriage, which can be found in some form in both prescriptive systems and those immediately derived from them (at least 4, if not 5). However, where this form of exchange cannot be repeated in the following generations between the same groups or lines, a different terminological pattern is likely to emerge in descending levels, starting with ego’s and then proceeding to −1. Whatever the merits of this particular argument (which I partly derive from Viveiros de Castro 1998), certainly once the +1 level ceases to be bifurcate we can expect that the terminology has lost all systemic association with marriage, which is now likely to be a ‘complex structure’ in Lévi-Strauss’s terms and not directed by the terminology – that is, with open choice of spouse apart from incest prohibitions and constraints through, for example, property, status or opportunity. Stage 8, with its zero-equation pattern and possible periphrastic phrases for at least some kin types, is probably relatively unstable and tends to drift into stage 9, as has certainly happened repeatedly in Europe. To repeat, stage 9 really does seem to represent a certain end point: just as no attested system seems to evolve into TLP, so nothing seems ever to have evolved away from stage 9. Indeed, it is hard to envisage any further development, not least because most if not all of the logically possible patterns of kin-type equation and distinction are accounted for in earlier stages. Stage 9 cannot evolve directly from any of the genuinely prescriptive systems (1 to 3), nor from Iroquois or C-O, but may do so successively through 7 and 8, or else directly from stage 6 through the introduction of lineal– collateral distinctions (this is a formal possibility, though it may also explain the cognatic features of Murdock’s ‘Eskimo’ type; this hypothesis needs further testing). Generally, the assumption is that all these changes are irreversible (cf. Kryukov 1998: 313 on all these points; note that he uses Lowie’s ‘lineal’ in place of Needham’s ‘cognatic’; see also Godelier et al. 1998: 406). As already suggested above, derivation of stage 9, cognatic, from stage 8, the zero-equation, many-term scheme, would involve the deletion of some terms; for example, of FB in favour of the term for MB becoming a global term for ‘uncle’ (indicated for Polish; Parkin

Kinship as Classification

1995). The argument that TLP cannot be derived from anything else because it would involve deleting terms, which is less likely than their expansion (Allen et al. 2008: 108), therefore seems not to be applicable to the transition between stages 8 to 9. In fact, the ‘Hawaiianization’ of terminologies (stage 6), in which intra-cognate distinctions within a genealogical level are removed, also involves a process of deletion. In practice, of course, redundant terms simply fall out of use, rather than being deliberately or consciously deleted. Some caveats should be made at this point in the chapter. First, as should by now be apparent, these nine stages are not an inevitable series through which all societies either have gone or will go. In any case, that would be close to impossible. Once some changes have taken effect, other, alternative ones appear to be ruled out. Thus it is hard to conceive of a society changing from stage 1 (TLP) to stage 3 (Aranda) via stage 2 (Kachin), for reasons already given; similarly, Hawaiianization may rule out change into Crow-Omaha and vice versa because either change means making a definite choice between lateral and lineal extension. Secondly, while with few exceptions (mainly urban, literate South Indians) prescriptive systems involving cross-cousin marriage are found in relatively small tribal societies with simple traditional technologies and minimal or no division of labour, many societies that can be defined in this way do not have such systems and may indeed share the same terminological principles with a society in the more technologically and economically developed world, as with the famous near-match of Inuit (‘Eskimo’) terminologies with English. One striking instance of this is hunter-gather societies, which, outside Australia, usually lack such systems, despite both they and the systems being seen as having an evolutionary priority (often falsely as far as specific present-day hunter-gather populations are concerned). This suggests that grand paradigms of change are less likely to be useful than specific histories of change in limited regions of the world. This recalls Allen’s observation (2008), already cited above, that the reasons for change are likely to be local. It also suggests that work in this area is more likely to give rise to middle-range theories of the sort Jack Goody adopted into anthropology from Robert Merton’s sociology, not grand theories in the manner of Radcliffe-Brown’s functionalism, Lévi-Strauss’s structuralism or David Schneider’s cultural relativism. Nonetheless, Allen himself sees grand theories of world-historical change as still a possibility (personal communication).



47

48

How Kinship Systems Change

III In the context of terminological change, the connection between prescriptive terminologies and prescriptive marriage systems based on cross-cousin marriage requires we address the problem of what makes a society abandon cross-cousin marriage, with the possible implications of this for the associated terminology (see also Ch. 4, this volume). As just noted, Allen has argued (2008) that such explanations are likely to be local rather than global or universal and need investigating separately in each case. Outside elite pressure is certainly one factor, as with missionary influence condemning the marriage of cross cousins or upper-caste dislike of any cousin marriages in India (especially in the north) influencing the lower castes and tribes, many of the latter especially having cross-cousin marriage in some, often attenuated form. However, that does not explain how change started in the first place – that is, how these elite ideas emerged from a hypothetical situation of universal cross-cousin marriage possibly even in prehistory. As regards Native North American huntergatherers, one relatively recent hypothesis is that TLP terminologies and cross-cousin marriage are found in small bands that maintain band endogamy, while Iroquois crossness appears with larger bands that tend towards band exogamy, since exogamy permits, or forces, the search for a wider range of potential spouses than cross cousins, who in this scenario are apparently found within the band (Ives 1998). Trautmann and Barnes (1998: 54–55) make a similar claim, also for North America, including Crow-Omaha systems with Iroquois for this purpose. This may be true of North America, but it can hardly be a universal explanation, as many societies around the world with the TLP complex are agricultural and divided into exogamous descent groups as the main operators of cross-cousin marriage, meaning that one’s cross cousins are in other descent groups. The debate between Dole (1969) and Basso (1970) mentioned earlier similarly revolved around the question of whether the Hawaiianization of ego’s level among the Xingu reflects a situation in which population decline has led to group endogamy (Dole) or to group exogamy (Basso). Again, whatever the case, it can only be a local explanation. Lastly here, Kronenfeld (1989: 96 ff.) has speculated on the possible reasons for a change from Dravidian (or TLP) to Iroquois, claiming that an Iroquois terminology is easier to handle when it comes to tracing kin ties genealogically and that the

Kinship as Classification

imperative to trace them would arise if the fit between genealogy and social category he claims to accompany Dravidian terminologies broke down. It is not entirely clear what he means by ‘social category’ at this point in his article: Dravidian-type terminologies are not necessarily associated with sociocentric groups like marriage sections, though they do typically distinguish lineal/parallel kin from cross kin/affines. In any case, Kronenfeld does not suggest why such changes might take place. His reliance on social morphology in explaining terminological types has been briefly criticized by Trautmann and Barnes (1998: 57, n. 6). As Needham was at pains to point out in several of his analyses of systems of affinal alliance (e.g. 1967), there is no necessary association ethnographically between marriage practice and the terminology one would expect from it. While there may be limits to such variation, a prescriptive terminology can occur with a ‘complex’ system of marriage, a preference for cross cousins as spouses may occur with a non-prescriptive terminology, and a symmetric or twoline prescriptive terminology may occur with an asymmetric pattern of affinal alliance based on MBD/FZS marriage (though not, it seems, the reverse). These combinations of inconsistent features themselves suggest change. As a general hypothesis, therefore, we might argue that human populations may start as tetradic, then move to TLP, before changing into one of the other stages set out above. This may be to another form of prescription, such as stages 2 and 3: there is evidence of stage 1 to stage 2 transformations in Southeast Asia, especially eastern Indonesia (i.e. from symmetric to asymmetric or twoline to three-line), and evidence of stage 1 to stage 3 transformations in Australia, where the latter appear to be concentrated (i.e. from two-line to four-line). Reports (e.g. by Turner 1977: 40; McConvell 1997; Dousset 2002) indicate that different Native Australian societies have influenced one another in the adoption of different affinal alliance systems, even since contact with Europeans – that is, very recently, given the 60,000-year history of human settlement in Australia – and that these influences have been indigenously recognized for what they are and discussed as such, suggesting that people are conscious of them. Alternatively, transition from TLP may be out of prescription altogether, certainly to Iroquois, and possibly to C-O. In the case of Iroquois, one can trace the breakdown of prescription logically in the very differences this type displays from Dravidian or TLP. C-O, by contrast, seems remoter from TLP, and as we have seen,



49

50

How Kinship Systems Change

direct transformations between stages 1 or 2 and stage 5 are uncertain and controversial. In principle, one possible reason for the abandonment of crosscousin marriage may be connected with an emerging dissatisfaction with the alleged closeness of ego and alter as spouses. While potentially relevant, this hypothesis is in need of serious qualification. It has long been recognized that cross-cousin marriage does not typically apply to first cross cousins alone but also to second etc. cross cousins and even alters who are not clearly linked to ego at all genealogically, but only as members of a marriageable category. This is the classificatory principle at work, in which a term applies not just to the logically most reductive or closest kin type but to their remoter equivalents too on the principle of same-sex sibling equivalence (and therefore not only involving cross cousins). This principle goes right back to Morgan, was dismissed by Kroeber (1909) and was revived in White’s defence of Morgan (White 1958). It is distinct from the prescription of second cousins as spouses (stage 3 above), although the classificatory principle applies to kin in these cases too. In any case, demographically it would be impossible for a system of cross-cousin marriage to work if it applied solely to first cross cousins (Needham 1959: 127; 1960: 105), who may even be banned in particular prescriptive systems; for example, the Wikmungkan prohibition of the genealogical FZD (Needham 1962; McKnight 1971). The dislike of close kin as marriage partners is definitely an aspect of how many societies look at marriage but so is a preference for them, and this distinction cannot be simply mapped onto that between prescription and its absence. Perhaps the most that can be said is that the stability of systems of cross-cousin marriage relies on first cousins being included in the prescription: where they are not, terminology and marriage practice may begin to part company. Examination of genealogies, where available, can reveal the extent to which genealogical cross-cousins or other prescribed spouses are actually married, a good example being Rivière’s study of the Trio of Guyana (1969). Certainly, Lévi-Strauss’s notion of semi-complex structures, in which close kin are banned in marriage with reference to their membership of a range of lineages, does have ethnographic validity. One way of describing the formal features of systems of cross-cousin marriage is to say that (assuming patrilineal descent) a man repeats the marriage of his father, in that he takes his wife from his mother’s kin group, although this is a formal property of the model and not nec-

Kinship as Classification

essarily an ethnographically universal practice. Semi-complex structures, by contrast, may well prohibit one’s mother’s lineage or other kin group as a source of a spouse for oneself, as well as other lineages (e.g. those of MM, FM). This is often linked to C-O terminologies but not necessarily, as the case of north India indicates (with its zeroequation rather than C-O terminologies), and it usually means that one should not repeat a marriage into a particular kin group within a certain number of generations, often specified as a norm or rule. Thus in north India, alongside the so-called four-got rule (a got being a patrilineage), which prohibits ego from marrying a woman from any of the got of his four grandparents (incidentally ruling out firstcousin marriage), there is also the sapinda rule (sapinda = ‘relations’), banning marriages to anyone related to ego within seven generations on the father’s side and five on the mother’s (often reduced in practice). One additional feature that is often found in such cases, though it is less discussed in the literature, is a preference or tendency to remarry into such kin groups after the prohibited number of generations has passed. In other words, it is good to marry into a previously related kin group, but not too closely (e.g. Parry 1979; Parkin 1992: Ch. 8). Even less remarked is the possibility that alliances between sibling groups within the same generation may be intensified, as among the Munda, where many tribes in this language family of east-central India describe their marriage practices in terms of marrying a sibling’s spouse’s sibling (equivalent to a cross-cousin with prescription; Parkin ibid.); while current marriages may not repeat those of previous generations, or be repeated in the following generations, they may follow one another quite intensely within the same generation. The delay in immediate alliance renewal between groups also has the effect of dispersing the alliances of one’s kin group among a number of other kin groups over time, with possible implications for power, wealth or prestige. This dispersal is not itself unique to this sort of situation and can also be found with prescriptive crosscousin marriage of all sorts, in which, sometimes, it is sufficient for one sibling to marry according to the rule. However, in the situation I am currently discussing, exemplified by north India, it does mean that while one’s spouse may conceivably be definable as a remote cousin, marriage practices are not going to be defined by marriage to cousins, including cross cousins. In north India, indeed, cousin marriage of any sort is widely disliked, in contrast to the south, which is famous for its canonical examples of TLP terminologies and bilateral



51

52

How Kinship Systems Change

cross-cousin marriage. In north India, closeness of relationship is by definition a restrictive factor in choosing a spouse in a way it is not with prescriptive systems, even though particular societies with prescription, such as the Wikmungkan of Australia, already mentioned, may ban a genealogical first cross cousin (here FZD) despite their bilateral prescriptive system (Needham 1962; McKnight 1971). The desirability of marrying closely related kin is, of course, a practice of elite groups the world over, culminating in sibling marriages in some rare but frequently cited cases but also being an aspect of upper-middle-class marriages in nineteenth-century Britain (cf. Kuper 2008 on the tangled relations of the Wedgewoods and Darwins). None of this typically involves terminological prescription, much though it may involve cousins. It is therefore difficult to see questions of the genealogical closeness or distance of kin as spouses as an immediate factor in terminological change, nor is it the exclusive property of a certain type of society. As for the emergence of zero-equation and especially cognatic terminologies, this may be linked with a situation in which lineal distinctions or those between patrikin and matrikin cease to be operationally relevant, despite a residual preference, perhaps, for the former as more salient than the latter in people’s identity formation (e.g. transmission of surnames). The indication in Polish that stryj as FB is giving way to wuj (formerly just MB) as uncle is one example of this trend (Parkin 1995). Marriages to close kin may therefore be a bit of a red herring in this context, though they have occasionally had a peculiar impact on anthropological theory. For instance, it is just possible that the oddity that Morgan married his own cross cousin prevented him from associating Dravidian terminologies with cross-cousin marriage in order to preserve his own self-image as a civilized gentleman who could not conceivably marry like a Native American (Trautmann 1984). More fundamentally, Morgan linked both Iroquois and Dravidian terminologies with the very nineteenth-century notion of group marriage, not with cross-cousin marriage, which the Iroquois lack. This involved downplaying the differences Morgan himself recognized between the two types of terminology in favour of their similarities, also made necessary by his desire to use terminological types to prove his theory of the Asiatic origins of Native Americans (Dravidian was seen as an Asian version of North American Iroquois in this argument; cf. Trautmann ibid.; 1981: 71–72). It was left to Rivers to make the link between Dravidian terminologies and cross-cousin marriage



53

Kinship as Classification

(Trautmann ibid.: 74). However, neither Morgan nor Rivers was very concerned with changes in the terminologies themselves, which they used mainly as a proxy to suggest change in other parts of the kinship system, especially forms of marriage (cf. Parkin 1997: 162). There is another way of looking at the impetus for change. By definition, terminological change is a matter of alterations in how kin are defined. Many if not most societies that are prescriptive or immediately post-prescriptive in type (i.e. tetradic society, as well as stages 1–4 and possibly 5) define everyone within them as kin, regardless of the possibility or otherwise of tracing genealogical links between different individuals. However, other societies restrict the circle of kin, very often on the basis of genealogical traceability, defining other members of the society as lying outside that circle and using other sorts of category (professional, ethnic, friendship etc.) to place them in one’s social universe. This is typically the case for modern societies with a well-developed division of labour, so that for many people one is defined by one’s profession or other non-kinship role rather than a kinship status, though it is not restricted to them. South India is an example where this situation goes along with terminological prescription and prescribed cross-cousin marriage, but more generally it is probably the case that the emergence of non-kin-based categories of significant others leads to a situation in which one typically marries someone not previously defined as related to oneself. This in turn makes terminological prescription redundant. Contemporary reports (Fuller and Narasimhan 2008; Kapadia 1993) of changing attitudes to marriage in South India indicate that this may be happening here too, as cross-cousin marriage is falling out of favour in at least some circles of what nonetheless remains a complex and stratified social system in which class is emerging as an influence alongside caste. This situation invites a longitudinal study to confirm this change as well as to test the proposition that change will ensue to the terminology as well.

IV Returning to the start of this transformational paradigm, it is reasonable to ask at what stage of prehistory TLP terminologies emerged and became established and whether the first stage was indeed something like tetradic society. Apart from some eighteenth-century precedents, usable historical ethnographic data on this hardly go back

54

How Kinship Systems Change

beyond Morgan’s researches in the middle of the nineteenth century, but we can nonetheless assume that such terminologies have a long history before that, possibly going back millennia, though in the absence of direct evidence we can hardly know for certain. However, Hocart long ago sung the praises of circumstantial evidence as every bit as useful and reliable as direct evidence (1936: Ch. 1). Tetradic theory uses precisely circumstantial evidence in that it extrapolates backwards from attested terminologies and marriage systems to a putative ancestor. Moreover, it also postulates that the four categories designated by the hypothetical four kin terms need not be labelled linguistically but could be marked in some other way such as by colour differences (Allen 2008), suggesting a further step backwards in time, before humans had language (60,000 years BP?). Other social anthropologists have noted Allen’s initiative and commented critically on it. Barnard (2008), who sees phases in the development of language as more or less paralleling phases in the development of kinship – the latter phases represented respectively by the ideas of Morgan, McLennan and Lévi-Strauss – would seem to regard language as a more likely means of expressing the tetradic form. He also recognizes the potential of tetradic theory but does not see it as the only possible ancestor. Layton (2008) goes a step further to deny that the simplicity of tetradic society gives it a historical priority but does not develop the point. At the very least, whether one accepts tetradic society as prior to TLP or not, humans with either must have developed the ability both to classify kin through a terminology, whether linguistic or not, and to formulate rules, or at least recognize norms, regarding marriage practice, as Meyer Fortes pointed out in what was probably his very last work (1983). As Gamble has pointed out (2008: 33), this clearly distinguishes non-human primates from at least later hominins: ‘Negotiation to form alliances and cliques is a primate trait . . . while formal recruitment to pre-existing categories is not [i.e. is human].’ Allocation to such categories is also at root mandatory, despite their potential for negotiation and metaphor. Gamble’s insight goes along with oft-repeated emphases that primate kinship can only be studied biologically – only among humans is kinship also social, in the sense of requiring both rules and classification, and only living humans can discuss it with one another and with the anthropologist. If the study of human biology can put dates to the emergence of these intrinsically human features (and there still seems to be wide-

Kinship as Classification

spread disagreement regarding the dating of this emergence, as well as what form or forms it took; see Allen et al. 2008), it will give us some idea of how far back kinship terminologies as classifications may go, even if we can never be entirely certain of the form they took originally. Recent developments in cognitive science and the links being forged between it and social anthropology, evolutionary biology, evolutionary psychology and linguistics (on the latter, see Jones et al. 2010) are increasing the potential for this to happen, as some of the chapters in the recent volume Early Human Kinship (Allen et al. 2008) and the Per Hage memorial volume ( Jones and Mililic 2010) indicate. However, the basic distinctions in this form of evolutionism are those between distinct hominin and by extension primate species, not those between the stages in terminological change I have postulated in this chapter: they therefore do not constitute a direct model for terminological changes. The latter all apply to modern humanity – that is, to homo sapiens in the categories of evolutionary biology; they therefore assume the same cognitive and classificatory faculties on the part of all the different speech communities that produce and use them. Many anthropologists, after all, have openly marvelled at the ability of certain ethnic groups in places like Australia to devise complicated systems of cross-cousin marriage that they (the anthropologists) have great difficulty in grasping. How such systems arose and developed is as legitimate a question as their principles and practice in the here and now. It is nonetheless a question that recognizes difference and diachronic development only within the parameters of a common humanity and humanity’s common cognitive, classificatory and rule-generating abilities.

Notes * Originally published in the Journal of the Anthropological Society of Oxford Online 4(2) (2012): 183–211. Reproduced with permission.

References Allen, Nicholas J. 1975. ‘Byansi Kinship Terminology: A Study in Asymmetry’, Man 10: 80–94. ———. 1976. ‘Sherpa Kinship Terminology in Diachronic Perspective’, Man 11: 569–87.



55

56

How Kinship Systems Change

———. 1986. ‘Tetradic Theory: An Approach to Kinship’, Journal of the Anthropological Society of Oxford 17(2): 87–109. ———. 1989a. ‘Assimilation of Alternate Generations’, Journal of the Anthropological Society of Oxford 20(1): 45–55. ———. 1989b. ‘The Evolution of Kinship Terminologies’, Lingua 77: 173–85. ———. 1998. ‘The Pre-history of Dravidian-Type Terminologies’, in Maurice Godelier et al. (eds), Transformations of Kinship. Washington DC and London: Smithsonian Institution Press, pp. 314–31. ———. 2004. ‘Tetradic Theory: An Approach to Kinship’, in Robert Parkin and Linda Stone (eds), A Reader in Kinship and Social Organization. Malden: Blackwell Publishing, pp. 314–31 (slightly revised version of Allen 1986). Allen, Nicholas J., Hillary Callan, Robin Dunbar and Wendy James (eds). 2008. Early Human Kinship: From Sex to Social Reproduction. Malden and Oxford: Blackwell, for the Royal Anthropological Institute. Barnard, Alan. 2008. ‘The Co-evolution of Language and Kinship’, in Nicholas J. Allen et al. (eds), Early Human Kinship: From Sex to Social Reproduction. Malden and Oxford: Blackwell, for the Royal Anthropological Institute, pp. 232–43. Barnes, R.H. 1976. ‘Dispersed Alliance and the Prohibition of Marriage: A Reconsideration of McKinley’s Explanation of Crow-Omaha Terminologies’, Man 11: 384–99. ———. 1977. ‘Alliance and Categories in Wailolong, East Flores’, Sociologus 27(2): 133–57. ———. 1980. ‘Comment on Blust 1980’, Current Anthropology 21(2): 231. ———. 1984. Two Crows Denies It: A History of Controversy in Omaha Sociology. Lincoln and London: University of Nebraska Press. ———. 2012. ‘What is Left Out in Kinship’, in Laurent Dousset and Serge Tcherkézoff (eds), The Scope of Anthropology: Maurice Godelier’s Work in Context. New York and Oxford: Berghahn, pp. 191–205. Basso, Ellen. 1970. ‘Xingu Carib Marriage and Kinship Terminology: Another View’, Southwestern Journal of Anthropology 26: 402–16. Beck, Brenda. 1972. Peasant Society in Konku. Vancouver: University of British Columbia Press. Berlin, Brent, and Paul Kay. 1969. Basic Color Terms: Their Universality and Evolution. Berkeley, CA: University of California Press. Blust, Robert. 1980. ‘Early Austronesian Social Organization: The Evidence of Language’, Current Anthropology 21(2): 205–47. Coult, Alan. 1965. ‘Terminological Correlates of Cross-Cousin Marriage’, Bijdragen tot de Land-, Taal en Volkenkunde 121(1): 120–39. D’Andrade, Roy. 1995. The Development of Cognitive Anthropology. Cambridge: Cambridge University Press. Dole, Gertrude. 1957. ‘The Development of Patterns of Kinship Nomenclature’, PhD thesis. Ann Arbor: University of Michigan. ———. 1960. ‘The Classification of Yankee Nomenclature in the Light of Evolution in Kinship’, in Gertrude Dole and Robert Leonard Carneiro (eds), Essays in the Science of Culture in Honor of Leslie A. White. New York: Thomas Y. Crowell, pp. 162–78. ———. 1969. ‘Generation Kinship Nomenclature as an Adaptation to Endogamy’, Southwestern Journal of Anthropology 25(2): 105–23.

Kinship as Classification

———. 1972. ‘Developmental Sequences of Kinship Patterns’, in Priscilla Reining (ed.), Kinship Studies in the Morgan Centennial Year. Washington DC: Anthropological Society of Washington, pp. 134–66. Dousset, Laurent. 2002. ‘On the Misinterpretation of the Aluradja Kinship System Type (Australian Western Desert)’, Social Anthropology 11(1): 43–61. Dyen, Isadore, and David F. Aberle. 1974. Lexical Reconstruction: The Case of the Proto-Athapaskan Kinship System. Cambridge: Cambridge University Press. Ehret, Christopher. 2008. ‘Reconstructing Ancient Kinship in Africa’, in Nicholas J. Allen et al. (eds), Early Human Kinship: From Sex to Social Reproduction. Malden and Oxford: Blackwell, for the Royal Anthropological Institute, pp. 200–31. Elmendorf, W.W. 1961. ‘System Change in Salish Kinship Terminologies’, Southwestern Journal of Anthropology 17: 365–82. Eyde, David B., and Paul M. Postal. 1961. ‘Avunculocality and Incest: The Development of Unilateral Cross-Cousin Marriage and Crow-Omaha Kinship Systems’, American Anthropologist 63: 747–71. Fêng, Han-yi. 1937. ‘The Chinese Kinship System’, Harvard Journal of Asiatic Studies 2: 141 ff. Fortes, Meyer. 1983. Rules and the Emergence of Society. London: RAI (Occasional Paper no. 39). Forth, Gregory. 1985. ‘Layia (FZS, ZH, m.s.): The Evolutionary Implications of Some Sumbanese Affinal Terms’, Sociologus 35(2): 120–41. ———. 1988. ‘Prescription Gained or Retained? Analytical Observations on the Relationship Terminology of Ndao, Eastern Indonesia’, Sociologus 38(2): 166–83. ———. 1990. ‘From Symmetry to Asymmetry: An Evolutionary Interpretation of Eastern Sumbanese Relationship Terminology’, Anthropos 85(4–6): 373–92. Fox, James. 1984. ‘Possible Models of Early Austronesian Social Organization’, Asian Perspectives 26(1): 36–43. ———. 1987. ‘“Between Savu and Roti”: The Transformation of Social Categories on the Island of Ndao’, in Donald Laycock and Werner Winter (eds), A Word of Language. Canberra: Australian National University, pp. 195–203. Fuller, C.J., and Haripriya Narasimhan. 2008. ‘Companioniate Marriage in India: The Changing Marriage System in a Middle-Class Brahman Subcaste’, Journal of the Royal Anthropological Institute 14: 736–54. Gamble, Clive. 2008. ‘Kinship and Material Culture: Archaeological Implications of the Human Global Diaspora’, in Nicholas J. Allen et al. (eds), Early Human Kinship: From Sex to Social Reproduction. Malden and Oxford: Blackwell, for the Royal Anthropological Institute, pp. 27–40. Godelier, Maurice, Thomas R. Trautmann and Franklin Tjon Sie Fat (eds). 1998. Transformations of Kinship. Washington DC and London: Smithsonian Institution Press. Good, Anthony. 1980. ‘Elder Sister’s Daughter’s Marriage in South Asia’, Journal of Anthropological Research 36: 474–500. ———. 1981. ‘Prescription, Preference and Practice: Marriage Patterns among the Kondaiyankottai Maravar of South India’, Man 16: 108–29. ———. 1991. The Female Bridegroom: A Comparative Study of Life-Crisis Rituals in South India and Sri Lanka. Oxford: Clarendon Press. Goodenough, Ward. 1956. ‘Componential Analysis and the Study of Meaning’, Language 32: 195–216.



57

58

How Kinship Systems Change

Guermonprez, Jean-François. 1998. ‘Transformations of Kinship Systems in Eastern Indonesia’, in Maurice Godelier et al. (eds), Transformations of Kinship. Washington DC and London: Smithsonian Institution Press, pp. 271–93. Héritier, Françoise. 1981. L’exercice de la parenté. Paris: Gallimard. Hicks, David. 1981. ‘A Two-Section System with Matrilineal Descent among the Tetum of Eastern Indonesia’, Sociologus 31(2): 180–84. ———. 1983. ‘A Transitional Relationship Terminology of Asymmetric Prescriptive Alliance among the Makassai of Eastern Indonesia’, Sociologus 33(1): 73–85. ———. 1985. ‘A Transitional Two-Section System among the Mbae-Speakers of Manggarai, Eastern Indonesia’, Sociologus 35(1): 74–83. ———. 1986. ‘The Relationship Terminology of the Ema’, Sociologus 36(2): 162–71. Hocart, Arthur M. 1936. Kings and Councillors: An Essay in the Comparative Anatomy of Human Society. Cairo: Paul Barbey. Hornborg, Alf. 1998. ‘Serial Redundancy in Amazonian Social Structure: Is There a Method for Post-structural Comparison?’ in Maurice Godelier et al. (eds), Transformations of Kinship. Washington DC and London: Smithsonian Institution Press, pp. 168–86. Ives, John W. 1998. ‘Developmental Processes in the Pre-contact History of Athapaskan, Algonquian and Numic Kin Systems’, in Maurice Godelier et al. (eds), Transformations of Kinship. Washington DC and London: Smithsonian Institution Press, pp. 94–139. Jones, Doug et al. 2010. ‘Human Kinship, from Conceptual Structure to Grammar [plus comments section]’, Behavioural and Brain Sciences 33: 367–416. Jones, Doug, and Bojka Milicic (eds). 2010. Kinship, Language, and Prehistory: Per Hage and the Renaissance in Kinship Studies. Salt Lake City: The University of Utah Press. Kapadia, Karin. 1993. ‘Marrying Money: Changing Preference and Practice in Tamil Marriage’, Contributions to Indian Sociology 27(1): 25–51. Korn, Francis. 1973. Elementary Structures Reconsidered: Lévi-Strauss on Kinship. London: Tavistock Publications. Krengel, Monika. 1989. Sozialstruktur in Kumaon: Bergbauen im Himalaya. Wiesbaden: Franz Steiner Verlag. Kroeber, Alfred. 1909. ‘Classificatory Systems of Relationship’, Journal of the Royal Anthropological Institute 39: 77–84. Kronenfeld, David B. 1989. ‘Morgan vs. Dorsey on the Omaha Cross/Parallel Contrast: Theoretical Implications’, L’Homme 29(1): 76–106. Kryukov, Mikhail V. 1998. ‘The Synchro-diachronic Method and the Multi-directionality of Kinship Transformations’, in Maurice Godelier et al. (eds), Transformations of Kinship. Washington DC and London: Smithsonian Institution Press, pp. 294–313. Kuper, Adam. 2008. ‘Changing the Subject – about Cousin Marriage, among Other Things’, Journal of the Royal Anthropological Institute 14: 717–35. Lane, Robert, and Barbara Lane. 1959. ‘On the Development of Dakota-Iroquois and Crow-Omaha Kinship Terminologies’, Southwestern Journal of Anthropology 15: 254–65. Layton, Robert. 2008. ‘What Can Ethnography Tell Us About Human Social Evolution?’ in Nicholas J. Allen et al. (eds), Early Human Kinship: From Sex to Social

Kinship as Classification

Reproduction. Malden and Oxford: Blackwell, for the Royal Anthropological Institute, pp. 113–27. Leach, Edmund. 1945. ‘Jinghpaw Kinship Terminology’, Journal of the Royal Anthropological Institute 75: 59–72. Lévi-Strauss. 1966. ‘The Future of Kinship Studies’, Proceedings of the Royal Anthropological Institute 1965: 13–22. McConvell, Patrick. 1997. ‘Long-Lost Relations: Pama-Nyungan and Northern Kinship’, in Patrick McConvell and Nicholas Evans (eds), Archaeology and Linguistics: Aboriginal Australia in Global Perspective. Melbourne: Oxford University Press, pp. 207–36. McKinley, Robert. 1971a. ‘A Critique of the Reflectionist Theory of Kinship Terminology: The Crow/Omaha Case’, Man 6(2): 228–47. ———. 1971b. ‘Why Do Crow and Omaha Terminologies Exist? A Sociology of Knowledge Interpretation’, Man 6(3): 408–26. McKnight, David. 1971. ‘Some Problems Concerning the Wik-mungkan’, in Rodney Needham (ed.), Rethinking Kinship and Marriage. London: Tavistock, pp. 145–80. Needham, Rodney. 1959. ‘An Analytical Note on the Kom of Manipur’, Ethnos 24: 121–35. ———. 1960. ‘Alliance and Classification among the Lamet’, Sociologus 10(2): 97–119. ———. 1961. ‘An Analytical Note on the Structure of Sirionó Society’, Southwestern Journal of Anthropology 17: 239–55. ———. 1962. ‘Genealogy and Category in Wikmunkan Society’, Ethnology 1: 223–64. ———. 1967. ‘Terminology and Alliance, II: Mapuche, Conclusions’, Sociologus 18(1): 39–53. ———. 1969. ‘Gurage Social Classification: Formal Notes on an Unusual System’, Africa 39: 153–66. ———. 1971. ‘Remarks on the Analysis of Kinship and Marriage’, in Rodney Needham (ed.), Kinship and Marriage (ASA Monographs 11). London: Tavistock, pp. 1–34. ———. 1980. ‘Principles and Variations in the Structure of Sumbanese Society’, in James J. Fox (ed.), The Flow of Life. Cambridge MA and London: Harvard University Press, pp. 21–47. Parkin, Robert. 1992. The Munda of Central India: An Account of their Social Organization. Delhi: Oxford University Press. ———. 1995. ‘The Contemporary Evolution of Polish Kinship Terminology’, Sociologus 45(2): 140–52. ———. 1997. Kinship: An Introduction to Basic Concepts. Oxford and Malden: Blackwell. ———. 1998. ‘Dravidian and Iroquois in South Asia’, in Maurice Godelier et al. (eds), Transformations of Kinship. Washington DC and London: Smithsonian Institution Press, pp. 252–70. Parry, Jonathan. 1979. Caste and Kinship in Kangra. London: Routledge & Kegan Paul. Read, Dwight. 2001. ‘What is Kinship?’ in Richard Feinberg and Martin Ottenheimer (eds), The Cultural Analysis of Kinship: The Legacy of David Schneider and



59

60

How Kinship Systems Change

its Implications for Anthropological Relativism. Urbana: University of Illinois Press, pp. 78–117. Rivière, Peter. 1969. Marriage among the Trio: A Principle of Social Organisation. Oxford: Clarendon Press. Scheffler, Harold. 1971. ‘Dravidian-Iroquois: The Melanesian Evidence’, in L.R. Hiatt and C. Jayawardene (eds), Anthropology in Oceania: Essays Presented to Ian Hogbin. Sydney: Angus and Robertson, pp. 231–54. Scheffler, Harold, and Floyd Lounsbury. 1971. A Study in Kinship Semantics: The Siriono Kinship System. Englewood Cliffs: Prentice-Hall. Service, Elman. 1971. Cultural Evolutionism: Theory in Practice. New York: Holt, Rinehart and Winston. Tjon Sie Fat, Franklin E. 1998. ‘On the Formal Analysis of “Dravidian”, “Iroquois”, and “Generational” Varieties as Nearly Associative Combinations’, in Maurice Godelier et al. (eds), Transformations of Kinship. Washington DC and London: Smithsonian Institution Press, pp. 59–93. Trautmann, Thomas R. 1981. Dravidian Kinship. Cambridge: Cambridge University Press. ———. 1984. ‘Decoding Dravidian Kinship: Morgan and McIlvaine’, Man 19: 421–31. ———. 2012. ‘Crossness and Crow-Omaha’, in Thomas R. Trautmann and Peter M. Whiteley (eds), Crow-Omaha: New Light on a Classic Problem of Kinship Analysis. Tucson: The University of Arizona Press, pp. 31–50. Trautmann, Thomas R., and R.H. Barnes. 1998. ‘“Dravidian”, “Iroquois” and “CrowOmaha” in North American Perspective’, in Maurice Godelier et al. (eds), Transformations of Kinship. Washington DC and London: Smithsonian Institution Press, pp. 27–58. Trautmann, Thomas R., and Peter M. Whiteley (eds). 2012. Crow-Omaha: New Light on a Classic Problem of Kinship Analysis. Tucson: University of Arizona Press. Turner, David. 1977. ‘The Concept of Kinship: Some Qualifications Based on a Re-examination of the Australian Data’, Bijdragen tot de Land-, Taal en Volkenkunde 133: 23–43. Viveiros de Castro, Eduardo. 1998. ‘Dravidian and Related Kinship Systems’, in Maurice Godelier et al. (eds), Transformations of Kinship. Washington DC and London: Smithsonian Institution Press, pp. 332–85. Wheeler, C.J. 1982. ‘An Enquiry into the Proto-Algonquian System of Social Classification and Marriage: A Possible System of Symmetric Prescriptive Alliance in a Lake Forest Archaic Culture during the Third Millennium BC’, Journal of the Anthropological Society of Oxford 13: 165–74. White, Leslie. 1958. ‘What is a Classificatory Kinship Term?’ Southwestern Journal of Anthropology 14: 378–85. Witowski, Stanley. 1972. ‘Guttman Scaling of Semantic Distinctions’, in Priscilla Reining (ed.), Kinship Studies in the Morgan Centennial Year. Washington DC: Anthropological Society of Washington, pp. 167–88.

2 

Terminology and Alliance in India Tribal Systems and the North-South Problem

I It has become increasingly clear that the so-called ‘tribals’ of middle India do not fall neatly into the usual division between south India and north India as regards kinship.* This division has itself long been recognized and is, inter alia, one between the prescriptive and the non-prescriptive (cf. Needham 1973), or between areas that value and do not value cross-cousin marriage as a mode of affinal alliance. But the tribal kinship systems of Madhya Pradesh, southern Bihar and Orissa also have a number of distinctive features, so much so that at first sight they suggest the existence of a third kinship model for the subcontinent (cf. Parkin 1985: 720). One purpose of this chapter is to draw attention to these features and to locate them properly in the context of Indian kinship generally, especially the divide between north and south just mentioned. Another is to argue that previous attempts to explain away this divide have failed because kinship systems in the two main regions have been compared synchronically – that is, as static systems – and the gulf has proved too wide to reduce either to the other. Instead, a diachronic, historical or developmental approach is advocated – that is, one that postulates a cline of transformation between the two poles represented by

62

How Kinship Systems Change

north and south. This, too, is not possible through direct comparison, but it becomes feasible once some of the tribal systems mentioned above are taken into account. The significance of the latter is not so much that they constitute a distinct third kinship model for the subcontinent as that their configurations show a mix of features drawn from both north and south. This allows them to be seen as a typological compromise between the two regions and, moreover, it is very possible that they form a point on the postulated cline of transformation between them. The emphasis throughout will be on kinship terminology and patterns of affinal alliance rather than the influence of caste or status. This is not to deny the significance of such factors, and they will be brought into the discussion on occasion (see Milner 1988 for a recent assessment): but they stand high in the consciousness of the social actor and are accordingly subject to considerable variation, not to say manipulation, especially in their interaction with other phenomena. This makes any reductionist comparison of the sort that is usually applied to kinship in north and south India difficult if not impossible (by reductionist, I mean comparison designed to minimize and not merely account for these profound differences). It seems preferable to concentrate instead on factors that are less explicit and used largely unconsciously, or at least without reflection, especially kinship terminology, with which there are definite and more predictable limits as to variation. It is this that facilitates the tracing of diachronic change and so gives it greater explanatory potential in the present case. To illustrate what is involved, I will begin by describing one of the tribal systems in detail, that of the Juang, a Munda-speaking group of perhaps 20,000 individuals living in northern Orissa. For this, I will be relying on data collected by Charles McDougal in the late 1950s and early 1960s and published in two articles (1963b, 1964) and a microfilmed edition of his thesis (1963a). Thanks to his work, we have one of the fullest and most perceptive accounts of any tribal group in this part of India, one which is refreshingly free from any theoretical pre- or misconceptions and is content to allow the data to speak for itself. Following this, I will make some briefer comparative comments on some neighbouring groups to show that in its main characteristics this system is not at all unique to the Juang but is widespread in the immediate area. I will then go on to argue its



63

Terminology and Alliance in India

significance for the study of kinship in the subcontinent as a whole along the lines sketched above.

II In Keonjhar, the area in which McDougal worked, the Juang are divided into eighteen patriclans or bok (literally ‘brother’), which are totemic and have two significant levels of segmentation. At the lower level are the unnamed kutumali (an Oriya word), which consist of four or five agnatically related families and are important at life-crisis rites, as the preferential unit of inheritance and as the main collective contributors to brideprice. At the village level are the thirty-eight local descent groups (LDGs), one for each Juang village in Keonjhar, which are more extensive and the operational units in spouse exchange between villages. Only for the kutumali are genealogical connections at all traceable: clan identity is a matter of having the same totem, LDG identity mostly a matter of living in the same village. This means that villages are not only conceived of as agnatic units but as spouse-exchange groups too. The only non-agnates present in the village should be inmarrying spouses (though there may nonetheless be some other affines present). Relations between villages-cum-LDGs are limited to spouse exchange and otherwise border on the hostile. From the point of view of any one village, other villages with whom there are or potentially could be relations of alliance through marriage are designated bondhu. Those with whom such relations are impossible are called kutumb (both are words of Indo-European origin and have probably entered Juang through Bengali or Oriya). Kutumb means basically ‘unmarriageable’ in Juang but in two different senses: one is the recognition of agnatic relationship (some villages are associated with the same clan and are therefore agnatically, if rather distantly, related); the other is the existence of a quarrel between two villages that were once in a bondhu relationship. This dichotomy is purely relative: there are no exogamous moieties or dual organization, with particular agnatic groups belonging absolutely in one division or the other. Just as significant as the kutumb-bondhu dichotomy is the horizontal division of Juang society into two sets of alternating generations. Each is known as a bhaiguli or ‘group of brothers’, but they are not

64

How Kinship Systems Change

given specific names. Ego’s genealogical level is thus associated with +2 and −2 (i.e., the levels of the grandparents and grandchildren) in opposition to the set formed by +1 and −1 (i.e., the levels of the parents and children). As well as being important from the point of view of affinal alliance, this dichotomy takes other forms. Generally, joking relationships are appropriate only with members of one’s own bhaiguli, avoidance relationships with members of the opposing one. This dichotomy is also stressed in ritual ways and is paralleled in the division of the village community into ritual age grades: the kongerki or unmarried youths and kamando or old men cooperate in the running of the majang or communal house and are opposed to the bontay or married men, who, though they may enter the majang, are less involved in its affairs. Finally, there is evidence that the Juang connect the +2 level with ego’s as regards the recirculation of soul substance and in naming (grandparents and grandchildren are also connected in the Juang myth explaining the origin of death; see Parkin 1992: Ch. 9). Affinal alliance among the Juang is described by McDougal as involving ‘classificatory sister exchange’ (1963b: 185–86) or ‘prescriptive symmetric alliance’ (1964: 319). In other words, it can be related to that large body of anthropological theory that concerns itself with ‘cross-cousin marriage’ or, better, positive marriage rules, in that ego is directed towards particular terminologically defined categories in seeking a spouse. We shall see that while in the Juang case the latter label is appropriate, the former certainly is not. The terminology is basically prescriptive, in that it expresses the continuous operation of positive marriage rules. Affinal alliance among the Juang should only take place between members of LDGs and villages that are bondhu to each other and members of the same bhaiguli. Marriage with anyone in another kutumb group would be bogodung or incest; marriage with anyone in the opposite bhaiguli (i.e. from the +1 and −1 levels) would be jenkani, literally ‘adultery’ but covering all sexual improprieties that are not bogodung. The superimposition of the horizontal dichotomy over the vertical one does not amount to a four-section system, due to the relativity of the designations kutumb and bondhu. Spouse selection is usually made at one of the periodic mutual dancing visits between bondhu villages: the bride is simply captured and married almost immediately to the groom, who himself may not be aware beforehand

Terminology and Alliance in India

of what is intended. Such matters are generally left in the hands of the adult males of the village, and the principals to the match themselves are rarely consulted. Even within the collection of kinship categories that are both bondhu and members of one’s own bhaiguli, one category is formally prohibited, namely ajikar (which includes eBWeZ, eZHeZ, WeZ, WFFZ, FMeZ, SSWeZ, FFZHeZ, FFZHBSD, FFBWBSD). In first marriages, the preference is for a woman in the category saliray (which includes eBWyZ, eZHyZ, WyZ, WBSD, FMyZ, SSWyZ, FFZHyZ), but those in the categories na (PM, MFZ, FMBSD) and bokosini (CD, FFZSD) are also acceptable, being regarded as virtually equivalent to each other and to saliray. We see immediately, therefore, that the preferred categories do not include first cross cousins. The main preferred category, saliray, does not include any cross cousin at all and is most usefully seen as designating ‘elder sibling’s spouse’s younger sister’ (eGEyZ) instead. First cross cousins are terminologically equated with siblings and parallel cousins in such a way as to suggest that matrilateral cross cousins are of higher status: MBD is aji, also elder sister, while FZD is bokoray, also younger sister. These two categories are possible ones as marriage partners, but they are less preferred, and of course actual siblings are banned. Other preferences exist for second marriages (polygynous ones, or any following divorce or widow[er] hood), but we need not go into details here (but see McDougal 1964). McDougal offers some statistics on the actual categories that are involved in marriage, and they accord to quite a high degree with the above preferences and prohibitions. Sixty-two per cent of the marriages in his sample were with a saliray, only about 4% with a saliray of the +2 or −2 levels; only another 4% were with the supposedly equivalent na or bokosini. Twenty-three per cent were with a classificatory cross cousin but only 3% with a genealogical cross cousin (all with MBD). There were no marriages with an ajikar, the one bondhu category of ego’s own bhaiguli not allowed him, but there were a considerable number into the other bhaiguli, despite this being formally barred to him too (18%, mostly with a classificatory MyZ or ZD). Such deviant marriages have to be tolerated, says McDougal, because of demographic difficulties in finding marriageable women (women marry earlier than men). They are simply redefined according to what the rules should be, as is characteristic of all prescriptive systems.



65

66

How Kinship Systems Change

In addition to these rules, there is evidently an imperative for each LDG-cum-village to disperse alliances among several others and not to ally with the same one all the time. This accords with the ban on first cross cousins, and it also explains why the preference is better expressed in terms of eGEyG categories instead. According to McDougal, ego should not marry where his father married (which would indicate continuous cross-cousin marriage) but should only repeat the marriage of his father’s father; in other words, no alliance between two bondhu descent groups should be repeated immediately, but only after the lapse of at least one and sometimes three generations. In the intervening period, the alliance is still maintained, not through repeated marriages but in other ways – that is, through prestations and periodic mutual dancing visits between bondhu villages. It is significant that one’s father’s father is an agnatic ascendant who falls into the same terminological category and bhaiguli as one’s elder brother, an agnate of one’s own genealogical level. The specification eBWyZ that is included in saliray may therefore be interpreted as really expressing the requirement to repeat the marriage of a classificatory elder brother. Even those saliray who are genealogically in the +2 or −2 level may in fact be of about the same age as ego, despite the genealogical distance. This accords with the rule that marriages between two particular bondhu groups should be repeated within the same generation set or bhaiguli, since the only ban is on their repetition in the ensuing one or three. The specification eBWyZ expresses, of course, an asymmetric exchange of spouses between bondhu groups: Figure 2.1. GEG marriage (asymmetric) same-sex sibling pair (1) / same-sex sibling pair (2) Brother and male ego

BW and BWZ

(brother marries BW, ego marries BWZ)

(respective spouses: ego’s brother and ego)

Note: Double marriage (ego marries BWZ). Direction of spouse transfer asymmetric.

However, McDougal emphasizes that exchange is symmetric, and the prescribed category saliray also includes the specification eZHyZ, which in itself expresses this symmetry:



67

Terminology and Alliance in India

Figure 2.2. GEG marriage (symmetric) opposite-sex sibling pair A / opposite-sex sibling pair B sister and male ego

ZH and ZHZ

(sister marries ZH, ego marries ZHZ)

(respective spouses: ego’s sister and ego)

Note: Double marriage (ego marries ZHZ). Direction of spouse transfer symmetric.

Thus, what we have among the Juang are symmetric and asymmetric exchanges between two groups of siblings who are in a relationship of actual or residual affinity with each other. Only a moment’s thought is needed to see that this is also characteristic of cross-cousin marriage. But there is an important conceptual difference. With crosscousin marriage, the analyst, and sometimes the people themselves, can and does trace ego’s links with his or her cross cousin through genealogical links passing through previous generations. Analytically, this cannot be the case where marriage is between eGEyG specifications, since these are affinal by definition and therefore do not entail any such links. It is probably not how the Juang view things either, since their genealogical memory is shallow, barely extending past two generations. Thus it can hardly cope with second cousins, let alone the fourth cousins that would be involved in cases where there is a three-generation delay before alliances are renewed (for example, FFFFZSSSD). Instead, reference is made first to village-cumLDG, then to category, in choosing spouses – genealogy apparently does not come into it at all. This is also expressed in the Juang terminology. This is basically a symmetric prescriptive one, which means that it has a pattern recognizably similar to those of south India, but there are some significant differences. The prescriptive equations cover consanguineal specifications in +1 and −1 (e.g. MB = FZH, and FZ = MBW), but there are no equations linking cognates and affines, since these will not actually coincide if the rules are followed. For example, MB and WF have separate terms, which is consonant with the fact that if classificatory MBD is rarely married, classificatory MB will rarely be one’s WF (similarly, FZ ≠ WM). Ego’s level is terminologically generational as regards consanguines: as we have seen, all cousins, cross as well as parallel, are equated with siblings – with whom sexual relations would be incestuous – as if to signal that cross cousins, too,

68

How Kinship Systems Change

are unmarriageable. The other terms in this level are normally translatable as affinal ones, the specifications involved not being traceable through genealogical (i.e. consanguineal) connections with ego at all. In view of the survival of prescriptive features in the +1 and −1 levels, it is quite probable that ego’s level once had them too, perhaps with the same basic lexis. This would mean that the distinction in this level would have shifted from cross/parallel to affinal/cognatic: in other words, the original cross-affine equations have been broken, as in the other two medial levels, but then merged with parallel kin types rather than left distinct. The Juang thus show very forcefully the inadequacy of ‘crosscousin marriage’ as a way of describing the sort of affinal alliance system that consists of positive marriage rules plus a prescriptive terminology: to talk of sibling’s spouse’s siblings in the present case is certainly more accurate. But ultimately, both interpretations are simply instances of the process of translation that anthropology necessarily entails, however much it is desired to stick close to the indigenous categories. A Juang in reality marries neither a cross cousin or even a sibling’s spouse’s sibling but a saliray – that is, a marriageable woman from an affinal village; and even if he marries someone in a terminologically wrong category, she will be redefined as a saliray.

III The Juang affinal alliance system is clearly not unique in this part of India, just exceptionally well reported. I shall briefly review the situation here; fuller details can be found elsewhere (Parkin 1992, especially Chs. 7 and 8, and Appendix B). The Santal of Keonjhar, neighbours of the Juang who also speak a Munda language, are almost equally well reported, thanks to a book by Bouez (1985). Their alliance system specifies a sangat or eZHyZ as the required category of spouse and is therefore one of symmetric spouse exchange. In addition, while exchanges are repeated within the same generation, they are avoided for at least another three (i.e. a longer minimum period than among the Juang). As for the Munda of the Ranchi area, the data indicate a similar system of non-repetition so long as the mutual dancing visits arising out of previous alliances



69

Terminology and Alliance in India

are kept up, which may be several generations (characteristically three). There is much less detail here, and we are not even certain of the prescribed kin term (though, here and elsewhere, it may turn out to be goi, usually GEG, opposite sex). However, we do have the useful additional information that the repetition of alliances after this period ‘has rather the nature of a rule’ (Yamada 1970: 385), for it has to be sought where a previous alliance can be traced – that is, no marriage should take place with a previously unrelated exchange group, despite the three-generational delay. Again, the agnatically defined village rather than the totemic descent group is the initial reference point, since knowledge about the villages of origin of inmarrying women is much more complete than knowledge of their natal descent groups. There is also evidence (of varying quality) of such systems among other Munda speakers, such as the Birhor, Korwa, Kharia and Korku, as well as the Dravidian-speaking Oraon and Indic-speaking Hill Bhuiya and Hill Kharia, all of the Bihar–Orissa area (Parkin 1992: Ch. 8). The Mon-Khmer-speaking Khasi, away to the north-east in Meghalaya, are another instance, and further examples may well be found in the Himalayas, among Tibeto-Burman-speaking groups. The rule sometimes reported against marrying a cross cousin in the lifetime of his or her parents (Birhor, Khasi) may be a cultural expression of the same system, in that it tends to interfere with the immediate repetition of alliances. Another frequently reported rule against marrying someone previously unrelated even where alliances may not be repeated in the immediately following generation (e.g. the Munda) is a more evident expression of it.

IV So what is the significance of systems like that of the Juang in the context of the usual division of kinship in India between north and south? Let us start by assessing the matter typologically. Clearly, the terminologies are prescriptive, like those of south India, though the latter usually differ in being prescriptive in all levels, in having full cognate-affine equations and in lacking alternate generation equations. Alliance in the south generally consists of a preference for a particular (often unilateral) cross cousin within the terminological

70

How Kinship Systems Change

category of bilateral cross cousin, rather than for direct exchange between two groups of siblings whose villages were allied in some previous (though not the previous) generation, though both south and central India have positive marriage rules. The relationship with north India does not seem close at first sight. North India terminologies are, of course, non-prescriptive, and there is no class of spouse identifiable through kin term prior to the marriage itself. They typically resemble the Juang ones in identifying both cross and parallel cousins with siblings, so that as far as cognatic specifications are concerned, both sorts of terminologies are generational in ego’s level. But in the +1 and −1 levels, north Indian terminologies have individualizing terms for cognatic kin, whereas Juang has symmetric prescriptive equations. Affinal terms are separate in both cases, though the pattern differs in ego’s level, and in Juang but not north Indian this level also has equations with the alternating levels (+2 and −2). In north Indian but not Juang, wife-givers are distinguished from wife-takers, reflecting the fact that, especially among high-status groups in north India, direct exchange is strongly devalued. Alliances are certainly dispersed through many exchange groups. In the north-west, this is the result of the wellknown four-got rule, which disallows any marriage in which any of the four grandparental got of the prospective bride and groom coincide. This rules out the immediate repetition of the alliance up until the third generation from ego’s, and it accords with the absence (as with the Juang) of cognate-affine equations in the terminology (it affects only ego’s direct descendants; his collateral descendants are not barred from marrying where he had married). Once one alliance by marriage has been made into a particular village, there is often a desire to reinforce it with other marriages between members of the same sibling groups or perhaps same age groups in the two villages (villages are often got-specific). In Kangra, BWZ marriages are an approved means of achieving this, especially if they involve actual siblings, but ZHZ marriages are less desirable, owing to the dislike of exchange marriages (Tiemann 1970; Parry 1979). But there is no suggestion of any positive marriage rule enjoining such renewal. Indeed, Parry (1979: 261–62) claims that the tendency is merely a statistical one that is hardly recognized explicitly at all by the people of Kangra themselves. Nor is there any category of potential spouse identified by kin term, as in a truly prescriptive system.



71

Terminology and Alliance in India

V This typological comparison is enough to show systems like that of the Juang to be compromises between north and south India. How can a diachronic approach improve on this demonstration? First, it is necessary to show just what is wrong with most reductionist comparisons so far. Previous attempts to compare kinship systems in India have concentrated exclusively on north and south and have tried to compare them synchronically. The search has been for common factors ‘underlying’ both, despite their ‘surface’ differences. But the differences involved are actually too radical and profound – much more so than those between either region and systems like that of the Juang – to be reconciled by direct comparison. Foremost in these attempts has been Louis Dumont, who, after demonstrating the affinity inherent in certain terms in a typical Dravidian terminology, and the oppositional, alliance nature of the whole pattern (1953), went on to seek the same qualities in north India, despite acknowledging the difficulties involved (1962, 1966). This attempt to establish an underlying pattern uniting the two regions accorded with his original postulate that ‘sociologically, India is one’ (1957: 9, original emphasis). However, he did not always compare like with like in making these attempts. First, although there were surface similarities between north and south in gift exchanges and the use of particular kin types in funerary ritual, the data was in both cases drawn only from single areas in each region. Second, his comparison of the repetition of intermarriage in the two regions was only possible through bringing in much wider kin groupings in the north than in the south. Third, nothing was able to reduce the fundamental differences between north and south Indian kinship terminologies in the synchronic analyses to which Dumont restricted himself, and he was forced to resort to behavioural stereotypes to conclude his argument. Finally, he claimed that north Indian terminologies were as classificatory as south Indian ones but admitted that to argue this in the case of the north, one was forced to rely on the address terminologies. All in all, the search for underlying or background similarities, though not necessarily misdirected, was ill-argued and vague. Dumont himself accepted at the outset that ‘the postulate of the unity of India is seriously challenged in the field of kinship’ (ibid.: 18),

72

How Kinship Systems Change

and he is now content to propose merely that ‘the India of caste is sociologically one’ (1983: 106, emphasis added; i.e. the uniformity concerns cultural values rather than whole social systems). Other attempts have tried to explain south India in terms of north India and vice versa by stressing the ubiquity of caste and claiming its relationship to kinship to be the same throughout India (e.g. Carter 1973; Barnett 1976; the latter supposition was rightly challenged by Dumont himself, 1983: Ch. 4). In addition, Carter deliberately minimized the significance of ‘cross-cousin marriage’ in the south and relied heavily on Maharashtra, really an area of mixed systems, in trying to compare the two regions synchronically. All these attempts must be regarded as insufficient not only because of the difficulties involved but also because of their uniform reliance on slender and suspect evidence, their forced reasoning, and their complete neglect of diachronic possibilities. This neglect was no doubt due in part to a theoretical or analytical approach that devalues history in anthropology as anything more than a charter for the present in the minds of the people. Certainly, anthropology should not be a slave to history but should exploit it, as it exploits any other source of information, on its own terms. Much depends on exactly what is being examined. Dumont, Carter, Barnett and others mostly concentrated on institutional factors (marriage rules, prestations, ritual action etc.) where change, if it takes place at all, may take a number of largely unpredictable forms and directions. Such changes are often explicit and recognized by the people themselves and may even be deliberate, especially in India: caste assemblies often discuss reforms in the lifestyles of their members as well as being concerned to maintain standards, and many lowstatus groups have outlawed cross-cousin marriage as well as communal dancing or the consumption of alcohol in an effort to rise in status. However, they are less likely to discuss changes in terminological usage. Terminology is an aspect of language, and although like language generally, the exact form it takes is a product of society, its use becomes, through social conditioning, almost automatic and taken for granted, defining the categories in which all thought ultimately has to be expressed. As Ardener puts it, ‘words set up by categories bear all the signs of materiality to the untutored human being’ (1982: 12); ‘once the classification exists . . . it is part of the total experience



73

Terminology and Alliance in India

of unsuspecting individuals’ (ibid.: 6). Thus, change in terminology, as in language, far from being deliberate, is often unperceived by (though this does not mean imperceptible to) the ordinary social actor. Moreover, terminological principles are really very few – it is only the various ways in which they may be combined that seem to make the terminologies themselves so endlessly variable. This in its turn makes both cross-cultural comparison and the tracing of diachronic developments easier. In short, a kinship terminology is not just another social institution but a semantic field, which although certainly of present significance to its users is also subject to longterm and unnegotiable changes in a limited number of possible directions. It is indisputably one area of anthropology where a diachronic approach may tell us as much as, if not more than, a purely synchronic one such as Carter’s rather arid componential analysis of the Marathi terminology. Let us now examine the potential of this approach in more detail both theoretically, and with reference to India specifically.

VI Symmetric prescriptive kinship terminologies, especially those with regular alternate generation equations, are characterized by their lexical economy and structural simplicity and by the fact that they contain relatively few categories, each covering or equating a wide range of genealogical positions. It is a pattern often felt to be at or near the starting point for any diachronic development that may have taken place. A reduced model of it has recently been put forward (Allen 1986) that exploits to the full the logic contained in the basic features associated with such terminologies generally, namely the crossparallel distinction, the principle of same-sex sibling equivalence and, in many cases, the equivalence of alternate generations. The result is a structure that, assuming gender distinctions to be absent or subsidiary (e.g. indicated only grammatically), consists of just four terms, cross and parallel in each of two genealogical levels; it is accordingly named tetradic. As such, it is not actually attested, though systems like that of the Kariera in Australia are close approximations to it. Nonetheless, it is the simplest conceivable model of a kinship system (i.e. has both distinction of generations and incest rules) and

74

How Kinship Systems Change

is therefore something to which real kinship systems can be related in analysis – an anchor, as it were, for kinship theory generally, especially its diachronic aspects. One of the basic assumptions of the model is that any move away from it would entail the breach of at least some of the key equations, with a consequent need for the coining of new terms, each covering a smaller number of genealogical positions. Indeed, this is so of any transformation of a symmetric prescriptive terminology: for example, a category originally uniting MB and FZH may be split into two categories so that MB no longer has the same term as FZH, a new term then being required to cope with this new situation (this would signal a change to an asymmetric terminology such as that of the Purum; see Needham 1962). Changes in the reverse direction are theoretically not inconceivable but they are far less likely, for they would involve not merely simplification but the confusion of existing categories through the removal of certain categorical boundaries, the deletion of some terms, and the merger of genealogical positions hitherto distinguished (Allen 1986). What evidence there is – and some of it is documentary (e.g. Fêng 1937; Spiro 1977; cf. Parkin 1988b) – supports the former possibility only – that is, of change in the direction of increasing complexity and individuation of categories. Even where there is no documentary evidence, a case can sometimes be made by comparing existing terminologies directly, especially where these are linguistically related. The assumption here is that this relationship indicates a greater uniformity in the past, in terminology as in other aspects of language: we know in general that languages change, and they always seem to change irreversibly. Sometimes change can be discerned simply by reflecting on apparent inconsistencies in patterning (see further below). Of course, change is not inevitable, nor need it all take place in the same direction: evolution is not at all unilineal, as the Indian situation itself demonstrates. A final point is that seeing terminology primarily as an aspect of language obviates the search for particular sociological factors prompting change. The reasons for terminological changes then become those for linguistic change in general, especially in so far as there is a general tendency towards increasing discrimination in many semantic domains – colour classification, for instance (see Berlin and Kay 1969), or classification of the natural world (see Brown 1984).



75

Terminology and Alliance in India

VII In India, the terminological systems closest to the tetradic model are those of the Koraput Munda in southern Orissa and the central Dravidian-speaking groups that virtually surround them. All these groups have symmetric prescriptive terminologies with alternating generation equations, the latter being absent in Dravidian terminologies further south. Tetradic theory gives reasons for postulating that terminologies with alternate generation equations are historically prior to those without them (Allen 1986: 99–100), and I have accordingly argued elsewhere (Parkin 1988a) that the central Dravidian ones are therefore typologically more basic and closer to Proto-Dravidian. This, then, is one possible cline of transformation in kinship terminologies in India, though the distinction it has produced is not very profound: all the terminologies involved are still fully symmetric prescriptive. A second possible cline is initiated by the Koraput Munda, whose terminologies are very similar to the central Dravidian ones but linguistically related to those of the north and central Munda, including the Juang, Santal and Munda discussed above. Even though the terminologies of the latter groups are now somewhat different in type, this linguistic relationship indicates a greater past uniformity. In this case, alternate generation equations have been retained despite the other changes. These have regularly resulted in ego’s level ceasing to be prescriptive and cognate-affine equations being breached in all levels, though the equations involving cognatic kin remain prescriptive in + 1 and −1. This seems to be the first step in this second cline, culminating in the typical north Indian terminology. This is a more radical transformation, in that it crosses the boundary between prescriptive and non-prescriptive. There is no known documentary evidence showing directly that a non-prescriptive north Indian terminology has been derived directly from one with a prescriptive pattern. However, in ego’s level in both north Indian and Juang, all cousins share terms with siblings in a generational – that is, lexically very economical – pattern (cf., for example, Parry 1979; Vatuk 1969). This is perhaps in north India as much as among the Juang a sign of the established fact that there can be no cross-cousin marriage and no formal maintenance of alliance through continually repeated marriages. It is also a sign that the pos-

76

How Kinship Systems Change

sibility is still regarded sufficiently seriously to be guarded against. This provides the most obvious link between the two systems. To see how the + 1 and −1 north Indian patterns may have come about, we need to examine certain other tribal terminologies, which seem to be transitional in a different way. The typical prescriptive pattern in Dravidian shows only two or at most three categories: e.g., F = or ≠ FB = MZH ≠ MB = FZH = EF. North Dravidian only retains prescriptive features in the pattern for +1 parallel kin (there are none in any other level), and even these are compromised by the accretion to them of some cross-kin specifications. Here, for example, is the + 1 pattern for cognatic kin in the terminology of the Malpahariya, of Santal Parganas, Bihar: jetha kaka jethi kaki mosi mosa mama mami pisi pisa

FeB, MeZH; but also FeZH FyB, MyZH MeZ, FeBW; but also FeZ FyBW MyZ MyZH MB MBW FyZ FyZH

And among the neighbouring Malto, similarly: pipo dada peni kale moma momi chacho pinso, bnarha

FeB, MeZH; but also FeZH in some dialects FyB, MyZH; but also FyZH and FeZH in some dialects MeZ, FeBW MyZ, FyBW; but also FyZ in some dialects MB MBW FeZ; also FyZ in some dialects FeZH in some dialects

These patterns seem to be the result of the collapse of the characteristic two-line cross-kin equations. As can be seen, some cross kin tend to be isolated with their own terms, as in north India. Both these groups appear to have affinal alliance systems like those of the Juang, though details are sketchy, and it is not clear whether there are any positive marriage rules (see Parkin 1992: Appendix II; see

Terminology and Alliance in India

Parkin 1988b for similar examples of this sort of pattern from outside India). The typical, completely non-prescriptive north Indian terminology is even more prone than north Dravidian to isolate a large number of categories, each with its own term. This is especially so in the +1 and −1 levels and among affines in ego’s level. For example, MB usually has a term of its own, as do FZH, FB and MZH – they are not conflated into possibly just two categories, and two terms, as in south India. This may be the ultimate fate of the north Dravidian terminologies, especially Malpahariya. Here, most terms are individualizing; kaka may soon become so too, since its specification of MyZH is redundant to mosa, which is the usual Indic term for this kin type, as kaka is for FyB. Only jetha and jethi will then conflate different specifications in the +1 level of this terminology. One apparent complication still has to be resolved; namely, that north Indian terminologies are conventionally associated with one language family (Indo-European or IE), south Indian with another (Dravidian), and the Juang-type system with a third (Munda). This means that the postulated changes seem not to have taken place within one and the same language family. In fact, however, the distinction between terminological systems does not coincide with that between language families, in India or anywhere else: the logical types of patterning are exhausted much sooner than the variety of linguistic forms that may give them expression. We have seen that the ‘Juang’ system implicates members of all three language families, not to mention Mon-Khmer and possibly Tibeto-Burman too. Similarly, there are a couple of lexically Dravidian terminologies that are non-prescriptive (Nayar, Brahui), while conversely, quite a large number of lexically IE terminologies are at least partly symmetric prescriptive: in the south, they include Konkani, Marathi, Sinhalese and some dialects of Gujarati, while in the north there are other examples, such as Pahari in the Himalayas, Shina in the Karakorum, and Hill Kharia and Hill Bhuiya in Orissa. These IE prescriptive terminologies are the only ones definitely known to be, or to have been prescriptive throughout the whole geographical and linguistic range and history of this, the most extensive and well-documented language family of all. They must have come into existence as part of a process of linguistic transfer in which a particular population switched from a non-IE to an IE language but



77

78

How Kinship Systems Change

retained as regards kinship terminology the pattern of the original language (Trautmann 1981; Parkin 1987: 167–68). This may have happened far more extensively in the past as part of a general trend towards the adoption of IE languages by the population at large. It means that many more IE languages might have had prescriptive terminologies at one time, perhaps over a much greater area of India than today – that is, including north India, which is virtually surrounded by areas of prescriptive kinship, not only to the south and east but also in the Himalayas and Karakorum. Later, the development of notions of relative purity and impurity, and of achieving merit and removing sin through unsolicited and unreciprocated gift-giving to superiors, all of which originated outside the realm of kinship, led to a move towards asymmetric exchange and away from repeating marriages as a way of maintaining alliances. In short, they caused the attrition of the original systems without effacing all traces of a definite structure in the separation of wife-givers from wife-takers in both terminology and alliance. This, of course, is the process of caste impinging on kinship. Caste may have affected kinship differently in different parts of India, as Dumont has often argued (e.g. 1966, 1983: 165 ff.), but this has been a historical process as well as an aspect of the system in operation at any one time (as with the spread of ‘Hinduism’ through the subcontinent generally; cf. Parkin 1992: Ch. 1). To neglect the diachronic dimension, as has persistently been done in the present case, is to be forced into comparing the incomparable. Yet, as I have tried to argue, this incomparability is really only the result of systems having grown apart in history. Notes * Originally published in Contributions to Indian Sociology 24(1) (1990): 61–76.

Reproduced with permission.

References Allen, N.J. 1986. ‘Tetradic Theory: An Approach to Kinship’, Journal of the Anthropological Society of Oxford 17(2): 87–109. Ardener, Edwin 1982. ‘Social Anthropology, Language and Reality’, in David Parkin (ed.), Semantic Anthropology. London: Academic Press, pp. 1–14.

Terminology and Alliance in India

Barnett, Steve. 1976. ‘Coconuts and Gold: Relational Identity in a South Indian Caste’, Contributions to Indian Sociology (n.s.) 10(1): 133–56. Berlin, Brent, and Paul Kay. 1969. Basic Color Terms. Berkeley: University of California Press. Bouez, Serge. 1985. Réciprocité et hiérarchie: l’alliance chez les Hos et les Santals de l’lnde. Paris: Société d’Ethnographie. Brown, Cecil H. 1984. Language and Living Things: Uniformities in Folk Classification and Naming. New Brunswick: Rutgers University Press. Carter, Anthony T. 1973. ‘A Comparative Analysis of Systems of Kinship and Marriage in South India’, Proceedings of the Royal Anthropological Institute for 1973: 29–54. Dumont, Louis. 1953. ‘The Dravidian Kinship Terminology as an Expression of Marriage’. Man 53: 34–39. –––––. 1957. ‘For a Sociology of India’, Contributions to Indian Sociology 1: 7–22. –––––. 1962. ‘Le vocabulaire de parenté dans l’Inde du nord’, L’Homme 2(2): 5–48. –––––. 1966. ‘Marriage in India, the Present State of the Question: III. North India in relation to South India’, Contributions to Indian Sociology 9: 90–114. –––––. 1983. Affinity as a Value: Marriage Alliance in South India, with Comparative Essays on Australia. Chicago: The University of Chicago Press. Fêng, Han-yi. 1937. ‘The Chinese Kinship System’, Harvard Journal of Asiatic Studies 2: 141 ff. McDougal, Charles. 1963a. The Social Structure of the Hill Juang. Ann Arbor: University Microfilms. –––––. 1963b. ‘The Social Structure of the Hill Juang: A Précis’, Man in India 43(3): 183–91. –––––. 1964. ‘Juang Categories and Joking Relationships’, Southwestern Journal of Anthropology 20(4): 319–45. Milner, Murray J. 1988. ‘Status Relations in South Asian Marriage Alliances: Toward a General Theory’, Contributions to Indian Sociology (n.s.) 22: 145–69. Needham, Rodney. 1962. Structure and Sentiment. Chicago: The University of Chicago Press. –––––. 1973. ‘Prescription’, Oceania 43: 166–81. Parkin, Robert. 1985. ‘Munda Kinship Terminologies’, Man 20(4): 705–21. –––––. 1987. ‘Kin Classification in the Karakorum’, Man 22(1): 157–70. –––––. 1988a. ‘ Reincarnation and Alternate Generation Equivalence in Middle India’, Journal of Anthropological Research 44(1): 1–20. –––––. 1988b. ‘Prescription and Transformation in Mon-Khmer Kinship Terminologies’, Sociologus 38(1): 55–68. –––––. 1992. The Munda of Central India: An Account of their Social Organization. New Delhi: Oxford University Press. Parry, Jonathan. 1979. Caste and Kinship in Kangra. London: Routledge & Kegan Paul. Spiro, Melford E. 1977. Kinship and Marriage in Burma: A Cultural and Psychodynamic Analysis. London and Los Angeles: University of California Press. Tiemann, Gunther. 1970. ‘The Four-Got-Rule among the Jat of Haryana in Northern India’, Anthropos 65(1–2): 166–77.



79

80

How Kinship Systems Change

Trautmann, Thomas. 1981. Dravidian Kinship. Cambridge: Cambridge University Press. Vatuk, Sylvia. 1969. ‘A Structural Analysis of the Hindi kinship Terminology’, Contributions to Indian Sociology (n.s.) 3: 94–115. Yamada, Ryuji. 1970. Cultural Formation of the Mundas. Tokyo: Tokai University Press.

3 

From Tetradic Society to Dispersed Alliance

Introduction In the previous chapter, I introduced the practice of marriage to GEG (sibling’s spouse’s sibling) categories among certain Munda- and Dravidian-speaking tribes in India.* I suggested that they could be seen as a step away from cross-cousin marriage in that they envisage an intensification of marriages within a single generation between two groups of siblings that are also spouse-exchange groups. Also associated with this, however, are rules against renewing marriages between the same groups for one or more of the following generations, though at the end of this period alliances usually become possible again. These delays effectively rule out prescribed cross-cousin marriage, which formally, though not always in practice, is predicated on the assumption that all egos in society will marry according to the prescription and that therefore such marriages are repeated from generation to generation.1 Such delays in renewing alliances are also the outcome of prohibitions on marriage into the descent groups of, for example, ego’s parents or some or all of ego’s grandparents, which are rather more frequently reported in the literature and also in effect rule out crosscousin marriage. What is more, they have been associated, primarily by Lévi-Strauss, with Crow-Omaha terminologies, or what, more broadly, he calls ‘semi-complex structures’ (1966). In fact, the practice of breaking up the intergenerational continuity of alliances between specific spouse-exchange groups, however expressed, does

82

How Kinship Systems Change

not need Crow-Omaha terminologies, as the example of north Indian marriage practices shows (see also below), nor are any South Asian kinship terminologies at all Crow-Omaha in type.2 The main aim of this chapter is to describe briefly other examples of this practice in other parts of the world. I am especially concerned to stress cases where spouse exchange between two sibling groups is intensified within one generation but is not permitted in the immediately following generation(s). I believe this aspect of the intensification of alliances in these cases of intermarriage between groups of siblings (GEG marriages) has not been given its due weight so far in the literature, despite its potential relevance as a step away from cross-cousin marriage in evolutionary terms. This alliance practice is close to what one finds in many Munda societies, though the practice of renewing the alliances of previous generations after a gap is also found in the caste society on occasion; for example, in Kangra (Parry 1979). From the indigenous point of view, it may be related to a feeling in north India that, while close kin marriage is wrong (unlike cross-cousin marriage in south India), it is nonetheless desirable to marry people with whom you can trace some past alliance. In India, with its deep concern for status in a highly hierarchical society – status that is partly negotiated through marriage alliance – this is an important consideration in choosing suitable spouses for one’s children. This set of circumstances is also reflected in two mechanisms for dispersing alliances across generations in India, the four-got rule and the sapinda rule (e.g. Tiemann 1970). The former in particular bans any marriage where any of the four grandparents of the bride and groom are the same, very like many reported Crow-Omaha prohibitions on alliance. The desire not to marry into unrelated groups is not restricted to north India: Faron (1962) mentions that a fear of sorcery is behind such practical restrictions among the Mapuche of Chile, though such alliances do occur. The Mapuche case is interesting as a possible example of a shift from asymmetric prescriptive alliance to a mere preference for MBD/ FZS marriages, which the recorded genealogies showed to be rare by the early 1960s (Faron 1962; also discussed in Needham 1967). These generational delays in repeating alliances also occur elsewhere in the world; for example, among the Eastern Abelam, Iatmul and Gnau of New Guinea (Forge 1971: 137, 142, 143 n. 9). In the latter case, the relationship created by a marriage is either ended in the fourth generation or renewed through marriage between FFMBSSD

From Tetradic Society to Dispersed Alliance

and FFFZSSS (Forge 1971: 143 n. 9, after G.A. Lewis). Other examples include the Gumuz (James 2012: 140–46) and most famously the Samo (Héritier 1981)3 and in Kumaon District in north-west India, where 40% of marriages can be defined as being between groups of siblings, though no generational rule of delay is mentioned (Krengel 1989). Another possibility are the Xingu Carib (or Kalapalo), for whom Basso says that ‘“sibling exchange” marriage . . . is considered highly desirable’ (1970: 410) but also that marriages with nonrelatives may be preferred in the next generation, though previous relations of marriage are also exploited in finding spouses (ibid.: 411, 413). Like, apparently, the Gumuz (James ibid.), Xingu Carib parents actively negotiate what Basso (ibid.: 411) calls ‘sibling obligations’ to obtain spouses for their children. The Xingu Carib terminology is bifurcate merging in +1 and −1 but generational in ego’s generation, and it is clear from Basso’s description that there is no prescriptive alliance here. One consideration in negotiating marriages is the potential of new alliances to provide support and refuge should an individual be accused within the group of witchcraft, accusations that could lead to one’s murder (Basso 1984). The specification ‘GEG’ for such marriages can only apply to second etc. marriages in such a series, as it assumes that at least one marriage has already taken place between the two groups of siblings. No doubt in most cases ‘sibling’ must be treated as a classificatory category, as is clearly the case among the Gumuz ( James ibid.). Yet other examples are discussed by Viveiros de Castro (1998) and Tjon Sie Fat (1998), in a volume dedicated to the differences between Dravidian and Iroquois (Godelier et al. 1998). Both authors tend to think in terms of marriage to remote cousins rather than to GEG categories, meaning that the relevant passages have to be read using some lateral thinking, but both mention sister exchange as the basic marriage rule, which apparently often takes place quite intensively. Both authors show that, if the generational delays in repeating alliances are taken into account, and assuming repetition does reoccur as soon as it becomes formally possible, one can conceive of these cases as marriages between remotely related cousins, though whether people’s genealogical memory allows them to see things this way is another matter – hence the perpetual use of GEG categories instead. In this context, Viveiros de Castro distinguishes three-generation cycles like those of many GEG-marrying Munda, which he calls ‘concentric’, from the two-generation cycles of the Aranda system, one of expected



83

84

How Kinship Systems Change

marriage to second cross cousins, which he calls ‘diametric’ (1998: 356; the Aranda system is prescriptive, the Munda cycles not). The examples he goes on to discuss (ibid.: 356 ff.) include the Umeda and Gnau, studied by Alfred Gell; the Yafar, studied by Bernard Juillerat (all of Papua New Guinea); the Ngawbe of Panama, studied by Phillip Young; and the Kandoshi of western Amazonia, studied by Anne-Christine Taylor (e.g. Taylor 1998: 205, described in more detail below). Tjon Sie Fat also discusses the Ngawbe (1998: 86–87), pointing out that while they have a three-generation exchange cycle based on marriage to FFZDD, a two-generation cycle based on marriage to MMBDD is also possible. While FFZDD is Iroquois cross and Dravidian parallel, the reverse is the case for MMBDD (ibid.: 69, Table 3.3). Another possible example is the Kuma of Papua New Guinea, studied by Marie Reay, though the rule of delay, if any, is not specified in Tjon Sie Fat’s description (1998: 89–90). Finally, Tjon Sie Fat describes both the Mundumugor, studied by Nancy McDowell, and the Yafar of Papua New Guinea as having a four-generation rule of delay and sister exchange modelled on marriage to FFFZDDD (ibid.: 88–89). To repeat, while these examples are ethnographically rooted, one can well imagine that the use of long genealogical formulae in how they are formally presented goes way beyond informants’ actual genealogical knowledge or memory – hence their use of more cognitively manageable GEG categories instead. Also of interest is another chapter in the same volume on the Cree and Ojibwa of Ontario and eastern Manitoba, Canada, studied by Désvaux and Selz (1998). Referring to Hallowell’s 1932 sample, according to which only 22% of marriages within a single kindred were with a first cross cousin, the authors suggest that this low figure was not likely to be due to missionary influence discouraging the practice but to the fact that cross-cousin marriage is not very salient in the indigenous view. Instead, the marriage preference is couched rather in terms of the repetition of existing marriages, evidently within the same generation rather than repeating those of previous generations. Almost double the percentage of the authors’ sample of 244 marriages were of this type (17%) than the percentage of cross-cousin marriages (9.5%), and alliances were about equally symmetric (sister exchange or ZHZ marriage) and asymmetric (groups of brothers marrying groups of sisters or BWB marriage) (1998: 155–56, 166 n. 18). No rule of generational delay is noted here, but the data imply it, and the possibility that the marriage system is changing away from

From Tetradic Society to Dispersed Alliance

strict cross-cousin marriage is obvious. For the authors, population increase combined with community or settlement exogamy, giving more choice in marriage partners, is a more likely reason for this apparent change than missionary influence, which was not very strong in their field site. Reasoning surrounding exogamy and endogamy, apparently of communities, appears elsewhere in this same volume (Godelier et al. 1998; see also Ives 1998; Asch 1998) in discussions of broadly similar cases, but it does risk confusing residential closeness, genealogical closeness and categorical closeness: cross-cousin marriage does not rule out choice of partner.4 Also in the same volume (Godelier et al. 1998) is a chapter by Taylor (1998) comparing the Achuar, Aguarana and Kandoshi, Jivaroan-speaking groups of western Amazonia. The main contrast is between the first and last of these peoples. The Achuar have a more or less straightforward system of bilateral cross-cousin marriage with a terminology to match, and with a clear preference for the genealogical first cross cousin. The one peculiar practice is what Taylor calls ‘quasi-“prescriptive”’ sororal polygyny (ibid.: 188), which has the effect of dispersing the alliances of a group of brothers, as the sisters of the wife of one brother are then not available to the others. The Kandoshi, by contrast, permit classificatory sister exchange within the same generation in a manner that is clearly intensive: in Taylor’s own words (ibid.: 201), ‘marriage here tends to take on a collective character, with most “brothers” of each generation taking wives in the same local group’. She also finds that, statistically, ‘a high proportion of marriages’ (ibid.) result from WB–ZH ties –that is, ties between two men who have exchanged sisters, which can also be read as the GEG category of ZHZ. Concomitantly, marriages into the immediately succeeding generations between the same groups are prohibited. Statistics, however, ‘attest the existence of a few marriages between kin from the same local group three generations down the line, and above all a proliferation of marriages between residential groups having intermarried as a whole four generations before’ (ibid.). Basically, this goes along with a dislike of but tolerance for distant kin in marriage, as well as a decided preference for alliances with non-kin. However, due to genealogical amnesia, Taylor says, the real situation is that kin ties are no longer remembered after three generations, making the descendants of previous affines non-kin.5 Taylor also gives genealogical specifications for the third cousins, who would be marriageable in realliances after three generations,



85

86

How Kinship Systems Change

though given the possibility of genealogical amnesia it is doubtful that they mean anything to the Kandoshi themselves: ‘MMMZSSD, MMFBSSD (cross by Iroquois reckoning, parallel by Dravidian accounting), or FFFZSSD, FFMBSSSD (cross in Dravidian, parallel in Iroquois)’ (ibid.: 205). Two further examples are described in an earlier volume on marriage in Papua New Guinea entitled Pigs, Pearlshells, and Women (Glasse and Meggitt 1969). Cook’s chapter on the Manga mentions a basic rule that second cousins are considered marriageable, provided they do not reside close to ego. However, it seems that this rule applies to second cousins who are simultaneously related to either male or female ego as FMBSC and FFZSC, while MMBDC and MFZDC are banned (see especially Cook 1969: 115, Fig. 4). Although the Manga apparently have an Iroquois terminology (called Seneca-type by Cook, after Pospisil; cf. Cook ibid.: 109, 111), Cook also decides that this is a prescriptive system, not a mere preference, evidently relying here on a misreading of Needham (Cook ibid.: 109). In fact, only 5 out of 186 cases in a sample followed this alleged rule, many more being free matches (29 in number) or straightforward cases of sister exchange between two unrelated men (32 in number). This low figure is possibly because it is only male ego’s eldest daughter that has to follow the ‘prescriptive’ rule, backed up by the fear of sorcery from her MB. Also, previous marriages are important, in the sense that they ‘establish kinship ties which ideally last for a specified two generations, after which the descendants are again regarded as non-kinsmen’ (ibid.: 100); this ties in with the ‘prescription’ for second cousins. And further, sibling groups are recognized as the main spouse-exchange groups: ‘Such affinally linked units are referred to as “brother-brother” units.’ From the point of view of the marriages of any one unit with another, ‘no additional marriages may be conducted with that unit for at least one more generation’ (ibid.: 107), though it is not clear whether marriages can be intensified within a generation. Finally, for a woman, FMBSS marriage (reciprocally FFZSD marriage for a man), at any rate, is also conceived as male ego giving his daughter to his MBS as a bride for the latter’s son – that is, MBSS (ibid.: 108). Cook’s Figures 3 and 4 (ibid.: 111, 115) combine the ‘prescriptive’ rule with the practice of sister exchange, indicating that marriage practices fundamentally involve symmetric exchange between groups. This is an interesting case, since it is clearly a society on the cusp of abandoning cross-cousin marriage entirely, and mar-

From Tetradic Society to Dispersed Alliance

riage practices already have most of the features of the other cases described in this chapter. Another example from the same volume are the Daribi, studied by Roy Wagner (1969). Here the basic exchange group is again a group of full siblings, called a zibi, who share incoming and outgoing obligations, meaning that although sister exchange is licit, it is disliked, as only a woman is obtained in exchange for a sister, a woman who cannot be shared out between the recipient brothers in the way a bridewealth can. In general, indeed, a zibi should not be both wifetaker and wife-giver to one’s own zibi. One way of getting round this is an arrangement ‘whereby a man gives his daughter by one wife in marriage to the brother of another of his wives’. This obviously represents an exchange, but the women involved are in different genealogical levels. Even so, ‘as in sister exchange, this kind of ongoing relationship is precluded by the fact that the lines have exchanged women in both directions’ (ibid.: 61).6 Marriage should not take place with second cousins or any closer kin, ruling out repeat marriages between any two zibi. However, as far as same-generation marriages between zibi are concerned, ‘once a woman has passed from one zibi to another in marriage, the further giving of women in the same direction is enjoined by the kinship system’ (ibid.: 61); and further, each zibi can ‘continue to take as wives the sisters of those women it has already married’ (ibid.: 63). This is clearly connected with the coming together (but not fusion) of clans into what Wagner calls a ‘superclan’ or ‘community’. The main example he gives is of two clans, Pobori and Wazo, living in a single community: ‘Pobori and Wazo clans were formed from a single ancestral unit by segmentation, but have since intermarried heavily’ (ibid.: 66), in the proportion of 26% of all the marriages of both clans over a four-generation period, though presumably still without repeating marriages between the same sub-clan units in subsequent generations. This may be why the remaining 74% of the marriages of both clans were contracted elsewhere, and there is certainly a discernible degree of alliance dispersal. Wagner sums this up by saying that ‘the concentration of marriages, and therefore of alliances, is an optative element in Daribi social structure, subject to conscious manipulation. . . . The technique of concentrating alliance ties by making many marriages in one place is a tactic within the system of “multiple and distributive opposition”7 obtaining among clans’ (ibid.: 65–56).



87

88

How Kinship Systems Change

Other possible examples of intense repetition of alliances between sibling groups within a generation include the Hopi and related Hano of the south-western United States.8 According to Eggan (1950: 35; also ibid.: 57), ‘two brothers occasionally marry two sisters or marry into the same clan, though this is not a definite rule’, though ‘sister exchange between households is a chance affair’ (ibid.: 57). Among the Shoshone as well, it is said that groups of siblings often intermarry, in circumstances that suggest that here too there are bans on repeating the marriages of the previous generation in the direct line (ibid.: 193). Moreover, at Laguna, Eggan remarks, ‘[Elsie] Parsons noted a tendency for brothers and sisters to marry into the same clan’ (ibid.: 265). Among the Tzotzil of Chiapas state, Mexico, Hopkins (1969: 101) reports a tendency for sibling groups to intermarry, either by direct exchange (ZHZ/BWB marriage) or by two brothers marrying two sisters (BWZ/ZHB marriage). In Ethiopia, the Gurage permit the latter but not the former (Shack 1966). Among the Mkako of eastern Cameroon, however, even this is banned (Copet-Rougier 1987), as it is among the Muyu of Oceania, it being wrong for two brothers to take wives from the same village (Schoorl 1993: 44). In the island of Euboea, Greece, direct exchange by groups of siblings was also evidently banned formerly: ‘. . . once a woman had married, her male collateral kin could not after that point take a woman back from her husband’s collateral kin; and any man from her own kin, if he did so, was said to have “turned back” . . . to take his wife, and to be courting disaster’ (Du Boulay 1982: 229–30). This conforms to a basic principle in this society that movement should continue around a circle and not turn or be passed back along it, else disaster will strike. Both the delays and the intensification of alliances within a generation are likely to be reported more widely in world ethnography, though they have not really been theorized as such, nor even been widely noticed on the comparative level.9 Nonetheless, the ultimate explanation for them is probably to be sought in strictly local reasons for the abandonment of prescriptive alliance or cross-cousin marriage. More broadly, while they do not need Crow-Omaha terminologies to function, the specific way in which they disperse alliances can also be seen in Lévi-Strauss’s ‘semi-complex systems’ (1966), of which Crow-Omaha systems are the archetype (for more on these systems, see this volume, Chs. 7, 8).

From Tetradic Society to Dispersed Alliance

Appendix Other examples are given here of the phenomena dealt with in this chapter, though in many cases the description relies on sources that are rather sketchy and/or ambiguous.10 First, however, I should point out that Jean-Claude Muller (1980) has anticipated me in part by writing about what he calls ‘exclusive straight sister exchange’, which he describes as sister exchange without cross-cousin marriage. To all intents and purposes, this is clearly the same thing as what I am calling GEG marriages, although Muller interprets the phenomenon rather differently, without referring to siblings-in-law, for example. Also, his whole article is an attack on Lévi-Strauss for ignoring this practice as a possible intermediate form in the transition from elementary to complex structures. Finally, Muller’s examples are mostly drawn from West Africa, especially northern Nigeria, where he builds on the earlier work of the administrator-anthropologist C.K. Meek. However, as I have already shown, this is not the only part of the world to have such preferences and practices. Thus Goodenough mentions that on Truk ‘in any one generation a majority of the members of a given lineage will frequently be married into one lineage or ramage’ (1951: 120). Similarly, BWZ/ZHB marriages are possible in Pukapuka in the Cook Islands, where otherwise only third cousins are marriageable, though about 20% of marriages defy this rule by marrying someone more closely related (Hecht 1977: 190– 92). Goodwin reports a similar preference for BWZ/ZHB marriages among the Western Apache (1942: 315). Richard Feinberg supplies a detailed explanation of this preference and its implications for Anuta Island, in Oceania (1979: 343): Anuta does not have marriage classes, and the closest thing to a positive marriage rule is a mild preference for a man to take his taina papine [or] ‘sister-in-law’ for his spouse. The closer the ‘sister-in-law’ the more appropriate the choice, and the best possible marriage partner from a kinship point of view is the sister of one’s brother’s wife. Such a marriage has the effect of reinforcing an alliance which has already been established, and it maintains the social and economic solidarity of sisters by bringing them into the same patongia [patrilateral extended families, RP]. However, a man does not have a ‘sister-in-law’ until someone of his generation marries. Up to that time all female relatives in a male ego’s generation are uniformly classed as kave [or]



89

90

How Kinship Systems Change

‘sibling of opposite sex’, and no positive preferences may be inferred, therefore, from the alliance pattern of the parents’ generation. More important for Anuta are proscriptions specifying whom one may not marry. The rule is that close kin must be shunned. Incest prohibitions and their derivative exogamic regulations are calculated on a cognatic basis. Closeness, however, is relative, and consequently the rules of incest and exogamy are phrased in terms of degrees of impropriety rather than absolute prohibition. Marriage to a third cousin or anyone more distant usually does not arouse particular concern, although in the abstract, informants sometimes say that it should not be done. Marriage to a second cousin is looked upon with disfavour, marrying a first cousin is definitely tapu [tabooed, RP], and marriage to a full biological sibling is deemed too absurd for serious consideration.

The Foi of New Guinea allow both ‘the marriage of two brothers to women who are of the same clan’ (Weiner 1988: 105) and what is in effect classificatory sister exchange, though the Foi dislike sister exchange in principle. However, classificatory sister exchange overrides this dislike: ‘A man can give his sister to one man and can marry the sister of a classificatory brother of that man, but the two relations remain entirely distinct’ (ibid.: 106). In the case of true sister exchange, they would be more firmly embedded in one another. Bernard Juillerat, writing about Richard Thurnwald’s work among the Banaro of Papua New Guinea conducted from 1913 to 1915 ( Juillerat 1993: 38–40), even describes a double exchange of spouses involving four couples as the ideal marriage arrangement here, though it is evidently only possible if there is a roughly equal gender ratio both within and between the various parties, something achieved by manipulating the gender ratio through adoption and allegedly, and more drastically, infanticide. This seems to be an example of the intensification of GEG marriages within a single generation mentioned earlier. Cross-cousin marriage, by contrast, is not allowed (ibid.: 46– 47). Writing of Ambryn (1970: 58), Scheffler remarks that there is classificatory sister exchange because ego’s ZH has an obligation to find ego a wife, though she will not necessarily be ZHZ. However, in other cases, such marriages are disliked or rejected. Thus while the Shavante of the Amazon favour groups of brothers marrying groups of sisters, the neighbouring Kayapó ban such marriages (Rivière 1980: 535). For the Melpa of Papua New Guinea (PNG), Andrew Strathern reports that there is no ZHZ/BWB mar-

From Tetradic Society to Dispersed Alliance

riage (or sister exchange) but some repetition of marriages within a generation takes place, presumably asymmetrically – that is, BWZ/ ZHB marriages (1972: 132). In an old text on the Bemba of what is now north-east Zimbabwe, Audrey Richards remarks that BWZ/ZHB marriages are allowed but ‘thought not to be very satisfactory’ (1951: 225). Among the Anlo Ewe, Nukunya states that there are no ZHZ/ BWB or sister exchange marriages, nor may two related men marry two women who are themselves related as sisters or in other ways. He adds that ‘In the past these prohibitions [i.e. marriage prohibitions generally] extended as far as third cousins of whatever line’ (1969: 65, 66). A similar situation is reported by Elizabeth Colson for the Plateau Tonga of northern Zimbabwe (1962: 110) and by Teitelbaum for the Jokwele Kpelle of Liberia and Guinea (1980: 33). Esther Goody rules out the marriages of two brothers to two sisters among the Gonja of northern Ghana (1973: 219), as does Hallpike among the Konso (1972: 118).

Notes * Full version originally published in the Journal of the Anthropological Society of Oxford-online 5(2) (2013): 194–206. Reproduced with permission. 1. The account in Chapter 2 was an initial attempt to get to grips with the situation that has just been described within the compass of an article, and it also tried to situate the ‘tribal’ systems of central India within the wider context of marriage practices in India, with its strong contrast between north and south. It also relied heavily on a single, particularly welldescribed tribe, the Juang. For a more detailed account of the situation among Munda-speakers generally, see Parkin 1992, Ch. 8, especially pp. 153 ff. 2. I return to the case of Crow-Omaha in Chapters 7 and 8. 3. James hints that other groups may qualify, such as the Umeda and Baruya of New Guinea and the Mkako of Cameroon (ibid.: 137–38), while Forge suggested that many New Guinea societies permit the renewal of alliances after the lapse of the requisite number of generations (1971: 142, 144 n. 9). Haenen (1988: 474) mentions that the Moi of Irian Jaya, who have asymmetric alliance, permit the reversal of alliances after three generations, the exact possibilities being calculated by specialists called nè foolus, literally ‘history men’. 4. Chantal Collard (personal communication) collected data from a village near Quebec City indicating that 8% of marriages involved sibling dyads (the practice did not extend to third siblings and beyond), preferably involving two brothers marrying two sisters but almost as many involving sister exchange. Something similar may occur in Brittany (Segalen 1985). These examples show that the salience of sometimes intensified sibling ex-



91

92

How Kinship Systems Change

5.

6.

7. 8. 9.

10.

change can characterize more developed and/or sociologically complex European and Europeanized societies too. The Kandoshi terminology is non-prescriptive. The general similarity of the Kandoshi case to the Munda one is striking, and Taylor herself remarked on it at the conference at which we respectively presented these cases (see Godelier et al. 1998). This arrangement resembles the situation described by Ian Walker for the island of Ngazidja and is similarly a bit of a makeshift, though it is not identical with it, due mainly to the difference in descent mode (patrilineal among the Daribi, matrilineal on Ngazidja). See Parkin (2013). Citing R.F. Salisbury. Paragraph added in 2020. One exception is Paul Henley, who has posited an ‘Amazonian’ model of kinship that ‘is similar to the canonical dravidianate insofar as the general distribution of terminological categories in the three medial generations is concerned, but it is very different in three other crucial and related respects: the absence of a positive rule of marriage, the absence of a category of cross-relative in Ego’s own generation and the presence of a set of exclusively affinal terms’ (1996: 62). Henley sees this at once as a semi-complex system and as the basis from which elementary structures have derived in Amazonia (rather than vice versa), which taken literally seems like a contradiction in terms and certainly goes against much received wisdom on this matter. He also makes it clear that the cross-parallel distinctions involved are Dravidian, not Iroquois (the latter is a further possible derivation from his ‘Amazonian’ type). Although there is evidence of both the intensification of sibling exchanges within a generation without crosscousin marriage and the repetition of alliances after the elapse of a number of generations in the Amazon, Henley does not include either among the characteristics of his ‘Amazonian’ model. As I hope to have shown here, it is anyway not restricted to the Amazon but occurs in many other parts of the world as well. New data not in the original publication.

References Asch, Michael. 1998. ‘Kinship and Dravidianate Logic: Some Implications for Understanding Power, Politics, and Social Life in a Northern Dene Community’, in Maurice Godelier, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 140–49. Basso, Ellen B. 1970. ‘Xingu Carib Kinship Terminology and Marriage: Another View’, Southwestern Journal of Anthropology 26: 402–16. ———. 1984. ‘A Husband for His Daughter, a Wife for Her Son: Strategies for Selecting a Set of In-laws among the Kalapalo’, in Kenneth M. Kensinger (ed.), Marriage Practices in Lowland South America. Urbana and Chicago: University of Illinois Press, pp. 33–44.

From Tetradic Society to Dispersed Alliance

Colson, Elizabeth. 1962. The Plateau Tonga of Northern Rhodesia: Social and Religious Studies. Manchester: Manchester University Press for the Rhodes-Livingstone Institute. Cook, E.A. 1969. ‘Marriage among the Manga’, in Robert M. Glasse and Mervyn J. Meggitt (eds), Pigs, Pearlshells, and Women: Marriage in the New Guinea Highlands. Englewood Cliffs: Prentice Hall, pp. 96–116. Copet-Rougier, Elisabeth. 1987. ‘“L’antilope accouche toujours de l‘éléphant” (devinette Mkako): étude de la transformation du marriage chez les Mkako du Cameroun’, in David Parkin and David Nyamwaya (eds), Transformations of African Marriage. Manchester: Manchester University Press for IAI, pp. 76–92. Désvaux, Emmanuel, and Marion Selz. 1998. ‘Dravidian Nomenclature as an Expression of Ego-Centred Dualism’, in Maurice Godelier, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 150–67. Du Boulay, Juliet. 1982. ‘The Greek Vampire: A Study of Cyclic Symbolism in Marriage and Death’, Man 17(2): 219–38. Eggan, Fred. 1950. Social Organization of the Western Pueblos. Chicago: The University of Chicago Press. Faron, Louis C. 1962. ‘Matrilateral Marriage among the Mapuche (Araucanians) of Central Chile’, Sociologus 21(1): 54–66. Feinberg, Richard. 1979. ‘Kindred and Alliance on Anuta Island’, Journal of the Polynesian Society 88(3): 327–48. Forge, Anthony. 1971. ‘Marriage and Exchange in the Sepik: Comments on Francis Korn’s Analysis of Iatmul Society’, in Rodney Needham (ed.), Rethinking Kinship and Marriage. London: Tavistock, pp. 133–44. Glasse, Robert M., and Mervyn J. Meggitt (eds). 1969. Pigs, Pearlshells, and Women: Marriage in the New Guinea Highlands. Englewood Cliffs: Prentice Hall. Godelier, Maurice, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds). 1998. Transformations of Kinship. Washington: Smithsonian Institution Press. Goodenough, Ward H. 1951. Property, Kin, and Community on Truk. New Haven: Yale University Publications in Anthropology, no. 46. Goodwin, Grenville. 1942. The Social Organization of the Western Apache. Chicago: The University of Chicago Press. Goody, Esther. 1973. Contexts of Kinship: An Essay in the Family Sociology of the Gonja of Northern Ghana. Cambridge: Cambridge University Press. Haenen, Paul. 1988. ‘Marriage Alliance among the Moi of Irian Jaya (Indonesia)’, Bijdragen tot de Land-, Taal- en Volkenkunde 144(4): 464–77. Hallpike, C.R. 1972. The Konso of Ethiopia: A Study of the Values of a Cushitic People. Oxford: Clarendon Press. Hecht, Julia. 1977. ‘The Culture of Gender in Pukapuka: Male, Female and the Mayakitanga Sacred Maid’, Journal of the Polynesian Society 86(2): 183–206. Henley, Paul. 1996. South Indian Models in the Amazonian Lowlands. Manchester: University of Manchester (Manchester Papers in Social Anthropology, no. 1). Héritier, Françoise. 1981. L’Exercise de la parenté. Paris: Gallimard, Le Seuil. Hopkins, Nicholas A. 1969. ‘A Formal Account of Chalchihuitán Tzotzil Kinship Terminology’, Ethnology 81: 85–102.



93

94

How Kinship Systems Change

Ives, John W. 1998. ‘Development Processes in the Pre-contact History of Athapaskan, Algonquian, and Numic Kin Systems’, in Maurice Godelier, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 94–139. James, Wendy. 2012. ‘A Tetradic Starting Point for Skewing? Marriage as a Generational Contract: Reflections on Sister-Exchange in Africa’, in Thomas R. Trautmann and Peter M. Whiteley (eds), Crow-Omaha: New Light on a Classic Problem of Kinship Analysis. Tucson: The University of Arizona Press, pp. 135–52. Juillerat, Bernard. 1993. La revocation des Tambaran: les Banaro et Richard Thurnwald revisités. Paris: CNRS Editions. Krengel, Monika. 1989. Sozialstruktur in Kumaon: Bergbauern im Himalaya. Wiesbaden: Franz Steiner Verlag. Lévi-Strauss, Claude. 1966. ‘The Future of Kinship Studies’, Proceedings of the Royal Anthropological Institute 1965: 13–22. Muller, Jean-Claude. 1980. ‘Straight Sister Exchange and the Transition from Elementary to Complex Structures’, American Ethnologist 7(3): 518–29. Needham, Rodney. 1967. ‘Terminology and Alliance: II, Mapuche; Conclusions’, Sociologus 17(1): 39–54. Nukunya, G.K. 1969. Kinship and Marriage among the Anlo Ewe. London: Athlone Press. Parkin, Robert. 1992. The Munda of Central India: An Account of their Social Organization. New Delhi: Oxford University Press. ———. 2013. ‘Note on Oblique Exchange in a Matrilineal Society in the Comoro Islands’, Journal of the Anthropological Society of Oxford 5(2): 207–11. Parry, Jonathan. 1979. Caste and Kinship in Kangra. London: Routledge. Richards, Audrey. 1951. ‘The Bemba of North-Eastern Rhodesia’, in Elizabeth Colson and Max Gluckman (eds), Seven Tribes of Central Africa. London: Oxford University Press, pp. 181–225. Rivière, Peter. 1980. ‘Dialectical Societies’, Man 15: 533–40. Scheffler, Harold W. 1970. ‘Ambryn Revisited: A Preliminary Report’, Southwestern Journal of Anthropology 26: 52–66. Schoorl, J.W. 1993. Culture and Change among the Muyu. Leiden: KITLV Press. Segalen, Martine. 1985. Quinze Generations de Bas-Bretons. Paris: Presses Universitaires Françaises. Shack, William A. 1966. The Gurage: A People of the Ensete Culture. Africa London: Oxford University Press for the IAI. Strathern, Andrew. 1972. One Father, One Blood: Descent and Group Structure among the Melpa People. London: Tavistock. Taylor, Anne-Christine. 1998. ‘Jivaro Kinship: “Simple” and “Complex” Formulas: A Dravidian Transformation Group’, in Maurice Godelier, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 187–213. Teitelbaum, Michele. 1980. ‘Designation of Preferential Affinity in the Jokwele Kpelle Omaha-Type Relationship Terminology’, Journal of Anthropological Research 36(1): 31–48. Tiemann, Gunther. 1970. ‘The Four-Got Rule among the Jat of Haryana in North India’, Anthropos 65: 166–77.

From Tetradic Society to Dispersed Alliance

Tjon Sie Fat, Franklin. 1998. ‘On the Formal Analysis of “Dravidian”, “Iroquois”, and “Generational” Varieties as Nearly Associative Combinations’, in Maurice Godelier, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 59–93. Viveiros de Castro, Eduardo. 1998. ‘Dravidian and Related Kinship Systems’, in Maurice Godelier, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 332–85. Wagner, Roy. 1969. ‘Marriage among the Daribi’, in Robert M. Glasse and Mervyn J. Meggitt (eds), Pigs, Pearlshells, and Women: Marriage in the New Guinea Highlands. Englewood Cliffs: Prentice Hall, pp. 56–76. Weiner, James F. 1988. The Heart of the Pearl Shell: The Mythological Dimension of Foi Sociality, Berkeley etc.: University of California Press.



95

4 

Why do Societies Abandon Cross-Cousin Marriage? I

Professor Harold Scheffler’s prominence in the anthropology of kinship is well established by virtue of his enormous corpus of written work on the subject, the logical rigour and scholarly care he evinced in his arguments, and his influence over American anthropology in particular in promoting semantic analyses of kinship terms and terminologies.* His studies of the latter were no doubt in part stimulated by his fruitful association with Floyd Lounsbury, but his independent work was nonetheless strikingly consistent in all these respects, and his other main claim to fame in kinship studies, his rethinking of the topic of descent, seems to have been entirely his own. There have, of course, been controversies and debates, some of them quite pointed, especially with various kinds of structuralist, who see things pretty much in a diametrically opposed fashion, though occasionally also with scholars who in general terms can be located within the Scheffler-Lounsbury ‘camp’ itself (see Scheffler and Lounsbury 1971: 73 ff.). At this point, I should admit that for the most part my own training and long-standing views oppose me intellectually to that camp. I was trained at Oxford in the heyday of structuralism – with which Oxford anthropology at the time was widely associated elsewhere – and I was supervised for my doctorate on Austroasiatic kinship by

Why do Societies Abandon Cross-Cousin Marriage?

N.J. (‘Nick’) Allen, who once identified himself to me in conversation as an ‘evolutionary structuralist’. He himself had been a student of one of Scheffler’s main adversaries, Rodney Needham, whom I knew personally and under whose influence I also fell, and who at the time counted as a leading British structuralist – less explicitly evolutionist than Allen but still nonetheless interested in how both kinship terminologies and marriage practices change.1 Given the structuralist aspect, there was also an explicit tendency to see kinship more as a matter of category words than of genealogy. Although genealogy was not dismissed entirely, it was not ordinarily seen as having much to do with how various indigenous peoples saw kinship, a position I adopted too at the time. More recently, however, I have moved away from this structuralist orthodoxy sufficiently to appreciate more how peoples the world over do think genealogically some of the time and that they are not as ignorant of the biological aspects of reproduction and parenthood as they are sometimes made out to be.2 This has still not made me an extensionist, as I made clear in an article published in 1996 on the contextual uses of genealogy and category (Parkin 1996). Among other things, that article compared Scheffler’s considerations of Tamil kinship with Dumont’s more ethnographically grounded studies, to the definite advantage of the latter. I suggested that genealogy and category are not mutually exclusive ways of interpreting kinship terminologies but rather two different forms of knowledge entirely, the latter being a form of classification, the former involving stepwise calculation to specify the details of a relationship more precisely; their difference is therefore ultimately a matter of the different contexts in which they are likely to be used.3 As I still adhere to this position, I do not want to return to this particular debate here but instead will focus on the applicability of Scheffler’s ideas to questions of change in kinship terminology and affinal alliance that have concerned Needham, Allen and myself for a number of years. That applicability is called into question first by Scheffler’s downplaying of the importance of affinity in kinship terminologies associated with prescriptive alliance in favour of his focus on consanguinity; and secondly, by his scepticism, amounting in effect to a denial, that kinship terminologies reflect social morphology in any significant sense. These positions were largely established in Scheffler’s debates with the structuralists, but among other things they minimize the potential significance of the most convincing starting



97

98

How Kinship Systems Change

point for theories of change, namely the existence of some form of cross-cousin marriage.4 More generally, Scheffler’s fundamentally synchronic approach cannot satisfactorily account for change almost by definition. I shall therefore return to essentially structuralist arguments in favour of both the existence of cross-cousin marriage (or ‘prescriptive alliance’ in Needham’s early terminology) and systemic change by asking first why societies drift away from cross-cousin marriage; and secondly, what changes when they do so – only ‘the system’ of cross-cousin marriage or rather how key relatives are classified (especially, how they cease to be classified as cross cousins)? First, however, I will briefly review Scheffler’s main arguments and targets, and also indicate my main sources, before setting out my own understandings of Scheffler’s two positions just outlined above. Section II will discuss the central matter in this article with examples; namely, the circumstances in which a society may abandon crosscousin marriage. Section III will examine the proposition that kinship terminologies also have a role to play here through the different ways in which they classify key relatives in systems of cross-cousin marriage and the practices that immediately evolve from them. Section IV provides a brief conclusion. Throughout his career, Scheffler has rejected the structuralist position that kin terms should be seen as category words that do not primarily have genealogical referents. Instead, he has consistently advocated the semantic theory of kin terms he worked out with the anthropological linguist Floyd Lounsbury. For Scheffler, kin terms are precisely genealogical denotata focused on a single kin type of a sort found within the nuclear family. This focus on genealogy is tied to Scheffler’s doctrine linking polysemy by sense generalization to the extension of meaning of some terms outwards from such foci, as well as his rejection of structuralist and structural–functionalist assumptions that kinship terminologies reflect certain aspects of social morphology. For the Lévi-Straussian structuralist, the latter typically means cross-cousin marriage or prescriptive alliance in its various forms, which are seen precisely as a reflection of such category words and the systems they form. While Scheffler certainly recognizes the facts of affinal alliance, his focus on genealogy also has the result that it renders affinity secondary to genealogy, whereas the structuralists do the reverse in making affinity primary in their models of prescriptive alliance. Thus, as I shall argue, the disjuncture that Scheffler

Why do Societies Abandon Cross-Cousin Marriage?

tends to see between terminology and social morphology must be severely qualified in relation to prescriptive alliance, although it becomes more evident the more a society moves away from prescriptive alliance as the basis for its regulation of marriage. Although Scheffler does deal with other structuralists such as Lévi-Strauss and Leach on occasion, his key targets would seem to be Dumont and Needham.5 His attack on Needham is particularly associated with a single book, A Study in Structural Semantics: The Siriono Kinship System, written with Floyd Lounsbury and published in 1971.6 I have relied on it greatly here in setting out Scheffler’s views as I see them, though I have also used other sources (Scheffler 1972, 1977, 1984) without (re-)reading the whole of Scheffler’s vast corpus. I now return to the two main aspects of Scheffler’s work mentioned above that I am dealing with in this chapter. I shall then address the key topic of this chapter, namely the circumstances in which a population may abandon cross-cousin marriage for a non-prescriptive form of affinal alliance that is no longer reflected in the kinship terminology and can therefore no longer be considered prescriptive. The overall arguments are first that Scheffler’s theories make more sense in respect of societies that do not pursue cross-cousin marriage than those that do; and secondly, that, even so, they cannot adequately account for systemic change between what are very definitely two different situations or stages.

Consanguinity and Affinity For Scheffler, kinship terms and terminologies are just that: they are predicated on local cultural views of procreation and its concomitant forms of consanguineal relatedness based on primary links of the sort one finds in the nuclear family (Scheffler and Lounsbury 1971: 63). These are generally considered ‘focal’ for Scheffler, and other consanguineal relationships and their denotata are treated as ‘extensions’ outwards from these foci. This is usually demonstrated in his analyses by a variety of ‘extension’ or ‘equivalence’ rules (they are apparently the same; see Scheffler and Lounsbury 1971: 51) expressed through a chain of genealogical symbols starting with a focal kin type (say, F) and leading to those kin types at a further genealogical distance from ego (e.g. FB, FFBS etc.) with which the focal one



99

100

How Kinship Systems Change

is equated in the terminology. This is in effect the ‘classificatory’ idea, which, of course, has a long history of use in the anthropology of kinship. Scheffler also invokes the notion of polysemy a great deal, one of his criticisms of other schools of the anthropology of kinship being that they ignore it and assume that any kin term only has one intrinsic meaning – that it is monosemic. For him, this is connected, inter alia, with the structuralist doctrine that category is prior to genealogical position and that categories have not a focal meaning (do not mean F, therefore, in our example) but an intrinsic, monosemic meaning along the lines of ‘male patrikin of the previous generation’ (i.e. F, FB, FFBS etc.). For Scheffler, when structuralists break down categories genealogically, they are relying on polysemy by sense specialization while at the same time denying the possibility of polysemy by sense generalization – that is, the extensionism described above from F to FB, FFBS etc. Formally speaking, both are possibilities, but it is clear that for Scheffler the latter is more important because he thinks that this is the way most kinship terminologies are structured, as well as accounting for – and making manageable, for both the anthropologist and the native informant7 – all the denotata of an individual term. It is, in short, the way most people think most of the time: ‘. . . polysemy within the domain of kin classification is really what it is “all about”’ (1972: 325).8 Three other points ought to be made here for purposes of clarification. First, while Scheffler certainly recognizes change and history, he sees his analyses as basically synchronic in kind (1972: 313–14). This is another way in which he departs from the structuralists, especially perhaps Needham, who regularly invoked terminological change to explain internal inconsistencies in patterning between terminology and alliance (e.g. Needham 1966–67, 1974), without making that central to his work. Secondly, Scheffler is quite clear that his extensionism is not intended to explain how children learn kin-term use. However they learn, they end up learning the adult classification, and that is what the analyst must focus on (Scheffler and Lounsbury 1971: 62). Thirdly, Scheffler is clear that polysemy does not necessarily introduce metaphor (e.g. 1972: 318 ff.). Kin terms certainly have metaphorical uses, but Scheffler conducts his semantic analyses on the basis that polysemy by sense generalization is a matter of relating the genealogical denotata of kin terms to ordered sets of rules. The possible metaphorical aspects of kin-term use should, in his scheme, be subject to a separate study.

Why do Societies Abandon Cross-Cousin Marriage?

Since Scheffler sees kin terms as having primarily consanguineal significance, he is apt to view affinal denotata as secondary. It is not clear whether he thinks that affinal denotata can ever be focal denotata, though he appears to make an exception for H and W. However, extensions from H or W as focal kin types would not work in the same way as F > FB > FFBS links in the normal classificatory sense: one can get from H to HB, but then what? Logically, HBFBS etc., but these are not specifications the kinship analyst is ever likely to have to deal with. Trautmann was later (1981: 59–60) to identify another problem in deciding, for example, the focal specification in equations like FZD = HZ, since neither is obviously prior, and positing such equations leads to a circularity in which each specification implies the other. In the case of prescriptive terminologies, which may well lack terms that are solely affinal, Scheffler chooses to view terms denoting affinal relatives as basically consanguineal. For a structuralist like Dumont, by contrast, they, like the terminology as a whole, may primarily express affinity, as he found was the case for ethnographically Dravidian terminologies like Tamil (Dumont 1953). For Scheffler, probably, the very invention of affinal specifications to attach to such terms is merely a function of the way the western analyst interprets and analyses non-western classifications; consanguinity and affinity are analytical concepts, but the former has more relevance than the latter because it is linked to genealogical ties connected with locally valid but still very general cultural ideas of conception and birth. While accepting the social uses of kinship terms and categories, Scheffler rejects the structuralist notion that kin terms must be considered primarily as ‘social’ categories both because of the potentially affinal denotations of some of them and because of their infinite extension from a specific genealogical base through the classificatory idea. That is, he disputes the idea that they are anything more than expressions purely of kinship in the narrow genealogical sense that he has always been keen to stress as primary. Thus, ‘The spouse relationship is essential to any system of consanguinity and affinity . . . but not to systems of consanguinity, i.e. systems of kin classification per se’ (Scheffler and Lounsbury 1971: 81 n. 11, emphases in the original). And further, ‘Relations of genealogical connection, or kinship proper, are fundamentally different from and are logically and temporally prior to any social relations of kinship’ (Scheffler and Lounsbury 1971: 38) – of a jural kind, more especially (Scheffler and Lounsbury 1971: 39).



101

102

How Kinship Systems Change

Do Systems of Prescriptive Alliance Exist? While Scheffler does sometimes discuss the potential social morphology correlates of specific terminologies (see Scheffler and Lounsbury 1971: Chs. 8, 9), his analyses of the latter do not require this and can be, and normally are, conducted in a sociological vacuum. In general, he is sceptical of views that particular types of terminology ‘reflect’ aspects of the social structure. One of his targets here is Radcliffe-Brown’s suggestion that Crow and Omaha terminologies reflect the principle of descent-group unity – matrilineal and patrilineal respectively – on the basis that some of their internal equations map this out to some extent (ibid.: 15–18; also 63–64). Because they only do this ‘to some extent’, and because of the considerable variety of Omaha terminologies especially, Scheffler denies that this is a significant correlation; in addition, he has also deconstructed the idea of descent totally, clearly preferring the notion of filiation – that is, parent–child ties – as being more significant.9 Another target of Scheffler’s is the argument that cross-cousin marriage in Australia is invariably linked to section systems, which he dismisses because of the variation in both cross-cousin marriage and section systems that populations there exhibit (Scheffler 1977). One general principle invoked here (e.g. Scheffler and Lounsbury 1971: 152–53) is the wellknown fact that similarly classified kin are not necessarily treated alike jurally: in particular, our obligations to a closely related individual in a particular kin class may be followed more diligently than in respect of someone in the same kin class who is more distantly related to us. To an extent, then, this is also a matter of behaviour as well as jural rights and obligations. More important here, however, is Scheffler’s dismissal of the notion of prescriptive alliance as a system reflected in particular forms of terminology. Scheffler does not deny that ego may have a claim in marriage on an alter who is usually going to be a cross-cousin of some description, but he also notes that such claims are rarely enforced across the whole society, that they may be evaded without detriment to the way the society defines itself, and that as a result the statistical count of such marriages may be very low indeed. In addition, societies united in their possession of a prescriptive terminology are scarcely similar in any other respect. For Scheffler, therefore, attempts by Needham and his followers to treat such societies as holistic in the Maussian sense, typically correlating prestations and

Why do Societies Abandon Cross-Cousin Marriage?

dual symbolic classification with the principles of affinal alliance, are ultimately hollow.10 While we may have a marriage rule predicated on cross-cousin marriage and the expected terminological equivalences, systems of asymmetric affinal alliance – or any other form of affinal alliance, for that matter – do not exist for Scheffler: ‘. . . the most distinctive feature of those [terminological] systems which do employ the MBD-FZS spouse equation rule is that rule itself’ (Scheffler and Lounsbury 1971: 220). Further, the reciprocal rights associated with affinal alliance exist between individuals, not social groups (ibid.: 223): in Scheffler’s view it is therefore not correct of Needham to view prescriptive alliance as the cement of society, or as holistic in the sense that the whole of society is ordered by a simple relation of either symmetry or asymmetry between spouse-exchanging groups, expressed in symbolic values, as well as in actual exchanges of spouses and wedding prestations.11 Also, for Scheffler the failure of all egos to marry, say, MBD/FZS, despite a rule enjoining them to do so, is to be seen as intrinsic to the way such rights and obligations are pursued, not interpreted as breaches of the rules requiring redefinitions of the relatives involved (ibid.: 223–24), the latter being a key property of prescription for Needham (see further below). Scheffler also evidently feels that Needham himself has caused confusion by first positing prescriptive alliance between groups (especially descent groups), then being forced to deny, in the face of contrary evidence, that groups of any sort were necessary for prescriptive alliance to be pursued. Scheffler and Lounsbury make much of this in relation to the Sirionó of north-east Bolivia, where a rule of asymmetric cross-cousin marriage apparently occurs in the absence of any social groups like descent groups regularly exchanging alliance partners. As a result, ‘. . . it is not necessary to posit a system of affinal alliance between descent groups to give a reasonable and satisfactory account of the matrilateral cross-cousin marriage prescription of the Sirionó’ (Scheffler and Lounsbury 1971: 178). There is also a clear tendency in their discussions of some of the structuralist ‘classics’ for Scheffler and Lounsbury to stress the Crow-Omaha-type equations some of their terminologies make and to underplay the alliance aspects. Thus the Kachin are described as ‘just an Omaha-type system with an overlaid MBD-FZS-spouse equation rule’ (Scheffler and Lounsbury 1971: 199), which itself merely ‘has the status of a corollary of a more fundamental rule’ (in this case an Omaha-skewing rule; Scheffler and Lounsbury 1971: 178).



103

104

How Kinship Systems Change

Elsewhere, Scheffler (1971) also denies that ‘Dravidian’ terminological patterns reflect the practice of bilateral cross-cousin marriage of which they are a logical expression. Indeed, he claims that there is no fundamental difference between Dravidian and Seneca-Iroquois terminologies, despite his close collaborator Lounsbury (1964) already having proved that Morgan had been wrong to treat them as identical. However, treating them as identical is precisely what Scheffler does on the basis of their focal kin specifications being identical, and despite Seneca-Iroquois terminologies not being considered prescriptive (unlike Dravidian terminologies), partly because SenecaIroquois terminologies typically have separate affinal terms, and partly because they are less consistent in how they treat the cross-parallel distinction; for Scheffler, these differences are secondary to what unites them, namely their focal kin types. As in the case of asymmetric alliance, Scheffler also pointed out (ibid.) that even in the south Indian region that gave them their name Dravidian terminologies can coexist with a preference for marriage to MBD but not FZD, or vice versa, or no declared preference at all, as well as one for the bilateral cross cousin that the terminology expresses. His position here was rejected by Trautmann (1981: 60–62) after the latter had spent a dozen or so pages of his monumental study of Dravidian kinship submitting a logically consistent Dravidian-type terminology to semantic analysis in the manner of Scheffler and Lounsbury and then deciding that it was essentially a circular procedure: for instance (and as noted above), it could not handle consanguineal–affinal equations of the type FZD=HZ, as there was no way of deciding which of these kin types was the focal one (if any, indeed, as neither occurs within ego’s nuclear family; see Trautmann 1981: 48–62). As for the lack of fit between terminology and marriage patterns, Trautmann sought to deal with this by bringing in history and the probability that the situation has changed over time: ‘I hypothesize that bilateral cross cousin marriage is ancestral to all particular cognate Dravidian systems we find in the ethnographic present’ (ibid.: 62, emphasis in the original), regardless of whether such systems have that form of marriage at the present day. This recalls a similar demarche made on a number of occasions by Needham in relation to asymmetric alliance. Thus in one double article (1966–67) Needham shows that MBD marriage can occur in three societies with a symmetric prescriptive, an asymmetric prescriptive and a nonprescriptive terminology respectively and postulates historical change

Why do Societies Abandon Cross-Cousin Marriage?

to account for this. Elsewhere, he makes a similar historical argument in the case of the Warao of Venezuela (Needham 1974), who have moved away from bilateral cross-cousin marriage while retaining much of the terminology which hypothetically originally went with it. The same idea of change also underpins Allen’s theory of the earliest human kinship system as tetradic, which he has put forward as a starting point for change of this type, involving a form of bilateral cross-cousin marriage closest to the Kariera system, though it is not attested ethnographically.12 In fact, it is still possible to find societies with prescriptive terminologies where classificatory cross-cousin marriage clearly takes place on a regular basis, nullifying Scheffler’s and others’ objections to their existence on the basis of low rates of actual cross-cousin marriage and inconvenient unilateral preferences accompanying bilateral terminologies. Nevertheless, the argument that there is frequently a lack of fit between the patterns respectively of terminology and alliance has long been realized. Here Scheffler converges just slightly with his antagonists: the question is really what to do about it. Needham’s answer (1973) was eventually to decide that, despite the word ‘prescribe’ logically being applied to the marriage rules, the notion of prescription was really to be located in the pattern of the terminology, not in that of the alliances themselves. As a categorical system, any kinship terminology defines how people, or rather the categories they belong to, are related, but a prescriptive terminology also redefines the kin involved in a marriage that breaks the rules: further, a terminology cannot be broken in the way that rules can. Moreover, one expects rules to be broken as a normal part of the operation of social life anywhere, therefore to expect anything approaching 100% observance of them is simply unrealistic, just as it is equally naïve not to accept that, even while being observed, rules may well be manipulated to satisfy particular interests or simply to accommodate what is possible for particular egos.13 As a result, one needs a third level of analysis, namely actual behaviour – that is, the extent to which people obey the rules.14 At this point, therefore, we can see that there is a wide gulf between the positions Scheffler has adopted and those of his structuralist opposite numbers, myself included. In what follows, I will seek to take the matter forward both ethnographically and theoretically. Section II is more concerned with rules and behaviour; Section III with terminology or the level of classification. Also, Section II is much more rooted in ethnography, whereas the arguments in Section III are more



105

106

How Kinship Systems Change

general. This means that, apart from the occasional stray remark, I am not linking the data in Section II with an examination of any associated terminology. Rather, to repeat, my aim is to show that having cross-cousin marriage and not having it should be seen as two separate situations, possibly linked as stages; and that, despite Scheffler’s argument to the contrary, they are not invariably to be interpreted simply in terms of a failure to observe marriage rules 100%. If that were routinely the case, why have the rules and the associated terminology in the first place? Nonetheless, in Section III I return to the terminology as a possible explanation for, or at least associated feature of, such changes. Discussion here revolves around the insight – drawn partly from how prescriptive terminologies work in redefining kin who have married ‘wrongly’ – that classifications do not just reflect the world but determine in large measure how the world will be perceived. This insight can, in principle, be applied to kinship terminologies as much as to anything else.

II In pursuing such lines of enquiry myself, I have adopted structuralist paradigms in their essentials, as well as preferring to focus on cases where it is feasible, even necessary, to bring in a diachronic perspective to explain the synchronic analysis. Synchronic analysis itself often reveals inconsistencies in the logical patterning of the terminology,15 which also involves relating actual kinship terminologies to a set of types against which ethnographic data can be measured. One can then test the basic hypothesis that change has taken place in order to account for the logical inconsistencies, possible redundancies, etc. in the terminology. There is sometimes resistance to such methods: for Ellen Basso (1970), in her debate with Gertrude Dole (1969), one needs to seek the reasons for a terminology having the pattern it does in present-day social practice, not – as she alleged Dole was doing – rely on either theories of social change or alleged deviations from neat typologies to explain that pattern.16 However, this does not always lead us very far, and when, for example, one finds evidence of prescription in the terminology but not the actual alliances or marriage rules that would logically correspond with it, then it is reasonable to posit that change has taken place in the latter but has not started or not been completed in the former.

Why do Societies Abandon Cross-Cousin Marriage?

One factor to be taken into account here is the persistence of theories that cross-cousin marriage, especially in its symmetric form, represents the original form of marriage in human history and that all other forms have derived historically from it. This was the underlying assumption of much of Needham’s comparative work on kinship, but at the present day it is more usually associated with Allen’s tetradic theory, already mentioned in passing in the previous section (see note 12 for references). This theory postulates a particular variant of bilateral cross-cousin marriage as the starting point for human kinship. It has become increasingly influential in recent years though also controversial, in the sense that its argument that it accounts for early human kinship has been questioned by others on ethnographic grounds (e.g. Barnard 2008; Layton 2008). Nonetheless it forms a reasonable starting point for theories of the evolution of kinship systems. Next, therefore, I consider what steps may lead from a system of cross-cousin marriage (situation 1) below) to the ‘open’ arrangements of semi-complex and complex societies (Lévi-Strauss 1949) that lack them: 1) Cross-cousin marriage, with a sociocentric terminology to match, whether tetradic – i.e. with only four terms (Allen 1986), or not, and with everyone following the prescription generation after generation. 2) Possible evolution from symmetric to asymmetric prescriptive, or alternatively to an eight-section system assuming marriage consistently (and symmetrically) between genealogical second cross cousins, or to some other system that can be described as prescriptive. These options may well be mutually exclusive. 3) Only one member of a group of siblings is required to follow the prescription; other siblings may or must marry into other families or kin groups (this is more often noted of asymmetric prescription than symmetric). (Examples: Hicks 1985: 77–78 on the Mbae of Manggarai, Indonesia; Lindell et al. 1979: 64, 66 on the Kammu (or Khmu) of Thailand; and Kunstadter 1966 on the Lawa (or Luaq) of Thailand.) Possible emergence around this point of a class of non-relatives not covered by the terminology. I am treating this as a separate stage from that in which everyone is expected to marry a classificatory cross cousin.



107

108

How Kinship Systems Change

4) Abandonment of cross-cousin marriage seen as a repeated practice of exchange between kin groups but still a tendency, even a rule, for groups of siblings to marry exclusively and sometimes intensively with one another; e.g. through direct exchange within that generation. Such marriages may be expressed as taking place between siblings-in-law (GEG categories), and they are often associated with a ban on repeating alliances between any two groups in the immediately following generation(s). Some Munda peoples of central India provide examples (see Parkin 1992: Ch. 8, also, Ch. 2). Terminological change of some sort (e.g. towards Iroquois or Hawaiian, or at least the emergence of specifically affinal terms) is likely as a result. 5) Though there are no longer any marriage prescriptions, marriage prohibitions continue to be framed in part by referring to kin categories and/or social groups, such as clans related in specific ways to ego. Often associated with Crow-Omaha terminologies (especially by Lévi-Strauss) but not only or necessarily. 6) Even social groups like clans cease to be relevant, or even to exist, and only certain categories named as unmarriageable are left of previous situations in which kin categories governed marriage options. Inter alia, the situation in most western societies, where for many people marriage partners are not supposed to be related prior to the marriage at all. There is a certain tendency for later stages of this sequence to be associated with first a greater dispersal of marriages between groups, and secondly, greater individual freedom of choice unrestricted by social obligations to marry in particular ways. Both assumptions have to be qualified. Cross-cousin marriage may appear excessively restrictive of choice to the western mind, but in fact first cross-cousin marriage could not work for simple demographic reasons, as not every ego will have a referent in that category. As a result, anthropologists soon realized that a wider range of equivalent kin is involved, such as second and remoter cross cousins, or persons placed in the same categories but without traceable genealogical links to ego, and they developed the notion of the classificatory cross cousin to cope with this. Later, especially after the structuralist revolution in kinship studies, it began to be recognized that suitable spouses need not be defined genealogically at all but should rather be treated as members of a category that may be based, for example, on long-term inter-

Why do Societies Abandon Cross-Cousin Marriage?

group relations – that is, ego seeks a spouse where other members of his or her group have done so already in the recent past without needing to determine exact links genealogically with that spouse. In other words, choice is still possible with cross-cousin marriage, as ‘cross cousin’ may actually be quite a large category in the indigenous view. Classificatory cross-cousin marriage also makes possible the dispersal of alliances between groups, though this may be ruled out or restricted where there is a set number of marriage classes (e.g. four or eight) or just two moieties. Situation 2), above, involves such dispersal in that more than two spouse-exchange groups are involved. Situation 3) enjoins dispersal in subsequent generations by virtue of the ban on repeat alliances. Situation 4) ensures it by banning, for example, male ego from marrying into his mother’s clan, which is where his father had sought a spouse, thus preventing the intergenerational repetition associated with cross-cousin marriage. With situation 5), we arrive at the abandonment of any influence of category over marriage apart from the incest taboo. Even with cross-cousin marriage (situation 1), but also more generally, the relevant categories may frequently be manipulated to justify technically ‘wrong’ marriages. Even if that is not the case, there may be more than one genealogical path linking ego to a desired spouse, which may provide a way of justifying an otherwise questionable match or rejecting a perfectly sound one. There is nonetheless the possibility that cross-cousin marriage still restricts alliances in ways that come to be seen as unacceptable and that this induces change. Why should this be? Despite Scheffler’s scepticism, it has long been recognized that in most societies marriages are not just a matter of individual choice but are attached to social obligations between groups – indeed, individual choice may mean nothing, as in the very many societies where children are betrothed or bestowed on their spouses and the latter’s social groups by their parents or other senior relatives, often at or even before birth (cf. Needham 1986). These social obligations may be a matter of politics, especially as they may be manipulated to suit a particular political strategy – this is the concept of alliance, linked to marriage especially by French writers such as Lévi-Strauss (French allié = affine). To pursue political strategies effectively, one needs flexibility – that is, the ability to choose from among several partners, whether the chosen partner is a spouse or a bestower of spouses. As already indicated, a system of strict cross-cousin marriage does not necessarily rule out



109

110

How Kinship Systems Change

choice, and the categories even of these systems are perfectly capable of manipulation in the interests of pursuing a political strategy, though the choice may not be ego’s and alter’s. However, as already noted too, choice may be limited if the number of social groups (classes or moieties) associated with cross-cousin marriage is also restricted. This may lead initially to such groups being abandoned but cross-cousin marriage continuing (also to the abandonment of tetradic society). However, even this situation will have restrictions that may eventually be seen as irksome, and the resulting tensions in the system may then set in train the sorts of change listed above. Political strategies will, of course, tend to be pursued by the leaders of society, if any. However, ordinary members of society may also be led to pursue them in their own private interests or those of their immediate family. The interests involved may be relatable to notions of romantic love, as ideally in most western societies, but more usually they will be connected to considerations of one’s future economic survival or appropriate social positioning – in some societies, ego may have closer relationships with siblings-in-law, who may also be cross cousins, than with siblings (e.g. Sinhalese in Sri Lanka, Leach 1960). There is also a gender aspect here: given that most societies have distinct regimes of male dominance, women are more likely overall to be the pawns in male games of marriage politics than vice versa. Age is yet another factor: even males, as boys, may become pawns in the same game, while in societies where older men tend to monopolize marriages through polygyny, as among the Tiwi in Australia (Hart and Pilling 1979), younger men may be at the mercy of their elders’ political arrangements well into middle life. The individual’s desire for a perfect partner, often interpreted as a matter of romantic love, is certainly a factor demanding a degree of choice, but it is clearly more relevant in the west and societies that have been significantly influenced by it. Worldwide, it is of less salience than the importance given to marriage, and to the use of women in marriage, as a mode of political alliance between social groups.17 Missionary influence may also have an impact in drawing many societies away from cross-cousin marriage, as with at least some Lamaholot villages in eastern Indonesia (Barnes 1977: 137, after Raymond Kennedy). However, this is not always the case, despite the presence of missionaries (e.g. Désvaux and Selz 1998 deny this happened among the Cree and Ojibwa of Ontario and eastern Manitoba, Canada); similarly, cross-cousin marriage apparently survives among

Why do Societies Abandon Cross-Cousin Marriage?

south Indian Christians, despite the Catholic Church not approving of it (Kapadia 1993: 46).

The Americas Many sources on the Americas explain the retreat from cross-cousin marriage in terms of the physical expansion of hitherto small communities and/or their greater contact with neighbouring groups etc. One example is Paul Henley, in a brief but wide-ranging comparative study of Amerindian kinship in the Amazon (1996).18 He points out that where different populations are scattered along river systems, it tends to be those living up the headwaters that pursue cross-cousin marriage, while those downstream do not do so. The latter are more in contact with other groups, partly because of trade or political conflicts or alliances with such groups. However, Henley doubts that the usually suggested trajectory from cross-cousin marriage to its absence works in the Amazon, and indeed he seeks to reverse it, seeing his ‘Amazonian type’ as being more fundamental. This type is similar to the canonical dravidianate in so far as the general distribution of terminological categories in the three medial generations is concerned, but it is very different in three other crucial and related respects: the absence of a positive rule of marriage, the absence of a category of cross-relative in Ego’s own generation and the presence of a set of exclusively affinal terms. (1996: 62)

It is also evident that both the intensification of sibling exchanges within a generation without cross-cousin marriage and the repetition of alliances after the elapse of a number of generations also occur in the Amazon, though Henley does not list these as features of his type (cf. Parkin 2013b). He argues that the cross-cousin marriage of groups up the headwaters is an adaptation of his Amazonian type to cope with the consequences of population decline and/or small populations, which have fewer options than those downstream. He also identifies groups on middle stretches of water that have Iroquois crossness, not Dravidian. The geographical transitions thus correlate with the typological ones. In particular, the upstream groups may be remoter, more peripheral, being those that went furthest into the interior, assuming, as is likely, that these river systems were the main means of access and movement to and in these areas.



111

112

How Kinship Systems Change

More specific examples include the Urarina of lowland Peru (Walker 2009) and the various Jivaroan groups dealt with by Taylor (1998), also discussed in this context by Walker (2009: 65). Other anthropologists, too, have tried to connect geographical isolation of any sort with cross-cousin marriage. Needham (1974: 27–29) noted that the Warao, who do not have or no longer have cross-cousin marriage, have been exposed to outside influences and contacts for centuries through their occupation of part of the Orinoco delta, whereas the Yanomamo, although perhaps closely connected with them in prehistory, have been much more isolated until recent times up the headwaters of the Orinoco and other rivers. Needham only notes (ibid.: 28) that they also have a symmetric prescriptive terminology. His point would have been strengthened had he consulted their main ethnographer at the time, Chagnon, who makes it clear that, unlike the Warao, they also have bilateral cross-cousin marriage (Chagnon 1968: 125 ff.). In the literature on North America, the explicit focus tends to be more on notions of endogamy and exogamy (e.g. Ives 1998; Smith 1974 on the Ojibwa), though arguments for their significance tend to resemble those made for the South American examples. While the type of group or unit to which the endogamy and exogamy apply is not always specified, a local residential community composed largely or entirely of recognized kin usually seems to be intended. The usual argument appears to be that cross-cousin marriage is pursued in such communities as a system of close-kin marriage. Conversely, in areas where both population and food resources are thinly distributed, cross-cousin marriage may be less viable as a basis for cooperation between widely dispersed groups. Also, the system becomes less and less attractive as individual communities expand demographically and/or geographically; for example, by moving into new hunting and foraging territories, whether in the plains or in forests, or expanding trade relations. Under these circumstances, a greater range of political and trading alliances with other groups becomes necessary, and as these are partly pursued through affinal alliance, close-kin marriage becomes a constraint, as cross cousins tend not to be found in these other groups. Cross-cousin marriage is therefore progressively abandoned, although there may nonetheless be a tendency to marry known kin or affines from previous alliances in a more general sense. The terminologies are also modified, perhaps in only one or two levels to begin with, such as the Hawaiianization of ego’s level,

Why do Societies Abandon Cross-Cousin Marriage?

but there may also be a shift from Dravidian to Iroquois crossness, though this is difficult to document. For the Ojibwa (Smith 1974), one imperative historically may have been to use affinal alliance more widely than cross-cousin marriage to unite all Ojibwa groups in a single federation. However, contrary to Hickerson (1962), who saw cross-cousin marriage as dying out among the Ojibwa in the late seventeenth century for reasons that apparently had little to do with contact with Whites, Smith argues that the abandonment of this form of affinal alliance came much later, in the nineteenth century. It also had reasons very much associated with contact: a decline in hunting possibilities, itself a consequence of the excessive demands to supply the fur trade, which drove out food sources; the US government’s reservations policy and the consequent sedentarization of Ojibwa and their dependence on government handouts; a population explosion; and the rise of individual trading and other forms of employment, with less sharing across kin groups and the dropping of traditional obligations to kin and affines. However, Smith is more inclined to accept the possibility that it was cross-cousin marriage itself that linked bands in the pre-contact period. He also argues that, in modern conditions, groups became more endogamous as relations between groups became less important. However, he does not suggest that cross-cousin marriage was reverted to because of this. Cross-cousin marriage may be linked to endogamy in this theory, but the indications are that once this form of marriage has been abandoned there is no way back, whatever the circumstances. Similar models of small, isolated or dispersed populations practising cross-cousin marriage and larger, more consolidated ones abandoning it can be found in Ives’s extensive comparative studies of Native American populations (e.g. 1998 on Athapaskan, Algonquian and Numan populations). Ives makes much of the modalities of what he calls ‘sibling cores’ and their residence patterns. In the case of bilateral cross-cousin marriage, where men exchange their sisters, it is claimed that they will live together and that their children may then marry in the same fashion. However, where the rule is that two brothers marry two sisters, it is linked in this theory with residential exogamy. And this seems to reflect deliberate decisions:19 writing of the Wrigley Slavey (after Asch 1998), Ives says that they ‘deliberately fashioned same-sex sibling cores that enforced local-group exogamy in the first descending generation. The entire logic of this framework is to keep potential affines outside the local group. . . . Asch



113

114

How Kinship Systems Change

found a distinct tendency to call even cross cousins by sibling terms, widening the field for marriages’, as any categorical sibling would be banned in marriage (Ives 1998: 100). Yet there is no intrinsic link between residence rules and marriage rules. Ives’s theory may reflect local conditions (and his reconstructions are meticulous and detailed), but it cannot apply universally.20 The more general theory that cross-cousin marriage disappears with increases in group size and consolidation may have something to it, but it does rely on a notion of cross-cousin marriage as necessarily close-kin marriage, which, as argued above, it may not be. One countervailing theory, generally dismissed as unlikely by other anthropologists,21 was put forward by Gertrude Dole (1969): she saw endogamy caused by population decline among the Kuikuru as forcing a change away from cross-cousin marriage, as cross cousins would live in the same residential cluster and ultimately see each other as non-marriageable kin. The terminology reflects this in its Hawaiianization of ego’s level but not as yet the adjacent levels, a pattern Dole called ‘bifurcate generation’. This was immediately rejected by Basso (1970) from her work on the nearby Kalapalo, where she noted that although here too cross-cousins could live in the same cluster they still married one another. However, they also glossed over the fact so as not to draw attention to the reality of affinity (as elsewhere in this region, the Kalapalo seek to define affines as consanguines wherever possible).22 Another possible cause is hinted at very briefly by Elmendorf (1961) in his comparison of the interior and coastal Salish in northwest North America. While the interior retains a bifurcate collateral terminology reminiscent of cross-cousin marriage, the coastal areas have a ‘lineal’ (or ‘cognatic’, to use Needham’s more exact term) terminology and no cross-cousin marriage. The coastal areas also have more social and political stratification, with a stratum of chiefs. Elmendorf does not elaborate further on the implications of this, but he may well have in mind an idea that the constraints of cross-cousin marriage were found to restrict the sorts of political alliances in which the chiefs presumably indulged. And as Sahlins remarked, writing on Fiji, ‘when differences in community-wide political rank become very great, kin terms and usages are likely to be dropped altogether’, it being ‘somewhat improper for low-ranking people’ to use them of or to their superiors (cited in Service 1971: 111). However, south India is replete with stratified polities pursuing cross-cousin marriage:

Why do Societies Abandon Cross-Cousin Marriage?

in any such society, people in lower social strata will refer to and address each other in kin-related ways, even if they do not do so when looking above them in the social scale.

South India This has long been recognized as an area of extensive cross-cousin marriage, basically bilateral or symmetric prescriptive, but often with a preference for one or other cross cousin without that upsetting the symmetric prescriptive terminology.23 It has also given the name ‘Dravidian’ to this type of affinal alliance, as this is also an area of Dravidian speech communities, and there is indeed a high though by no means invariable association between language and type of affinal alliance here (see Trautmann 1981). However, some more recent ethnography (Fuller and Narasimhan 2008; Kapadia 1993) indicates clearly that cross-cousin marriage is no longer being followed as consistently as in the past for reasons ultimately connected to changes in attitudes to arranged marriages. Before discussing these cases, we should turn briefly to the Nayar, historically a strongly matrilineal group of sub-castes in Kerala with a very attenuated system of marriage not involving cross cousins per se. In this system, Nayar women first attained marital status through a ritual involving a man from another lineage – a ritual marriage that is not necessarily consummated24 – before being impregnated through sexual relations with a series of other, so-called sambandham partners in a relationship that is perhaps most suitably described as concubinage. The usual explanation for the emergence of this system is that, while some Nayar were rulers, others formed a military caste, and it was a way of protecting the taravad or matrilineal extended family from outside interference when Nayar men were absent on military service. Indeed, it had the effect of doing this more generally: the taravad was based on brother–sister ties, not husband-–wife ties, the role of the sambandham partners merely being to impregnate Nayar women and nothing more.25 Despite the peculiarities of this system, which has been treated as a test of the proposition that marriage is universal, both Dumont (1983 [1961]) and Trautmann (1981: 208–14, 417–25) manage to show convincingly that it can be fitted into pan- or south Indian norms respectively. Other reasons for thinking that it may represent a shift away from a more original sys-



115

116

How Kinship Systems Change

tem of bilateral cross-cousin marriage are its existence within an area that is strongly associated with the latter – the Nayar are also Dravidian speakers, after all – and the possibility that the Nayar only became matrilineal in historical times, perhaps in the tenth century AD (Moore 1985: 526).26 Under legal changes introduced originally by the British, most taravads have been dissolved, and kinship now basically consists of bilateral nuclear families. Chie Nakane, in a study carried out in the mid−1950s (1963: 24–25), indicated that some cross-cousin marriage was briefly being revived among the Nayar as the taravads broke up, since it was a way of keeping property together that might otherwise have been dispersed between competing kin groups. However, she adds that ‘the present younger generation strongly avoid cross-cousin marriage, as they think it is not good biologically’ (ibid.: 25), perhaps a reference to folk theories of inbreeding emerging under western influence. I turn at this point to more recent material on this region. Fuller and Narasimhan (2008) discuss the case of another high-status Brahman sub-caste, the Vattima of Tamilnadu, a state neighbouring Kerala. Here, the prevailing influences leading away from cross-cousin marriage appear to be exclusively modern, especially the intrusion of class values into caste practice. More specifically, these are middle-class values that are construed locally somewhat differently from their supposed Euro-American models, for example, in that they insist on withdrawing women from non-domestic labour. The Vattima are influenced by north Indian values – in their case by the kanya dan ideology of giving a virgin daughter to a family of higher status within the sub-caste as a supreme gift to one’s superiors – but they have traditionally pursued cross-cousin marriage as well.27 However, there has recently been a shift in the ideal criteria adopted in seeking spouses, with traditional emphases on the importance of a suitable alliance, regardless of one’s daughter’s wishes, tending to be replaced by a greater stress on the personal characteristics and compatibility of prospective spouses. Added to this is a focus on education and employment prospects in India’s modern economy, as well as the use of global networks, global forms of advertising etc. to find the right match. Fuller and Narasimhan are careful to point out that these are not love marriages and that there is no conflict with the concept of arranged marriages; it is simply that those actually getting married are more likely to be involved in the arrangements themselves. The authors accordingly call these ‘companionate mar-

Why do Societies Abandon Cross-Cousin Marriage?

riages’, to stress this new focus on the compatibility of spouses. There are now also more marriages to non-Vattima Brahmans, though much more rarely to non-Brahmans. There seems to have been a greater stress traditionally on patrilineal descent, as a bride would stay in her natal home until the birth of at least one child, though children were also fostered long-term to their mother’s brothers later. Now conjugal nuclear families are established immediately after marriage. The authors also state that there are now fewer close-kin marriages, only 10% of their sample being with a cross-cousin or sister’s daughter:28 ‘The last close-kin marriage in our genealogies occurred in 2002; we know of only two others since 1990’ (ibid.: 742). This is said to reflect modern concerns about inbreeding and the greater concern for the compatibility of partners, but the imperative of keeping land together through cross-cousin marriage is no longer so strong now that so much land has been sold outside the community. Furthermore, it is said that a daughter-in-law now comes under less pressure from her natal family to knuckle under to her new affines if the two families are not already related through previous affinal ties. Conversely, north Indian influence may be reflected in the circumstance that the groom’s family is now expected to pay less of the wedding expenses. This entails a shift away from the rough balance of marriage prestations associated with cross-cousin marriage in the direction of an absolute imbalance between the bride’s side giving everything as a dowry while the groom’s side gives nothing, as in the classic north Indian practice of kanya dan (literally ‘the gift of a virgin’ in marriage). An earlier article by Karen Kapadia (1993) on non-Brahman castes in Tamilnad also notes a decline in close-kin marriages, including between cross cousins. As with the previous case, this is felt to reduce the influence of the bride’s natal family over that of the groom, and it also means that women are left much more to their own devices in disputes with their husbands’ families. Again, there is a shift away from a rough balance in marriage prestations to a north India-style dowry, aggravated by the large-scale out-migration of men and a concomitant shortage of husbands. Kapadia is much more concerned with the consequences of these changes for women, and she stresses in particular the greater seclusion of women under the apparently mistaken local assumption that withdrawing them from their traditional labour activities reflects western middle-class values – as if this were upper-class Victorian Britain rather than contemporary British society.



117

118

How Kinship Systems Change

As with the previous case, Kapadia mentions education and employment prospects as more important considerations than the traditional emphasis on kin as marriage partners. The claims of such kin on one another are now being ignored in the pursuit of other, more modern interests. One of these is certainly prospective husbands seeking the largest dowry, concomitant with their own rise of status in the labour market as they abandon traditional agricultural labour for more comfortable jobs in the government service. However, Kapadia’s statistics (1993: 44 ff.) indicate that it is the lower castes that are most likely to practise non-kin marriage, though still with considerable percentages marrying a close or classificatory cross cousin or sister’s daughter, and despite the latter being regarded as the ideal forms of marriage. However, this is not a new situation: ‘it . . . is clear that there has been a striking discrepancy between marriage preference and marriage practice in the non-Brahman lower castes for at least five decades’, ‘right through the three generations considered’ (ibid.: 46). Only the wealthier Vellan Chettiar caste of landowners lived up to this preference, with over 97% of marriages with cross kin. Despite the changes in lower-caste practices, however, there has been no change in the pattern of terminology (ibid.: 48–49), which is probably the case for the previous example also. This is hardly surprising in itself, given received wisdom that changes in terminology lag behind changes in alliance practice, though in Kapadia’s case the changes in practice were evidently already in train around the time of the Second World War, fifty years before she wrote. In both these cases, modern changes are obviously having an impact. Among these are the development of the modern Indian economy, the sale of land, the out-migration of men especially, urbanization and modern lifestyles generally, and the influence of ideas of class drawn ultimately, if in modified form, from the west. However, the spread of north Indian values relating to kinship specifically, especially dowry marriage and marriage to cousins and other close kin, can also be detected. A very recent book by the late Isabelle Clark-Decès (2014) confirms the sudden and rapid move away from close-kin marriages currently among Tevar castes in Tamil Nadu. Like Anthony Good (1980), she sees elder sister’s daughter marriage as traditionally more important than cross-cousin marriage generally in this area, but in respect of both forms she also rejects the alliance perspective deriving from Dumont (especially Dumont 1953), seeing Tamil attitudes to marriage much more as a matter of like marrying like, of marriage

Why do Societies Abandon Cross-Cousin Marriage?

between status equals, not of structured oppositions between consanguines and affines who, potentially at least, are status unequals. This is especially true of elder sister’s daughter marriage, which, she argues, represents a closed marriage within the kin group formed of opposite-sex siblings and their descendants, whereas cross-cousin marriage has at least the potential for extending links through marriage and negotiating status in the usual Indian fashion. Clark-Decès also offers three explanations for the decline in all close-kin marriages in these castes (2014: 123–28). First, improved living conditions, public health campaigns and presumably (though not mentioned by Clark-Decès) ‘modern’-style aspirations for education and steady employment have combined to lower family size and increase age differences between the generations, thus reducing the number of close kin one may marry while at the same time making it less likely that a mother’s brother and an elder sister’s daughter will be of roughly the same age at marriage, despite the difference in genealogical level. Secondly, attitudes have been changed by the somewhat distorted ideas of the genetic damage caused by closekin marriages that now circulate freely across India, whether in the media or through official public health campaigns. Thirdly, considerations of the financial standing and educational levels of both bride and groom are replacing the generally very strong claims (Tamil urimai) that close kin formerly had on one another as spouses. Contra Kapadia (1995), however, Clark Decès does not attribute this change solely to the growing practice of dowry payments, which, unlike Kapadia, she does not see as a solely modern innovation any more than are the negotiations over status with which the practice is intimately connected. Rather, it seems Clark-Decès is arguing that young women are catching up with young men in the educational stakes, making them a more valuable asset in the marriage market and thus contributing to changes in traditional marriage attitudes and practices generally. I should mention one other case, or series of cases, here, namely the Munda speech communities further north in India, which were the subject of part of my doctoral thesis. I have discussed them on many occasions before in this context (see especially Parkin 1992: Ch. 8; see also above, Ch. 2) and will only repeat here that, in so far as they have abandoned cross-cousin marriage, the impetus has almost certainly been the influence of the surrounding caste society, which typically marries in north Indian fashion (i.e. no cousin mar-



119

120

How Kinship Systems Change

riage at all). It is, in short, an attempt to rise in the local hierarchy by imitating elite practice in what is a very status-conscious society, even in remote areas.29

III In accounting for change in kinship systems, these examples indicate that we can only think in terms of local ethnographic reasons, not global or universal ones, for the abandonment of cross-cousin marriage. However, because many such changes are accompanied by changes in terminology, we may find that the more general explanations lie in theories of classification, rather than of marriage per se. This conclusion is supported by the observation that categorical patterns are limited in number in a way that the details of actual marriage practices do not seem to be. To quote Needham, in the course of his examination of the Warao case, ‘very unlike social factors can produce like forms of classification’ (Needham 1974: 40). In other words, while potential changes in social morphology and attitudes and the reasons for these changes are many, the logical possibilities in which a kinship terminology can be constructed are few,30 meaning that exact correlations between these levels of analysis, though possible, are not inevitable (cf. Good 1981). Needham continues (ibid.), ‘The decisive factors, I suggest, have been, not particular empirical circumstances or legislative motives, but general possibilities and constraints of a purely formal nature.’ Methodologically, the restricted range of terminological possibilities makes it easier to control for variation, as well as to trace possible changes themselves, the future direction of which can to some extent be predicted. One possible approach is that of the lexical universalists, who might also be called lexical evolutionists. An early such work was Berlin and Kay’s famous and influential study of colour terms (1969), which set out a predictive model of change with respect to the order in which some colour terms appear in evolutionary time. This methodology was followed in other work by, for example, Cecil Brown on life-form terms (1984) and Stanley Witowski on kin terms (1971, 1972). The latter in particular applied Berlin and Kay’s insights to kinship terminologies regarding the predictability and order with which certain features disappear in circumstances of change: for example, prescriptive equations generally disappear before classif-

Why do Societies Abandon Cross-Cousin Marriage?

icatory ones, such that prescriptive equations imply the presence of classificatory ones as well but not necessarily vice versa. Yet kinship terminologies are different from the sorts of classification studied by Berlin and Kay, and by Brown. The latter grow in number of categories over time, each category shrinking in its semantic range as other, more specific categories emerge, as in the Linnaean classification of the natural world. This reflects the growth in knowledge about that world, or, for example, the range of colours recognized by a colour terminology.31 Kinship, conversely, is not subject to such growth in knowledge – at a basic level, relationships and alters (‘relatives’) have always been the same, though classified differently – so that change can only be effected in the form of how this finite knowledge is expressed by each emerging set of categories.32 Another possibly significant factor is the way in which categories and classifications can assume a very real concreteness in people’s minds, despite their variations in form over both time and space. As Ardener pointed out (1982: 12), ‘worlds set up by categories bear all the signs of materiality to the untutored33 human being’. Earlier in the same paper (ibid.: 6), he suggests: ‘Once the classification exists . . . it is part of the total experience of unreflecting individuals.’ Finally, as Needham remarks, ‘In a prescriptive system especially there is an absolute categorical determination which is hard to evade or change [and] which tends towards conservatism’ (1974: 41; also 1973). In other words, rules and behaviour are more labile than classifications: as already noted, rules can be broken, behaviour manipulated, but a classification is fixed, at least synchronically, and also diachronically within certain limits. When one adds to this the consideration, already noted, that because rules can be broken and behaviour manipulated complete uniformity between these three levels is not to be expected and is rare in practice, then there is almost bound to be a time-lag between changes in rules and practices and changes in terminology. But as Needham also says (1974: 41), ‘Yet prescriptive systems do change, and the problem is how precisely they do so. The crucial issue is the extent to which individuals make conscious alterations and adjustments; for the more deliberately they are supposed to act the more striking it is that their cumulative decisions should result in a common type of transformation.’ This reference to the possible impact of ‘cumulative decisions’ is itself striking in a paper by such a committed structuralist. However, it suggests that the cumulative impact of what are basically the same decisions being taken because individuals in a



121

122

How Kinship Systems Change

society are repeatedly faced with essentially the same circumstances may eventually make the lack of fit between terminology on the one hand and alliance rules and practice on the other intolerable. At this point, the terminology may begin to change. It is grasping these circumstances that is difficult: the necessary evidence for change is often circumstantial – the very fact of a mismatch between the respective patterns of terminology and alliance – while evidence for what may have caused it is even more often non-existent or irrecoverable. However, there are exceptions, as in Europe and much of Asia, where there are written records. While, for example, there is no evidence that historical or pre-historical Indo-European speech communities had cross-cousin marriage,34 it is a reasonable hypothesis, based on Fêng’s careful study of kin terms (1937), that China had bilateral cross-cousin marriage into the early historical period. Historical records helped Trautmann immensely in fixing the limits of Dravidian kinship in South Asia (1983). In the Americas, finally, we do have some, often rather patchy written sources on Native American kinship patterns – dictionaries and word lists, travellers’ and missionaries’ accounts etc. – going back in some cases to the sixteenth century, of which anthropologists have made quite extensive use.35 However, the importance of classification ultimately lies in how kinship terminologies are articulated with marriage choices. While there has been a tendency since Morgan to see change in the former as lagging behind change in the latter, we have also seen that where they are congruent the terminology will guide ‘wrong marriages’ (e.g. with a parallel cousin or younger aunt instead of a cross cousin) into the right classificatory channels. Similarly, the terminology can be used to rule out all cousin marriage by the simple device of classifying cousins as siblings, as in north Indian terminologies such as Hindi, Bengali or Gujarati. This can also have a knock-on effect on other parts of the terminology, in which the cross-parallel distinction may be modified or abandoned. Certainly, the terminology may well be reacting to change elsewhere in the system, of the sort discussed in Section II, above. Nonetheless, we are justified in asking just what is meant by ‘abandoning cross-cousin marriage’ and whether it might not take the form of how genealogical cross cousins are reclassified as kin prohibited in marriage, typically as siblings. Genealogically (i.e. analytically), therefore, cross cousins do not disappear, but in the classificatory sense they are taken into other categories.

Why do Societies Abandon Cross-Cousin Marriage?

It is also in circumstances of flux and uncertainty that genealogy might become more important, since more exact calculations might have to be made regarding the suitability of potential spouses, for example, excluding genealogical cross cousins but allowing classificatory ones. Both the terminology itself and genealogical reckoning by virtue of it have their own dynamics, not just the rules or practices of marriage; and this may even be reinforced by the limited number of patterns the terminology can assume, as well as by the propensity of any classification to appear concrete and ‘natural,’ when in fact it is subject to cultural variation.

IV To conclude, Scheffler’s disinclination to see in cross-cousin marriage a ‘system’ is based on the inevitable failure of any society to reach 100% observance of the marriage rule, and to that extent is understandable. However, although his approach is rooted in analysis of the terminology, it fails to recognize the extent to which classifications may be articulated in changes in how people marry, as well as in marriage practices at a particular point in time. In this respect, his approach is quite different from the position Needham eventually adopted (i.e. in Needham 1973), namely that the classification or terminology was where prescription should be identified, not the pattern or rule of marriages. Coupled with the synchronic bias in Scheffler’s approach, which did not permit effective consideration of change, it can readily be seen how his debates with the structuralists could become largely a matter of the two sides talking past each other, with little hope of reconciliation.

Notes * First published in Warren Shapiro (ed.), Focality and Extension in Kinship: Essays in Memory of Harold W. Scheffler, Acton: Australian University Press (2018). Reproduced with permission. 1. I may therefore be said to represent the ‘Needhamite’ tendency in this volume [i.e. Shapiro 2018], many of the other contributions to which are closer to the honorand’s intellectual concerns. While I acknowledge Prof. Scheffler’s recent sad loss, in the remainder of this chapter I will use the present tense to discuss his work and its impact. 2. E.g. Parkin (2013a). I realize the importance of distinguishing biology and genealogy in making this statement. The claim that genealogy does not mat-



123

124

How Kinship Systems Change

3.

4.

5.

6.

7.

8.

ter to many peoples in the world and that it simply reflects western thinking is ironically the one major feature shared by both structuralist approaches to kinship and the Schneiderian, cultural approaches that both rejected and replaced them. Note added 2020: For example, an English speaker may mention that someone is his cousin in ordinary conversation as a category but resort to genealogy in order to spell out the exact route of the relationship for his interlocutor. I have briefly returned to these issues more recently (Parkin 2009). This has long been a controversial term for what is recognized to be a form of marriage into a category of kin with a wide range of possible genealogical referents, and not only first cross cousins – assuming they can even be traced genealogically, or indeed will be, neither of which is necessarily the case. Nonetheless, I choose it here for reasons of standardization and also because some of the data and arguments I am using assume a genealogically defined referent of this type. I will occasionally refer to alternatives (especially Rodney Needham’s ‘prescriptive alliance’) where relevant. Needham mainly used other people’s work, not his own, to provide the evidence for his views. However, Robert Barnes, one of his students, did provide an example of a prescriptive system from his own ethnography on the east Indonesian domain of Kédang (Barnes 1974), a structuralist ethnography in every way, and he also contributed articles on prescriptive alliance based on his own ethnography (e.g. 1973, 1977), a path followed later by one of his own students, Penelope Graham (1987). Other students of Needham’s included Peter Rivière (1969) on the Trio of Guyana, David Hicks, who wrote many works on Timor (see especially Hicks 1978, 1990), and Gregory Forth (e.g. 1985, 1988, 1990). Later, Needham himself was to catch up with a book on the Sumban domain of Mamboru, based on fieldwork carried out some years before (1987). The fact of co-authorship means we can never really know who wrote what in this book, but that still should not detract from its use here in a chapter dedicated to Scheffler specifically. The book appears under Scheffler’s name as well, so one can assume that he agreed with its contents. Also, he and Lounsbury were obviously close intellectual allies in general, even though Scheffler (1971) did seek to re-establish the identity between Dravidian and Iroquois that Lounsbury (1964) had previously been at such pains to refute. See also Trautmann (1983: 85–88), who follows Lounsbury in respect of this disagreement, though otherwise he is critical of semantic analysis. Above all, the book on the Sirionó handily brings together key Schefflerian perspectives on the categorical. Predictability is an aspect of this process – that is, by following the rules, both the anthropologist and the indigenous ego can predict what term will be used to address or refer to a particular alter. The words ‘all about’ were originally Needham’s and are clearly being used ironically by Scheffler here. Unfortunately, I have not been able to find the original reference in Needham’s copious writings.

Why do Societies Abandon Cross-Cousin Marriage?

9. One might suggest that as with extensionism related to the study of kin-

10.

11.

12.

13.

14.

15.

16.

17.

ship terminologies the focus on filiation reflects a preference for stepwise thinking, as well as, here, a recognition of the shallowness of genealogical memory that is frequently encountered and an appreciation that descent groups rarely act together, whence ‘descent category’ may be more useful. Scheffler’s deconstructions of the notion of descent (e.g. 1966, 1985, 2001) are ultimately more cogent than, say, Kuper’s exaggerated dismissal (1982) of the whole idea of descent as an academic myth that has had to be deconstructed not just once but twice in anthropology’s intellectual history. Good examples are Needham’s analyses of the Purum (Needham 1958) and Lamet (Needham 1960). Needham did in fact recognize that societies united by prescriptive alliance could vary in other ways. For him, principles of organization – symmetry, asymmetry, transitivity – were more important than typologies of societies. See, for example, Needham 1971. Needham was following Lévi-Strauss here but also seeking to go beyond him by listing key oppositions that expressed the structure (usually asymmetric). Examples of such ‘total structural analysis’ include Needham 1958 on the Purum and 1960 on the Lamet. The original statement is Allen 1986 (republished with revisions 2004). More recent versions include Allen 1989 (written particularly with linguists in mind), 1998 and 2008. Allen reviews and answers critiques of his theory at 2008: 108–9. I return to his theory briefly below. (Note added 2020: I have retained the present tense in discussing Allen’s work, despite his death in 2020.) Prescriptive systems in particular, though not exclusively, often make it possible to trace ties with another relative down more than one pathway, enabling one to find a reason for marrying alter, as well as for not doing so. See Needham 1973. One example where this three-level analytical model is adopted in full for a particular ethnographic case is in Good’s description (1981) of Tamilnad in south India, where there is a logically very consistent and ‘pure’ terminology of symmetric prescriptive type and where 95% of the population marry someone in the prescribed category, though only 25% of spouses are first cross cousins, with a slight preference for FZDy/MBSe. North India has a number of examples, e.g. the Himalayan district of Kumaon (Krengel 1989) and the Malpahariya of Bihar (Parkin 1998; cf. Ch. 5). (Note added 2020: the Malpahariya data were taken from Sarkar 1937). In the article Basso was criticizing, Dole was suggesting yet another terminological ‘type’, namely bifurcate generational – that is, Hawaiian in ego’s level and bifurcate merging or bifurcate collateral in +1 and −1. It is not that notions of romance are absent elsewhere but they may well be placed in a different category than marriage, which is seen politically and socially as a more serious matter. The difference between the two domains may relate to stages in the life cycle, as with the Muria ghotul or youth dormitory in central India, where relationships between the genders formed in the youth dormitory are broken off by their parents when it comes to marriages, which should proceed in different directions (Elwin 1947).



125

126

How Kinship Systems Change

18. 19.

20.

21.

22.

23.

24.

Other examples may relate rather to the different contexts of changing social events. For example, among the Miao of south-west China, courting is placed in a different category than marriage, and even married men may take part in ritual events focused on courting (Chien 2013). See also the excellent discussion of Henley’s views in Walker 2009: 65–66. This itself is not unusual – that is, populations thinking about how they marry and asking themselves whether they could do it differently. See Layton 2008: 122. Theories of change in kinship sometimes forget this point and are often presented as if everything happens without those who are affected being aware of it. Nonetheless, similar arguments are occasionally made in relation to other parts of the world. Thus Barnes doubts that change took place from symmetric to asymmetric prescription on Alor and Pantar in eastern Indonesia, where both forms occur: ‘The symmetric systems of Alor and Pantar are found amongst trading populations situated along the coast and culturally distinct from the more anciently indigenous groups of the interior. They give the impression of being less permanently settled, and for a shorter time, than most Lamaholot communities. It would obviously demand a rather difficult historical argument to explain how they managed to retain the original form of social organization, while the agricultural communities to the west of them shifted to asymmetric alliance’ (1977: 153). The implication is that change, if change there was, proceeded in the opposite direction (this from a student of Needham’s otherwise generally under the latter’s influence at that time!). The theories of Ives etc. for North America may also find resonance in the distinction in India between the Dravidian south, with bilateral cross-cousin marriage and village endogamy, and the Indo-European north, with no cousin marriage and village exogamy, the latter area also featuring dispersal of alliances and thus a greater distance between spouses, both genealogically and geographically. One exception is Shapiro, who uses Dole’s theory to support an argument that the Sirionó may have changed from a lineal (Lowie) or cognatic (Needham) terminology to a prescriptive one (1968: 52). Basso clearly prefers an explanation in terms of synchronic analysis and rejects the evolutionary implications of Dole’s account. Both papers are summarized by Needham (1974: 32–35). See also Dole’s reply to Basso (Dole 1984). To find fully asymmetric systems in South Asia, one has to look at the Himalayas, stretching from the Indo-Burmese borderlands, with groups such as the Purum and Garo (the latter a mixed system, however; see Needham 1958, 1966), to as far west as the Kham Magar in Nepal (cf. Oppitz 1988). At the heart of the ritual was the tying of a tali or silver or gold token around the neck of the bride by a man of a different lineage (or even caste in some cases, e.g. Nambudiri Brahmans), as is done in ordinary marriages across south India as well. These rituals linked so-called enangar relations between different matrilineages, likened by Dumont to the sort of inherited affinal relationships that one also finds with the regular relations of cross-cousin marriage, despite the latter’s absence here. See Dumont 1983 [1961]: 117 ff.

Why do Societies Abandon Cross-Cousin Marriage?

25. Often conjugal visits were very brief, not even sleepovers being necessary.

26.

27.

28.

29.

30.

31. 32.

33.

34.

35.

At this extreme, matrilineal systems do seem to go along with a devaluation of marriage, let alone affinity: another example are the Moso of south-west China (Nongbri 2010). Moore prefers to see the taravad as a purely residential unit and downplays the matrilineal aspects accordingly, but she cannot deny them entirely. More conventional accounts of the Nayar system include Gough (several works, but especially 1959) and Fuller (1976). As Kapadia points out (in an earlier article, 1993: 28), ‘by consistently emphasising a “patrilateral” preference (with FZS as ideal spouse) [Tamilnad] Brahmins have made the cross-kin system hypergamous.’ Sister’s daughter’s marriage is a widespread practice accompanying bilateral cross-cousin marriage in many communities in south India. It can be construed as male ego taking his sister’s daughter as a wife for himself, not for his son, as with cross-cousin marriage. A useful account is Good 1980. One should therefore add this to, for example, changing bride price for dowry, and burial for cremation, as well as giving up alcohol, youth dormitories, mixed-sex dancing etc. E.g. for +1 male consanguines, only the following four patterns are attested: F = FB ≠ MB, F ≠ FB ≠ MB, F = FB = MB and F ≠ FB = MB. A fifth logical possibility, F = MB ≠ FB, is not. Sometimes traceable; for example, ‘orange’, ‘lilac’ and ‘purple’ have known origins as loans into English. One of the characteristics of prescriptive terminologies is that they are closed systems of classification, meaning that one can give any alter within them a term through a recursive process, however long the chain of genealogical symbols. With non-prescriptive systems, this recursive process does not apply, but one can still locate any alter by using the chains of symbols themselves, however long. This has always conditioned knowledge of kin ties and it always will: it is not to be compared to the biologist continually finding new species in, say, the Amazon forest or the Mariana trench to add to Linnaeus’s classification. In Ardener’s mind, this potentially objectionable word probably meant little more than that in ordinary social practice ordinary human beings take their classifications for granted and are unaware both of this fact and of possible alternatives. There are exceptions in South Asia, most prominently Sinhalese (see Trautmann 1981: 153–55), but these are most probably due to a population retaining its kinship system on changing its language, terms in the new language being invented or modified to suit; indeed, this is strongly indicated in the Sinhalese by the circumstance that most of its kinship vocabulary is Dravidian, even though the Sinhalese language itself is Indo-European. Cf. Bruner (1955) on a lexically English but structurally Crow-Omaha terminology in North America. This can be compared to Africa, Australia or Oceania, where contact has been much more recent on the whole and the time depth of such sources (if they exist at all) is far shallower. Kryukov (1998: 298–99; see also,



127

128

How Kinship Systems Change

Ch. 1) lists other techniques to which the analyst may have recourse in reconstructing the past.

References Allen, N.J. 1986. ‘Tetradic Theory: An Approach to Kinship’, Journal of the Anthropological Society of Oxford 13: 139–46. ———. 1989. ‘The Evolution of Kinship Terminologies’, Lingua 77: 173–85. ———. 1998. ‘The Pre-history of Dravidian-Type Terminologies’, in Maurice Godelier, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 314–31. ———. 2004. ‘Tetradic Theory: An Approach to Kinship’, in Robert Parkin and Linda Stone (eds), Kinship and Family: An Anthropological Reader. Oxford: Blackwell, pp. 221–35. ———. 2008. ‘Tetradic Theory and the Origin of Human Kinship Systems’, in Nicholas J. Allen, Hilary Callan, Robin Dunbar and Wendy James (eds), Early Human Kinship: From Sex to Social Reproduction. Oxford: Blackwell, pp. 96–112. Ardener, Edwin. 1982. ‘Social Anthropology, Language and Reality’, in David Parkin (ed.), Semantic Anthropology. London: Academic Press, pp. 1–14. Asch, Michael. 1998. ‘Kinship and Dravidianate Logic: Some Implications for Understanding Power, Politics, and Social Life in a Northern Dene Community’, in Maurice Godelier, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 140–49. Barnard, Alan. 2008. ‘The Co-evolution of Language and Kinship’, in Nicholas J. Allen, Hilary Callan, Robin Dunbar and Wendy James (eds), Early Human Kinship: From Sex to Social Reproduction. Oxford: Blackwell, pp. 232–44. Barnes, R.H. 1973. ‘Two Terminologies of Symmetric Prescriptive Alliance from Pantar and Alor in Eastern Indonesia’, Sociologus 23(1): 71–89. ———. 1974. Kédang: A Study of the Collective Thought of an Eastern Indonesian People. Oxford: Clarendon Press. ———. 1977. ‘Alliance and Categories in Wailolong, East Flores’, Sociologus 27(2): 133–57. Basso, Ellen B. 1970. ‘Xingu Carib Kinship Terminology and Marriage: Another View’, Southwestern Journal of Anthropology 26: 402–16. Berlin, Brent, and Paul Kay. 1969. Basic Color Terms: Their Universality and Evolution. Berkeley: University of California Press. Brown, Cecil. 1984. Language and Living Things: Uniformities in Folk Classification and Naming. New Brunswick: Rutgers University Press. Bruner, Edward M. 1955. ‘Two Processes of Change in Mandan-Hidatsa Kinship Terminology’, American Anthropologist 57(4): 840–50. Chagnon, Napoleon. 1968. Yanomamo: The Fierce People. New York: Holt, Rinehart and Winston. Chien, Mei-Ling. 2013. ‘Tensions between Romantic Love and Marriage: Performing “Miao Cultural Individuality” in an Upland Miao Love-song’, in James Wilkerson and Robert Parkin (eds), Modalities of Change: The Interface of Tradition and Modernity in East Asia. New York and Oxford: Berghahn Books, pp. 93–116.

Why do Societies Abandon Cross-Cousin Marriage?

Clark-Decès, Isabelle. 2014. The Right Spouse: Preferential Marriages in Tamil Nadu. Stanford: Stanford University Press. Cook, E.A. 1969. ‘Marriage among the Manga’, in Robert M. Glasse and Mervyn J. Meggitt (eds), Pigs, Pearlshells, and Women: Marriage in the New Guinea Highlands. Englewood Cliffs, NJ: Prentice-Hall, pp. 96–116. Désveaux, Emmanuel, and Marion Selz. 1998. ‘Dravidian Nomenclature as an Expression of Ego-Centered Dualism’, in Maurice Godelier, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 150–67. Dole, Gertrude. 1969. ‘Generation Kinship Nomenclature as an Adaptation to Endogamy’, Southwestern Journal of Anthropology 25: 105–23. ———. 1984. ‘The Structure of Kuikuru Marriage’, in Kenneth M. Kensinger (ed.), Marriage Practices in Lowland South America. Urbana and Chicago: University of Illinois Press, pp. 45–62. Dumont, Louis. 1953. ‘The Dravidian Kinship Terminology as an Expression of Marriage’, Man 53: 34–39. ——. 1983 [1961]. ‘Nayar Marriages as Indian Facts (with comment)’, in Louis Dumont, Affinity as a Value: Marriage Alliance in South India, with Comparative Essays on Australia. Chicago and London: The University of Chicago Press, pp. 105–44. Elmendorf, William W. 1961. ‘System Change in Salish Kinship Terminologies’, Southwestern Journal of Anthropology 17: 365–82. Elwin, Verrier. 1947. The Muria and their Ghotul. Bombay: Oxford University Press. Fêng, Han-yi. 1937. ‘The Chinese Kinship System’, Harvard Journal of Asiatic Studies 2: 141 ff. Forth, Gregory. 1985. ‘Layia (FZS, ZH, m.s.): The Evolutionary Implications of some Sumbanese Kin Terms’, Sociologus 35(2): 120–41. ———. 1988. ‘Prescription Gained or Retained? Analytical Observations on the Relationship Terminology of Ndao, Eastern Indonesia’, Sociologus 38(2): 166–83. ———. 1990. ‘From Symmetry to Asymmetry: An Evolutionary Interpretation of Eastern Sumbanese Relationship Terminology’, Anthropos 85(4–6): 373–92. Fuller, C.J. 1976. The Nayars Today. Cambridge: Cambridge University Press. Fuller, C.J., and Haripriya Narasimhan. 2008. ‘Companionate Marriage in India: The Changing Marriage System in a Middle-Class Brahman Subcaste’, Journal of the Royal Anthropological Institute (n.s.) 14: 736–54. Good, Anthony. 1980. ‘Elder Sister’s Daughter Marriage in South Asia’, Journal of Anthropological Research 36: 474–500. ———. 1981. ‘Prescription, Preference and Practice: Marriage Patterns among the Kondaiyankottai Maravar of South India’, Man 16: 108–29. Gough, Kathleen. 1959. ‘The Nayars and the Definition of Marriage’, Journal of the Royal Anthropological Institute 89: 23–34. Graham, Penelope. 1987. ‘East Flores Revisited: A Note on Asymmetric Alliance in Lebola and Wailolong, Indonesia’, Sociologus 37(1): 40–59. Hart, Charles W.M., and Arnold R. Pilling. 1979. The Tiwi of North Australia. New York: Holt, Rinehart and Winston. Henley, Paul. 1996. South Indian Models in the Amazonian Lowlands. Manchester Papers in Social Anthropology, 1.



129

130

How Kinship Systems Change

Hicks, David. 1978. Structural Analyses in Anthropology: Case Studies from Indonesia and Brazil. St Augustin bei Bonn: Anthropos-Institut. ———. 1985. ‘A Transitional Two-Section System among the Mbae-Speakers of Manggarai, Eastern Indonesia’, Sociologus 35(1): 74–83. ———. 1990. Kinship and Religion in Eastern Indonesia. Gothenburg: Acta Universitatis Gothoburgensis. Hickerson, H. 1962. The Southwestern Chippewa: An Ethnohistorical Study. Washington: American Anthropological Association, Memoir 92. Ives, John W. 1998. ‘Developmental Processes in the Pre-contact History of Athapaskan, Algonquian, and Numic Kin Systems’, in Maurice Godelier, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 94–139. Kapadia, Karen. 1993. ‘Marrying Money: Changing Preference and Practice in Tamil Marriage’, Contributions to Indian Sociology (n.s.) 27(1): 25–51. ———. 1995. Shiva and Her Sisters: Gender, Caste, and Class in Rural South India. Boulder, CO: Westview Press. Krengel, Monika. 1989. Sozialstruktur in Kumaon: Bergbauen im Himalaya. Wiesbaden: Franz Steiner Verlag. Kryukov, Mikhail V. 1998. ‘The Synchro-Diachronic Method and the Multidimensionality of Kinship Transformations’, in Maurice Godelier, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 294–313. Kunstadter, Peter. 1966. ‘Residential and Social Organization of the Lawa of Northern Thailand’, Southwestern Journal of Anthropology 22(1): 61–84. Kuper, Adam. 1982. ‘Lineage Theory: A Brief Retrospect’, Annual Review of Anthropology 11: 71–95. Layton, Robert. 2008. ‘What Can Ethnography Tell Us About Human Social Evolution?’ in Nicholas J. Allen, Hilary Callan, Robin Dunbar and Wendy James (eds), Early Human Kinship: From Sex to Social Reproduction. Oxford: Blackwell, pp. 113–27. Leach, Edmund. 1960. ‘The Sinhalese of the Dry Zone of Northern Ceylon’, in George P. Murdock (ed.), Social Structure in Southeast Asia. Chicago: Quadrangle, pp. 116–26. Lévi-Strauss, Claude. 1949. Les structures élémentaires de la parenté. Paris: Presses Universitaires Françaises. Lindell, Kristina, Rolf Samuelsson and Damrong Tayanin. 1979. ‘Kinship and Marriage in Northern Kammu Villages: The Kinship Model’, Sociologus 29(1): 60–84. Lounsbury, Floyd G. 1964. ‘The Structural Analysis of Kinship Semantics’, in Horace G. Lunt (ed.), Proceedings of the Ninth International Congress of Linguists. The Hague: Mouton, pp. 1073–93. Moore, Melinda A. 1985. ‘A New Look at the Nayar Taravad’, Man (n.s.) 20: 523–41. Nakane, Chie. 1963. ‘The Nayar Family in a Disintegrating Matrilineal System’, International Journal of Comparative Sociology 3: 18-45. Needham, Rodney. 1958. ‘A Structural Analysis of Purum Society’, American Anthropologist 60: 75–101. ———. 1960. ‘Alliance and Classification among the Lamet’, Sociologus 10: 97–118.

Why do Societies Abandon Cross-Cousin Marriage?

———. 1966–67. ‘Terminology and Alliance, I and II’, Sociologus 16: 141–57, and 17: 39–53. ———. 1971. ‘Remarks on the Analysis of Kinship and Marriage’, in Rodney Needham (ed.), Rethinking Kinship and Marriage. London: Tavistock, pp. 1–34. ———. 1973. ‘Prescription’, Oceania 43: 166–81. ———. 1974. ‘The Evolution of Social Classification: A Commentary on the Warao Case’, Bijdragen tot de Land-, Taal, en Volkenkunde 130: 16–43. ———. 1986. ‘Alliance’, Oceania 56(3): 165–80. ———. 1987. Mamboru: History and Structure in a Domain of Northwestern Sumba. Oxford: Clarendon Press. Nongbri, Tiplut. 2010. ‘Family, Gender and Identity: A Comparative Analysis of Trans-Himalayan Matrilineal Structures’, Contributions to Indian Sociology 44(1– 2): 155–78. Oppitz, Michael. 1988. Frau für Fron: die Dreierallianz bei den Magar West-Nepals. Frankfurt am Main: Suhrkamp. Parkin, Robert. 1992. The Munda of Central India: An Account of their Social Organization. Delhi: Oxford University Press. ———. 1996. ‘Genealogy and Category: An Operational View’, L’Homme 139: 85–106. ———. 1998. ‘Dravidian and Iroquois in South Asia’, in Maurice Godelier, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 252–70. ———. 2009. ‘What Shapiro and McKinnon Are All About, and Why Kinship Still Needs Anthropologists’, Social Anthropology 17: 158–70. ———. 2013a. ‘Relatedness as Transcendence: On the Renewed Debate Over the Meaning of Kinship’, Journal of the Anthropological Society of Oxford Online 5(1): 1–26. ———. 2013b. ‘From Tetradic Society to Dispersed Alliance: Notes Arising from a Chapter by N.J. Allen’, Journal of the Anthropological Society of Oxford Online 5(2): 194–206. Rivière, Peter. 1969. Marriage among the Trio: A Principle of Social Organization. Oxford: Clarendon Press. Sarkar, S. 1937. ‘The Social Institutions of the Malpahariyas’, Journal of the Asiatic Society of Bengal, Letters 3: 25–32. Scheffler, Harold W. 1963. Choiseul Island Social Structure. Berkeley: University of California Press. ———. 1966. ‘Ancestor Worship in Anthropology: or, Observations on Descent and Descent Groups’, Current Anthropology 7: 541–48. ———. 1971. ‘Dravidian-Iroquois: The Melanesian Evidence’, in Lester R. Hiatt and Chandra Jayawardena (eds), Anthropology in Oceania: Essays Presented to Ian Hogbin. Sydney: Angus and Robertson, pp. 231–54. ———. 1972. ‘Kinship Semantics’, Annual Review of Anthropology 1: 309–28. ———. 1977. ‘Review Article: Kinship and Alliance in South India and Australia’, American Anthropologist 79: 869–82. ———. 1984. ‘Markedness and Extensions: The Tamil Case’, Man 19: 557–74. ———. 1985. ‘Filiation and Affiliation’, Man 20: 1–21. ———. 2001. Filiation and Affiliation. Boulder CO and Oxford UK: Westview Press.



131

132

How Kinship Systems Change

Scheffler, Harold W., and Floyd G. Lounsbury. 1971. A Study of Structural Semantics: The Siriono Kinship System. Englewood Cliffs, NJ: Prentice-Hall. Service, Elman R. 1971. Cultural Evolutionism: Theory in Practice. New York: Rinehart and Winston. Shapiro, Warren. 1968. ‘Kinship and Marriage in Siriono Society: A Re-examination’, Bijdragen tot de Taal-, Land- en Volkenkunde 124(1): 40–55. ———. 2018. Focality and Extension in Kinship: Essays in Memory of Harold W. Scheffler. Acton: Australian National University Press. Smith, James G.E. 1974. ‘Proscription of Cross-Cousin Marriage among the Southwestern Ojibwa’, American Ethnologist 1(4): 751–62. Taylor, Anne-Christine. 1998. ‘Jivaro Kinship: “Simple” and “Complex” Formulas: A Dravidian Transformation Group’, in Maurice Godelier, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 187–213. Trautmann, Thomas R. 1981. Dravidian Kinship. Cambridge: Cambridge University Press. Walker, Harry. 2009. ‘Transformations of Urarina Kinship’, Journal of the Anthropological Society of Oxford Online 1(1): 52–69. Witowski, Stanley. 1971. ‘A Universalist Account of Kinship Semantics’, Ph.D. thesis. University of Iowa. ———. 1972. ‘Guttman Scaling of Semantic Distinctions’, in Priscilla Reining (ed.), Kinship Studies in the Morgan Centennial Year. Washington DC: Anthropological Society of Washington, pp. 167–88.

5 

Dravidian and Iroquois in South Asia

Introduction This chapter discusses terminological data from the South Asian region from a historical point of view.* It continues a series of articles whose overall intention has been to argue that a recourse to history might obviate the difficulties encountered in using a purely synchronic perspective to compare the very dissimilar terminological systems of south and north India (see Parkin 1990, 1992b, 1992c). The present concern is not to go into these difficulties again (q.v. Parkin 1990; see also Ch. 2) but to improve the model of change presented earlier with the aid of some interesting new evidence, some of which shows that at least one typologically ‘Iroquois’ system exists in this region. Such systems were originally assumed to be identical to typologically Dravidian ones, a conflation which even entered one of the textbook typologies, that of Murdock (1949). But in 1964, Lounsbury showed that although Iroquois terminologies do resemble Dravidian ones in their treatment of the nearest collaterals in the medial three levels, they differ from them as regards further collaterals, whom they treat on an absolute-sex basis rather than on the relative-sex basis that the cross-parallel distinction depends on (see Trautmann and Barnes 1998 for a more detailed account). The significance of the discovery of an Iroquois system in South Asia for arguments relating to terminological change in the region will be demonstrated here by comparing the Burushaski terminology with the Dravidian systems

134

How Kinship Systems Change

of the subcontinent, and also with north Indian, to which it also bears some similarities. A further link in the chain of transformations is provided by yet other data from the Himalayas. Ethnographers have always associated the South Asian region in part with Dravidian systems but seldom with Iroquois ones. This may not reflect local realities so much as the bias of local traditions of ethnographic reporting. The relatively late identification, in the 1960s, of the Iroquois system as separate from Dravidian and its predominant association with North America in both the ethnographic and the theoretical literature may have discouraged the search for it in the subcontinent. Very recently, however, I have realized that one example does exist on the periphery of the region. This evidence may also contribute something to our understanding of the ways in which the non-prescriptive north Indian terminology differs from prescriptive south Indian or ‘Dravidian’, as well as their possible historical links.

Burusho Kinship Terminology Several years ago, I suggested that the kinship terminology of the Burusho of the Karakorum region of northern Pakistan was basically two-line symmetric prescriptive (Parkin 1987). I am grateful to Dr Peter Parkes for recently directing my attention to a text (Ali 1983) I was back then ignorant of, one that contains data confirming my diagnosis so far as it went but that also permits a more complete picture of the present-day nature and possible history of the terminology to emerge. Ali is more interested in the descent and status systems of the Burusho than in marriage, and his treatment of the kinship terminology is a formal one, after the style ultimately of Harold Scheffler and more immediately of Anthony Carter, his doctoral supervisor. Nonetheless, his list of kinship terms (1983: 45) can still be seen to be modified two-line prescriptive by those sympathetic to that particular approach, at least as regards the core kin types. He worked in the Hunza valley, so that the terminology he records is in that dialect rather than either the closely related Nagir one, spoken in an adjacent valley, or Yasini, a remoter dialect spoken in the Yasin valley further east. It thus corresponds most closely to the dialect and terminology I marked ‘Bu’ in the earlier article, which in fact stood for the dialects of Hunza and Nagir taken together.

Dravidian and Iroquois in South Asia

Before examining the terminology, however, it is important to look at the Burusho marriage system. Ali’s data here are sketchy, but they are supplemented to some extent by another, slighter work that has also only come to my attention recently, by Hamid (1979: 82). In this strongly Islamic environment, an ever-present possibility is the breakdown of clan exogamy in favour of patrilateral parallel cousin marriage. For the most part, the Burusho appear to have resisted this, despite their long having been Shiite or Ismailite (cf. Parkin 1987: 160). According to Ali (1983: 73, 74 n. 27), clan exogamy is the norm, only royal and vizier clans indulging in FBD-FBS marriage to any extent, and Hamid (ibid.) confirms him as far as the rule of exogamy is concerned. The latter also tells us that divorce and the remarriage of divorcees and widows are allowed. As to whether there is a positive rule of marriage, there are still no clear statements, though there are some interesting indications of a system not dissimilar to certain affinal alliance systems in central India. Ali makes a bare mention of cross-cousin marriage (1983: 44), but it is not clear whether this would include first cross cousins, especially since earlier sources appear to rule this out (see Parkin ibid.). This interpretation would appear to be confirmed by Hamid (ibid.), who says that close kin are avoided in marriage. A further statement, one of particular significance in the present context, is that affines are separated from matrikin in relation to normative behaviour as well as terminologically. The former are considered to be remoter than the latter and engage in less intimate and less long-term links with ego and his or her group (Ali ibid.: 291).1 The kin terms and their specifications are shown below in the spellings given by Ali. Although slightly different from those given in other sources (see Parkin 1987: 162), these spellings are easily recognizable. As for the core specifications – that is, those nearest to ego, the data given above accord remarkably well with other sources on the whole, being fuller but not really contradictory. There are immediate differences to note in ego’s level (cf. Parkin 1987: 164). Ali’s data continue to leave Adam Nayyar isolated as the only authority to give any relative-age terms. The latter’s reporting of acho as EGE also conflicts with Ali, who places such specifications under riik. Acho no longer appears to have the affinal specifications recorded by earlier authorities, which I had earlier viewed as being problematic because they were redundant to riik. There are two other observations of a more general kind to be made that have a bearing on what I published



135

136

How Kinship Systems Change

Table 5.1. Burushaski kinship terminology pi

PP, PPP etc.

u

F, FB, MZH, MH, FFBS, FMZS, MMZDH, FFFBSS; also FMBS, FFZS

mi

M, MZ, FBW, FW, MMZD, MFBD, FFBSW, MFFBSD; also MFZD, MMBD

ngo

MB, FZH, MFBS, MMZS, FFBDH, FMZDH; also MFZS, MMBS, FMBDH, FFZDH, HMB, WMB, HFZH, WFZH

ntso

FZ, MBW, FFBD, FMZD, MFBSW, MMZSW; also FFZD, FMBD, MFZSW, MMBSW, HFZ, WFZ, HMBW, WMBW

skir

EF, EFB, EMZH

skus

EM, EMZ, EFBW

tcho s

sG, PCss, FFBSCss; also PGCss

lus

Bws, PSws, ZHws, HBws, FFBSSws; also PGSws

yas

Zms, PDms, BWms, WZms, FFBSDms; also PGDms

yar

H

yus

W

riik

WB, ZHms, HZ, BWws; also EGE

sildir

CEF, CEFB, CEMZH

silgus

CEM, CEMZ, CEFBW

i

S, ssGS, WS, HS, WZS, HBS, FBSSms, MZSSms, MMZDSSms, MZDSws, FBDSws, MMZDDSws; also FZSSms, MBSSms, FZDSws, MBDSws

ei

D, ssGD, WD, HD, WZD, HBD, FBSDms, MZSDms, MMZDSDms, MZDDws, FBDDws, MMZDDDws; also FZSDms, MBSDms, FZDDws, MBDDws

saghun

osGC, WBC, HZC, FBDCms, MZDCms, FBCws (sic FBSCws?); also osGCE, FZDCms, MBDCms, FZSCws, MBSCws

rar

DH, BDHms, ZDHws, WZDH, HBDH, FBSDHms, FBDDHws

khakin

SW, BSWms, ZSWws, WZSW, HBSW, FBSSWms, FBDSWws

mis

CC, CCC etc.

Dravidian and Iroquois in South Asia

previously. First, all cousins, not merely parallel ones, and apparently of whatever degree (certainly first cousins), are equated with ego’s siblings. Secondly, there is the regularity with which terms for primary affines are separated from those for cognates and the affines of cognates. The separation of affinal terms was already realized earlier, but the generational treatment of cousins in Burushaski had to be left a supposition on the evidence then available (Parkin 1987: 166–67). Ali’s data places the existence of both these important features in the Burushaski terminology beyond all doubt. Neither circumstance should surprise us any longer in South Asia. While the Dravidian systems of south India are famous for being reasonably pure examples of two-line symmetric prescriptive, at least in the three medial levels, elsewhere in India one frequently encounters apparently transitional forms that have a clear prescriptive basis but lean in a non-prescriptive direction. They are typically found in areas of contact with the non-prescriptive north Indian system that dominates the central area of South Asia. As I have often pointed out before (e.g. Parkin 1990, replicated as Ch. 2; also 1992a, Chs. 7, 8), certain tribal terminologies of central India exemplify these tendencies perfectly, as do some terminologies in the Himalayan region. One can almost certainly talk of an evolutionary trend in these cases. The Burushaski terminology, of the Karakorum further west, can now be added to their number on the basis of these two features, though it possibly differs in respect of remoter kin types in the levels adjacent to ego’s, as we shall see later. These similarities are particularly evident among the Mundaspeaking groups of southern Bihar2 and parts of Orissa. At least in the north of the Munda area, regular systems of affinal alliance exist that are conceived indigenously not as systems of cross-cousin marriage but as ones involving exchange among sibling pairs defined as being eGEyG to one another (i.e. eZHyZ and eBWyZ for male ego, eZHyB and eBWyB for female ego). This goes along with a system that does not allow the immediate renewal of alliances between the same two alliance groups but does allow renewal subsequently, after the lapse of at least one and, more typically, three generations. It also corresponds to the separation of affines from cross kin: if immediate renewal, through cross-cousin marriage, is not allowed, WF cannot also be MB (i.e. he cannot be the father of a cross cousin), nor can WB also be a cross cousin. It is the fact that the inter-cognate equations



137

138

How Kinship Systems Change

of formerly prescriptive terminologies nonetheless remain in the +1 and −1 levels of the terminology that suggests typological change and thus that the terminology is itself transitional when viewed from the viewpoint of global evolutionary trends. The generational pattern of ego’s level, on the other hand, as well as the rule of delay before an alliance can be renewed, draws these tribal systems nearer to north Indian systems. These, though non-prescriptive terminologically and normally with little sense of system as regards affinal alliance, sometimes have either an expectation of or a statistical tendency towards the repetition of alliances between the same groups within the same generation, though there is never any hint of a rule of renewal in the longer term. The Jat are an example of the former (Tiemann 1970), the people of Kangra in north-west India an example of the latter (Parry 1979). The significance of GEG (siblings’ spouses’ siblings) as categories in a system of affinal alliance that exhibits some of the regularity of prescriptive systems should be explored a little further. A major question of interest is how far this can be seen as indicating a prescriptive system. Traditionally, writers on these groups have tended to be so concerned to establish the presence or absence of ‘crosscousin marriage’ that they have not considered the possible existence of other formulations. However, there is evidence in some cases of a specific category (Munda, Korwa goi, Santal sangat, Juang saliray) isolating GEG specifications within the terminology, together with some indications, at least in the latter two cases, that the terms involved stipulate a marriage preference (McDougal 1963; Bouez 1985). Cross cousins are explicitly ruled out as marriage partners, but is this the insuperable problem it might seem? Certainly even the most radical structuralist approaches to such systems tend to assume the existence of cross-cousin marriage in practice, even if modified through the notion of classificatory relationships, and they would also insist on regular cognate-affine equations in the terminology. Some, indeed, would locate ‘prescription’ in the terminology itself rather than see it as a property of the rules, despite the fact that this is what the word logically suggests (most obviously Needham 1973; this is my own practice too). However, another aspect of this approach is to emphasize category over genealogy and to stress sensitivity to the indigenous representation over the requirements of analysis. One inference from this is that the Juang, for example, do not marry siblings’ spouses’ siblings, much less cross cousins, but saliray

Dravidian and Iroquois in South Asia

(their term for the stipulated category). In other words, we are possibly only restricting ourselves in arriving at their own view of their marriage system when we translate it into apparently more readily understandable terms through the use of genealogical or other analytical denotata. The real problem in such cases is to translate terms like saliray adequately into analytical language. The solution is not difficult in all cases, but in some it certainly is. GEG categories are appropriate to the model of any situation in which groups of siblings intermarry. This is so where the marriages are taking place as a one-off event, not to be repeated in later generations. But they can also be applied to the marriage of cross cousins, even where they are expected to be repeated, generation after generation; what is different about the former is the absence, perhaps actually the non-recognition, of vertical ties linking the referents genealogically. Given all these considerations, it is not immediately clear that the northern Munda systems should be considered non-prescriptive. Personally, I am content to remain agnostic for the time being, on the basis that adequate descriptions of kinship systems are at least as important as fitting them into analytical typologies. In any case, the potential interest of these groups for questions of terminological transformation in South Asia in no way depends on this question being decided.

Iroquois in South Asia As already indicated, the evidence of the similarity of the Burushaski terminology to those of some Munda groups does not exhaust the interest of Ali’s data. While the first group of specifications under each term in his list fits the two-line symmetric or ‘Dravidian’ scheme (i.e. those for cognates), those that appear after the word ‘also’ (where this is given) do not. Apart from the fact that the primary affines have separate terms from cross kin, the core – that is, the genealogically closest specifications – adhere to the formal two-line prescriptive model, though the extensions on the whole do not. Ali himself calls the pattern ‘Iroquois’ (1983: 47), his simultaneous reference to Scheffler indicating that he does not intend by this term a Dravidian or two-line prescriptive system. A closer examination reveals that the pattern as regards these extensions actually corresponds to the ‘Cheyenne’ variant of ‘Iroquois’



139

140

How Kinship Systems Change

rather than what might in this context be called ‘pure Iroquois’. In this variant, while the +1 and −1 levels have Iroquois-type crossness, ego’s level is generational (see Trautmann 1992). However, Burushaski also has the equations MB = FZH and FZ = MBW, a feature that despite the lack of affinal terms betrays a prescriptive origin for the terminology regardless of all its other changes. It appears from Trautmann and Barnes (1998) that such equations do occur in Iroquois and its variants as well as in Dravidian and that they are compatible with all of them. This is welcome information, given the fact that traditional North Americanist discourse on the subject tends to ignore the treatment of PGE specifications, despite their importance in allowing us to make some crucial distinctions between terminological types (e.g. those between prescriptive and non-prescriptive, or between symmetric and asymmetric prescription).3 The discovery that the Burushaski terminology proves to be an Iroquois variant was something of a surprise as regards South Asia. Is it the only example? It is significant in this context that Trautmann and Barnes (1998) give a second test of the difference between Dravidian and Iroquois, namely whether cognate-affine equations occur in quantity or not. In their view, they have to be present for a terminology to be Dravidian; if they are not, Iroquois becomes a possibility. They do not say, rightly, that their absence as such defines a system as Iroquois: many other sorts of terminology that have nothing else in common regularly separate affines from cognates (CrowOmaha, Hawaiian, cognatic). But how far are they really a necessary feature of a Dravidian terminology? They are always present in the ideal type, but is this always true ethnographically? One can certainly conceive of a terminology that is Dravidian and not Iroquois in its cognatic terminology but that regularly has separate affinal terms (e.g. MB=FZH≠EF). In fact, such terminologies are to be encountered quite frequently in South Asia, appearing inter alia in Munda (Parkin 1992a: Ch. 7; see also Ch. 2), in Dumont’s own field data from south India (1957), and in a terminology recorded by Stirrat from Sri Lanka (1977; discussed in Parkin 1992b). Whether more of them, apart from Burushaski, actually prove to be Iroquois rather than Dravidian would have to be tested specifically as regards their treatment of the remoter cousins. This seems a surer criterion of the difference than the treatment of affines alone. The north and central Munda terminologies would, on the face of it, be a good place to start looking for further Iroquois features in

Dravidian and Iroquois in South Asia

South Asia. The generational nature of ego’s level in most Munda terminologies, in which all first cousins and seemingly remoter ones too share terms with siblings, would rule out strict Iroquois, though it would allow the Cheyenne variant. However, although a large number of terminologies have been recorded for these two branches of the language family, the sources available rarely give sufficient data on the sorts of extensions that would be relevant in deciding whether a particular terminology were or were not Iroquois. There are some very piecemeal exceptions.4 First, there is Santal kaka, FyB, MyZH, FPGyS, step-F. Second, Juang atir is FeB, MeZH, FMBS, though the feminine equivalent atirae, MeZ, FeBW, FMBD, FZHeZ is consistent only with the Dravidian pattern. Third are Ho kaka, FyB, MyZH, FFGS, and hatom, FZ, MBW, EM, FFGD, plus closely related Bhumij kaka, FyB, MyZH, FFBSy, FFZS, and hatom, FZ, MBW, FFZD. Finally, Bhumij putara, eBS, FBSS (assuming ms), and bhagina, FBDS (assuming ms) – perhaps connected with bhanja, osGS – are consistent with both Iroquois and Dravidian patterns. Apart from the single example above, McDougal’s data on the Juang (1963, 1964), which do give some cognatic extensions, clearly place them in a Dravidian, not in any sort of Iroquois pattern.5 As for possible Iroquois features in South Asian terminologies that are lexically Dravidian, again the poor recording of wider collateral specifications is an impediment to arriving at a decisive view. Only two of the terminologies listed by Trautmann have data on extensions: Marathi non-Brahman (1981: 117, after Carter) and Coorg (ibid.: 162, after Emeneau). Data on the first are defective and therefore inconclusive as regards +1 and −1, but those concerning ego’s level are just sufficient to indicate a Dravidian, not an Iroquois pattern. In the Coorg case, useful data are restricted to the +1 and −1 levels, but they also indicate a Dravidian, not an Iroquois pattern as regards the relevant categories. This is also true of the Tamil terminology recorded in Konku by Beck (1972: Appendix F). The paucity of data with respect to Dravidian is, of course, to be laid not at Trautmann’s door but at those of his sources, but it does show how a specific search for Iroquois patterns in South Asia has rarely been considered either necessary or interesting. This reflects the fact that the cousin terminology will normally be expected to be generational here when it is not obviously Dravidian. The Burusho terminology conforms to this rule in being generational in ego’s level, but its adjacent levels have proved to be an exception to the norms



141

142

How Kinship Systems Change

of the region. This has only come to light through the ethnographer’s (Ali’s) decision to record the terms for wider collateral kin types in these levels, presumably following the example of his supervisor Carter. Given Ali’s relative lack of interest in the terminology, we can only regard this as fortuitous.

North Indian Kinship Terminology There is one other terminological pattern of special interest in South Asia, namely north Indian. I want to consider this system in the light of the foregoing, using data supplied by Vatuk (1969) and Parry (1979), the former being one of the standard texts on the Hindi terminology, the latter providing good comparative evidence, arranged explicitly in the same manner, from Kangra, north-west India. Table 5.2. Hindi kinship terminology as example of north Indian pattern baba

FF

FFB, FMB, FFZH, FMZH, FZHF, FZHFB, FZHMB, FZHFZH, MZHF, MZHFB, MZHMB, MZHFZH, MZHMZH

dadi

FM

Baba’s wife

nana

MF

MFB, MMB, MFZH, MMZH, MBWF, MBWFB, MBWMB, MBWFZH, MBWMZH, FBWF, FBWB, FBWMB, FBWFZH, FBWMZH

nani

MM

nana’s wife

dadasara

EFF

spouse’s baba

dadas

EFM

spouse’s dadi

nanasara

EMF

spouse’s nana

nanas

EMM

spouse’s nani

tau/kaka

FeB/FyB

FPGS (e/y), also FZHBWB (e/y)

tai/kaki

FeBW/FyBW

tau’s wife/kaka’s wife

mama

MB

MPGS, also MBWB, MBWBWB, FBWB, FBWBWB, FBWBWZH, MZHBWB, MBWZH, MBWBWZH

mami

MBW

mama’s wife



143

Dravidian and Iroquois in South Asia

bua

FZ

FPGD, also FZHZ, FZHBW, MZHZ

phupha

FZH

bua’s husband

mausi

MZ

MPGD, also FBWZ, MZHBW, HBWM, ZHBWM, WZHM, BWZHM

mausa

MZH

mausi’s husband

mavsa

BWF, ZHF

father of bhabhi or jija

mavsi

BWM, ZHM

mother of bhabhi or jija

ma, amma, mata ( ji)

M

bap, pita ( ji)

F

sasur

EF

sas

EM

pitasara

EFyB

spouse’s kaka

pitas

EFyBW

spouse’s kaki

tayasara

EFeB

spouse’s tau

tayas

EFeBW

spouse’s tai

maulasara

EMB

spouse’s mama

maulas

EMBW

spouse’s mami

phuphasara

EFZH

spouse’s phupha

phuphas

EMZH

spouse’s bua

mausasara

EMZH

spouse’s mausa

mausas

EMZ

spouse’s mausi

bhai

B

PGS, also HBWB, ZHBWB, WZHB, BWZH; BWBWZH (ws)

bhabhi/ bhababu

BW/yBW

W and WZ of bhai; also BWBW (ws), BWBWZ (ws)

bahen

Z

PGD; also HBWZ, ZHBW, WZHZ, BWZHZ; ZHZHBW (ms)

jija, bahenoi

ZH

H and HB of bahen; also ZHZH (ms), ZHZHB (ms), ZHZHZH (ms) (continued)

144

How Kinship Systems Change

sala

WB

WPGS; also WBWB, WBWZH, BWB (ms), BWBWB (ms), BWBWZH (ms)

salhaj

WBW

sala’s wife

sali (WyZ in WZ Parry 1979, which gives jethal for WeZ)

WPGD

sarhu

WZH

sali’s husband

jeth/devar

HeB/HyB

HPGS (e/y)

jethani/daurani HeBW/HyBW wife of jeth/devar nanad

HZ

HPGD; also HZHZ, HZHBW; ZHZ (ws), ZHZHZ (ws), ZHZHBW (ws)

nandoi

HZH

nanad’s husband

samdhi

CEF

F of jamai or bahu

samdhin

CEM

M of jamai or bahu

beta

S

contextually also GS

beti

D

contextually also GD

bhatija

BS

S of bhai, sala, jeth, devar; BWB (ws), BWBWB (ws)

bhatiji

BD

D of bhai, sala, jeth, devar

bhanja

ZS

S of bahen, Sali, nanad; ZHZ (ms), ZHZHZ (ms)

bhanji

ZD

D of bahen, Sali, nanad

jamai

DH

H of beti, bhatiji, bhanji, also poti, dhevti (see G −2)

bahu

SW

W of beta, bhatija, bhanja, also poti, dhevti (see G −2)

pota

SS

S of bhatija, bhanja

poti

SD

D of bhatija, bhanja

dhevta

DS

S of bhatiji, bhanji

dhevti

DD

D of bhatiji, bhanji

Source: After Vatuk (1969). Specifications in the middle column are the minimal genealogical designations for each term. Remoter kin types covered by the same terms are given in the right-hand column.

Dravidian and Iroquois in South Asia

The north Indian terminological system has proved difficult to pin down, whether in its own terms or in comparison with the more clearly defined prescriptive systems of south India. I have suggested elsewhere (e.g. 1990, reproduced here as Ch. 2; also 1992a, 1992b) that a diachronic approach might prove more fruitful than the synchronic comparisons that are more usually attempted. We certainly have the conceptual resources with which to describe this terminology, if not actually to define it. It is clearly non-prescriptive: affinal terms are normally separate, and there are not even the inter-cognate equations (MB=FZH etc.) that one encounters in the north and central Munda terminologies, for example. The only exception to this statement, as Trautmann has pointed out (1981: 98 ff.), is that certain affines who can be described as wife-giver’s wife-takers and wifetaker’s wife-givers of ego’s level may be classed as siblings, though in a manner that is non-prescriptive (cf. Vatuk ibid.). Ego’s level is generational as regards cognates, but the affinal area of this level usually exhibits a series of individual terms for each primary kin type. This is replicated in +1 and −1 for both cognates and affines, though gender differences, especially between married couples, are often marked by what a linguist might consider morphological variation (e.g. kaka/ kaki, FyB, FyBW) rather than terminological separation. In +2 and −2 also, each of the four minimal kin types receives a separate term. It is tempting to call this last pattern ‘individualizing’, but this expression is more usually applied to situations in which the kin types that are genealogically closest to ego (P, G, C) have separate terms in a terminology that is otherwise prescriptive or at least classificatory. Allen prefers to call the principle involved ‘zero-equation’ – that is, one in which terms and kin types, at least in parts of a terminology, have a one-to-one correspondence (1989: 178); Kryukov’s ‘Arabic’ type is also relevant here (1972), as is Murdock’s ‘Sudanese’. A further distinction may be necessary, according to whether some ‘terms’ do or do not take a descriptive – that is, circumlocutary – form, such as Arabic bint amm, literally ‘father’s brother’s daughter’. However, as this too is a type of zero equation, that is not immediately relevant here. In any event, it is not easy to conceive of this principle being applied right through a terminology, assuming that this covered a wide range of kin and did not, as with us, recognize the largest number of individuals in the society as non-kin: otherwise, the number of terms required would have to be impossibly large. And if we examine the North Indian terminologies recorded by Vatuk and Parry more



145

146

How Kinship Systems Change

closely, we see that there are indeed terminological equations even for these supposedly individualizing terms. These equate primary kin types not with each other, as in a prescriptive or even to a limited extent a classificatory terminology (perhaps we should say here, with Needham, ‘lineal’ terminology), but with a range of wider kin types that can clearly be seen as equivalents in some sense. I deliberately avoid using the term ‘classificatory’ to describe them for the moment. At the +1 level, there are six terms for male cognates – F, FeB, FyB, MB, FZH and MZH – which are therefore all terminologically distinct from one another. For the purposes of this demonstration, we can disregard relative age and reduce the set to five terms. Each term except that for F, which is a purely zero-equation or individualizing term, has equations not with any other primary kin type (not even the equation F = FB obtains here) but with the same-sex cousins of the parent who provides the initial link from ego. In other words, the equation G=PGC always applies in the formula used for these kin types, as it does in ego’s level, which is generational. It applies equally in the affinal part of the terminology in all levels, in the sense that the terms for WB or EMZ, for instance, also cover those cousins of W, EM etc. who are of same sex as their siblings. Thus in many areas of the terminology, this principle of siblingcousin equivalence is applied in a manner that is internally coherent but that cannot be considered conventionally Hawaiian or generational. What one sees is its thorough-going application not only to ego’s immediate siblings and cousins but also to those of his parents and same-level affines. This means that although the +1 level is not generational as regards primary kin types – quite the reverse, with separate terms for each, not just one for all – it can be considered generational as regards kin types at a wider remove. This pattern is unique, as far as I know, and cannot be considered Iroquois or any of its variants any more than it can be considered Dravidian. However, it can logically be seen as a further development of the former, one that retains the equivalence of, for instance, PG and PPGC on a relative-sex basis (ego’s parent to alter) but that no longer makes the inter-cognate equations that survive in Iroquois. At this stage, I am remarking only on the logical potential and am making no reference to historical possibilities. One consideration is whether we can call these particular sorts of equivalences classificatory. The most usual definition of this term is that it describes lineal equations of the sort F = FB, with the assumption that equivalent kin types at a wider collateral distance are

Dravidian and Iroquois in South Asia

also included – in this instance FFBS, FFFBSS etc. But this does not actually cover all its uses. One also talks freely of the ‘classificatory MBD’, for example, in order to indicate that not only the actual MBD is being referred to but all her collateral equivalents too. However, not even the actual MBD is likely to be equated with a primary kin type (which here could only be a sibling); if she were, she would be unlikely to be distinguished from parallel cousins either in practice, and one would call the equation Hawaiian or generational rather than classificatory. This shows that lineal-collateral equations need not actually be present – and therefore that they cannot form part of the definition of the term ‘classificatory’. Despite the usual definition, equations of the type FB = FFBS = FFFBSS etc. where F has a separate term are, according to this logic, as classificatory as those where F is also a part of the equation. On this basis, then, the north Indian terminology is zero equation as regards the lack of equivalence between primary kin types themselves but classificatory in the sense that each of the latter (except F) is terminologically equated with a distinct set of equivalent kin types. But this ignores the usual assumption that not all collateral kin types are equivalent in a classificatory equation: the principles of relativesex designation and of same-sex sibling equivalence come into play. This applies no less to the idea of a classificatory MBD, since the actual MBD and her collateral equivalents who make up this category, although all cross kin to ego, are parallel kin to one another. But from the egocentric view of a terminology, parallel kin are linked with lineal kin and opposed to cross kin, who will also be affines if the terminology is prescriptive. This means that while FFBS (father’s parallel cousin) is considered to be a classificatory equivalent of FB, FMBS (father’s cross cousin) is not. An equation that covered all three specifications would not, in this view, be classificatory. This renders Iroquois and its sub-types, and also north Indian, not simply non-prescriptive but non-classificatory, since they all equate (for instance) a parent’s sibling with all that parent’s same-sex cousins, cross as well as parallel.

Transformation of South Asian Kinship Terminologies Thus the Burushaski terminology is the sole clear representative known of typological Iroquois in South Asia. To this fact, some other interesting and relevant evidence can be introduced. To begin with,



147

148

How Kinship Systems Change

note the ways in which Burushaski is distinguished from north Indian. First, there is the fact that the former does not distinguish primary kin types from one another in the way that the latter does. Secondly, a related point, Burushaski has classificatory equations of the sort defined above: not even parental terms are individualizing (F = FB = FFBS = MZH etc.). Thirdly, although Burushaski has no cognate-affine equations, it does have the inter-cognate equations and distinctions for both cross and parallel kin that one would expect to find in a prescriptive terminology. Neither of these latter two features is characteristic of north Indian. What the two terminologies share, however, is the generational nature of ego’s level and the treatment of second cousins in the +1 level. The second comparison depends on making allowance for the distinction from one another in north Indian but not in Burushaski of the primary kin types with whom these second cousins are equated. Further east, in the upland area of Kumaon, north-west India, a terminology has been recorded with a pattern that is clearly intermediate between these two. The data come from a German publication (Krengel 1989: 196–97) in which the terminology is deliberately set out on the model used by Parry (who had himself adopted it from Vatuk) to facilitate comparison with other north Indian terminologies (see Table 5.3.). The treatment of first and second etc. cousins is the same as in these two earlier texts – that is, generational. What is different is the treatment of cognates in +1 and −1. In +1, parallel kin are mostly equated with one another in the manner of prescriptive systems, even though there are no cognate-affine equations: FeB = MeZH (thulbaujyu ‘big father’); FyB = MyZH (kaka or kasabaujyu ‘little father’); MeZ = FeBW (thullja ‘big mother’), though MyZ is distinguished from FyBW (kaij, kaki) in conformity with north Indian. In −1, although the term for BS (ws) is not clearly recorded (presumably bhandya and bhandye, just given as BS, BD), the equations C = ssGC = EssGC (cyal, celi) are, and ZS (ms) and ZD (ms) (bhanja, bhanji) have separate terms equating them with HZC, like a symmetric prescriptive terminology. +1 cross kin, however, have separate terms not only from parallel kin but also from one another (i.e. mam, MB ≠ bhin, FZH; didi, FZ ≠ mami, MBW), as normally in north Indian. The Kumaon terminology can thus be described as having a basically north Indian pattern with some, perhaps residual, features consistent with prescription but not north Indian. Certain north Dra-



149

Dravidian and Iroquois in South Asia

Table 5.3. Kumaon kinship terminology bubu

FF, MF

FFB, (FFZH), FMB, FMZH, MFB, (MFZH), MMB, MMZH, Z of any bubu

ama

FM, MM

W of any bubu

bur caur

HFF, WFF, HMF, WFM, FZHF any bubu of H/W

bur sasu

HFM, WFM, HMM, WMM, FZHM

ija, ij

M

baujyu, babu

F

thulbaujyu, kaka

FeB/FyB

FFBS, FFZS, FMBS, FMZS (e/y)

kasabaujyu

MZH

MFBDH, MMBDH, MFZDH, MMZDH

thulija/kaki

FeBW, MeZ/FyBW

W of any thulija/kaki

kaij

MyZ

mam

MB

MFBS, MFZS, MMZS, MMBS, FWBW

mami

MBW

W of any mam

didi, pusyani

FZ

FFZD, FFBD, FMBD, FMZD (classified with eZ)

bhin

FZH

H of any didi

caur

WF, HF

HFB, WFB, HMB, WMB, HMZH, WMZH, ZHF, ZHFB, BWF

sasu

WM, HM

W of any caur

dad/bula

eB/yB

FBS, FZS, MBS, MZS, H of any sai, jethau, nand, pauni

boji/bvari

eBW/yBW

(yBW classified with SW)

didi/bhuli

eZ/yZ

FBD, FZD, MBD, MZD, W of any sau, jethu, javai, bhin (eZ classified with FZ, yZ classified with BDws)

bhin/javai

eZH/yZH

BWBfs, ZHB (eZH classified with FZH, yZH classified with DH)

any ama of H/W

(continued)

150

How Kinship Systems Change

jethu/sau

WeB/WyB

WFBS, WMBS, WFZS, WMZS, BWBms (WyB classified with WBS)

didi/bhuli

WeBW/WyBW

W of any jethu/sau

jethau (pauni)/ WeZ/WyZ, ZHZ sai

any Z of W; any Z of ZH

dad/bhuia

WeZH/WyZH

any H of jethau, sai

jithana/dyor

HeB/HyB

HFBS, HMBS, HFZS, HMZS

jithani/dyorani HeBW/HyBW

W of any jithana/dyor

pauni/nand

HeZ/HyZ

HMBD, HFBD, HFZD, HMZD, ZHZws, HFZ

dad/bhuia

HeZH/HyZH

ZHZH, HFZH

samdi

SWF, DHF

F of any javai or bvari

samdyani

SWM, DHM

W of any samdi

cyal

S

BSms, WZS, ZSws, HBS

celi

D

BDms, WZD, ZDws, HBD

bhadya

BSws?

bhadye

BDws?

sau

WBS

sai

WBD

bhanja

ZSms

HZS

bhanji

ZDms

HZD

javai

DH (yZH)

H of any celi, bhanji, natini, bhuli, including SWB, DHB, yZHB

bvari

SW (yBW)

W of any cyal, bhanja, nati, including SWZ, DHZ

nati

SS, DS

natini

SD, DD

Source: Krengel (1989: 196–97). Specifications in the middle column are the minimal genealogical designations for each term. Remoter kin types covered by the same terms are given in the last column.

Dravidian and Iroquois in South Asia

vidian terminologies of Bihar indicate an even further and later stage of transition from prescriptive to non-prescriptive in India. They have equations consistent with prescription only in the +1 level, and only in the most vestigial form; in addition, they exemplify what I have suggested elsewhere (Parkin 1988b: 62 ff.) about terminological redundancy indicating possible change. Since this point is still likely to be somewhat obscure, let us look at the Malpahariya +1 consanguineal terminology (Sarkar 1937). Table 5.4. Malpahariya kin terms (+1 consanguines) jetha

FeB, MeZH; FeZH

mosa

MyZH

kaka

FyB, MyZH

mama

MB

jethi

MeZ, FeB; FeZ

mami

MBW

kaki

FyBW

pisi

FyZ

mosi

MyZ

pisa

FyZH

Source: Sarkar (1937).

One can see equations consistent with prescription here for the terms jetha and jethi, but contaminated, as it were, by extra specifications that would not belong there in any purer prescriptive system. Kaka and mosa especially should be compared. Both are MyZH, but it is mosa that one can expect to prevail in the struggle over this specification, for which it represents the standard north Indian term (e.g. Hindi mausa). This would mean the disappearance of another residual prescriptive equation, an assumption strengthened by the fact that kaka, FyB, MyZH is normally only FyB in north Indian. In general, this part of the terminology appears to be moving towards the zero-equation structure of the typical north Indian terminology, in which each primary kin type has its own term. Here, in effect, one is able to predict terminological change because of the rich comparative evidence we have from elsewhere in the region.



151

152

How Kinship Systems Change

The terminologies discussed here together with others that are typical of various parts of the south Indian region suggest two different transformational paths. I have already broached this question in two earlier articles (Parkin 1990, reproduced as Ch. 2; also 1992b) but take this opportunity to make some revisions here. Table 5.5. Trajectories of change in South Asian kinship terminologies 1.

Each path starts from Central Dravidian / Koraput Munda (two-line symmetric prescriptive terminologies in five levels and with alternating generation equations, thus differing in the last respect from Dravidian systems further south; see Parkin 1988a, 1992a). These lead to either:

2.

South Dravidian (two-line symmetric prescriptive in three levels, with a non-prescriptive pattern in +2 and −2 and no alternating generation equations); or:

3.

North Munda (+1 and −1 cognatic terminology consistent with two-line symmetric prescriptive but with separate affinal terms; generational in ego’s level)

4.

Burushaski (similar to north Munda, but with Iroquois pattern for +1 and −1 cognates)

5.

Kumaon (differs from Burushaski in having only residual equations consistent with prescription for +1 and −1 parallel kin types; primary cross kin in these levels have individualizing terms)

6.

Malpahariya (evident break-up of remaining prescriptive features)

7.

Standard north Indian (total removal of such features)

Proving the historicity of these transformations is the real challenge, something that will ultimately require much careful reconstructive work. But there are already some pieces of plausible evidence. The splitting of cognate-affine equations in Munda terminologies has generally been effected with the aid of loans from Indo-European (IE) languages (see Parkin 1992, Ch. 7, for details), whose adoption was clearly a historical event or series of such events. The exact circumstances in which the Malpahariya terminology seems to be losing its remaining prescriptive features is another example (see Parkin 1992b; 1992c). There is more to discuss concerning the remains of such features in Kumaon, a terminology that is lexically IE and that exists in a frontier area between the prescriptive systems of the Himalayas and the non-prescriptive systems of the plains. This raises

Dravidian and Iroquois in South Asia

the possibility that it falls into the class of South Asian terminologies that appear to have kept a partially or wholly prescriptive pattern but to have switched lexis from another language – Dravidian, Munda or Tibeto-Burman – into IE (e.g. Marathi, Konkani, Sinhalese, Shina; see Trautmann 1981: 147 ff.; Parkin 1987: 167). Alternatively, there may simply have been some direct influence from neighbouring prescriptive terminologies on some parts of it for reasons that are at present obscure. Another difficulty is the paucity of evidence at present for the existence of more terminologies similar to the Burushaski one in South Asia. As already demonstrated, there are some small indications of a Cheyenne pattern in certain Munda terminologies, and a dedicated search might reveal that these are not as isolated as they now appear to be. Conversely, we can observe that possible transformational paths are neither identical nor preordained, and that history will not always have responded everywhere in the same manner. Historically, there are two possible hypotheses as regards South Asia. One is that the transformations I have suggested took place wholly within South Asia. Sufficient evidence exists within the region to support such a hypothesis now that the Burushaski evidence is clearer and can be linked to at least one terminology, Kumaon, which is mostly but not entirely north Indian in structure. Against this hypothesis is the fact that north Indian terminologies are overwhelmingly if not entirely Indo-European lexically and that this language family – South Asian examples such as those listed above excepted – cannot be identified with prescriptive systems at any point in its history, a history that is the deepest we have for any language family. If one were to chart a shift from prescription to non-prescription in South Asia, one would need to find a means of making IE terminologies prescriptive in the region. This is not impossible through the process of combining lexical transfer with the retention of prescriptive terminological patterning, at least in part, I have just mentioned.6 However, there is no evidence that this took place everywhere, only locally, and in regions of contact between IE and other language families. The second hypothesis is therefore more likely; namely, that however IE speakers arrived in the subcontinent they brought with them terminologies that were already non-prescriptive. It may nonetheless be the case that the prescriptive systems they encountered there arrested their further development towards a more cognatic structure, in contrast to what appears to have been the development followed



153

154

How Kinship Systems Change

by linguistically cognate terminologies in Europe. It may well be that IE terminologies generally were at one stage in their history characteristically zero equation (as regards primary kin types, at least), a pattern that can be discerned in the Latin terminology and that still survives in Scandinavian ones.7 By its very nature, a zero-equation structure can logically only move towards the production of equations, and logically in more than one direction. One such direction would be to introduce classificatory and possibly also prescriptive equations. This, however, offends against the usual hypothesis that prescriptive precedes non-prescriptive (into which latter class zero equations fall), and classificatory precedes cognatic, in any transformational sequence. The other is to produce a cognatic pattern like most modern West European terminologies. As already noted, this seems to have been the most likely development in Europe itself. In South Asia, however, it did not take place, either because of the internal dynamics of Indo-European north Indian terminologies themselves or because of influence from the prescriptive terminologies of the region. In some cases, this influence clearly extended to producing prescription through lexical transfer, but in others it might have been restricted to preventing the further development of north Indian terminologies in a cognatic direction. Conversely, the presence of north Indian in the subcontinent might itself have had an effect in drawing certain lexically Dravidian, Munda and Tibeto-Burman terminologies away from prescription, with its cognate-affine equations. This change may have been effected through the disappearance or diminishing importance of positive marriage rules, itself reflecting the impact of high-status marriage practices that disapproved of marriage with any sort of cousin.

Sanskrit and Hindi A final point in this chapter is the question of how the present-day Hindi terminology and its cognates arrived at their present form.8 The immediate ancestor of the terminologies listed by Parry (1979) and Vatuk (1969) is Sanskrit, described by Hocart (1928). Although the Sanskrit structure appears to be basically the same, at least in having a zero-equation pattern for primary kin types, many of the actual terms are quite different. In particular, many collaterals have clearly composite or descriptive terms that translate the respective

Dravidian and Iroquois in South Asia

genealogical specifications exactly; for example, matrsvasar MZ, maturbhratra MB, pitrsvasar FZ, svasriya ZC (from svasar Z), bhagineya ZS ms (from bhagini Z). The first and last of these are represented in the reduced Hindi forms mausi (also giving mausa MZH) and bhanja (also giving bhanji ZD) recorded by Vatuk, but the rest are not traceable in modern forms, nor is pitrvya FB (PGE forms are not given). The same is true of Sanskrit +2 terms, which are terms for parents plus maha/i ‘great’, though pautra/-i CC is represented in Vatuk’s pota/-i. However, affinal terms are mostly similar. Sanskrit also has special terms, though also composite, for certain cousins, which in Hindi are all classificatory siblings. Perhaps this indicates similar autonomy for them in Sanskrit, which they have lost in Hindi etc. Similar composite forms characterize the collateral parts of a number of modern European terminologies, including those of Scandinavian languages and Latin. They are featured in Chapter 6, this volume. The main point here, however, is to draw attention to these composite or descriptive terms, as they indicate that Sanskrit was a somewhat earlier stage in the development discussed in this chapter, in which the process of isolating a number of categories with their own terms was effected using composite or descriptive terms.

Notes * First published in Maurice Godelier, Thomas Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship, Washington and London: Smithsonian Institution Press (1998). Reproduced with permission. 1. Note added in 2020: this indicates that the terminology is not, in fact, prescriptive, or not any longer. As in the case of Ch. 2, therefore, I hereby repudiate my earlier tendency to call certain terminologies ‘prescriptive’ on the basis of the inter-cognate equations one expects of such terminologies even in the absence of cognate-affine equations, which are diagnostic of them. These remarks also apply to the discussion of this question at the end of this section, which is already more tentative than in Parkin 1990 (Ch. 2). 2. Note added 2020: since 2000, the separate state of Jharkhand. 3. Ali’s data also amount to a decisive refutation of Pfeffer’s (1984) hypothesis that the Burushaski terminology implies at least ‘four lines of regular exchange’, solely on the basis that it separates affines from consanguines. Nothing in Adam Nayyar’s data, on which Pfeffer actually based his analysis, suggests this hypothesis, and it must be rejected for similar reasons as was the case for middle India (see Parkin 1993). That is to say, there are insufficient terms to establish four lines in the terminology, and the terms have the wrong specifications to establish proper descent lines to begin



155

156

How Kinship Systems Change

4. 5.

6.

7.

8.

with. Ali’s data, which were almost certainly too new to have been known to Pfeffer at the time of his study, are sufficiently full to put his hypothesis beyond all salvation. Data are from Parkin 1984, Ch. 6, tables 8, 9; original sources listed at ibid.: 349. (Note added 2020: last sentence of original paragraph omitted as no longer appropriate.) There are also two examples from the related Mon-Khmer languages of Southeast Asia, namely Khasi (of Meghalaya, north-east India) and Bahnar (of the highlands of south-central Vietnam). I give the essential details below. Data are from Parkin 1984, tables 14, 19; original sources list at ibid.: 350–51: Khasi kmie, M, MZ, MPGD, FBW; kñi MB, MPGS, FZH; kpa kha F, FB, FPGC (kpa = MZH) Bahnar (western dialects) duc FyZ, MyBW, FyPGD, MyPGSW; yang MyZ, FyBW, MyPGD, FyPGSW; ka, me kra FeZ, FePGD; ma, ñô MyB, FyZH, MyPGDH, FyPGSDH (also MyZH, MyPGDH). Note added in 2020: I now think this must have happened quite regularly. It is pretty generally accepted that many if not most castes in India are of tribal origin, having made this shift for prestige reasons, alongside the adoption of the local Indo-European language or dialect. This process continues and is not only historical (e.g. the Bhuiya of Orissa; Parkin 2020: 74). Hypothetically, therefore, we can suggest that cousin marriage in any form would also disappear for prestige reasons, ensuring the dispersal of alliances among several spouse-exchange groups, as in north Indian marriages generally. Moreover, this would lead to the terminology sooner or later losing any vestiges of prescription and becoming north Indian too. The Polish terminology currently seems to be undergoing precisely this transformation from zero equation to cognatic (see Parkin 1995). Note added 2020: Scandinavian IE languages may actually also now be changing in this direction (see Ch. 6). Note added 2020: this section was not in the original publication but has been added especially for this volume.

References Ali, Tahir. 1983. The Burusho of Hunza: Social Structure and Household Viability in a Mountain Desert Kingdom. Ann Arbor, MI: University Microfilms. Allen, Nicholas J. 1989. ‘The Evolution of Kinship Terminologies’, Lingua 77: 173–85. Beck, Brenda. 1972. Peasant Society in Konku. Vancouver: University of British Columbia Press. Bouez, Serge. 1985. Réciprocité et hiérarchie: L’alliance chez les Hos et les Santals de l’Inde. Paris: Société de l’Ethnographie. Dumont, Louis. 1957. Une sous-caste de l’Inde du Sud: Organisation sociale et religion des Pramalai Kallar. Paris: Mouton.

Dravidian and Iroquois in South Asia

Godelier, Maurice, Thomas R. Trautmann and Franklin E. Tjon Sie Fat (eds). 1998. Transformations of Kinship. Washington DC and London: Smithsonian Institution Press. Hamid, S. Shahid. 1979. Karakorum Hunza: The Land of Just Enough. Karachi: Ma’aref. Hocart, A.M. 1928. ‘The Indo-European Kinship System’, Ceylon Journal of Science, 1: 179–204. Krengel, Monika. 1989. Sozialstruktur in Kumaon: Bergbauern im Himalaya. Wiesbaden: Franz Steiner Verlag. Kryukov, Mikhail. 1972. Sistema rodstva kinaitsev [The Chinese Kinship System]. Moscow: Nauk. Lounsbury, Floyd. 1964. ‘The Structural Analysis of Kinship Semantics’, in Horace G. Lunt (ed.), Proceedings of the Ninth International Congress of Linguists. The Hague: Mouton, pp. 1073–93. McDougal, Charles. 1963. The Social Structure of the Hill Juang. Ann Arbor, MI: University Microfilms. ———. 1964. ‘Juang Categories and Joking Relationships’, Southwestern Journal of Anthropology 20: 319–45. Murdock, George. 1949. Social Structure. New York: Free Press. Needham, Rodney. 1973. ‘Prescription’, Oceania 43: 166–81. Parkin, Robert. 1984. ‘Kinship and Marriage in the Austroasiatic-Speaking World: A Comparative Analysis’, DPhil thesis. University of Oxford. ———. 1987. ‘Kin Classification in the Karakorum’, Man 22(1): 157–70. ———. 1988a. ‘Reincarnation and Alternate Generation Equivalence in Middle India’, Journal of Anthropological Research 44(1): 1–20. ———. 1988b. ‘Prescription and Transformation in Mon-Khmer Kinship Terminologies’, Sociologus 38(1): 55–68. ———. 1990. ‘Terminology and Alliance in India: Tribal Systems and the North– South Problem’, Contributions to Indian Sociology 24: 61–76. ———. 1992a. The Munda of Central India: An Account of their Social Organization. Delhi: Oxford University Press. ———. 1992b. ‘Dispersed Alliance and Terminological Change in South Asia’, Journal of the Anthropological Society of Oxford 23(3): 253–62. ———. 1993. ‘Middle Indian Kinship: A Critique of Georg Pfeffer’s Interpretation’, Anthropos 88(4–6): 323–36. ———. 1995. ‘The Contemporary Evolution of Polish Kinship Terminology’, Sociologus 45(2): 140–52. ———. 2020. South Asia in Transition: An Introduction to the Social Anthropology of a Subcontinent. Lanham, MD: Lexington Books. Parry, Jonathan. 1979. Caste and Kinship in Kangra. London: Routledge & Kegan Paul. Pfeffer, Georg. 1984. ‘Kin Classification in Hunza’, Journal of Central Asia 7(2): 57–67. Sarkar, S. 1937. ‘The Social Institutions of the Malpahariyas’, Journal of the Asiatic Society of Bengal, Letters 3: 25–32. Stirrat, R.L. 1977. ‘Dravidian and Non-Dravidian Kinship Terminologies in Sri Lanka’, Contributions to Indian Sociology 11: 271–93.



157

158

How Kinship Systems Change

Tiemann, Gunther. 1970. ‘The Four-Got Rule among the Jat of Haryana in Northern India’, Anthropos 65: 166–77. Trautmann, Thomas. 1981. Dravidian Kinship. Cambridge: Cambridge University Press. ———. 1992. ‘Dravidian Kinship as a Structural Type and as a Cultural Type’, paper presented at the Conference on Kinship in Asia, Moscow: Russian Academy of Sciences. Trautmann, Thomas, and R.H. Barnes. 1998. ‘“Dravidian”, “Iroquois”, and “CrowOmaha” in North American Perspective’, in Maurice Godelier, Thomas R. Trautmann and Franklin E. Tjon Sie Fat (eds), Transformations of Kinship. Washington DC and London: Smithsonian Institution Press, pp. 27–58. Vatuk, Sylvia. 1969. ‘A Structural Analysis of the Hindi Kinship Terminology’, Contributions to Indian Sociology 3: 94–115.

6 

Indo-European Kinship Terminologies in Europe Trajectories of Change

Introduction This is a study of kin terms and kinship terminologies in IndoEuropean (IE) languages.* There is, of course, no shortage of such studies already (e.g. Delbrück 1889; Hocart 1928; Galton 1957; Friedrich 1966; Szemerényi 1977; Kullanda 2002), which go back to the nineteenth century. By and large, however, most of them are concerned with reconstructing terms historically, right back to proto-IE, and/or studying specific terminologies or groups of terminologies either synchronically or diachronically, or both. Rather less attention has been given to changes in IE languages in Europe generally following the break-up of IE into its component families, which for this chapter means Baltic, Slavonic, Greek, Latin/Romance, Germanic and Celtic. The present chapter seeks to develop understanding of this topic. More specifically, I hypothesize that there has been a trend – starting over two thousand years ago in some cases, but only happening now in others, and not made at all in yet others – for IE terminologies in Europe to shift from a zero-equation pattern (i.e. with separate terms for most kin types) to a cognatic one (in Rodney Needham’s sense of the term – i.e. involving a broad distinction between lineal and collateral relatives, but not among the latter in any generation).1

160

How Kinship Systems Change

In 1998, I published a chapter in a volume devoted to the differences between the Dravidian and Iroquois variants of kinship terminology in many parts of the world and possible transformations between them.2 While the Dravidian type has been well known from south India since the nineteenth century, and indeed takes its name from the language family that predominates in this area (though similar terminologies are found worldwide), firm recognition that the Iroquois type was different had to await a study by Floyd Lounsbury in 1964 (1964a) showing that the cross-parallel distinction worked differently in such cases for kin beyond first-cousin range. My own chapter in that volume (Parkin 1998) identified an Iroquois-type terminology in South Asia (the non-IE Burusho terminology of northern Pakistan). However, my chapter also went on to describe certain features of the very different kinship terminology of Hindi, an IE language of north India, pointing out its tendency to have a separate term for each kin type, a pattern that has been described as ‘zero equation’ (Allen 1989), though it is also represented by Murdock’s ‘Sudanese’ (1949: 224, 238–39). Much more recently (2012), I have suggested that zero-equation terminologies are likely to emerge from bifurcate collateral ones: for example, in the +1 pattern for male referents, F ≠ FB = MZH becomes F ≠ FB ≠ MZH, and MB = FZH becomes MB ≠ FZH. In fact, in principle any terminology with characteristic equations (prescriptive, bifurcate merging, bifurcate collateral, Crow-Omaha) might break down those equations to produce a zero-equation pattern, whether through separate terms or relative product-type phrases (i.e. descriptive phrases). Many African terminologies have this kind of pattern, as does Arabic. However, it is not claimed that the pattern necessarily occurs throughout a terminology, and indeed the Hindi terminology is only partially of this type. Nonetheless, very broadly speaking we have zero-equation terminologies, at least partially, in the east of the IE speech area and cognatic ones in the west, with intermediate or mixed patterns in between. Since this is a single language family with a substantially common linguistic inheritance, it is reasonable to assume that changes have taken place historically that can mostly be contained within that inheritance – that is, they involve little or no borrowing from non-IE language families. It is the aim of this chapter to try and trace at least some of these changes, without seeking to account for them linguistically or sociologically apart from some very brief comments where relevant.

Indo-European Kinship Terminologies in Europe

To return to the second of my articles just mentioned (Parkin 2012; reproduced here as Ch. 1), this is a general theoretical article setting out the most likely trends in changes in kinship terminologies, though, barring the occasional brief comment, it presents typological sequences rather than strictly historical ones. It argues inter alia that while zero-equation terminologies by definition lack terminological equations between kin types or only have a few of them, terminologies affected by this principle would have to reintroduce equations in changing to a cognatic terminology. What was lacking at the time of writing the earlier article on Burusho and Hindi (Parkin 1998) was hard evidence of this happening: Hindi is related linguistically to other Indo-European languages in Europe, but there was no obvious way of tracing the change from the Hindi zero-equation kinship terminology to a cognatic terminology like English in the absence of intermediate patterns. This situation has now changed following a closer examination of the varying patterns of IE terminologies in modern Europe. Sources for this exercise include dictionaries, though these rarely give enough detail on their own, and, where available, previous studies of kin terms in languages such as Greek, French and Catalan.3 However, in the case of the Slavonic and Baltic branches especially, but also Norwegian, I have also drawn on oral communications from a number of students and colleagues who are mother-tongue speakers of some of these languages, as well as my own field enquiries some years ago in Poland (see Parkin 1995).4 While this informant base is, of course, extremely small, and while many of the written sources have limitations of their own, I am confident I have enough material to make a reasonably sound assessment of the variations in these terminological patterns and possible pathways of change between them.

Lexical Universalism/Evolutionism An important inspiration for this and other such studies is the work of the so-called ‘lexical universalists’, especially the famous and at times controversial study of the growth of colour terminologies by Brent Berlin and Paul Kay (1969). The major claim in this approach is that, largely but not entirely, the order in which certain colour terms emerge historically can be predicted. Thus, starting with terms for white, black and red – the latter the only so-called ‘hue’ colour



161

162

How Kinship Systems Change

in this initial paradigm – either yellow or ‘grue’ (green/blue) emerges next as a separate term, then the other one of this pair, then ‘grue’ separates into green and blue, and so on. As a result, as Allen has pointed out (1984), this makes these scholars ‘lexical evolutionists’ as much as ‘lexical universalists’. The methods they use are based on a scrutiny of a large number of ethnographic texts drawn from around the world. The results are more a typological sequence than a truly verifiable historical one, but there are some exceptions where there is evidence for change, such as the introduction of English ‘orange’ from Arabic naranja (both fruit and colour), this possibly having been defined as a shade of yellow or red earlier. Regarding how colour categories change, Edwin Ardener showed how the Welsh colour terminology, once very different in its distribution of categories across the colour spectrum than the English one, has converged with the latter in modern times (1971: xxi). Similar studies were made in the lexical universalist camp of how life-form terms emerge (Brown 1984) and on developmental sequences in kin categories (Witowski, his unpublished thesis, 1971, but also 1972). In the case of life-form terms, thanks to the biological sciences the inventory of terms is constantly and apparently endlessly expanding as more and more discoveries are made. The same might be said of colour terms, expansion of separate terms for different shades only being limited by the imagination of paint manufacturers and their marketing specialists. However, these are now shades rather than colours: another of Berlin and Kay’s observations is that certain colours – primarily those that emerge as separate at the earlier stages of the sequence – can be considered ‘focal’ because the physiology of the eye responds to them more directly, implying that the terminological expansion of colour terms cannot go on indefinitely. This is even more the case with kin terms, as the number of theoretically discriminable kin types is limited by the facts of kinship, and so, therefore, are the paradigms into which kin terms can be ordered by equating or distinguishing kin types in different ways. For example, there are only so many ways of ordering terms for +1 male kin (F = FB = MB; F ≠ FB ≠ MB; F = FB ≠ MB; F ≠ FB = MB), and not all of those are found, or likely to be found, in reality (in this example, F = MB ≠ FB). We therefore already have a limit on variation within this particular semantic domain – that is, kinship terminologies. One more point to be made here, following Witowski, is that the presence of a particular pattern in one part of a terminology may imply its

Indo-European Kinship Terminologies in Europe

presence in another part but not vice versa. For example, although Hawaiianization is often present in ego’s generational level but absent from the +1 level, the reverse situation (i.e. absent from ego’s level, present in +1) does not seem to occur ethnographically. Its presence in +1 thus normally suggests it will be present in ego’s level as well but not necessarily vice versa. Similarly, prescriptive equations expressing the operation of repeated cross-cousin marriage suggest cross/parallel distinctions but not vice versa, given the appearance of the latter in non-prescriptive Iroquois and Crow-Omaha terminologies as well. This principle can therefore be used to explain and even predict terminological change, much as Berlin and Kay argued for colour terms. Finding historical evidence for change, however, is less easy and is the real challenge for work in this area.5 I now turn to discuss separately the situation in each branch of Indo-European languages in Europe, apart from Albanian, due to insufficient data.6

Baltic A study trip I made to Latvia and Lithuania in the spring of 2013 indicated that the Latvian and Lithuanian terminologies may represent a kind of missing link both typologically and historically between IE languages of India like Hindi and the European branches of IE. Lexically, the links between them are recognized to be close, though in fact many of them are reflexes of Proto-IE. As already indicated, however, systemically there is more of an affinity. In particular, the patterns of the historical Lithuanian and possibly also Latvian terminologies have a close similarity to Hindi without being identical, while the present-day terminologies in these two languages appear to be undergoing or to have undergone developments similar to those already made centuries ago by West Germanic (English, Dutch, German) and Romance (French, Italian, Catalan, Spanish, Portuguese). In this sense, they are changing in parallel with North Germanic or Scandinavian languages (Swedish, Norwegian, Danish, Icelandic), though there are also differences between the latter and the Baltic group. The Baltic languages of interest here are Lithuanian, Latvian and Old Prussian, the latter no longer a living language. Published sources for Lithuanian are primarily Buivydiene˙ 1997 (in Lithuanian



163

164

How Kinship Systems Change

but with an English summary),7 and for Old Prussian Mažiulis 1988 etc., but for Latvian I have had to rely on dictionaries and personal enquiries. Mažiulis provides some cognates of common Baltic terms and etymologies but nothing of structural significance except for a now redundant Latvian term (see below). There are frequent lexical similarities with Slavonic languages, representing direct loans in some cases but indicating a common heritage in others, reflecting the hypothesis that both branches formed a single branch of IE in the past.

Lithuanian We start with Lithuanian, which has clearly become increasingly cognatic over time, historically being more zero equation in pattern. PP terms now are senelis PF, senele˙ PM (from a root for ‘old’; cf. Latin), though formerly the latter appeared as mocˇiute˙.8 Note also prosenelis PPF, prosenele˙, promocˇiute˙ PPM, proprosenelis PPPF, propromocˇiute˙ PPPM. Descending reciprocals include dukraite˙ or vaikaite˙ CD, but vaikaitis or a loan from Slavonic, anu¯kas, are used for CS. The latter reappears in both male and female forms in −3: proanu¯kas for CCS, proanu¯ke˙ for CCD, with provaikaitis and provaikaite˙ as alternatives respectively. Collaterals in +2 include de˙de˙ PPB (also ‘uncle’; very similar to forms in Russian). The prefix pro- is clearly cognate with pra- etc. in Polish and other Slavonic languages. In +1, te˙vas is F, motina or mócˇia M. FB and MB are both now de˙de˙ (cf. PPB above), though MB was formerly avynas, MBW avà, avýniene˙.9 Terms for PZ also seem to have changed: today both FZ and MZ are teta (cf. PPZ, above) or de˙diene˙ (grammatically a feminine form of de˙de˙), but in the past FZ was de˙dina, MZ móša (cf. Hindi mausı¯ ). The resemblance of de˙de˙ and teta to old Russian forms (see Friedrich 1964) indicates that the PG area of the Lithuanian terminology may have become cognatic through their borrowing. Also recorded historically are tet(ul)e˙nas PZH and de˙diene˙ PBW.10 In −1 su¯nus is S, dukte˙ D. There seems to have been a similar re-sorting of GC terms as of PG terms: today GS is su¯ne˙nas, GD duktere˙cˇia, obviously derived from the terms for S and D. However, formerly there were four terms, brole˙nas or brolaitis BS, brolaite˙ BD, sesere˙nas ZS and sesere˙cˇia ZD, the two pairs derived from the terms for B and Z respectively. Sesere˙nas and sesere˙cˇia are also given as PGC, the only cousin terms recorded apart from tetule˙nas (literally ‘aunt’s children’; but cf. the PZH term above). Thus these areas of the Lithuanian terminol-

Indo-European Kinship Terminologies in Europe

ogy have clearly moved from a zero-equation or bifurcate collateral pattern to a cognatic one. There are similar derivations for HBS, dievere˙nas or dieveráitis, and HBD, dievere˙cˇia; and WGS, svaine˙nas, WGD, svainiene˙ derived respectively from a now redundant dieveris HB and a still current sváine˙ WZ. SW is marti, DH žéntas, terms also formerly covering BW, ZH respectively – that is, they linked the adjacent 0 and −1 generations (a feature also widely found in IE languages in India, NB). In +1, terms for EP are also undergoing a change similar to that in the consanguineal +1 terminology: EF is now úošvis, EM úošve˙, though formerly these were WP terms only, HF being šešuras, HM anýta. The only terms recorded for CEP categories are svo¯tas CEF and svocˇià CEM, dialect words from the east and south Aukstaicˇiai regions of Lithuania, and very similar to Slavonic forms. In ego’s level, B is brolis, Z sesuo, W mote˙, žmonà or pati, H výras (also redundant diedynas, pats). The affinal terminology of ego’s level again shows a change from a zero-equation pattern to a cognatic one. Formerly, there were separate terms for all specifications, but many have now become obsolete. Apart from bróliene˙, all those that follow immediately have now fallen out of use in the meanings given, if not absolutely: marti (cf. SW, above) or bróliene˙ BW; žéntas, ZH (cf. DH, above); láigonas WB; láigoniene˙ WBW; dieveris HB; jente˙ HBW; móša HZ (cf. MZ, above); and moše˙nas HZH. Now the terms originally confined to WZ, sváine˙, and WZH, sváinis are used for all these specifications, according to gender.11 The affinal terminology in ego’s level in Lithuanian thus indicates a historical pattern similar to present-day Hindi etc., with separate terms for each kin type in this field. Subsequently, there has been a change to just two of these terms being used for all these kin types, according to gender. As Buivydiene˙ points out, this means that descriptive terms now have to be used to specify these kin types more precisely. In fact, most of the terminology has made similar changes, making Lithuanian a link (at least typologically) between Indic languages and IE languages in Europe, as well as between the zeroequation and cognatic patterns.

Latvian The Latvian KT appears to represent an intermediate position between these two patterns, as very many kin types, especially for con-



165

166

How Kinship Systems Change

sanguines, are designated by compound terms based on the primary set one would find in the nuclear family: te¯vs F, mate M, bra¯lis B, ma¯sa Z, vı¯ rs H, sieva W, de¯ls S, meita D. In +2 PP terms are generally vecte¯vs PF, veca¯ma¯te PM, combining parent terms with vecs ‘old’; similarly, −2 terms are mazde¯ls CS, mazmeita CD, incorporating maz ‘little’ (also mazberns CC). However, compounds are recorded alongside these in some cases, e.g. meitade¯ls DS, but not *de¯lsde¯ls SS. In +1, PG terms are compounds, apart from te¯vocis alongside te¯vabra¯lis for FB. However, the term krustte¯vs, literally ‘cross-father’ or ‘step-father’, does not mean FB or ‘uncle’ in the modern language.12 In addition, the loan term onkulis (a loan ultimately from French oncle, perhaps introduced via German Onkel) is now being used for PB and PZH. Similarly, tante has evidently been borrowed from French or German for PZ, PBW, alongside traditional compounds. Conversely, no terms corresponding lexically to ‘nephew’, ‘niece’ or ‘cousin’ have been borrowed: the former two are essentially covered by modified compound terms, while cousin terms are similarly derived from those for siblings, according to gender, viz. bra¯lens PGS and ma¯sı¯ca PGD. In the −1 affinal field, SW is vedekla, DH znots (cf. Lith. žéntas) alongside the compound meitasvı¯rs. Only compounds are recorded for EP terms. As in contemporary Lithuanian, just two terms cover the affinal field in ego’s level, according to gender, namely svainis (masc.) and svaine (fem.).13 Compound terms are also recorded in this area, but assuming the terminology has evolved in a similar way to Lithuanian, these would appear to have followed, not preceded, the reduction to just two primary terms in this field, being used genealogically to specify relationships more precisely rather than as primary terms in their own right; they thus resemble the descriptive phrases mentioned by Buivydiene˙ for Lithuanian. In Lithuanian, the process has clearly gone from i) a zero-equation pattern consisting of primary terms for each kin type, to ii) a concentration of all these kin types on two gendered terms, to iii) compounding as a reaction to ii), adopted in order to describe relationships more clearly as a subsidiary feature (genealogical rather than categorical). It is at least likely that Latvian has undergone similar changes, and indeed Buivydiene˙ records some obsolete zero-equation-type terms for that language too, viz. ietere HBW (cf. Lith. jente˙), dievelis, dievainis HB, dievainı¯tis ‘das Schwägerlein’, dieverene HBD.14 Similarly, Mažiulis gives ma¯rša as BW in Latvian (cf. Lithuanian marti, above). Finally, Buivydiene˙ also records

Indo-European Kinship Terminologies in Europe

older Latvian ma¯se¯ns MZS, from ma¯sa MZ (cf. older Lithuanian moša, above; cf. Hindi mausı¯). However, in Latvian there has been another process in other parts of the terminology (especially for consanguines), of compounds becoming separate words in their own right. This resembles the opposite direction of change, one that can be identified for Scandinavian and indeed Germanic generally, namely that composite terms for collaterals precede rather than succeed the introduction of primary terms – usually through borrowing, e.g. of onkel and tante – arranged in a cognatic pattern. In Latvian, but apparently not Lithuanian, this latter process is represented by the contemporary introduction of the loans onkulis and tante for PG specifications to replace earlier compounds for this part of the terminology. The Latvian process could therefore have been similar to Lithuanian as regards affines in ego’slevel but was reversed in other parts of the terminology such as the PG field.

Slavonic KT As in the case of the Baltic branch of IE, terminologies in Slavonic languages are generally intermediate between zero equation forms and the cognatic pattern. However, there is considerable variation in this interesting group, examination of which demonstrates some of the changes I argue are taking place.

Serbo-Croat Some years ago, I published an article on changes in Polish KT (1995), which I shall use as a default in discussing other Slavonic terminologies here. I start with Serbo-Croat, for which Hammel 1968 is the leading published source. Since the break-up of Yugoslavia, what was formerly Serbo-Croat has become divided into separate Serbian, Croatian and Bosnian national languages for political reasons, though linguistically they remain essentially dialects of the same language. Any important variants will be noted as we proceed, but I shall continue to use the term ‘Serbo-Croat’ where data apply to both these languages, with or without Bosnian. In fact, most of the Serbo-Croat terms are recognizably Slavonic and clearly cognate with Polish. In the main, I shall only mention



167

168

How Kinship Systems Change

instances where the two terminologies diverge. In Serbo-Croat, the prefix pra- is not reduplicated for +/−4 levels and beyond as it is in Polish but restricted to +/−2 and +/−3; cˇukun- is used for +/−4. The PGE field is a little different from Polish, though lexically some terms are similar: thus tetka is FZ and MZ, tecˇa PZH, while stric, strina mean FB, FBW, ujak, ujna MB, MBW. By contrast, the pattern in Polish was formerly stryj FB, FZH, stryjna FZ, FBW; wuj, MB, MZH, ciotka MZ, MBW, though there was also evidence of the last two terms taking over the whole field in a cognatic pattern. In Serbo-Croat, this is so far only happening for tetka, now effectively meaning ‘aunt’ (consanguineal only, though, not affinal).15 Polish tes´c´ EF and tes´ciowa EM are represented in Serbian as tast, tašta, but only as WP; HF, HM are respectively svekar, svekrva, clearly from Germanic, though via which route is unclear. Cousin terms, in both ego’s level and +1, are essentially descriptive in a manner similar but not identical to Polish. However, affinal terms in ego’s level are quite different. Whereas in Polish there is a tendency to consolidate this field around just two gender-differentiated terms borrowed probably from German, namely szwagier and szwagierka, Serbian retains separate terms for most referents in this field, viz.: zaova HZ, svastika or svastica WZ, d(j)ever HB, šurak WB, zet ZH (also DH; also HZH in Hammel; cf. Polish zie˛c´, DH only), snaha or snaja BW, SW (Polish bratowa, synowa respectively) and jetrva HBW. A Turkish loan, pašenog, is used for WZH, probably related to pašanac, an alternative term for BW, SW, though Buivydiene˙ (1997: 180) gives svak for WZH. There are also separate terms for CEP, namely prijatelj CEF (also ‘friend’) and prija CEM (in Polish these are szwagier and szwagierka).16 Apart from the intergenerational mergers of BW and SW, and of ZH and DH, Serbo-Croat does not diverge very much from Polish in −1 or −2. Svastic´ WZS and svastic´ina WZD are recorded, clearly based on svastica WZ. Hammel (1968: 27) gives bratanac as BS ws, the ms term being sinovac, this relative-sex pattern, which is very unusual in a European language, also applying to BD (bratanica ws vs. sinovica ms). Also odd is the fact that while the former term is based on brat, B, the latter is based on sin, ‘son’. Hammel attributes the relative-sex feature here to the fact that in the traditional zadruga or extended family BC would have been co-resident, ZC not. ZS and ZD are therefore respectively sestric´, sestricˇina – that is, they display an absolute-sex pattern in both Serbian and Croatian. Apart from the

Indo-European Kinship Terminologies in Europe

relative-sex terms above, this pattern is very like Polish, and even Hammel adds the information (1968: 29 n. 24) that ‘in urban Serbia, the term sinovac is losing currency’. However, alongside these divided terms, Hammel (ibid.) also lists combined terms for nephew and niece, respectively nec´ak and nec´akinja, although, writing in the 1960s, he says they are heard ‘more rarely and usually among more educated persons; they are sometimes regarded as affected’. Further, he says that, ‘In Kotor, with its heavy Italian influence, one hears neput and neputa, nephew and niece’. These last two pieces of information indicate that in some areas the GC part of the Serbo-Croat terminology has a cognatic pattern either alongside or in place of an older zero-equation pattern, the latter being strongly represented in the terminology more generally.17

Bulgarian There is also considerable zero-equation patterning in the Bulgarian terminology, though we start with its cognatic aspects. Dyado and baba are respectively PF and PM, circumlocutions being resorted to in order to specify side ( po bashtina linia, ‘from the father’s side’; po maichina linia, ‘from the mother’s side’). The familiar pra- prefix is added to these terms for +3 kin and to uncle and aunt terms to provide the terms for great-uncle and great-aunt. The standard Slavonic roots vnuk, vnuchka are used for CS, CD. Great-nephews and greatnieces may use the terms for −1 niblings, plemennik GS, plemennichka or plemennitsa GD (also in Russian, from which they might have been borrowed), though this is not entirely clear. F is bashta, there being no reflex of Polish ojciec etc. here; M is maika. PZ is lelya, but the PB field is split between chicho FB and vuycho MB, the latter recalling Polish wuj etc. FZH and MZH are both svako, reflecting the situation for ‘aunt’, but the split between FB and MB is replicated with their wives, MBW being vuyna, clearly related to vuycho, FBW strinka, clearly related to Polish stryj FB, stryj(e)nka FBW, FZ etc. WP have separate terms from HP, viz. tast, tashta, vs. svekar, svekarva (cf. Serbo-Croat). The standard Slavonic roots for siblings occur, viz. brat B, sestra Z. Male cousin is bratovched, female cousin is bratovchedka, with the circumlocutions given above for side of family being added where necessary. H and W are saprug, sapruga respectively, though maj, literally ‘man’, and jena, literally ‘woman’, may be used as well. Sibling-in-law



169

170

How Kinship Systems Change

kin types mostly have individual terms, with familiar equations with −1 referents. Thus snaha is BW and SW, as well as being one possible term for HZ; zet is ZH and DH, as well as HZH. WB is shurei, WBW shurenaika, WZ baldàza, WZH badjanak. HB is dever, a familiar IE root for this kin type, HBW etarva (cf. Russian jatrov, Serbian jetrva). As well as snaha for HZ, this kin type is also zalva (cf. Serbian zaova). The only remaining −1 terms to note are sin S and dashterya D, both common Slavonic roots. Thus Bulgarian, too, retains some zero-equation features in key parts of its terminology, with some cognatic features in the PG field.

Russian The main published source for Russian is Friedrich (1964), who considers both history and the present-day terminology. While he does not do much to identify possibly obsolete kin terms, he does state: ‘By 1700, Russian terms for the avuncular and nepotic relationships were no longer bifurcate collateral and the kinship system had become technically lineal [sc. cognatic], lumping the parents’ brothers together as against the father, and so forth’ (ibid.: 141). As far as the latter are concerned, Friedrich was anticipated by P.A. Lavrovski, a nineteenth-century Russian philologist, who not only identified such changes but traced them back even further, to the twelfth to fourteenth centuries. Earlier, terms for ‘uncle’ distinguished FB as stryi or stroi from MB as (w)ui, before giving way to dyadya for both, coupled with tyotka for FZ and MZ, the latter attested as far back as 1178, possibly as a loan from Polish (Kryukov 1998: 301, who adds FBW and MBW to the specifications). Lavrovski saw this as constituting a simplification of the terminology consonant with the shift away from collective forms of social organization (the ‘clan’) to the nuclear family.18 Friedrich also mentions that BS and BD (bratánich, bratánna, from an older bratán B) were formerly distinguished from ZC (ZD term not given, but ZS is sestrich, from sestra Z),19 though both are now plemjannik/-itsa. It is clear from the table he gives (p. 140) that the present-day terminology is entirely cognatic as far as consanguines are concerned: grandparents,20 uncles, aunts, nephews and nieces are now distinguished by gender but not side; pra- (cf. Polish) is used as a marker for more distant lineals (ascending and descending),; and cousin terms are descriptive ones based on brat and sestra in a

Indo-European Kinship Terminologies in Europe

similar fashion to other Slavonic languages, though, possibly under French influence, kusin, kusina, though rare, can be found as far back as the nineteenth century (Y.S., p.c.). Otherwise first cousins are literally ‘second brothers/sisters’, second cousins ‘third brothers/sisters’ etc. (ibid.). Alongside djádja PB, PZH, in the north one found djadina PBW,21 though this is now also or instead tjótja, otherwise PZ (pp. 144–45). The PB term seems to be based on ded, PF. As we have seen, cognates of Polish stryj FB, wuj MB are not found here at the present day, but Buivydiene˙’s yu˘ for MB in earlier Russian (1997: 162; cf. end of my note 18) indicates a past bifurcation of MB and FB. The affinal terms listed by Friedrich (p. 144) have a more zeroequation pattern, at least in part. Though he does not make it clear whether these terms represent present-day or historical usage, merely saying that ‘the affinal terminology has been collated from the available documents’ rather than fieldwork, this same pattern is indicated by more contemporary data (Y.S., p.c.). In Friedrich’s account, HP (svjókor, svekróv) are distinguished from WP (test’, tjóshcha) as in Serbo-Croat but not in Polish, where cognates of the latter pair stand for all EP. In ego’s level, HZ, WB and WZ all have separate terms (respectively zolóvka, shúrin or shur’já,22 and svojáchenitsa or in dialect sves’). BW has the separate term bratanikha but it can also be nevéstka, a term also meaning SW (ws), HBW. A third term for BW, játrov’, is also EBW, though it is apparently the preferred term for HBW, in preference to WBW, and indeed had that meaning solely before about 1700 (Friedrich ibid.: 163 n. 10). However, more recent information (Y.S., p.c.) does not recognize bratanikha or játrov’, WBW in this paradigm being nevéstka. Both sources have snokhá as SW (ms); thus Russian has one example of a relative-sex pair, unusual for IE, though Friedrich also adds (ibid.: 155) first that a woman’s HM might use it of her too in some districts (i.e. it is ws too), and secondly, that an absolute-sex term, synóvka, also exists or existed (cf. Polish). For Friedrich, DH is covered by zjat’ or zja’já only, which also means ZH and HZH. However, there is an alternative term for the latter pair, svójak, also meaning WZH predominantly (only WZH for Y.S. p.c.). HZH is evidently also a subsidiary meaning of déver’ or dever’já (cf. Hindi), otherwise HB. Svat or svatov’já and svákha complete this field as CEF, CEM. Again, more recent information (Y.S., p.c.) did not recognize some of these terms as current, namely zja’já, dever’já and svatov’já. However, deleting these unrecognized terms mainly means simply removing alternative vocabulary from



171

172

How Kinship Systems Change

Friedrich’s list and does nothing in itself to dilute the zero-equation pattern of this part of the Russian terminology. Nonetheless, the recent changes in usage (i.e. 1960s or earlier) that Friedrich discusses briefly in a footnote (ibid.: 164, n. 11) indicate that there may have been some reduction of terms: ‘By 1963 most young, urban Russians . . . had forgotten the specific meanings of svat, svát’ja, had altogether lost játrov’, were using svoják for most close male affines, were using purely descriptive terms for all siblings-in-law, and were limiting the reference of zjat’ and nevéstka to children-in-law.’ Taken together with the information above, this suggests that terms for female affines of ego’s generation seem to have been less affected, but, nonetheless, this suggests a tendency to simplify this area of the terminology by moving towards more global terms, backed up by the use of descriptive terms and phrases when a relationship needs to be specified more narrowly, even if the process has been less complete here than in other cases. There are indications that this situation has been perpetuated into the following generations – that is, that descriptive terms are beginning to replace primary ones in this field – possibly alongside many of these older terms, which at present are also surviving (I.Z., p.c.; Y.S., p.c.). One other contemporary change in Russian is the gradual abandonment of otets for F, mat’ for M, ded for PF and baba for PM, all now being seen as excessively formal and even derogatory. Instead, what were originally the more affectionate address terms are becoming standard – that is, respectively, papa, mama, dedushka, babushka; Y.S., p.c..

Bielorussian Unsurprisingly, this has a closely similar terminology to Russian in both structure and vocabulary. Like Russian, Bielorussian has no cognate of Polish stryj FB; unlike Russian, it does retain stryechny brat, stryechnaya siastra for FBS, FBD; reflexes of Polish wuj are also lacking, however. While terms for HB and HZ are recognizably similar to those in Russian, the terms for WB and WZ are respectively shvager and shvagerka or svayachka, recalling Polish forms but presumably originally from German Schwager etc. Both Hammel (for Yugoslavia) and Friedrich (for Russia) suggest that the changes from zero-equation to cognatic reflect in part the decline and/or disappearance in these regions of traditional ex-

Indo-European Kinship Terminologies in Europe

tended families like the Yugoslav zadruga, with concomitant modern changes like the collectivization of agriculture and urbanization in these socialist societies (though potentially they apply to other sorts of society too). Thus a social system based on exogamous patrilineal extended families linked by marriage may well encourage co-resident paternal uncles to be terminologically distinguished from maternal uncles residing elsewhere. However, this becomes much less necessary with a shift to an urban environment based on the nuclear family having neither as resident, whereby both kin types eventually fall into the same category of ‘uncle’. One consequence of this is the progressive reduction in the circle of people recognized, or simply interacted with, as kin in modern conditions. The process from zeroequation to cognatic terminologies certainly requires the deletion of terms from the terminology. This goes against the usual assumption that classifications expand in size and complexity rather than decrease, though, to return for a moment to the lexical universalists, we do have reduction in the example of the loss of terms for flora and fauna in urban conditions that are more divorced from everyday contact with nature (cf. Brown 1984).

Czech The Czech terminology has many similarities to Polish, both lexically and systemically, but it is more definitely cognatic in form than any other terminology discussed thus far in the article. Interestingly, PB is strýc, PZ teta, comparable to Polish stryj, ciotka but without any cognate of Polish stryjna (FBW, FZ), wuj (MB, MZH) or wujna (MBW). WP = HP, (tchán EF, tchyneˇ EM), the prefix pra-, usually reserved elsewhere for generations remoter than +2 and −2, being added to these two terms for CEF, CEM (pratchán, pratchyné). Cousin terms are non-descriptive, unlike in other Slavonic, but they resemble the GC area of the latter in using terms based on bratr B and sestra Z, namely bratanec and sestrˇenice (cf. Polish bratanek BS, siostrzenica ZD). The term for H, manžel, recalls Polish ma˛z˙, but W is preferably a derived form, manželka, though žena ‘woman’ can also be used for this kin type (cf. Polish z˙ona). Affines in ego’s level all have a single term depending on gender: male švagr, female švagrová, as in many cases with such words probably from German. BS = ZS (synovec; cf. Polish synowa, SW), BD = ZD (neterˇ). The cognatic nature of the Czech terminology may reflect the long and intense Germanization



173

174

How Kinship Systems Change

of Bohemia in particular, which, unlike the case of Polish, severely restricted the use of Czech in the region and nearly drove it out of existence. More generally, taking the Czech example into account indicates that the Slavonic branch of IE spans the full range from something resembling broadly the Hindi zero-equation form to the west European cognatic form.

Ancient and Modern Greek This section describes changes in Greek kin terms from antiquity to the modern period, as well as within antiquity itself. Sources for antiquity are Hocart (1928), Miller (1953) and Szemerényi (1977); for the modern period, Andromedas (1957), Friedl (1962: 70 ff.), Herzfeld (1983) and Just (2000: Ch. 4).

Ancient Greek In +1, the terms given by Hocart are pate¯r F, me¯te¯r M, te¯t(h)is PZ (Szemerényi: linked with tethe PM) and me¯tro¯s MB, the latter Homeric according to Szemerényi. Mallory and Adams (2006: 216) state that this is the sole derivative in the daughter languages of PIE *méhatro¯us. It can be compared with the root *mehatruha~ MZ, yielding Gk metruia ‘step-mother’. The Homeric term for FB was patrokasigne¯tos, later giving way to patrádelphos and pátro¯s, the latter linked to Sk. patr. vya according to Hocart. From the time of Euripides, the term theîos appears for both MB and FB. According to Szemerényi, theia ‘aunt’ followed in the first century AD, presumably replacing te¯t(h)is. In due course, theîos and theia were borrowed by all the major western Romance languages except French and Romanian. The term for ‘brother’ also changed in the ancient Greek period itself. Originally phra¯te¯r, representing the standard IE root, this term came to be used for ‘clan brother’ alone in ancient times. Hocart says that it was first linked in a single phrase to adelphos, originally meaning ‘of the same womb (delphos)’, thus qualifying phra¯te¯r as ‘own brother’; later phra¯te¯r dropped out of this expression, leaving adelphos alone with the changed meaning of ‘brother’ and yielding a feminine form, adelphe¯, as ‘sister’. Szemerényi also gives sor for the latter, from IE *sweso¯r. Hocart gives kasigne¯tos and kasigne¯tai as the Homeric terms for male and female cousins, though strictly they

Indo-European Kinship Terminologies in Europe

meant uterine siblings (cf. patrokasigne¯tos FB, above; also FBS according to Wallis). Hocart mentions one other cousin term, anepsio¯s, FBS and FZS in Herodotos, and even DH and ZH, as well as ZS; later it was transferred wholly to the −1 level as GS, its meaning in modern Greek. According to Mallory and Adams (2006: 211), this is a regular development of PIE *népo¯ts, with CS as the most likely reconstructed meaning. This is also the opinion of Goody (1983: 266): ‘The Greek aneptios [sic], derived from grandson, stands for “cousin” and in modern Greek, “nephew”.’ Ancient Greek terms in −1 include paîs C, which Szemerényi derives from a root for ‘small’. Hocart also lists huiós S (a regular development of PIE *suhxnús) and thugate¯r D (Szemerényi dugater; Mycenean Greek tukate). Alongside anepsio¯s, discussed above, Hocart also gives adelphidéos as nephew (ZS in Herodotos), while Wallis has adelphide for niece. Szemerényi gives anepsia as GD and says that it and anepsio¯s mean ‘cousin’ in Ancient Greek, GC in modern Greek (see above). Ancient Greek terms in +2 also changed within the period of antiquity, from composite terms to terms derived from baby-speak. Thus Homer gives me¯tropátor MF, as well as Pindaric patropato¯r FF and me¯trome¯to¯r MM; Hocart also mentions patrome¯to¯r FM as a later term. These gave way to páppos PF and te¯the¯, later mámme¯, PM. In –2, huio¯nos is CS (Szemerényi: possibly SS because of derivation from huiós), but Hocart gives no term for CD. Szemerényi gives annis for PM as well as te¯the¯. Affinal terms in Ancient Greek include gambrós, primarily DH but also ZH (Homer), EF, as well as WB according to Wallis. However, Miller and Wallis also give pentheros for WF, ZH and DH,23 indicating flux in this area. Other terms are dae¯r HB (linked to dever etc. in other IE languages), gálo¯s HZ (Wallis adds BW to this term), aélioi WZH, núos SW (from PIE *snusós), hekurós EF (from PIE *swekuros) and hekurá EM, the latter two apparently older than pentheros. Szemerényi also gives enater HBW, presumably the same as the Homeric einateres recorded by Wallis.

Modern Greek Many of the above terms can be identified in those given by Andromedas for Modern Greek – that is, patéras F, mitéra M, adherfós B, adherfí Z, yos S, papús PF, thios PB, thia PZ, anepsyós GS, anepsyá GD,



175

176

How Kinship Systems Change

and ghambrós DH, ZH (also ‘groom’). New consanguineal terms in Modern Greek are yáya PM, kóri D, pedhí C (but on Karpathos yios; Vernier 1991: 112), ángonas CS, angóna CD, (e)xádherfos PGS and (e)xadhérfi PGD (both literally ‘from brothers’). It is the affinal terms that show the greatest change lexically from ancient Greek: nífi SW, BW (also ‘bride’), batzanákis WZH, sinifádha HBW (lit. ‘co-bride’), pentherós EF, pentherá EM, kunyádhos EB, kunyádha EZ, simpétheros (CEF) and simpethéra (CEM). Herzfeld (n. 1) notes that in western Crete kunyádhos is also ZH, ZHB, kunyádha also BW, BWZ, while Andromedas points out that both terms were borrowed from Romance (cf. Iberian languages). Just (2000: Ch. 4) does not discuss consanguineal terms very much but gives a full set of affinal terms in agreement with Andromedas, though with variant spellings. He also adds the information that cousin terms extend to the spouses of cousins and the cousins of spouses, and that simpétheros, simpethéra are also used for CEG, GEB, GEP, GCEPG – that is, any affine lacking a specific term listed above. Campbell gives some dialect forms (though recognizably based on standard modern Greek) for the Sarakatsani shepherds of Thrace (1964: 38, 42 n. 6, 110, 139, 174). Andromedas also draws attention to the fact that some Greek terms are loans (1957: 1088). These include Ancient Greek theíos and anepsiós from unspecified non-IE languages, as well as the element delphýs ‘womb’ (‘of uncertain origin’) in adelphós/-e¯. As just noted, kunyádhos/-a is from Iberian IE, and yáya also has (unspecified) Romance origins, while batzanákis is from Turkish.24 None of these loans or the other changes identified through a comparison between ancient and modern Greek sources has much in the way of structural significance, and they mostly just represent vocabulary replacement over time. Modern cousin terms are clearly derived from sibling terms, almost indicating a Hawaiianization of ego’s level, whereas in Ancient Greek they tended to be linked semantically with GC categories. In essence, though, the ascending part of the Greek terminology has been cognatic in type since the replacement of composite terms for +2 categories by baby terms and the introduction of theîos for both MB and FB in the post-Homeric period; similarly, the GC area appears to have been cognatic throughout most of recorded history. However, the sibling-in-law part of the terminology retains a largely zero-equation pattern, despite lexical replacement, and despite the consolidation of EB and EZ through Romance loans.

Indo-European Kinship Terminologies in Europe

Latin and the Romance Languages Sources for this branch include Wallis (1918) and Hocart (1928); Miranda (1974) on French; Iszaevich (1981) and Llobera (1997) on Catalan; and Callier-Boisvert (1968) on Portuguese.

Latin The Latin terminology has been much studied, and it is recognized that it is far from having had a cognatic pattern. In +1, pater F and patruus FB are etymologically linked in opposition to avunculus MB (‘little avus’), though without sharing terms exactly – that is, without entirely being bifurcate merging. Similar remarks apply to mater M and matertera MZ in opposition to amita FZ. In ego’s level, frater germanus B is distinguished from frater patruelis FBS (later just patruelis according to Hocart), soror Z from both soror patruelis FBD and consobrinus/-a MZC. However, Wallis says that FBC and MZC are sobrinius/-a to one another, while MBC and FZC are amitinus/-a, so that cross cousins and parallel cousins are distinguished in a way that is unusual in IE. Given also avus PF (avia PM) alongside avunculus MB, one can see why previous scholars (e.g. Lounsbury 1964b; Friedrich 1966) have thought they found Omaha features in this terminology. There is also the conflation of GC and CC under the terms nepos, neptis, retained in Italian and Romanian but not the more western Romance languages. Hocart says that nepos originally meant CC, becoming GC after Augustus. The main remaining consanguines are filius/-a S, D. Latin affinal terms are levir HB, glos HZ, socer EF, socrus EM, gener DH and nurus SW. Wallis adds janitrices HBW, fratria BW and enater ZH (where Z is deceased; possibly a loan of Greek enater HBW).

Italian, Catalan, Spanish and Portuguese The daughter languages of Latin, on the other hand, have become basically cognatic. I concentrate first on Italian, Catalan, Spanish and Portuguese (on the last three, some details here are from Entwhistle 1962: 64–65). All these terminologies have cognatic patterns for +1 kin, with Italian zio/-a and Spanish and Portuguese tio/-a for uncle and aunt (borrowed from Greek; cf. Catalan oncle and tia) alongside parent terms. Portuguese illustrates the pattern in ego’s level, with



177

178

How Kinship Systems Change

irmão B and irmã Z (Spanish hermano/-a; Catalan germá/-ano) being distinguished from primo irmão/hermano PGS and prima irmã/hermana PGD in Portuguese and Spanish (descending cousins being just primo/-a), though Catalan has cosí, cosine, either from Latin or loaned from French. Thus the Latin terms for siblings were replaced in all three Iberian languages, and those for cousin in Spanish and Portuguese. By contrast, Italian fratello, sorella, for B and Z, are clearly derived direct from Latin, like French fils, fille, which Miranda (1974) regards as exceptionally stable terms that emerged early in the development of French. Portuguese GC and GCC are sobrinho/-a (Spanish sobrino/-a), CC and CCC neto-/-a (Spanish nieto/-a). The Catalan equivalent of the latter is nét/-a, and all three forms are derived from Latin nepos, nepta. However, in Catalan GC are nebot, neboda. Portuguese cunhado is used of WB and ZH, cunhada of BW (also not derived from Latin); sogro/-a are EP, genro DH, nora SW (all derived from Latin; cf. Spanish suegro/-a). In Catalan, all cousins are cosí or cosine, DH gendre (are these French loans?) and SW nora. Thus these languages separate GC and CC categories (though the Catalan terms might have the same origin), while Italian retains the late Latin conflation of them in nipote.25

French Miranda (1974) is a useful source for the history of the French terminology. In +1, French distinguishes oncle PB and tante PZ from parent terms; the former is derived from Latin avunculus, the latter from amita (both have been borrowed by a number of Germanic languages as well). French cousin(e), Italian cugino/-a PGC are also distinguished from sibling terms; like the Catalan terms, they are clearly derived from Latin consobrinus/-a. Grandparent terms also differ among these languages, with Iberian terms deriving from Latin, Italian nonno/-a apparently being derived from baby language, and French grandpère, grand’mère being new developments (reflecting German influence? Cf. Grossvater, Grossmutter). French petit-fils and petite-fille for CC date from the fourteenth century, having replaced nibling terms in this field (petit/-e is a marker of distance here, like arrière, which expresses even greater distance). Until the sixteenth century, grandpère faced competition from aïeule in French (‘grand’ connotes ‘old’ here). French shows considerable deviation from other Romance languages in its affinal terminology. It retains gendre for DH but for SW

Indo-European Kinship Terminologies in Europe

discarded any reflex of Latin nurus and borrowed bru from Old High German, where it meant CE, later ‘bride’ (Modern German Braut).26 +1 and 0 level terms in beau-, belle- competed with Latin-derived terms from about the fifteenth century, which in most cases they replaced, only gendre surviving, and that with difficulty. Beau-père and belle-mère originally applied to step-P as well as EP; both entered Dutch in loan translation in the sixteenth century (schoonvater, schoonmoeder). In essence, some of the remaining Latin terms that did not fit a cognatic pattern (avunculus, amita, consobrinus) have expanded their semantic fields and supplanted others (patruus, matertera, other cousin terms) to produce a cognatic pattern in French. In the case of avunculus and amita, at any rate, it is significant that they, the cross-uncle and cross-aunt, have survived to mark the new contrast with parental terms and that their parallel equivalents have disappeared. While it is not as clear with Latin as in the case of Greek that these changes were made in antiquity, they had certainly taken effect by the time the daughter languages had reached the modern era. However, it is clear that the various new coinings, loans and the peculiar development of the French terminology within the Romance group are not linked to any of these structural changes, which appear to have taken place before the daughter languages of Latin had clearly become distinct.27

Germanic We have sufficient historical information on Germanic kin terms to be able to say quite a lot about changes from a zero-equation pattern to a cognatic one – indeed, the published sources are too numerous to discuss in any detail here, though Bjerke’s comparison of Old German and Old Norse should be mentioned (1969). However, these changes have only fully taken place in German, Dutch and English, and only to a limited extent in the Scandinavian languages, though there are indications that here too this development is underway.

Swedish, Norwegian and Danish In fact, Scandinavian languages are still partly zero-equation and have obviously been more so in history. Swedish, Norwegian and Danish



179

180

How Kinship Systems Change

are very similar in this regard. Thus Swedish has the primary ‘nuclear family’ terms far ( fader) F, mor (moder) M, bror (broder) B, syster Z, son S and dotter D, simply combining them to produce PP, PG, GC and CC terms; for example, morbror MB. However, Norwegian and Danish also have onkel and tante (also Swedish tant, onkel; Barlau 1981: 199) as cognatic-pattern terms for PG, as well as nevø, niese (Nw.; R.K., p.c.), niece (Dan.) for GC. In all three languages, cousin terms also follow a cognatic pattern, viz. Swedish kusin PGC, Nw., Dan. kusine PGD; Nw. fetter, Dan. fætter PGS.28 All these terms are clearly loans from French, possibly through German or English, except the last, which is also found in older German (Vetter) but not French. In Norwegian onkel, tante (both also used for PGE), nevø and niese are also compounded with grand-, which can only come from French or English, for great-uncle, great-aunt, great-nephew and great-niece, though the old compounds are still used for the four grandparents and grandchildren, the latter alongside barnebarn (lit. child’s child). R.K. adds the terms oldefar, oldemor for PPP, oldebarn for CCC. Barn is ubiquitous for ‘child’ in Scandinavian languages (cf. Scots English ‘bairn’, with the same meaning). Affinal terms also tend to follow the compounding principle; for example, Nw. systermann as ZH. W is kone (Nw., Dan.), fru (Sw.; cf. German Frau); H is mann (Nw.), mand (Dan.; cf. German Mann), hustru (Nw., Sw.). In Swedish, there is a whole series of related terms, viz. svåger EB, svägerska EZ, svärfar EF, svärmor EM, svärson DH, the latter alongside an alternative term of uncertain derivation, måg DH.29 Bjerke gives Old Norwegian variants or predecessor terms, thus mágr ZH (cf. Sw. måg, above) and versyster HZ, ver- being a bound morpheme for ‘husband’, or perhaps ‘spouse’ or even affines in general. Also conforming to the compounding principle was bród˜urkona BW. At one period púsa, from Fr. épouse, was used for ‘wife’, later replaced by húspreya; H was (hús)bóndi. More recent information (R.K., p.c.) indicates that the Old Norwegian bound affinal morpheme ver- (see above) reappears in varsyster BW and verbror ZH. However, there are also blanket terms for all sisters-in-law (i.e. BW, EBW, EZ) and brothers-in-law (i.e. ZH, EB, EZH), respectively svigersøster or svigerinne, and svigerbror or svoger. Sviger- and svoger are clearly Germanic in origin (cf. German Schwager), but due to their absence from Bjerke’s list of Old Norse terms it is unclear whether they are loans or cognates. They also appear in

Indo-European Kinship Terminologies in Europe

other levels, thus svigerfar, svigermor for EP, svigerdatter SW. Given the record of an alternative systermann for ZH, a clearly descriptive compound, we might speculate that any such terms isolating individual affines of ego’s generation is giving way to two blanket terms (with alternatives) distinguished only by gender, indicating a cognatic direction of change. This new information on Norwegian may provide a model for what is happening in Nordic languages generally – that is, the replacement of zero-equation terms with extensive loans in a cognatic pattern.

Icelandic Icelandic kin terms have been studied much more extensively that those in other Scandinavian languages – perhaps because they have been taken as being more conservative – and have given rise to a degree of controversy. A useful summary is Merrill 1964. Fundamentally another zero-equation terminology, the main differences are in the affinal terminology. Thus mágur (cf. måg, mágr above) is EB, ZH, mágkona EZ, BW. These terms were previously used for all in-laws according to Merrill, but in +1 and −1 they have now been replaced by consanguineal terms prefixed by tengd˜a-, from tengja, ‘tie together’ (e.g. tengd˜asonur DH). Rich (1976) also applies this prefix to affines of ego’s level as alternatives to mágur and mágkona, viz. tengd˜abród˜ir, tengd˜asystir; he also indicates that descriptive terms exist as alternatives for all affines, including svili EZH (cf. n. 28) and svilkona EBW. Nonetheless these affinal terms do not distinguish wife’s from husband’s kin, so are not as zero-equation as they could be. The cousin terminology also differs from other Scandinavian, formerly having a zero-equation pattern for parallel cousins (bræd˜rungur/-a FBC, systrungur/-a MZC),30 cross cousins apparently having only descriptive terms. However, Rich (1980: 476–77) indicates that these older terms are now giving way to two synonyms referring to all cousins (i.e. as glossing English ‘cousin’), namely tvímenningur and frœndsystkini. Similarly (ibid.: 477), he gives frœndi as a term equating FB and MB, frœnka as the female equivalent for FZ and MZ, stating that on occasion these are used of PGE referents too. In both these cases, therefore, a switch from a zero-equation to a cognatic pattern (what Rich calls a ‘centrifugal’ process) is indicated in part of the Icelandic terminology, too, countering Barlau’s argument (1981: 199) that this is not happening in Icelandic and that it is therefore different from



181

182

How Kinship Systems Change

other Scandinavian terminologies. In the PP area of the terminology, afi PF and amma PM now exist alongside the older compound terms, which they may be replacing; Rich (1976) says they only came into common use in the nineteenth century, the former derived from Latin avus, the latter possibly from the Latin root am-, as in amita. Informally, in address, the loan words pabbi and mamma are used instead of fad˜ir and mód˜ir (F, M). H is mad˜ur, W kona, often prefixed by eigen-, from eiga, ‘to own’ (cf. German eigen).31

German There are many sources for changes in German, but here I use mainly Mitterauer (2000). In about the tenth century, there were zeroequation terms for PG, namely Vetter FB, Base FZ, Muhme MZ and Oheim MB (the latter cognate with Latin avunculus according to Priebsch and Collinson 1948: 27). Earlier, FB had been fatureo, derived from or cognate with Latin patruus according to Wallis (1918). By about 1550, the former four terms were also being used for their respective children, depending on the gender of these cousins. By the end of the nineteenth century, these older terms were existing alongside the French loans Onkel, Tante and Cousin(e), which later replaced them apart from Vetter for male cousins (itself possibly now redundant, except in dialect). By the end of the nineteenth century also, only Muhme of the older terms was found in both −1 and ego’s level: Oheim was confined to the latter, Vetter and Base to the former. At some stage, these terms lost their zero-equation characteristics and were applied more to respective PG and PGE referents according to absolute sex: thus Oheim was PB, PZH, Muhme PZ, while Tante at this time was PBW. In −1, German has also adopted Neffe, Nichte as terms for GC, these being recorded as far back as Old High German (nevo, nift).32 Similar terms already occurred in Old English (nefa, nift) and Old Frisian (neva, nift). In fact, it is clear from Naroll (1958) that the set Vetter, Oheim, Base and Muome, as well as OHG, MHG neve, nift(el), were variously used of collaterals in all five medial levels of the terminology. Ultimately, while the latter have survived as GC terms, the former have been replaced by Onkel, Tante for +1, except possibly in dialect (Naroll worked in the Austrian Tyrol). Naroll also records descriptive terms for cousins in MHG, now all consolidated under the French loans Cousin and Cousine.

Indo-European Kinship Terminologies in Europe

Anglo-Saxon Lancaster’s list of Anglo-Saxon terms (1958) indicates a system of separate terms for all +1 cognates, with swor or geswiria for all firstcousin specifications (with the possibility of descriptive phrases for them individually), and nefa (masc.) and nefna, nift, genefa (fem.) for all descending lineals and collaterals – that is, GC, CC etc. (also with descriptive phrases as alternatives and geswiria as a synonym for ZD). The modern English cognatic terminology has clearly evolved from the Anglo-Saxon one. The modern English terms for GC, however, represent not a development of the Germanic terms but borrowings of their French originals directly into Middle English. ‘Niece’ may originally also have meant CD, while ‘cousin’ was borrowed from Old French, like ‘aunt’ and ‘uncle’ (see below). Thus English has borrowed cousin, nephew and niece from French, having become the archetypal cognatic terminology with their aid. Old English is, unsurprisingly, closer to Common Germanic, with sweor EF and modrige MZ. According to Barlau (1981), whereas Onkel first appears in German in the late eighteenth century, ante and uncle entered English from French in the thirteenth century – obviously the French influence was much stronger in England for political reasons connected with the Norman Conquest. Affinal terms in Anglo-Saxon given by Lancaster are tacor for HB, ad-um WB, ZH and DH, snoru SW, sweor EF and sweger EM; now, of course, all these terms have been replaced by consanguineal terms plus ‘in-law’. Husband already appears in the form (hus)bonda, recalling an Old Nw. form (see above). German affinal terms are today characterized by the word Schwieger- prefixed to terms for consanguines. An alternative to Schwiegerbruder for brother-in-law is Schwager, originally also EF and DH. Wallis (1918) gives older zeihhur HB, snura SW.

Dutch Dutch (cf. Barnard and Good 1984: 56–57) has also become cognatic, though neef GS and nicht GD also mean CS, CD and ‘cousin’, the only terms in the language uniting -2 and −1 with ego’s level (any cognate or loan of ‘cousin’ appears to be absent here). Dutch has retained oom (cf. German Oheim; also Old English e¯am PB) as both FB and MB rather than borrowing oncle from French. This is interesting



183

184

How Kinship Systems Change

because Dutch has borrowed tante PZ and also apparently taken over a number of loan translations from French; for example, grootvader (PF), grootmoeder (PM) from grand-père, grand-mère (unless these are Germanic: cf. German Großvater, Großmutter), kleinkind (CC; cf. Fr. petit-enfant), as well as a whole series of affinal terms consisting of consanguineals prefixed by schoon-, ‘beautiful’; for example, schoonvader EF (cf. French beau-père – a construction not found with the German cognate schön). One exception is zwager EB, ZH, cognate with German Schwager.

Celtic Irish The sources for this section include Charles-Edwards 1993 and 1970–72, both of which treat the subject historically.33 Early Irish has a number of descriptive terms, such as bráthair athar FB and siur athar FZ (athair = F, cognate with Latin pater; bráthair B and siur Z are also common IE roots). M is máthair, from which comes bráthair máthair MB; an alternative term for the latter, amnair, is ‘attested only three times and [is] of uncertain etymology’ (1993: 34). In +2, senathair is PF, usually FF (there is a separate term for MF, máthair athar), from sen ‘old’, while senmáthair, though marked ‘rare’, is PM. In −2, CS is ave, cognate with Latin avus, and also meaning ‘descendant’. For lower generations still, we have íarmue CCS, from íar ‘after’, and indue CCCS, from ind ‘end’, developing later into úa, ó, respectively. The author specifically says that there are no female equivalents for kin in these descending generations. In −1, ingen, ‘of uncertain etymology’, is D (ibid.). However, a recognizable IE root appears in Gaulish etc. duxtir, as well as, possibly, in Irish der, dar. Similarly, there is a term, nia, for ZS but nothing reported for ZD or BC. In fact, apart from nia, descriptive phrases are used for collateral kin types, indicating that early Irish had a zero-equation pattern in many parts of the terminology. Affinal specifications are dealt with similarly, based on fer H (cognate with Latin vir) and ben ‘woman, wife’; for example, fer a father ZH ms ( fer being the genitive singular of siur, Z). The only other term to note is mac S, with ‘no cognate outside Celtic’ (ibid.), though it is perhaps the most familiar term, as it frequently appears in Scottish and Irish surnames.

Indo-European Kinship Terminologies in Europe

Welsh and Breton Charles-Edward (ibid.) gives a few terms for early Welsh, with less explanation, though many of them are evidently cognates of early Irish terms; for Breton, my source is Edmonson (1957: 432). Of note are Welsh mab S and a female equivalent for D in march, with cognates in Breton, which also shares tad F and mam M with Welsh. Also, Irish bráthair B is seemingly represented by Welsh brawd and Breton breudeur. The Irish particle sen- ‘old’ here becomes hen-. Nai (cf. Irish nia) and ewythr are used in Welsh for the children of collaterals, alongside wyr, gorwyr GC. There are separate terms for PGC (not given for Irish). Here too, therefore, we have largely zeroequation terminologies at this point in the history of Celtic, achieved through the use of a number of descriptive terms based on the core terms for the nuclear family. This is more noticeable for Irish than for Welsh in Charles-Edwards’ sources. The modern Irish terminology, by contrast, has consolidated term usage into a cognatic pattern that is virtually identical to the present-day English terminology, with aunt and uncle terms apparently being borrowed while other specifications use the Irish language’s own resources (information from standard dictionaries).

Conclusion The present chapter has basically been concerned with tracing a development in IE kinship terminologies in Europe from the zeroequation pattern associated with Hindi and other north Indian languages to the cognatic pattern represented by English. It can be seen from the foregoing that it is the Latin or Romance branch of IE that has made this shift most comprehensively, the zero-equation features of Latin having evolved into the cognatic pattern of the daughter languages. A similar development has taken place in the Germanic branch (though not in all its languages), namely in English, German and Dutch (the West Germanic branch), though at least some Scandinavian languages (the North Germanic branch) show signs of a similar development. Similarly, the Slavonic branch has one basically cognatic terminology in Czech, while the other terminologies discussed here (Polish, Russian, Serbo-Croat and Bulgarian) have at least some zero-equation features, especially Serbo-Croat and



185

186

How Kinship Systems Change

Bulgarian, with Russian and Polish occupying a more intermediate position between the two patterns. There are indications of similar processes underway in Greek and more particularly in the Baltic branch, though the terminologies of the latter’s two living languages differ in detail. These processes may not have been completed, but the changes they do indicate are also reflections of the relationships between IE branches more generally: thus Slavonic and Baltic, which both show indications of these changes, are generally closer linguistically to Hindi and other Indo-Aryan languages than are either Germanic or Romance languages at the present day, as were also, however, the respective ancestors of Germanic and Romance, namely Common Germanic and Latin. Examination of Old Church Slavonic, a similar ancestor for modern Slavonic languages, might also reveal significant data about such developments in terminology, but this has yet to be undertaken. The significance of loans from other branches of IE should also be highlighted; for example, the number of terminologies that have borrowed oncle and tante, or neveu and nièce (I give the French forms as the most likely direct or indirect source) as a way of deleting lateral distinctions in the PG and GC fields. This is an alternative to one existing lateral term taking over from the other of the pair, as appears to be happening in Polish, where wuj may be taking over from stryj (i.e. wuj MB + stryj FB >>> wuj ‘uncle’). Among other changes, especially prominent is the shift of a zero-equation affinal pattern in ego’s level to just two gendered terms, whether loans, formerly existing affinal terms or other vocabulary drawn from the same language. This is strikingly evident in Lithuanian, thanks to Buivydiene˙’s careful historical study of that language, but it may also be happening in Latvian, Serbo-Croat, Russian and possibly Scandinavian. The material discussed in this article therefore indicates that terms can be deleted from terminologies (i.e. fall out of use) as well as be added to them through loans or new coinings. Indeed, the reduction of many ego-level affinal terminologies to just two gendered terms entails mass deletions of this sort. One other possibility is the lexical evolution of terms going along with a semantic shift; for example, from Latin avunculus (MB; zero equation) to French oncle (cognatic) and similarly Latin amita (FZ) to French tante. This process also entailed the deletion of terms, namely patruus FB and matertera MZ, as they gradually fell out of use with the development of French from Latin.

Indo-European Kinship Terminologies in Europe

What are the reasons for such changes? A persistent argument refers to the change from forms of collective and extended family organization such as the Yugoslav zadruga to the increasing predominance of the nuclear family in Europe, in which it is less important to distinguish, for example, side of family or other kin individually, especially as descriptive terms or other circumlocutions can always be brought in for reasons of greater precision. However, this is probably a local explanation at best and is in any case likely to be only part of the reason for change. In particular, cognatic terminologies are by no means confined to western Europe, as is indicated by the alternative term ‘Eskimo’ for them in the older literature.34 Indeed, in the Arctic there are definite indications that such terminologies may derive directly or indirectly from the very different symmetric prescriptive type of terminology that expresses repeated cross-cousin marriage (Ives 1998). Perhaps all we can say in general is to remind ourselves that particular terminological patterns reoccur in different parts of the world and in different societies that do not otherwise share much in common. In addition, societies are known to make changes to their ideas and practices concerning kinship, which may take very different forms: the more limited resources of the terminologies – that is, the domain of classification – are at the service of these changes, but they have to be fitted to them according to local circumstances. Whether we view things synchronically or diachronically, therefore, we should not always expect similarities in terminological pattern between different societies to reflect or be reflected in other ethnographic facts. Clarification of this question can only be produced, if at all, through the collection and examination of much more data.

Notes * First published in the Journal of the Anthropological Society of Oxford-online 7(2) (2015): 205–33. Reproduced with permission. 1. We owe the term ‘zero equation’ to N.J. Allen (e.g. 1989). Needham’s term ‘cognatic’ is equivalent to Robert Lowie’s ‘lineal’, but it expresses better the fact that Lowie’s ‘lineal’ terminologies do not mark out descent lines the way his term suggests. The paradigm case of cognatic is the English terminology, in which although there are distinctions between lineal and lateral relatives the latter are not further distinguished into patrilateral and matrilateral, nor into cross and parallel. Needham does use the term ‘lineal’ for certain types of terminology, basically those that do, by contrast, sort cate-



187

188

How Kinship Systems Change

2.

3. 4.

5.

6.

7. 8. 9.

gories into descent lines, and therefore including what Needham calls ‘prescriptive’ terminologies or those expressing regular cross-cousin marriage, which typically do not have lineal equations between adjacent generations, as well as Crow-Omaha ones, which do (Needham disliked the term ‘CrowOmaha’, NB). These uses are very different from Lowie’s ‘lineal’ or Needham’s ‘cognatic’. Reproduced with minor changes as Ch. 5. By these terms, I am referring to the formal typology, not any actual Dravidian or Iroquois terminology. The distinction also corresponds to Trautmann and Barnes’ Type A and Type B crossness (1998), which they prefer as avoiding ethnically specific labels for terminological patterns of wide cross-cultural relevance. Where I am silent about sources, the reader should assume that I have used dictionaries for want of anything more appropriate. Specifically, I am grateful to the following, who are all mother-tongue speakers of the languages that follow their names (listed in order of appearance in the article): Rasa Racˇiu¯naite˙-Paužuoliene˙ for Lithuanian; Ieva Raubisko for Latvian; Ana Ranitovic´ for Serbian; Yulia Savikovskaya for Russian; Ina Zharkevitch for Russian and Bielorussian; Iliyana Angelova for Bulgarian; Johana Musalkova for Czech; and Rosa Krogh for Norwegian (note added 2020: subsequently also Raluca Bianca Roman for Romanian). At certain points in the text below their information is referenced using their initials and ‘p.c.’ for ‘personal communication’. Kryukov (1998) has usefully listed the types of evidence one might have to resort to in order to assess whether terminological changes have actually taken place. I return to these and related issues, more from the point of view of marriage systems, in Parkin 2018, reproduced as Ch. 4. In this chapter, I am more concerned with changes in terminological patterns than with exact transcriptions of the terms themselves in relation to Greek and Slavonic languages that use Cyrillic, where I have generally followed my main sources in this regard. I have included diacritics elsewhere, however, including for Proto-Indo-European forms (in the latter case, h· and r· have been used for h and r with subscript circles, as the proper characters are not in Unicode). Also, while I do list and cite the necessary sources, I do not give an exact reference for every detail of what is inevitably a series of quite dense descriptions of words and their meanings. As already noted, some data come from checking in dictionaries, which are rarely anthropologically aware and usually need supplementary information from native informants; I have not given details of these works. As is conventional in linguistics, an asterix indicates unattested reconstructions or hypothetical forms of words. This work also comments usefully on terms in other IE languages in Europe. Mallory and Adams (2006: 216) give senmote˙ as PM, from PIE *seno-mehate¯´r, suggesting ‘old mother’. Mallory and Adams (2006: 214) give Old Lithuanian stru¯jus FB, clearly reminiscent of Slavonic forms, though it is a regular development of PIE *ph·atro¯us, and therefore cognate with Latin patruus.

Indo-European Kinship Terminologies in Europe

10. In fact -e˙ nas (masc.) and -iene˙ (fem.) can be used to indicate the spouses of

consanguines in general; for example, anu¯kiene˙ CSW. 11. Buivydiene˙ also gives žaláusis as WZ (also žila ausis, R.R.-P.), used in Lithu-

12. 13. 14. 15.

16.

17.

anian-speaking communities in western Belarus and derived from Bielorussian (R.R.-P.). As for the origin of the contemporary terms sváine˙ and sváinis, Buivydiene˙ gives a number of historical terms for affines of ego’s-level in sv-, indicating a Slavonic origin for this group. These changes were already underway by the seventeenth and 1 eighteenth centuries, when sváinis is recorded as a synonym of žéntas ZH. The contemporary terms are linked to a root meaning ‘own’; cf. Polish swoje. Indeed, no term in krust- means anything but ‘step-’ in modern Latvian, although it regularly appears for collateral kin in dictionaries (I.R. p.c.). Again they can be linked to roots meaning ‘own’; for example, savinieks, ‘one who belongs, relative’, possibly also represented in English ‘swain’. On the latter, see the explanation in note 10 for Lithuanian. Buivydiene˙ (1997) speculates that these and cognate terms may derive from dev ‘god’ etc. Baric´’s report (1967: 11) from western Croatia that stric is being extended to cover MB as well as FB is not confirmed by other sources, including A.R., p.c. Information on Serbian in this paragraph from A.R., p.c., unless otherwise stated. According to Hammel (1968: 27–28), snaja were women not senior to ego married into the zadruga or extended family cooperating economically, while jetrva is the term they used for one another (i.e. HBW). As for the terms for WZH, Hammel confirms that these are Turkish loans, namely pašenog or pašanac in Serbian, or badžanak in Bosnia and western Macedonia, the latter probably being borrowed in the fourteenth century or later, pašanac ‘very early’ (1968: 30 n. 26, after Filipovic´). Hammel (ibid., n. 25) also mentions svak as ‘an occasional term’ for HZH, as well as svojak as a possibly literary term for ‘relative’ but also meaning WZH, ZH ws and derived from svoj, ‘own’ (cf. Polish swoje). Here I offer a few remarks on marriage options in traditional Serbia, based on Hammel (1968: 31 ff.). Agnatic relations were avoided far more in marriage than matrilateral relations, which were especially pursued by Muslims, despite approval of FBD marriage, which hardly occurs here. Hammel also reports the practices of bride price, levirate and sororate in the past, the rationale for the levirate being the ‘preservation and accumulation of land’ (1968: 33), as well as privileged though canonically sinful sexual relations with brothers’ wives and sons’ wives on the part of co-resident men within the zadruga. Although sister exchange was disliked, it did occasionally occur, according to Hammel, having formerly been ‘long common among Bosnian Roman Catholics [i.e. Catholic Croats in Bosnia?] and is now found among Serbs as well’, while ‘Coon (1950) notes preferential direct exchange in Albania’ (Hammel ibid.: 35). However, Hammel reported unilateral marriages between a group of brothers and a group of sisters or between two agnatically related male cousins and two agnatically related female cousins (NB: not marriage between individuals who are themselves



189

190

How Kinship Systems Change

18.

19. 20. 21. 22. 23. 24.

25.

26. 27.

28.

cousins). Hammel attributes these details to the fact that ‘peasants think of families and even wider agnatic groups as playing unitary roles in marriage and sexual relationships [and] that there is some sense of corporacy and substitutability in the exercise of the roles . . .’, though also ‘a tendency toward unilaterality and a distaste for direct exchange’ (ibid.). The intensity of intermarriages between groups of siblings is found in other parts of the world (see especially Chs. 2, 3), but it is rarely reported in Europe. Not having access to Lavrovski’s original, nor being able to read Russian, I am relying here on Kryukov’s summary (1998: 295–96). However, Kryukov appears to have switched the meanings of the two pairs of terms stryi, stroi, and wui, ui, giving them respectively as MB and FB, whereas, where they do appear elsewhere (e.g. Polish, Serbo-Croat and vestigially Bulgarian), the meanings are reversed. My suspicion of error here is supported by the information that stryi is derived from PIE *ph·atro¯us, making it cognate with patruus, the Latin term for FB (Mallory and Adams 2006: 214; see also n. 9 above). Also, Buivydiene˙ (1997: 162) gives yu˘ as MB in older Russian. In pointing out this apparent mistake, I am, of course, mindful that Kryukov is himself a Russian speaker. These terms are now apparently redundant, except that bratán still exists as a very informal word for male friend; for example, ‘mate’ (YS, p.c.). Friedrich’s baba for PM is now said to be rather disparaging, dadushka being more acceptable (Y.S., p.c.). Not recognized as current by Y.S., p.c. Latter term not recognized as current by Y.S., p.c. Mallory and Adams (2006: 216) link this term with Indic bandhu, ‘relative’, especially affinal, from PIE *bhendr·ros. According to Just (2000: 109 n. 9), kunyádhos and badzanakis are not very familiar to young middle-class Greeks. In a passage more concerned with Turkish loans in Serbo-Croat, Hammel mentions badzanakis, badzanakia as Modern Greek terms for WZH and HBW respectively, both being of Turkish origin (1968: 30). This term applies to both male and female referents, though it is marked grammatically for gender in taking other words; for example, the definite article. Bru specifically from OHG bru¯Ƿ, via Old French bruz, brut (Ewert 1943: 291). Although located in the Balkans, the remaining major Romance language, Romanian, resembles Italian most closely in having a cognatic terminology, with similar GC/CC equivalence (nepot, nepoata˘) and two single gendered terms for siblings-in-law (cumnat, cumnata˘). SW is nora˘, DH ginere, EF socru, EM soacra˘ etc. However, reflexes of tio/-a etc. are lacking for PG(E) (information drawn from standard dictionaries). (Note added 2020: I am grateful to Raluca Bianca Roman, a native speaker of Romanian, for confirming the above account of the Romanian terminology.) Mallory and Adams (2006: 216) give swiri MZS in Old Swedish, a regular development of PIE *swesr(iy)ós ZS, while the similar looking Old Norse svili WZH is from a separate PIE root, *sweliyon id.

Indo-European Kinship Terminologies in Europe

29. Probably related to Anglo-Saxon mæg, a general word for kinsman given by

30.

31.

32.

32. 33.

Lorraine Lancaster (1958). Szemerényi (1977: 192) gives Old Norse mágr as DH, EF and brother-in-law. Although Merrill treats these as part of the modern terminology, Rich’s informants denied that they were still in use. In fact, they appear to have given way to descriptive terms, making the earlier distinction between cross and parallel cousins less significant. The historical and contemporary situations in Icelandic are actually quite complicated, with extensive debates and disagreements between earlier authors deserving of separate treatment. In addition to the references already cited, see also Pinson (1979). According to Priebsch and Collinson (1948: 123), Nichte derives from Middle High German niftel through a more general change from the consonantal dyad ft to ht dating back to Old High German (ca. late tenth century AD). This section has been written especially for this reissue and was not included in the original version. This was Murdock’s claim (1949: 226–27). However, at least some Inuit terminologies have classificatory equations in +2, while others merge +1 and +2 collaterals, and yet others have zero-equation patterns in +1 (Dole 1972: 140, Fig. 2, and 142–43). In Dole’s words, ‘thus the kin terms found among most Eskimos [i.e. Inuit] differ from the “Eskimo” pattern of Murdock’s classification’ (ibid.: 142). I am grateful to N.J. Allen for reminding me of this point, as well as for other useful comments on an earlier draft of this article. (Note added 2020: this note is an expanded version of the original.)

References Allen, Nicholas J. 1984. ‘Review of Brown 1984’, Journal of the Anthropological Society of Oxford 15(2): 169–72. ———. 1989. ‘The Evolution of Kinship Terminologies’, Lingua 77: 173–85. Andromedas, John. 1957. ‘Greek Kinship Terms in Everyday use’, American Anthropologist 59(6): 1086–88. Ardener, Edwin. 1971. ‘Introductory Essay’, in Edwin Ardener (ed.), Social Anthropology and Language. London: Tavistock Publications, pp. ix–cii. Baric´, Lorraine (née Lancaster). 1967. ‘Levels of Change in Yugoslav Kinship’, in Maurice Freedman (ed.), Social Organization: Essays Presented to Raymond Firth. London: Frank Cass, pp. 1–18. Barlau, Stephen B. 1981. ‘Old Icelandic Kinship Terminology: An Anomaly’, Ethnology 20: 191–202. Barnard, Alan, and Anthony Good. 1984. Research Practices in the Study of Kinship. London: Academic Press. Berlin, Brent, and Paul Kay. 1969. Basic Color Terms: Their Universality and Evolution. Berkeley: University of California Press. Bjerke, Robert. 1969. A Contrastive Study of Old German and Old Norwegian Kinship Terms. Baltimore: Waverly Press, for Indiana University (International Journal of



191

192

How Kinship Systems Change

American Linguistics Memoir 22, published as supplement to International Journal of American Linguistics 35(1): 1–172). Brown, Cecil. 1984. Language and Living Things: Uniformities in Folk Classification and Naming. New Brunswick: Rutgers University Press. Buivydiene˙, Ru¯ta. 1997. Lietuviu˛ kalbos: vedybu˛ giminyste˙s pavadinimai [Affinal Kin Terms in the Lithuanian Language]. Vilnius: Mokslo ir enciklopediju˛ leidybos institutas [Research and Encyclopaedia Publications Institute]. Callier-Boisvert, Colette. 1968. ‘Remarques sur le système de parenté et sur la famille au Portugal’, L’Homme 8(2): 87–103. Campbell, John K. 1964. Honour, Family, and Patronage. New York and Oxford: Oxford University Press. Charles-Edwards, T.M. 1970–72. ‘Some Celtic Kin Terms’, British Bulletin of Celtic Studies 24: 107–12. ———. 1993. Early Irish and Welsh Kinship. Oxford: Clarendon Press. Delbrück, Berthold. 1889. Die indogermanischen Verwandtschaftsnamen: Ein Beitrag zur vergleichenden Altertumskunde. Leipzig: Abhandlungen der Königlichen Sächsischen Gesellschaft der Wissenscheaften 11(v), pp. 380–606. Dole, Gertrude. 1972. ‘Developmental Sequences of Kinship Patterns’, in Priscilla Reining (ed.), Kinship Studies in the Morgan Centennial Year. Washington: The Anthropological Society of Washington, pp. 134–66. Edmonson, Munro S. 1957. ‘Kinship Terms and Kinship Concepts’, American Anthropologist 59(4): 393–433. Entwhistle, William J. 1962. The Spanish Language, 2nd edn. London: Faber & Faber. Ewert, Alfred. 1943. The French language, 2nd edn. London: Faber & Faber. Friedl, Ernestine. 1962. Vasilika: A Village in Modern Greece. New York: Holt, Rinehart and Winston. Friedrich, Paul. 1964. ‘Semantic Structure and Social Structure: An Instance from Russian’, in Ward H. Goodenough (ed.), Explorations in Cultural Anthropology: Essays in Honor of George Peter Murdock. New York: McGraw-Hill, pp. 131–66. ———. 1966. ‘Proto-Indo-European Kinship’, Ethnology 5(1): 1–36. Galton, Herbert. 1957. ‘The Indo-European Kinship Terminology’, Zeitschrift für Ethnologie 82(1): 121–38. Goody, Jack. 1983. The Development of the Family and Marriage in Europe. Cambridge: Cambridge University Press. Hammel, Eugene A. 1968. Alternative Social Structures and Ritual Relations in the Balkans. Englewood Cliff, NJ: Prentice-Hall. Herzfeld, Michael. 1983. ‘Interpreting Kinship Terminology: The Problem of Patriliny in Rural Greece’, Anthropological Quarterly 56(4): 157–66. Hocart, Arthur M. 1928. ‘The Indo-European Kinship System’, Ceylon Journal of Science 1: 179–204. Iszaevich, Abraham. 1981. ‘Corporate Household and Ecocentric [Sic] Kinship Group in Catalonia’, Ethnology 20: 277–90. Ives, John W. 1998. ‘Developmental Processes in the Pre-contact History of Athapaskan, Algonquian, and Numic Kin Systems’, in Maurice Godelier, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 94–139.

Indo-European Kinship Terminologies in Europe

Just, Roger. 2000. A Greek Island Cosmos: Kinship & Community on Meganisi. Oxford: James Currey, and Santa Fe: School of American Research Press. Kryukov, Mikhail V. 1998. ‘The Synchro-Diachronic Method and the Multidimensionality of Kinship Transformations’, in Maurice Godelier, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 294–313. Kullanda, Sergey. 2002. ‘Indo-European “Kinship Terms” Revisited’, Current Anthropology 43(1): 89–111. Lancaster, Lorraine. 1958. ‘Kinship in Anglo-Saxon Society––I’, British Journal of Sociology 9(3): 230–50. Llobera, Josep. 1997. ‘Aspects of Catalan Kinship, Identity, and Nationalism’, Journal of the Anthropological Society of Oxford 28(3): 297–309. Lounsbury, Floyd G. 1964a. ‘The Structural Analysis of Kinship Semantics’, in Horace G. Lunt (ed.), Proceedings of the Ninth International Congress of Linguists. The Hague: Mouton, pp. 1073–93. ———. 1964b. ‘A Formal Account of the Crow- and Omaha-Type Kinship Terminologies’, in Ward H. Goodenough (ed.), Explorations in Cultural Anthropology: Essays in Honor of George Peter Murdock. New York: McGraw-Hill Book Company, pp. 351–93. Mallory, James P., and Douglas Q. Adams. 2006. The Oxford Introduction to ProtoIndo-European and the Proto-Indo-European World. Oxford: Oxford University Press. Mažiulis, Vytautas. 1988 etc. Pru¯su˛ kalbos: etimologijos žodynas [The Prussian Language: Etymological Dictionary]. Vilnius: Mokslas. Merrill, Robert T. 1964. ‘Notes on Icelandic Kinship Terminology’, American Anthropologist 66: 867–72. Miller, M. 1953. ‘Greek Kinship Terminology’, Journal of Hellenic Studies 73: 46–52. Miranda, Pierre. 1974. French Kinship: Structure and Kinship. The Hague and Paris: Mouton. Mitterauer, Michael. 2000. ‘Die terminologie der Verwandtschaft: Zu mittelalterlichen Grundlagen von Wandel und Beharrung im europäischen Vergleich’, Ethnologika Balkanica 4: 11–44. Murdock, George. 1949. Social Structure. New York: Macmillan. Naroll, Raoul. 1958. ‘German Kin Terms’, American Anthropologist 60(4): 750–55. Parkin, Robert. 1995. ‘The Contemporary Evolution of Polish Kinship Terminology’, Sociologus 45(2): 140–52. ———. 1998. ‘Dravidian and Iroquois in South Asia’, in Maurice Godelier, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 252–70. ———. 2012. ‘Kinship as Classification: Towards a Paradigm of Change’, Journal of the Anthropological Society of Oxford Online 4(2): 183–211. ———. 2018. ‘Why Do Societies Abandon Cross-Cousin Marriage?’ in Warren Shapiro (ed.), Focality and Extension in Kinship: Essays in Memory of Harold W. Scheffler. Canberra: Australian National University Press, pp. 263–301. Pinson, A. 1979. ‘Kinship and Economy in Modern Iceland: A Study in Social Continuity’, Ethnology 18: 183–97.



193

194

How Kinship Systems Change

Priebsch, Robert, and W.E. Collinson. 1948. The German language, 3rd edn. London: Faber & Faber. Rich, George W. 1976. ‘Changing Icelandic Kinship’, Ethnology 15(1): 1–19. ———. 1980. ‘Kinship and Friendship in Iceland’, Ethnology 19(4): 475-93. Szemerényi, Oswald. 1977. ‘Studies in the Kinship Terminology of the IndoEuropean Languages, with Special Reference to Indian, Iranian, Greek and Latin’, Acta Iranica (Varia 1977), pp. 1–240. Trautmann, Thomas R., and R.H. Barnes. 1998. ‘“Dravidian”, “Iroquois”, and “CrowOmaha” in North American Perspective’, in Maurice Godelier, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 27–58. Vernier, Bernard. 1991. La genèse sociale des sentiments: aînés et cadets dans l’île grecque de Karpathos. Paris: École des Hautes Études en Sciences Sociales. Wallis, W.D. 1918. ‘Indo-Germanic Relationship Terms as Historical Evidence’, American Anthropologist 20(4): 419–31. Witowski, Stanley. 1971. ‘A Universalist Account of Kinship Semantics’, Ph.D. thesis. University of Iowa. ———. 1972. ‘Guttman Scaling of Semantic Distinctions’, in Priscilla Reining (ed.), Kinship Studies in the Morgan Centennial Year. Washington DC: Anthropological Society of Washington, pp. 167–88.

Part

II 

Crow-Omaha

7 

On the Origin of Crow-Omaha Terminologies

Introduction In this chapter, I focus on the question of where Crow-Omaha terminologies may be situated in evolutionary sequences of kinship terminologies from one ‘type’ or pattern to another. More particularly, I critically examine evidence supporting or refuting hypotheses that such terminologies derive from prescriptive terminologies expressing one of the two main forms of cross-cousin marriage: symmetric or bilateral (Dravidian type), and asymmetric or unilateral (specifically matrilateral – that is, MBD/FZS marriages; Kachin type).1 I shall also review the existing literature on some other speech communities, especially the Ungarinyin, Iatmul and Samo, which have been discussed from the point of view of prescription while actually having plentiful Crow-Omaha equations. Before I embark on a consideration of the evolutionary hypotheses mentioned above, I will emphasize two specific ideas that already exist in the literature, as together they set the framework for the discussions and demonstrations to come. One is Lévi-Strauss’s definition (1966) of Crow-Omaha terminologies as ‘semi-complex structures’, as they are situated between ‘elementary structures’, which have cross-cousin marriage and a terminology that expresses this, and ‘complex structures’ that lack such features. In this view, CrowOmaha terminologies are intermediate because they do not identify

198

How Kinship Systems Change

certain kin categories (normally cross cousins) that they expect as marriage partners in the way that elementary structures do, though they do use kin categories to prohibit certain relatives as marriage partners in connection with, for example, incest rules (in fact, both elementary and complex structures do this as well). The second idea is that, in the evolution of terminologies, and indeed of kinship systems as wholes, it is the elementary structures that have the priority – that is, that come first in any evolutionary sequence – in particular, terminologies and systems of symmetric or bilateral cross-cousin marriage (direct exchange). In other words, it is from this form, or in one influential version of this theory a now non-existent variant of it, that all other terminological types have derived, either directly or indirectly.2 The proof of this, it must be admitted, tends to be logical rather than ethnographic, but I find it no less plausible for all that, and it does give us guidance in deciding which, if any, evolutionary sequences are likely to have preceded which in discussing the history of kinship systems. In accordance with this claim, the assumption in the present case is that prescriptive systems come before Crow-Omaha ones, not the reverse, in both logical and chronological sequences. The hypothesis of a change from some type of prescriptive system to Crow-Omaha has been made more than once before. The favourite idea seems to have been that Crow-Omaha terminologies evolved from asymmetric prescriptive terminologies that express the operation of matrilateral cross-cousin marriage (or asymmetric prescriptive alliance, in this case MBD/FZS marriages). A basic reason for this is that some asymmetric prescriptive terminologies have vertical equations like Crow-Omaha ones. However, as other asymmetric prescriptive terminologies do not have such equations, or else differ in the kin types such equations unite, they are not in any way diagnostic of them. Most of these studies came out decades ago and were quickly criticized, even refuted, and are now virtually forgotten except by kinship aficionados. They include studies by Lane and Lane (1959), Eyde and Postal (1961) and Ackerman (1976), variously dismissed by Needham (1971: 14–16), Barnes (1984: 228–29, 242) and McKinley (1971: 240–42). Due to their age, hypothetical nature, obscurity of argument, own tentative approaches, lack of hard ethnographic evidence and the criticisms already made of them long ago, I do not intend to deal with them here. Despite this quite negative history of debate, however, the basic hypothesis

On the Origin of Crow-Omaha Terminologies

has been revived recently by Trautmann (2012), whose arguments I shall discuss. Other, somewhat fewer studies, conversely indicate that symmetric prescriptive or ‘Dravidian’ terminologies may be the ‘type’ from which Crow-Omaha terminologies have evolved. There is also a third ‘type’ that is relevant here, one that more uncontroversially derives from the Dravidian form, with which it shares many of its features, namely ‘Iroquois’. In a chapter written with Robert Barnes (Trautmann and Barnes 1998), Trautmann suggests that in North America, at least, it was from Iroquois terminologies that Crow-Omaha ones evolved. Since then (2012), Trautmann has also rejected the possibility that Dravidian terminologies themselves might have been the immediate source for Crow-Omaha ones, whether in North America or elsewhere. This raises the possibility of a three-term sequence from Dravidian to Iroquois to Crow-Omaha, which may well be relevant in the Americas though not necessarily elsewhere. These suggestions will also be addressed here.

From Asymmetric Prescriptive to Crow-Omaha One problem in deriving Crow-Omaha terminologies from asymmetric prescriptive ones is a lack of clear geographical proximity between the two types, as the latter mainly occur in Southeast Asia, where there are no examples of the former to the best of my knowledge. Conversely, places where Crow-Omaha terminologies tend to be found, such as the Americas, Africa and Papua New Guinea, lack asymmetric prescription almost entirely. Despite this, as already noted, it has long been recognized that many asymmetric prescriptive terminologies have very similar vertical equations to Crow-Omaha ones, a quite frequent example being MB = MBS = MBSS etc. In considering this hypothesis, one must content oneself with finding similarities rather than exact replication: the two types are acknowledged to be different for other reasons, not least the expectation, even injunction, to repeat marriages between the same groups or lines, generation after generation in the case of asymmetric prescription, while Crow-Omaha terminologies are widely associated with bans on such arrangements. Thus one would expect the two patrilines of MB and WB to be the same with asymmetric prescriptive alliance but separate in the Omaha type, as they would only coincide if MBD marriage



199

200

How Kinship Systems Change

were consistently followed generation after generation; I shall point out below, when possible, where these lines differ and where they coincide. However, other differences in vertical equations might not be explicable so easily and thus give grounds for suspicion that they are coincidental and do not indicate a past transformation between the two types. For example, skewing is one similarity between some but not all asymmetric prescriptive terminologies and Crow-Omaha ones, but the details are often quite dissimilar. In the former case, the usual assumption is that this reflects, at least in part, the superiority of wife-givers over wife-takers that is practically uniform in systems of asymmetric alliance, as such vertical equations in the terminology will then normally equate wife-givers with senior kin in ego’s line, and possibly in other lines as well, wife-takers with junior kin similarly. Lastly here, the hypothesis of a transition from asymmetric prescriptive to Crow-Omaha relies on a sufficient number of relevant vertical equations being transferred over, though what counts as sufficient will inevitably be a matter of debate in itself. I start with Trautmann’s own main example, the famous Kachin (or Jinghpaw) terminology studied by Leach (1961: Ch. 2), which Trautmann sees through, as it were, a Crow-Omaha lens (2012: 42 ff.). Below I list the Kachin terms with the vertical equations they make, also showing how they resemble or differ from the actual manspeaking Omaha terminology presented as Table 9 in Barnes (1984) and occasionally the woman-speaking terminology in Barnes’s Table 10.3 Clearly comparing these terminologies rules out any consideration of geographical proximity and mutual influence, as they are located on opposite sides of the world, respectively in Myanmar and the USA, but Barnes’s data for the Omaha case are exceptionally full and therefore most useful in making these comparisons. I use Leach’s original spellings for the terms in preference to Trautmann’s revisions, as the former are better known. Ji links male relatives in the line of wife-givers’ wife-givers (MMB, WMB etc.) with +2 males in male ego’s and his wife-givers’ lines. This pattern closely resembles Omaha term 1. Shu (ms) links female relatives in the line of ego’s wife-takers’ wife-takers with −2 male and female relatives of ego’s and the line of ZH. Omaha term 14 links CC, ZSC, ZDD ws, as does this Kachin term, but the first set of specifications above (FZHZHZ, FZHZD, ZHZD, ZHZHSD) are not mentioned anywhere in Barnes’ table, let alone equated.

On the Origin of Crow-Omaha Terminologies

Nam (ms) links male and female relatives of the wife-givers’ line in −1 and −2, as well as MBDy, the prescribed spouse for male ego. These specifications are scattered across four different Omaha terms (4, 6, 14, 17). Nam (ws) links BC and BCC, categories spread among Omaha terms 10, 11 and 14. Hkri (ms) links women in the wife-takers’ line, as well as ZS. Two specifications (FFZHZ, FZHZ) have no Omaha equivalents, while the other specifications are spread between Omaha terms 12, 13 and 14. Hkri (ws) links male relatives in the wife-takers’ wife-takers’ line in the medial three levels. Although the specifications FZHZH and ZHZH do not appear directly in the Omaha lists, they do appear in the form of what would be their equivalents in an asymmetric prescriptive terminology, namely HFZH and HZH (Omaha term 1). The remaining specification under this term, ZDH, appears in Omaha as terms 18 (ms) and 19 (ws). Gu links FFZH and FZH – that is, wife-takers – as in Omaha term 15, which, however, also maps out the descent line of WB etc. in Omaha: that is, it is symmetric in equating wife-givers and wife-takers. In both Omaha and Kachin, WB’s line only has descending vertical equations (WB, WBS, WBSS). Moi links the female equivalents FFZ, FZ, both in ego’s line; these are respectively terms 2 and 5 in Omaha. Ní (ws) links women of wife-givers’ wife-givers, closely resembling Omaha term 2. Ning (ws) links: a) women of the wife-givers’ line in ego’s and the −1 and −2 levels, therefore including MBD: this line closely resembles Omaha term 4; b) women of the wife-takers’ wife-takers’ line in the medial three levels, without a parallel in Omaha; and c) women of the wife-takers’ line in the medial three levels, also without an Omaha parallel, apart from the specification ZHZ appearing under Omaha term 14. a) would align partly with b) in a minimal three-line terminology, but the fit is not perfect because different generations are involved in the two cases. The last term, woi, has lateral, not lineal extension, linking +2 women in the lines of ego’s wife-givers and wife-givers’ wife-givers – that is, FM and MM. In Omaha, these are equated under term 2, though this term covers very many other specifications as well. Apart from ji and ní, the resemblance in vertical equations in Kachin and Omaha is not that close, and the equations many other Kachin terms make are either spread over several terms in the Omaha case



201

202

How Kinship Systems Change

or are not present in Omaha at all as recorded by Barnes. As already noted, one would not expect total consistency between the two terminologies, not least because this does not even exist among examples of either ‘type’ on its own viewed cross-culturally. It is obviously a matter of judgement how much resemblance there should be to make the hypothesis of a transformation from asymmetric prescriptive to Omaha reasonably plausible, but my own view is that the level of proof here is unsatisfactory. Below I focus on other examples – some of them much discussed – where this is even less evident. One such case is the Purum, examined by Needham (1962: 76– 77), and the subject of quite bitter controversies surrounding his interpretation of this example in the 1960s and early 1970s.4 Nonetheless, the fact that this is an asymmetric prescriptive terminology is clear enough for our purposes. There are two sets of equations strongly resembling those of Omaha terms 1 and 2 (Barnes 1984): pu, uniting FF with the line of male wife-givers; and pi, uniting FM with the line of female wife-givers’ wife-givers. The term tu unites junior wife-takers with ego’s −2 relatives, in rough approximation to Omaha term 14. Clearly, this can all be accounted for through the status difference between wife-givers and wife-takers in asymmetric prescriptive systems mentioned earlier and requires no further explanation. In Lamet, there are even fewer vertical equations: ta, FF, MF, MB, MBS; ya, FM, MM, MBW; and pe, FZH, ZH (Needham 1960a: 100), and the same explanation suggests itself. The same appears to be the case with two very rare examples of asymmetric prescriptive alliance in societies with matrilineal descent, for which the appropriate comparison is therefore with Crow equations rather than Omaha ones. One of these is the Sirionó of eastern Bolivia, also a subject of controversy surrounding Needham’s interpretation of them (1961: 243–44; 1964), on which I am nonetheless relying here. There is only one diagnostic Crow equation here, under the term ári, FZ, FZD, though it has a host of other specifications also linking MM and women in the +2 and +1 generations of ego’s wife-givers, wife-takers and wife-taker’s wife-takers. A further term with a large number of specifications, akwanindu, includes among them MBS and MBSS – unusually a diagnostic Omaha equation appearing in a terminology with a Crow one. Even less of a Crow pattern is found among the similarly matrilineal Mnong Gar of south-central Vietnam (Condominas 1960: 19–23), where again a Crow equation, F = FZS, exists alongside an Omaha one, MBC = MBSC.

On the Origin of Crow-Omaha Terminologies

In all these cases of asymmetric alliance, the number of CrowOmaha-type equations is low, casting doubt on the hypothesis of a transformation from the former to the latter. Certainly one could argue that the appearance of these fragmentary Crow-Omaha equations in what are otherwise terminologies with at least some of the features of asymmetric prescription indicates the start of a process of change between the two types.5 In no case, however, are the data for these cases as rich as those assembled by Barnes for the Omaha or Leach for the Kachin, this being a serious impediment to a decisive conclusion in respect of these other cases.

From Iroquois to Crow-Omaha in North America: Trautmann and Barnes (1998) The labels ‘Dravidian’ and ‘Iroquois’ were originally regarded as simple synonyms until Floyd Lounsbury (1964) showed that they were different in what is one of the main success stories of the method of formal semantics associated with both himself and his long-term intellectual ally and collaborator Harold Scheffler. The key difference between Iroquois and Dravidian is in their treatment of cousins beyond first-cousin range in terms of the relative-sex contrast between cross and parallel: Dravidian maintains the contrast consistently beyond this range, but Iroquois abandons it in favour of the allocation of kin to categories on an absolute-sex basis. This contrast can especially be seen in relation to second cousins, though it applies to third etc. cousins too, as well as second cousins once removed (i.e. parent’s cousins and cousins’ children in adjacent genealogical levels to ego’s) etc. For ego’s level, especially useful is Table 3-3 in Tjon Sie Fat (1998: 69), which sets out cross and parallel definitions for both second and third cousins. In other words, how cousins beyond firstcousin range are equated with other kin (including, but not only, first cousins) is one way of telling us whether a particular terminology is wholly or predominantly Dravidian or Iroquois in type. The (relative) presence or absence of cognate-affine equations is another significant difference, but that is not relevant here. One consequence of the discovery of these key differences is that the Iroquois pattern is not internally consistent, unlike the Dravidian pattern. This indicates that the former is a development of the latter rather than vice versa.6 Also, in testing the hypothesis of a change



203

204

How Kinship Systems Change

from one of these forms to Crow-Omaha, mentioned above, for Trautmann and Barnes (1998) it is Iroquois rather than Dravidian that is relevant as the source, at least as far as North America is concerned. One way of approaching the question of whether Crow-Omaha is derived from Iroquois rather than Dravidian, at least in North America, is to examine the specifications for second cousins and first cousins once removed in an actual Crow-Omaha kinship terminology and see to what extent they match the other specifications they are frequently equated with in terms of Iroquois crossness as opposed to Dravidian crossness. If a large number of the respective terms show Iroquois crossness for all or most of the specifications they denote, the hypothesis that Crow-Omaha terminologies derived from Iroquois ones will be considerably strengthened. For this purpose, I have again chosen the actual Omaha terminology, on which, thanks to Barnes’s meticulous work (1984), we have a great deal of information and discussion. I mostly use Barnes’s Table 9 of ms kin terms (ibid.: 132–34), though I also refer to his Table 10 of ws terms where appropriate. I start with second-cousin specifications in ego’s genealogical level, though it must be understood that, as the Omaha terminology blurs generational distinctions, quite a number of such specifications are equated with kin in other levels than ego’s. Similar considerations apply to first cousins once removed, to which I turn secondly: though they are genealogically +1 or −1 generation kin, many of these specifications are equated with kin in other levels in the actual Omaha terminology. In fact, inspection of the Omaha terminology reveals no consistent pattern in favour of either Iroquois or Dravidian crossness, especially in respect of the terms that equate second cousins in ego’s level with +1 or −1 level kin. In ego’s level, with regard to the terms that equate siblings and parallel first cousins with same-level second cousins, the latter are uniformly parallel kin in both Iroquois and Dravidian, as one might expect. In +1, regarding terms for P, PG and PGE categories, there are only two cases where the associated second-cousin specifications (MFBSS and MMZSS) align properly in terms of crossness with a PG specification, namely MB (ine’gi), both cousin specifications being both Iroquois and Dravidian cross. The other terms in +1 are completely non-aligned for crossness or only aligned in respect of Iroquois but not Dravidian or vice versa. In −1, in respect of the terms for ZS and ZD (iton’shka, iti’zhun), who are cross kin to male ego, the crossness of the associated second-

On the Origin of Crow-Omaha Terminologies

cousin specifications (FFBDC and FMZDC) is consistent in being cross in both senses, but in another case this is not so, this specification (MFZDC) being Dravidian cross but Iroquois parallel (note that these cases are the reciprocals of the three +1 examples involving MB mentioned above). The remaining −1 terms, for S and D (izhin’ge, izhun’ge), who are parallel kin to both male and female ego, show similar non-alignment with MFZSC, this specification being Iroquois cross but Dravidian parallel. In +2 and −2, there is extensive merging of cross and parallel, making the test irrelevant in these levels. Overall here, slightly more of the alignment in terms of crossness between these second-cousin specifications and core kin is Dravidian in type than Iroquois, though only by two cases. A similar situation occurs when we examine first cousins once removed. Parent’s cousins divide as follows: three specifications are equated with +2 kin in circumstances of an extensive merger of cross and parallel, so cannot contribute anything to this exercise; four specifications are parallel in both Dravidian and Iroquois, and a further four cross in both, aligning correctly with Omaha terms in all cases; and there is a conflict in the remaining four, three aligning correctly in Iroquois but not Dravidian, one vice versa. In the case of cousin’s children, four specifications are equated with +2 kin; four specifications are parallel in both Dravidian and Iroquois and a further four cross in both, aligning correctly with Omaha terms in all cases; and there is a conflict in the remaining four, three aligning correctly in Dravidian but not Iroquois, one vice versa. In other words, there is a rough balance between agreement and disagreement with the Omaha distribution of cross and parallel. Another approach is to take those second-cousin specifications that are involved in equations expressing the preferences for secondary marriage preferences with wife’s kin (WFZ, WZ, WBD, associated with Omaha equations generally and the actual Omaha specifically since Josef Kohler (1975 [1897]); cf. McKinley (1971), Barnes (1984)), and inspect them for similar alignments with other kin under the same term. Except that none of them involve sibling categories in ego’s level; they are confined to +1 and −1 – there is no discernible pattern. One further approach in both +1 and −1 is to consider the equated second-cousin specifications as a number of sibling pairs. In +1, MMZSS and MFBSS match the other specifications of their respective terms in both Dravidian and Iroquois, but MMZSD and MFBZD



205

206

How Kinship Systems Change

do not do so in either. FMBDS and MMBDS match in Dravidian but not Iroquois, FMBDD and MMBDD vice versa. In −1, FFBDC and FMZDC match in both Dravidian and Iroquois, MFZSC match in Dravidian but not Iroquois, MFZDC vice versa. There are also six third-cousin specifications in Barnes’s Tables 9 and 10, two in the former, the list of ms terms, three in the latter, the list of ws terms, and one in both. There is much more consistency of alignment here, only one ws specification showing a conflict between Dravidian and Iroquois. Both cases in the ms list show consistent crossness both ways, while all the others in the ws list are consistently parallel, as is the example in both lists. Overall, however, the results of these tests do little to confirm the hypothesis that Crow-Omaha terminologies are transformations of Iroquois ones. In the case of the actual Omaha terminology, at least as many of the expected alignments are actually based on Dravidian crossness, not Iroquois, and there many cases where individual terms mix the two types of crossness.

Iroquois to Crow-Omaha in North America: Dyen and Aberle (1974) Although Trautmann and Barnes (1998) postulate that Crow-Omaha derived from Iroquois (at least in North America), they rely on the geographical proximity of the two forms among Algonkian speakers in east-central North America and do not show how this happened by tracing actual changes in any terminologies, nor does Trautmann do so in his more recent return to the Crow-Omaha problem and the possible derivation of Crow-Omaha terminologies from other types, especially here asymmetric prescription or MBD/FZS marriage (2012). The only significant fact mentioned in this connection in this same volume (1998) is that the Apachean language subfamily has both an Iroquois terminology (Jicarilla Apache) and two terminologies combining Iroquois and Crow features (Navajo and Western Apache; Godelier et al. 1998: 17, after Donald and Tighe 1987). I have not been able to access Donald and Tighe’s collection, and Godelier et al. (ibid.) give no details of terms or specifications, merely remarking that the Crow features are few, though in conformity with Lounsbury’s Crow III rule.7 However, David Aberle, to whom the Donald and Tighe volume is dedicated, deals with the Navajo case in

On the Origin of Crow-Omaha Terminologies

two other publications to which I do have access. The earlier of these is a chapter on the Navajo in David Schneider’s and Kathleen Gough’s monumental collection of studies of matrilineal kinship (Schneider and Gough 1962). The later publication is Aberle’s study, with Isidore Dyen (Dyen and Aberle 1974), of the kinship terminologies of speakers of Athapaskan languages, of which Apachean forms a branch. The main aim of this latter, very comprehensive work, which evidently took seventeen years from conception to publication (ibid.: xiii), is to establish a proto-terminology for Athapaskan and also to draw conclusions from this exercise concerning aspects of the social organization of the proto-speech community, including mode of descent (which the authors suggest was originally matrilineal, as it still is in many of the daughter speech communities, though bilateral in others), residence (primarily matrilocal) and preferential marriage (sister exchange, and possibly bilateral cross-cousin marriage). As far as both the original and the devolved terminologies are concerned, they suggest a basic evolutionary trend from Iroquois to Hawaiian, and in some cases back again. The authors clearly recognize the distinction between Iroquois and Dravidian discovered by Lounsbury (1964; cf. Dyen and Aberle ibid.: 131–32) and specifically exclude the Pacific and Apachean branches of Athapaskan from being Dravidian in type, not least because not all Athapaskan terminologies have cognate-affine equations, though oddly, where these do exist, there tends to be a preponderance of G = EGE equations over those between cross cousins and EG/GE (ibid.: 69); the former are also characteristic of symmetric prescriptive terminologies, though they are not normally invoked in identifying such terminologies. For our purposes, it is enough to recognize that, with some exceptions, Athapaskan terminologies are not generally to be seen as having Crow-Omaha equations and tend to be Iroquois, not Dravidian in type, a finding that, for the authors, also fits the assumption that the related speech communities are primarily matrilineal.8 In Dyen and Aberle’s study, Aberle, the main author of Chapter 8 on Apachean, gives the equations M = MZ = FZ = FZD and MBS ws = BS ws = S ws for Navajo (Dyen and Aberle 1974: 224). On the same page, he also attributes some Crow equations to the related San Carlos terminology (included within Western Apache, above). However, neither of the above series of equations can be found in the lists of terms at the end of the book (ibid.: 444–49) for the San Carlos dialect, and only the former series, -má M, MZ, FZ, FZD, appears for



207

208

How Kinship Systems Change

Navajo. Even this is suspect at first sight, since the M = MZ part of the equation is taken from Edward Sapir’s notes, while the FZ = FZD part is taken from an earlier publication by Morris Opler (1936). The lists of terms at the end of Dyen and Aberle’s study (1974: 447, term 15) indicate clearly that the root -má is usually just M, MZ in Athapaskan languages. In fact, however, Opler himself derives the evidence for this longer list of specifications from Sapir’s own notes (Opler 1936: 631), the full list of specifications under this term in this version being M, MZ, FZ, FZD, ZD ms and MZDD ms. The reason for this is the influence which the recognition of lineages was beginning to exercise upon the terminological separation of generations. Dr Sapir’s genealogical data indicate that females of both the mother’s lineage and the father’s lineage, in spite of generation differences, can be classified together under the regular Navaho term for mother [-má, RP]. Relatives so classified included father’s sister, father’s sister’s daughter, and, besides mother’s sister, a man’s sister’s daughter and his mother’s sister’s daughter’s daughter. (Opler ibid.)

Thus what is indicated is a change of structure to this area of the Navajo terminology going on in the 1930s for unknown reasons. What may have happened, however, either sequentially or simultaneously, is that the term -má became Hawaiianized by the addition of FZ to M and MZ, together with or succeeded by the addition of FZD to form a Crow equation. The reason for the additions of ZD ms and MZDD ms is more obscure, except to note the very common classificatory equivalence between Z (a sibling) and MZD (a parallel cousin). Aberle’s remarks in his earlier account of the Navajo (1962: Ch. 2) do not agree entirely with the foregoing. Indeed, he remarks: ‘At certain points the unity of the unilineal group is expressed in [the] overriding of generations, but not at those points which would provide [a] Crow terminology’ (1962: 172). There is, in fact, one relevant Crow-type equation, shown on Aberle’s Figure 2.1 and, more clearly, Figure 2.4, between G (i.e. FC), PssGC, FMBC, FZSC and FZDSC. These terms unite the children of the male lineal ascendants and descendants of ego’s father’s matrilineage – that is, those in ego’s and the +2, +1 and −1 genealogical levels (the matrilineages of FMB, F, FZS and FZDS). However, the standard diagnostic equations of FZ = FZD, F/FB = FZS and MBC = ZC do not occur here, thus contradicting both Dyen and Aberle’s later statement regarding the equation M

On the Origin of Crow-Omaha Terminologies

= MZ = FZ = FZD = (above) and Sapir’s and Opler’s data on which Dyen and Aberle base their interpretation of the Navajo terminology. Other data in Dyen and Aberle are clearer regarding possible transitions from Iroquois/Dravidian to Crow-Omaha. Another Apachean language with, this time, an Omaha equation is Chiricahua. This has the term -ɣóyé MB, MZ, ZC, FZD, the last two specifications forming a standard Omaha equation. The closely related and neighbouring Mescalero have the same term for MB, MZ and ZC but not FZD, which is probably covered by Hawaiianized G/PGC terms in this dialect. Otherwise, this set of cognates classifies +1 and −1 specifications and affines in ego’s level, so the specification of FZD in Chiricahua looks as if it has been introduced to this term subsequently. There is no evidence of the reciprocal Omaha equation, MB = MBS etc. (Dyen and Aberle 1974: 223, 449). As far as preferential marriage is concerned among Apacheans, while sister exchange seems to be absent, the San Carlos dialect community has classificatory cross-cousin marriage, as the Navajo might have had in the recent past. Otherwise, the present-day Navajo, San Carlos, Kiowa Apache and possibly Chiricahua have what the authors call ‘sibling set’ marriage (ibid.: 229), which I take to be the practice of intensifying alliances between the same groups of siblings within a generation but not beyond it. Also of interest in this connection are two other branches of Athapaskan, namely Canadian and Pacific, as also identified and discussed copiously by Dyen and Aberle (1974). In the body of their text (ibid.: 224), the authors suggest that the Canadian branch of Athapaskan has three, possibly four languages with Crow-Omaha equations, namely South Tutchone, Kaska and Tahltan, all spoken just inland from the coast of western Canada, and (least certainly) Chipewyan, spoken well inland. In fact, the lists of terms at the end of their book do not indicate Crow-Omaha equations for Chipewyan and do not cover South Tutchone at all, but relevant information is available on Kaska and Tahltan, and in the body of their text the authors also indicate that the South Tutchone case has similar terms to Tahltan (ibid.: 268, 274). Thus in Tahltan we find Crow equations in the following terms: -sta FB, FZS (but F, -te˛, is separate), -tsiɂa MBD, BD, and also -siiya MBS, BS, though there is no indication of a term equating FZ and FZD. Since in other Athapaskan -sta is normally a +1 term (with a core meaning of F) and -tsiɂa and -siiya are both −1 terms, it is evi-



209

210

How Kinship Systems Change

dent that the cousin terms have been added subsequently in this language. This is the most comprehensive set of Crow-Omaha equations the authors record anywhere in Proto-Athapaskan. In the case of Kaska, two dialects are involved, one with a Crow equation, the other with an Omaha one. The Upper Liard dialect has -tsuu’ PM, FZ, FZD ms, EM, which incorporates the standard Crow equation FZ = FZD, but the male equivalent series is restricted to PF, FZS ms and EF, the terms for F and FB being separate. There is no indication of any terms for the reciprocal equation(s) MBC = C/ BC. The situation in the Ketchika River dialect is more complicated and involves Omaha equations, not Crow ones. Thus -uze is MB, MBS, a term usually restricted to MB and its −1 reciprocals in other Athapaskan, indicating that MBS is the introduced specification here. Finding the standard reciprocal Omaha equation FZC = ZC in the present case involves some lateral thinking. In this dialect, FZD is -tsiie, a term usually applied in other Athapaskan to −1 specifications such as D and BD ms (it is cognate with Tahltan -tsiɂa above), but not normally ZD, which in any case is absent from this term in this dialect. However, ZD does appear with D and BD in two closely related languages that are also geographical neighbours of Kaska, namely Slave -yä-dze or -ya˘-dze, and Sekani -tchewɂ (both further cognates of the Ketchika River dialect and Tahltan forms). Furthermore, in the Ketchika River dialect, the cross-cousin specifications MBD and FZS/ FZD are all equated with siblings and parallel cousins under terms that, in other Athapaskan, normally cover the latter alone, indicating Hawaiianization in this dialect. Again, therefore, the cross-cousin specifications seem to have been introduced, even though in these cases they are linked with terms in ego’s level, not −1, as in Omaha. Also, as in Omaha, the matrilateral and patrilateral cross cousins are categorized separately. Despite the patrilineal associations of Omaha equations, the authors state that no Canadian Athapaskans are known to be patrilineal (ibid.: 290), and indeed this form of descent is only found in the Pacific branch (ibid.: 353). Turning now to Pacific, three languages have Crow equations, as seen in the Kato, Wailaki and Lassik term at eZ, PssGDe, PosGDe ms (thus including FZD), FZ, MBW (Lassik also has WeBW for this term), though the reciprocal equations are nowhere evident (i.e. MBC = C/BC; cf. ibid.: 163). This cognate set is mostly associated with specifications in ego’s level, both siblings and cousins, so in this case it appears that it is the +1 specifications that have been introduced.

On the Origin of Crow-Omaha Terminologies

Dyen and Aberle spend a lot of time in the later portions of their book discussing, and mostly dismissing, the possibility of nonAthapaskan influences on the matrilineal and matrilocal aspects of Athapaskan social organization, but they grant a possible exception in the case of Kaska (ibid.: 367–68) and also speculate that the coastal and matrilineal Tlingit – whose language is coordinate with Athapaskan within the NaDene super-language family, rather than being part of Athapaskan (ibid.: 382, 388) – may have influenced the Crow features in Kaska and Tahltan (ibid.: 235). They also make some summary remarks about marriage prohibitions and preferences. It is evident that in none of the cases discussed above is crosscousin marriage prescribed, but some Kaska, including the Upper Liard group, have a preference for the matrilateral form, apparently without banning the patrilateral form, while the Southern Tutchone and Chipewyan have a preference for the latter, and the Chipewyan prohibit the former (ibid.: 282, Table 9.7, 284). To turn to the possible lessons of Athapaskan, one finding is that where Crow-Omaha equations do occur they nowhere form a consistent set: reciprocals are generally missing, except partially in Tahltan. Secondly, in most cases it is the cross-cousin specifications that appear to represent innovation, apparently having been transferred to terms whose cognates in other Athapaskan languages lack crosscousin specifications and indeed usually cover the adjacent generations to ego’s. This raises the possibility that other terms originally covered cross cousins as part of prescriptive equations with affinal specifications and that, on the breaching of these equations, the cross cousins ‘migrated’ to other levels in Crow-Omaha fashion, leaving the original terms with their affinal specifications alone. However, a perusal of Dyen and Aberle’s lists does not reveal any evidence of this, and in any case their reconstruction of Athapaskan terminological history provides very few indications of prescription. Otherwise, there are some fairly strong indications of Crow equations in the Navajo terminology, though the accounts are partially contradictory. Although it goes against the drift of Dyen and Aberle’s arguments denying outside influences on Athapaskan, the Navajo case may reflect the influence of the non-Athapaskan matrilineal peoples with which the Navajo have long been in contact, like the Hopi. One other consideration is that Dyen and Aberle often specify equations between siblings and siblings’ children as Crow-Omaha, in accordance with Lounsbury’s Crow II and IV variants (ZC ws = yG ws)



211

212

How Kinship Systems Change

and Omaha II and IV variants (BC ws = yG ws). These equations are not the standard or minimum diagnostics of such terminologies – they do not occur in actual Omaha, for instance (Barnes 1984) – and indeed Crow-Omaha terminologies normally equate siblings with parallel cousins, not kin in adjacent generations. Moreover, only Dyen and Aberle’s terms 34 (Proto-Athapaskan or PA *-k’elɂ BS ws = yB ws), 38 (PA *-deyeˇ-tce BD ws = yZ) and 59 (PA *s-titł yZ = BD ms) satisfy the criteria, but not, for example, their term 37 (PA *-deyeˇ BD ws = eZ ws). The equations under term 34 are found in some Canadian dialects (Sekani, Carrier, Sarsi) and Pacific dialects (Kato, Wailaki, Lassik, Sinkyone); those under term 38 in Carrier, Sarsi and Kato; and that under term 59 in Carrier, but none of them appear in those dialects and languages I discuss above. Nonetheless, Kaska (Ketchika River) -tca is CC = ZC, another possible Omaha equation (cf. Lounsbury III and IV, ZC ws = DC ws). It is notable that all these equations are Omaha, not Crow. Finally, there is one possible Crow equation in the Canadian dialect of Nabesna, viz. -tcai (same root as the foregoing) CC ws = BC ws (cf. Lounsbury III and IV, BC ws = SC ws). Equations between younger siblings and relatives of the −1 generation are not that uncommon worldwide and do not necessarily co-occur with diagnostic Crow-Omaha ones – that is, the minimum one needs to define a terminology as such. Conversely, Crow-Omaha terminologies can certainly occur without such equations as the information on the actual Omaha terminology assembled by Barnes (1984) also makes clear. Nonetheless, there is a clear impression, similar to the first set of equations, that where such equations occur in Athapaskan, the −1 specifications have been inserted into ego’s level under terms that do not generally appear in other Athapaskan. Neither set of equations has a predominant position in Dyen and Aberle’s lists of terms, indicating that they do not go back to Proto-Athapaskan but have a more recent origin – even perhaps in very modern times well after colonial-era contact, if Opler’s interpretation of Sapir’s note on Navajo is correct (see above). And given that Dyen and Aberle do not detect Dravidian or Type A crossness in Athapaskan but only attribute Iroquois or Type B crossness to it, including the proto-language, coupled with the evidence we do have of Crow-Omaha equations in the language family and their recency, Trautmann and Barnes’s hypothesis for Algonquian, namely that Crow-Omaha terminologies derive from Iroquois rather than

On the Origin of Crow-Omaha Terminologies

direct from Dravidian, would seem to apply to Athapaskan as well. Moreover, we do not have to rely on geographical proximity as the two authors did in the Algonquian case: for Athapaskan, we have sufficient linguistic evidence thanks to Dyen and Aberle’s careful reconstructions.

Symmetric Prescriptive or Dravidian Terminologies and Crow-Omaha The combination of Crow and Iroquois/Dravidian features within a single terminology – similarly Omaha with Iroquois/Dravidian – is not that unusual, as the actual Omaha terminology makes clear (Barnes 1984), and caution is needed in seeing a developmental process here. The ‘survival’ of inter-cognate equations of the sort that one regularly finds in prescriptive terminologies in the actual, nonprescriptive Omaha terminology can be explained in other ways. Thus, the actual Omaha terminology has the equations F = FB = MZH, M = MZ = FBW and FZ = MBW, though EM has a separate term, and the equation MB = FZH does not occur here. The relevant terms also have other specifications not usually found in prescriptive terminologies, though they are diagnostic of Crow-Omaha ones. While this coincidence of patterns may indicate change between them, it could also indicate a more stable situation reflective of the practice of two groups of (classificatory) siblings intermarrying intensely within the same generation, in which case the above +1 equations would apply as well. However, as the repetition of marriages to specific other groups in this generation are ruled out in later generation(s), they do not have the further affinal equations they would have in a prescriptive terminology. Nonetheless, what one might call ‘symmetric intercognate equations’9 like the above in Crow-Omaha terminologies should be recorded and assessed, as potentially they may be evidence of change in other cases. One example clearly indicating change from symmetric prescriptive or Dravidian to Omaha-type terminology is Allen’s study (1976) of the Sherpa terminology in Nepal. Allen compares two Sherpa dialects, one of which, Tsumje, is broadly symmetric prescriptive in type, suggesting bilateral cross-cousin marriage in the past, though this is no longer practised by the Sherpa. However, there are also Omaha equations in the related Khambu dialect, where the prescrip-



213

214

How Kinship Systems Change

tive equations have lost their characteristic affinal specifications. This suggests that in ego’s level original terms equating cross cousins with siblings-in-law have retained the latter specifications but discarded the former. As a result, MBC and FZC have migrated to the terms for MB/MZ and ZD respectively (+1/−1), producing Omaha equations. The relevant terms are ajang MB, MBS, MBSS but also (prescriptively) WyB; tsabyuk FZS, ZS ms; and tsabyung FZD, ZD ms.10 Another possible case is the Siane of the Bismarck Archipelago in Papua New Guinea, studied back in the 1950s by Salisbury. His chart of the Siane kinship terminology (1962: 19) has two quite different classifications for cross cousins. Thus nofonefo is PosGC, also PosGCC, uniting two generations, though not in Crow-Omaha fashion, as well as equating MBC with FZC in a symmetric inter-cognate equation. However, an alternative pair of terms distinguishes MBC from FZC because they both take part in Omaha-type equations, of which their distinction is also characteristic. Thus momonefo is MB, MBS, MBSS, komonefo FZC, FZSC, ZC. Moreover, nitofa is FZH, WB, WBC, WBW, so that there is a patrilineal equation uniting WB and WBS, who, with either bilateral or asymmetric prescriptive alliance, would also be MBS and MBSS. However, there is no marriage prescription, only a preference for classificatory FZD marriage, which appears to be very generally followed, but with a ban on the genealogical FZD (ibid.: 103; Salisbury 1956: 646). Salisbury’s further statement (1956: 646) that girls are encouraged to marry into their mother’s clan but not her lineage supports this, as it indicates marriage to MBS as the female preference: mother’s clan is also MB’s clan, and the latter’s son is the reciprocal of FZD, while the ban on the mother’s lineage indicates that those getting married should not be too closely related (specifying the clan suggests classificatory MBS). What might be behind the doubling of classifications for cross cousins among the Siane? On the usual assumption that symmetric prescriptive alliance is prior in historical terms, the supposition would be that the classification splitting the cross cousins from one another to produce vertical Omaha-type equations is the later one, while the fact that there are two alternatives, and therefore redundancy in relation to terms for cross cousins, indicates that terminological change may be going on right now. Alternatively, however, this is an example of Kronenfeld’s overlay theory (1980a, 1980b; see also below). This is supported by the preference for classificatory FZD/MBS marriage. This is clearly not prescriptive, especially as the

On the Origin of Crow-Omaha Terminologies

minimal genealogical referents are excluded, yet equally clearly it is a practice requiring an equivalent exchange of spouses over time between affinal exchange groups, as would occur with symmetric prescriptive alliance. This is a logical counterpart to what appears to be happening to the terminology. The Dani of Papua New Guinea also have Omaha equations in their terminology: akoja M, MZ, MBD (also FBW); ami MB, MBS; ejak C, ZC, FZC (all ws), ZHZ; akopak WF, WB, ZH, DH. Although the terminology is generally non-prescriptive, the ethnographer Haider (1974) links the terms for FZ and EM (‘mother-in-law of opposite moiety’) etymologically in prescriptive fashion, but it is possible that ‘EM’ should really be read as ‘HM’: oan (FZ) reappears in akhoan, in which akh- is an affinal marker, and, by analogy with HF as akhami, akhoan may reasonably be interpreted as HM (ibid.: 233–35). In that case, the equation would fit the pattern of asymmetric prescriptive alliance. Another patrilineal example of interest here are the Ungarinyin of Australia, initially studied by Radcliffe-Brown and Elkin (references in Livingstone 1959) but also used by Needham (1960b) as part of his attempt to prove that systems of prescriptive alliance based on repeated FZD marriage (patrilateral cross-cousin marriage) cannot exist. In his discussion and analysis of this case, he admits (1960b: 289 n. 38) the existence of Omaha equations in the Ungarinyin terminology but does not develop the point, already questioning their significance at this still relatively early stage in his career (cf. his more categorical statements in Needham 1971: 14–16). Certainly the Ungarinyin terminology has such equations in abundance. Thus the term ngadji unites MFZ, M, MBD and SW, the first three of which can be seen as women of ego’s and the +1 and +2 generations in the line of ego’s mother, to which SW would be added as the −1 representative if the terminology were prescriptive. The term kandingi gives us the classic Omaha series of MF, MB, MBS, MBSS and MBSSS, while waiingi denotes WFF, WF, WB and WBS, together with FMB, who would be equated with WFF were the terminology symmetric prescriptive or asymmetric prescriptive (that is, with MBD/FZS marriage, not FZD/MBS marriage). The fact that the MF and WFF series are separate in Livingstone’s data indicates that this is fundamentally not a prescriptive terminology of any sort, as does the lack of the more usual cognate–affine equations more generally. However, a much more recent source on this group, McConvell (2012), adds to



215

216

How Kinship Systems Change

waiingi the specifications FMB, MMBS, FMBSS, FMBSSS, which would be equated respectively with the corresponding members of the WFF series under this term with asymmetric alliance. This suggests that while a lot of these specifications are genealogically remote from ego, and most of them are not usually treated as diagnostic of prescription, the fact that they exist at all is intriguing in this Native Australian context, with its otherwise copious examples of prescriptive systems. Nonetheless, McConvell himself links them with the preference for FMBSD marriage here, in which ego formally repeats the marriage of his FF (he does not give the female equivalents), not his own father, as with prescriptive alliance. Other Native Australian groups, the Marra, Nungyuburu and Anindilyakwa, also have the MB, MBS, MBSS series, extending to MBSSS in the case of the Nungyuburu, as well as a series linking MMBS, MMBSS, MMBSSS. The Marra also have a term for FZS, FZSS and FZSSS, apparently not including ZS. The Anindilyakwa have a term for FZS and FZSS, though also covering MB as an alternative to the equation mentioned above. Returning to the Ungarinyin, yet other Omaha equations occur with malengi, inter alia FZC, ZC; and possibly in part wuningi FFZH, FZH, ZH (cf. Omaha ms term 15, Barnes 1984: 134); however, DH is also part of the latter series in the Ungarinyin case but a separate term 18 in the Omaha case. Lastly, the long series of specifications under wolmingi includes W, WBD and WFZ, a reprise of ngadji above from the perspective of the previous generation and a set of equations also found in the Omaha case, as we saw above. Needham, following Livingstone, points out that the Ungarinyin have to avoid the patrilines of FF, MF and MMB in marriage but can marry into the line of FMB in the shape of FMBSD (cf. McConvell, above), to which therefore ego’s wife, WFZ and WBD belong, ‘and these by category are permitted [but evidently not prescribed; RP] spouses’ as wolmingi (1960b: 289; also McConvell 2012: 246–47). This is also true in the Omaha case apart from FMBSD, who has a separate term in Omaha and is not an obvious spouse but who in Ungarinyin is also covered by wolmingi, a specification that would be equated with ‘wife’ were the terminology prescriptive. As just noted, marriage to this relative would represent male ego repeating the marriage not of his father but his FF, though none of these authors points this out. Thus, the Ungarinyin case has clear indications of a similar Omaha-type terminology and marriage preferences to North American groups halfway round the world like the Omaha and Osage.

On the Origin of Crow-Omaha Terminologies

Ackerman (1976) appears to have been the first to have noted the similarity between the actual Omaha case and that of the Ngarinyin, as he and later authors call this Australian ethnic group. In discussing the latter, he focuses particularly on FMBSD as a marriage possibility or preference, but he is also keen to see it as an evolved system of matrilateral cross-cousin marriage, in which case FMBSD occupies the same position as MBD, just as she is equated with FZD in the scheme of patrilateral cross-cousin marriage (see above). In fact, he constructs two minimal diagrams of these two forms of unilateral prescriptive alliance to show this, then another two, which, by adding a third line to each, ensures that these equations do not occur (ego’s first cross cousin and FMBSD being in different lines). In other words, his second pair of diagrams represent a hypothetical nonprescriptive development from a prescriptive situation, a development that is nearer the Omaha case and that of the Ngarinyin, too, in that ego no longer repeats the marriage of his father. In fact, the link he posits between first cross cousins and FMBSD only appears because of the nature of the diagram he chooses to show it, in which it is unavoidable, while the situation lacking that link is nearer the actual ethnographic situation, thus not providing the evidence he claims exists. As Barnes makes clear (1984: 242), Ackerman overinterprets the possibility of marriage into the clans of FM, MM and FMM among the Omaha into an injunction (or ‘prescription’) to do so, and Barnes also points out that Ackerman’s depiction of repeated marriage between the same lines in one of his diagrams is in conflict with the Omaha prohibition on marriage into mother’s clan. Moreover, in Barnes’s tables and charts MBD and FMBSD appear in completely different lines and have separate terms. Whatever proof there may be of the derivation of Omaha systems from asymmetric prescriptive alliance, or vice versa, Ackerman’s thesis certainly is not it.11 FMBSD marriage is also one of the ‘rules’ set out for the Iatmul, another Sepik River people, by Francis Korn (1973: Ch. 5) on the basis of Bateson’s original ethnography (1932a, 1932b, 1936). Korn’s reanalysis of Bateson’s material is clearly influenced by Needham, and in particular it studiously ignores the vertical (‘Omaha’) equations in the Iatmul terminology, which admittedly are not all continuous down the patriline because of a tendency to distinguish sets of alternating generations. Her interpretation of the Iatmul system of affinal alliance is that it employs a terminology in five patrilines in which each patriline takes spouses from each of two other patrilines



217

218

How Kinship Systems Change

in alternate generations. Although the latter feature is certainly found elsewhere (e.g. the Aranda/Arunta), a terminology composed of five lines out of structural necessity is not,12 nor is the system as a whole reported from anywhere else on earth that I know of. When one also adds to this the fact that other marriage options occur here, namely sister exchange (ZHZ marriage), FZD marriage and oblique marriage between FMBD and FZSS, as do at least some Omaha equations linking successive generations (e.g. FM, FMBD, FMBSD; FMB, FMBS, FMBSS; M, MZ, MBD, MBSD, MBSSD; FZH, HZH, DH), and not simply alternate ones, there is ample reason for questioning whether this is a prescriptive system at all. Although Korn dutifully lists all these marriage options, in making her actual analysis of what she feels is the Iatmul affinal alliance system she basically ignores all those but FMBSD. As already noted, FMBSD is marriageable in many societies, but either as a preference not reflected in any prescriptive features in the terminology – many of which nonetheless have vertical, usually ‘Omaha’ equations – or as a structural equivalent of other second cross cousins in four-line prescriptive systems like the Arunta (formerly Aranda).13 One other case we should consider is that of the Samo, a patrilineal people of Burkina Faso with an Omaha-type terminology who have become famous first because of Françoise Héritier’s detailed study of them (especially 1981) and secondly because of her claim – contested especially by Barnes (1984) – that they confirm LéviStrauss’s pronouncements (1966) concerning semi-complex systems (i.e. Crow-Omaha ones). In fact, a comparison reveals that the Samo terminology (Héritier 1981: Appendix 2), though not identical with the actual Omaha one, has many vertical equations in common with it, including the standard diagnostic ones of MB = MBS etc., and ZC = FZC etc., though MBD is equated with MZy rather than with M. The marriage rules and preferences, however, are quite different (ibid.: 86; cf. Barnes 1984: 198). First, and expectedly, male ego cannot marry into the lineages of any of his grandparents. Secondly, he cannot marry into a lineage into which either his father or his (i.e. ego’s) brother has already done so. Together with the first rule, the first part of the second rule excludes male ego from repeating the marriage of his father, while the second part of the second rule means that there is no replication or intensification of alliances within the same generation as there is for the Omaha. The third rule bans male ego from taking a wife polygynously from the lineages of his wife, WM,

On the Origin of Crow-Omaha Terminologies

WFM or WMM – that is, from the lineages of his own wife-givers, their wife-givers or the latter’s wife-givers. This resembles the chains of banned wife-giving lines in certain Indonesian societies I have discussed elsewhere (Parkin 2018, reproduced as Ch. 9), none of which have Omaha equations. But also, the ban on the wife’s lineage ensures that WZ, WFZ and WBD are also banned to male ego, whereas second marriages to one of these patrilineally linked women appears to be the key Omaha marriage preference, as Kohler showed and as we have already seen. Accordingly, the Omaha equations between these specifications, as well as between M, MBD and MFZ, do not obtain here. However, the ban on repeat marriages between the same two lines in the same generation (Héritier’s second rule) seems at variance with the terminology. Thus yεkwarε, F, FFBS, is also FMZS; nεkwarε, MZy, MBD, MBSD, MBSSD, is also MMZD, MMZSD; nere, MB, MBS etc., is also MMZS, MMZSS; and luli, ZC, FZC, is also FMZDC. This is the case when the Samo specifications are plotted against equivalent Omaha data, where very similar equivalences are found; if they are plotted against a template of asymmetric prescriptive alliance, the alignment of specifications with descent lines improves considerably. However, the Samo terminology is not prescriptive, and there is no prescribed spouse for either male or female ego. There are therefore some puzzles regarding the links between the Samo terminology and Samo marriage rules and preferences, making it difficult to read the latter off the former as can be done, in part, in the Omaha case. Barnes therefore seems justified in suggesting that ‘the Samo arrangement should be called a Samo, not an Omaha, system of alliance’ (1984: 198). Nonetheless, although the actual Samo and Omaha terminologies are not identical, they both belong to the formal Omaha ‘type’. Thus Samo nere is MB, MBS, MBSS, MBSSS, while nekware is MZy, MBD, MBSD, MBSSD, both with certain classificatory equivalents (M, MZe have separate individualizing terms). Again, however, there are additional specifications in different patrilines, namely MMBDC and MMFZC. Luli gives the standard equation ZC = FZC, as well as other kin types, mostly in different lines, as is to be expected here (Héritier 1981: 178). There are some indications of changes from a symmetric to an Omaha-type terminology in the case of the Mapuche, studied in Chile by Louis Faron (especially 1956). Although his data are confused, incomplete and contradictory, he presents what he calls a Dakota-type terminology (i.e. a patrilineal version of Iroquois?) for the start of



219

220

How Kinship Systems Change

the twentieth century (based on published sources) and himself recorded Omaha-type equations later in that century from the same people (especially M = MBD and MBW = MM).14 Also relevant here, finally, is a comparison of two closely related societies in central Brazil, the Shavante and Sherente, by David MayburyLewis (1979a), the sole survivors of the Akwe˜ branch of the central Gê-speaking peoples. Once probably a single ethnic group in the Rio Tocantins region of Goias state, it is thought that the Shavante split off from the Sherente at the start of the nineteenth century and migrated south-west to the region of the Rio Das Mortes in or near Mato Grosso state. Both groups ban marriage into male ego’s mother’s clan, therefore regarding MBD marriage as wrong and distinguishing MB from WF in their terminologies, and they both disallow sister exchange, but they both approve of marriages between groups of classificatory or real brothers and sisters. Both also have separate affinal terminologies, but while the Shavante terminology is Omaha in kind, the Sherente one is Dakota. At this point, I reintroduce the hypothesis of Trautmann and Barnes (1998) that in part of North America Crow-Omaha systems may have derived from Iroquois, rather than vice versa. The Dakota label is generally that given to the ‘Iroquois’ pattern in societies with patrilineal descent (the actual Iroquois of upper New York state being matrilineal), to which it can be considered an equivalent terminological pattern as there is no intrinsic difference between the two. The close relationship between the Sherente and Shavante on general cultural and linguistic grounds raises the interesting possibility that the Shavante derived their Omaha-type terminology from an earlier Dakota/Iroquois type that was perhaps common to both when they were living, virtually indistinguishably, in the same area – a development that would have had about 150 years to take effect by the time Maybury-Lewis studied them. The two consanguineal terminologies are very similar both lexically and structurally, the only structural difference being the extension of the Sherente equivalence of cross kin in ego’s and the −1 generations to the +1 generation as well in Shavante (see the two matrix diagrams in Maybury-Lewis 1979a: 225 ff.).15 The main conclusion from the foregoing accounts must therefore be that if Crow-Omaha terminologies did develop historically from prescriptive terminologies and/or the closely related Iroquois etc. variants, then the source is just as likely to have been a symmetric as an asymmetric terminology expressing MBD/FZS marriage. That is not to rule out the latter hypothesis, only to draw due attention

On the Origin of Crow-Omaha Terminologies

to both possibilities. Probably the local context should be taken into account in each case, as Trautmann and Barnes did in suggesting an Iroquois origin for the Crow-Omaha equations in Algonquian in North America, where both patterns coexist. Conversely, the absence of terminological types coexisting in this way may indicate a lack of similar change. Thus in Southeast Asia there are plentiful examples of asymmetric prescriptive terminologies but no Crow-Omaha ones, despite isolated examples of Crow-Omaha equations in one or two cases. Moreover, with respect to actual marriage practices, the evidence is that many societies in Southeast Asia have developed systems of what I call ‘evolved asymmetric alliance’ that do for affinal alliance something very similar to Crow-Omaha marriage systems but that lack the distinctive Crow-Omaha equations. In other words, in terms of affinal alliance, they have developed alternatives to CrowOmaha (see further, Ch. 9). There is clearly still much to be learned about the genesis of Crow-Omaha systems.

Notes 1. For an account of how I am defining Crow-Omaha terminologies, see intro-

duction to Chapter 8. 2. The originator of this latter version was N.J. Allen (1986, 1989a, 1989b;

most recent account Allen 2008), who died in 2020, while this collection was in preparation. He gave the name ‘tetradic society’ to the hypothesized non-existent form, as it bifurcates society both vertically, into moieties, and horizontally, into sets of alternating generations, producing four (‘tetra’) social groups. The two moieties govern marriage, the two sets of alternating generations the inheritance of one’s social identity. Although not proved to exist, tetradic society represents an extrapolation back from known features of kinship systems in the ethnographic record, including these two sorts of division. For Allen, the closest ethnographically verifiable kinship system to it was that of the Kariera in Australia, although at the present time it is doubtful that any speakers of the language still exist or that any Native Australians would unambiguously identify themselves as Kariera. 3. In the discussion below, the Omaha terms are identified by their numbers in Table 9, unless otherwise stated. 4. I am using Needham’s later and more economical account of this case (1962), rather than the earlier account (1958) that generated all the controversy. That controversy is summed up from Needham’s point of view in the introduction to Needham 1971. See also Ackerman (1976) for an anti-Needham position. 5. In an article published some time ago (1988: 59–62), I myself was rather more positive about the possibility of the Mnong Gar terminology evolving from asymmetric prescriptive to Crow-Omaha.



221

222

How Kinship Systems Change

6. I am talking about the ideal, formal models here, not actually terminolo-

7.

8.

9. 10. 11.

12.

13.

14. 15.

gies, which in both cases may differ in detail from these models. In order to avoid the use of ethnic labels for terminological patterns of widespread distribution, Trautmann and Barnes write of ‘Type A’ and ‘Type B’ crossness for Dravidian and Iroquois respectively (1998). While acknowledging their objection, I retain the ethnic labels here as they are better known, though on occasion I use Needham’s ‘symmetric prescriptive’ as a substitute for ‘Dravidian’. They actually seem to be restricted to Dyen and Aberle’s term 2 (1974: 445; this work is treated in detail below), where in many Athapaskan languages PM = FZ and/or EM. This also suggests the possibility that a Crow III equation, FM = FZ, has changed from one in which PM = EM, FZ potentially being EM through prescriptive alliance, or vice versa. What Dyen and Aberle say about the possibility of Crow-Omaha equations in Athapaskan in their text is not always borne out by the lists of terms at the end of their book, where they frequently cannot be found. Nonetheless, in general terms their findings are supported by later work, especially Ives 1998. To distinguish them from prescriptive – that is, cognate-affine, equations. Allen himself returned to this example subsequently; see Allen (2012: 55–57). The Ngarinyin have also been invoked by McConvell (2012, following Rumsey 1981) as an example of ‘overlay’ theory, following Kronenfeld’s work on the Fanti (especially 1980a, 1980b), according to which CrowOmaha terminologies share semantic space with other terminological forms in many of those societies that have them – that is, the different forms are subject to different contextual uses. This appears to have been anticipated by Mickey, writing on the Acoma, as an alternative explanation to change from a Crow terminology with alternating generation equations to a ‘bilateral’ or generational one (1956: 249); and also by Henderson, writing on the Onitsha Ibo (1967: 18, 49). A general argument here appears to be that, despite appearances, these are not examples of the familiar distinction between address and reference terminologies. These possibilities need to be probed further, but I do not propose to take up that task here. As Korn herself points out (1973: 97, n. 15), although the Kachin system has five lines in fact, this is not essential to the model of three-line asymmetric prescriptive alliance it otherwise belongs to. One other place in which FMBSD marriage occurs is on the island of Nias, off Sumatra, and here, too, it is not at all prescriptive but just one of a number of marriage preferences all designed to keep as close to conventional asymmetric alliance without actually marrying the genealogical MBD (Beatty 1990: 467–68; also Parkin 2018: 236–38). This case is discussed in more detail in Chapter 8. This indicates that neither terminology is a pure example of its supposed ‘type’, though as Maybury-Lewis mostly uses descriptions rather than genealogical denotata to translate kin terms, this is obscured at first sight and makes more global comparisons involving these two terminologies some-

On the Origin of Crow-Omaha Terminologies

what tricky and inexact. However, his mode of describing the terminologies is reflected in his broader belief that labels drawn from the conventional study of kinship do not help us understand Gê societies (especially Maybury-Lewis 1979b). Accordingly, although he begins his discussion of these two terminologies by suggesting the labels ‘Dakota’ and ‘Omaha’ for the Sherente and Shavante respectively (1979a: 240–41), he does not like either because the main difference between them comes down to how they treated cross cousins respectively, and he feels this was peripheral to how these two societies should be interpreted in the wider and more comprehensive sense. These labels do, however, suit my own arguments, and I have therefore considered it both convenient and instructive to retain them here.

References Aberle, David. 1962. ‘Navaho’, in David Schneider and Kathleen Gough (eds), Matrilineal Kinship. Berkeley and Los Angeles: University of California Press, pp. 96–210. Ackerman, Charles. 1976. ‘Omaha and “Omaha”’, American Ethnologist 3(4): 555–72. Allen, Nicholas J. 1976. ‘Sherpa Kinship Terminology in Diachronic Perspective’, Man 11: 569–87. ———. 1986. ‘Tetradic Theory: An Approach to Kinship’, Journal of the Anthropological Society of Oxford 13: 139–46. ———. 1989a. ‘Assimilation of Alternate Generations’, Journal of the Anthropological Society of Oxford 20(1): 45–55. ———. 1989b. ‘The Evolution of Kinship Terminologies’, Lingua 77: 173–85. ———. 2008. ‘Tetradic Theory and the Origin of Human Kinship Systems’, in Nicholas J. Allen, Hilary Callan, Robin Dunbar and Wendy James (eds), Early Human Kinship: From Sex to Social Reproduction. Oxford: Blackwell, pp. 96–112. ———. 2012. ‘Tetradic Theory and Omaha Systems’, in Thomas R. Trautmann and Peter M. Whiteley (eds), Crow-Omaha: New Light on a Classic Problem of Kinship Analysis. Tucson: The University of Arizona Press, pp. 56–66. Barnes, R.H. 1984. Two Crows Denies It: A History of Controversy in Omaha Sociology. Lincoln and London: University of Nebraska Press. Bateson, Geoffrey. 1932a. ‘Social Structure of the Iatmül People of the Sepik River’, Oceania 2: 246–89. ———. 1932b. ‘Social Structure of the Iatmül People of the Sepik River’, Oceania 2: 401–51. ———. 1936. Naven. Cambridge: Cambridge University Press. Beatty, Andrew. 1990. ‘Asymmetric Alliance in Nias, Indonesia’, Man 25: 454–71. Condominas, Georges. 1960. ‘The Mnong Gar of Central Vietnam’, in George Murdock (ed.), Social Structure in Southeast Asia. New York: Wenner-Gren Foundation for Anthropological Research, pp. 15–23. Donald, Leland, and Marion Tighe. 1987. ‘A Formal Analysis of Three Apachean Kinship Terminologies’, in Leland Donald (ed.), Themes in Ethnology and Culture History: Essays in Honor of David F. Aberle. Meerut: Folklore Institute, Archana Publications, pp. 34–80.



223

224

How Kinship Systems Change

Dyen, Isadore, and David F. Aberle. 1974. Lexical Reconstruction: The Case of the Proto-Athapaskan Kinship System. Cambridge: Cambridge University Press. Eyde, David B., and Paul M. Postal. 1961. ‘Avunculocality and Incest: The Development of Unilateral Cross-Cousin Marriage and Crow-Omaha Kinship Systems’, American Anthropologist 63: 747–71. Faron, Louis. 1956. ‘Araucanian Patri-organization and the Omaha System’, American Anthropologist 58(3): 435–56. Godelier, Maurice, Thomas Trautmann and Franklin Tjon Sie Fat (eds). 1998. Transformations of Kinship. Washington: Smithsonian Institution Press. Haider, Karl G. 1974. ‘Accounting for Variation: A Nonformal Analysis of Grand Valley Dani Kin Terms’, Journal of Anthropological Research 34(2): 219–62. Henderson, Richard N. 1967. ‘Onitsha Ibo Kinship Terminology: A Formal Analysis and its Functional Implications’, Southwestern Journal of Anthropology 23(1): 15–51. Héritier, Françoise. 1981. L’exercice de la parenté. Paris: Gallimard. Ives, John W. 1998. ‘Development Processes in the Pre-contact History of Athapaskan, Algonquian, and Numic kin Systems’, in Maurice Godelier, Thomas Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 94–139. Kohler, Josef. 1975 [1897]. On the Prehistory of Marriage: Totemism, Group Marriage, Mother Right. Chicago: The University of Chicago Press. Korn, Francis. 1973. Elementary Structures Reconsidered: Lévi-Strauss on Kinship. London: Tavistock Publications. Kronenfeld, David. 1980a. ‘A Formal Analysis of Fanti Kinship Terminology (Ghana)’, Anthropos 75: 586–608. ———. 1980b. ‘Particularistic or Universalistic Analysis of Fanti Kin-Terminology: Alternative Goals of Terminological Analysis’, Man 15: 151–69. ———. 2009. Fanti Kinship and the Analysis of Kinship Terminologies. Urbana and Chicago: University of Illinois Press. Lane, Robert, and Barbara Lane. 1959. ‘On the Development of Dakota-Iroquois and Crow-Omaha Kinship Terminologies’, Southwestern Journal of Anthropology 15: 254–65. Leach, Edmund R. 1961. Rethinking Anthropology. London: Athlone Press. Lévi-Strauss, Claude. 1966. ‘The Future of Kinship Studies’, Proceedings of the Royal Anthropological Institute 1965: 13–22. ———. 1949. Les structures élémentaires de la parenté. Paris: Presses Universitaires Françaises (Engl. edn. 1969). Livingstone, Frank B. 1959. ‘A Formal Analysis of Prescriptive Marriage Systems among the Australian Aborigines’, Southwestern Journal of Anthropology 15: 361–72. Lounsbury, Floyd G. 1964. ‘The Structural Analysis of Kinship Semantics’, in Horace G. Lunt (ed.), Proceedings of the Ninth International Congress of Linguists. The Hague: Mouton, pp. 1073–93. Maybury-Lewis, David. 1967. Akwe˜-Shavante Society. Oxford: Clarendon Press. ———. 1979a. ‘Cultural Categories of the Central Gê’, in David Maybury-Lewis (ed.), Dialectical Societies: The Gê and Bororo of Central Brazil. Cambridge MA and London: Harvard University Press, pp. 218–48.

On the Origin of Crow-Omaha Terminologies

———. 1979b. ‘Conclusion: Kinship, Ideology and Culture’, in David MayburyLewis (ed.), Dialectical Societies: The Gê and Bororo of Central Brazil. Cambridge MA and London: Harvard University Press, pp. 301–12. McConvell, Patrick. 2012. ‘Omaha Skewing in Australia: Overlays, Dynamism and Change’, in Thomas R. Trautmann and Peter M. Whiteley (eds), Crow-Omaha: New Light on a Classic Problem of Kinship Analysis. Tucson: The University of Arizona Press, pp. 243–60. McKinley, Robert. 1971. ‘A Critique of the Reflectionist Theory of kinship Terminology: The Crow/Omaha Case’, Man 6: 228–47. Mickey, Barbara H. 1956. ‘Acoma Kin Terms’, Southwestern Journal of Anthropology 12(3): 249–56. Needham, Rodney. 1958. ‘A Structural Analysis of Purum Society’, American Anthropologist 60(1): 75–101. ———. 1960a. ‘Alliance and Classification among the Lamet’, Sociologus 10: 97–119. ———. 1960b. ‘Patrilateral Prescriptive Alliance and the Ungarinyin’, Southwestern Journal of Anthropology 16: 274–91. ———. 1961. ‘An Analytical Note on the Structure of Sirionó Society’, Southwestern Journal of Anthropology 17: 239–55. ———. 1962. Structure and Sentiment: A Test Case in Social Anthropology. Chicago and London: The University of Chicago Press. ———. 1964. ‘Descent, Category and Alliance in Sirionó Society’, Southwestern Journal of Anthropology 20: 229–40. ———. 1971. ‘Remarks on the Analysis of Kinship and Marriage’, in Rodney Needham (ed.), Rethinking Kinship and Marriage, London etc.: Tavistock Publications, pp. 1–34. Opler, Morris. 1936. ‘The Kinship Systems of the Southern Athapaskan-Speaking Tribes’, American Anthropologist 38: 620–33. Parkin, Robert. 1988. ‘Prescription and Transformation in Mon-Khmer Kinship Terminologies’, Sociologus 38(1): 55–68. ———. 2018. ‘The Evolution of Kinship Terminologies: Nonprescriptive Forms of Asymmetric Alliance in Indonesia’, Journal of Anthropological Research 74(2): 232–51. Rumsey, Alan. 1981. ‘Kinship and Context among the Ngarinyin’, Oceania 51: 181–92. Salisbury, Richard F. 1956. ‘Asymmetrical Marriage Systems’, American Anthropologist 58(4): 639–55. ———. 1962. From Stone to Steel: Economic Consequences of a Technological Change in New Guinea. Melbourne: Melbourne University Press, for Australian National University. Schneider, David M., and Kathleen Gough. 1962. Matrilineal Kinship. Berkeley and Los Angeles: University of California Press. Tjon Sie Fat, Franklin. 1998. ‘On the Formal Analysis of “Dravidian”, “Iroquois”, and “Generational” Varieties as Nearly Associative Combinations’, in Maurice Godelier, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 59–93. Trautmann, Thomas R. 2012. ‘Crossness and Crow-Omaha’, in Thomas R. Trautmann and Peter M. Whiteley (eds), Crow-Omaha: New Light on a Classic Problem of Kinship Analysis. Tucson: The University of Arizona Press, pp. 31–50.



225

226

How Kinship Systems Change

Trautmann, Thomas R., and R.H. Barnes. 1998. ‘“Dravidian”, “Iroquois”, and “CrowOmaha” in North American Perspective’, in Maurice Godelier, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 27–58. Trautmann, Thomas R., and Peter M. Whiteley (eds). 2012. Crow-Omaha: New Light on a Classic Problem of Kinship Analysis. Tucson: The University of Arizona Press.

8 

Substitutability of Kin and the Crow-Omaha Problem

Introduction Crow-Omaha kinship terminologies have long posed a problem for those keen to associate terminological patterns with specific sociological features.* These attempts continue; witness a recent collection discussing this issue edited by Thomas Trautmann and Peter Whiteley (2012). In fact, the only terminological patterns that do show a logical fit with a definite sociological feature, namely cross-cousin marriage, are those described as prescriptive by Rodney Needham, whose language in these respects I general follow (e.g. Needham 1973) – and even then the expected correlation is by no means always found ethnographically. No such logical fit is so dramatically evident in the case of CrowOmaha terminologies, though explanations abound as to why they should exist at all. Most of the early theories were described by Robert McKinley in an article published in 1971, though he was later attacked for an aspect of his own interpretation by Robert Barnes (1976, 1984: Ch. 10). Nonetheless, both authors are useful to me in this chapter, though in their own ways: McKinley (q.v. 1971) by obviating the need to go through all the older theories, some of which are decidedly redundant; and Barnes because his book on the Omaha people themselves (1984) has been of immense value in guiding me through the maze of issues surrounding the actual Omaha ethnogra-

228

How Kinship Systems Change

phy and the reasons why the name ‘Omaha’ became one of the most famous in the anthropology of kinship. Perhaps the two most prominent explanations for Crow-Omaha terminologies, those that have survived the longest, are RadcliffeBrown’s theory that such terminologies are an expression of lineal unity (1952: Ch. 3) and Lévi-Strauss’s observation (1966) that they identify lineages related to ego into which ego may not marry, at least in the present and immediately following generation(s). Both theories have been much derided, especially the former, which has perhaps worn less well over time; but the latter, too, is vitiated by the fact that the marriage avoidances that are the subject of that theory by no means need a Crow-Omaha terminology to accompany them; witness north Indian kinship systems (Tiemann 1970). However, while Radcliffe-Brown’s theory is close to being tautological, it does at least focus on the terminologies themselves and their vertical, lineal equations. The fact that for Lévi-Strauss Crow-Omaha terminologies need marriage prohibitions to explain them, but not vice versa, means that his theory cannot be a universal one. It is, therefore, Radcliffe-Brown’s position that constitutes the launch pad for the arguments presented in this chapter, it being understood that my modifications of his position still do not constitute the universal theory we have all been searching for – so far in vain, and perhaps in perpetuity. Before proceeding further, I should make clear my own understanding of Crow-Omaha systems and equations, together with their implications for other aspects of kinship. Crow-Omaha equations appear first and foremost in the line of complementary filiation in Meyer Fortes’s sense (e.g. the mother’s patriline where there is patrilineal descent and patrilineal exogamy). Thus, in the case of Omahatype terminologies, one frequently finds terminological equations uniting MF, MB, MBS, MBSS, MBSSS etc. (minimally, perhaps, MB and MBS) – that is, equating the males of this line in successive generations. Omaha terminologies therefore express the operation of patrilineal descent, though not all societies with the terminology actually have the associated descent system. What has tended to be puzzling since Josef Kohler (1975 [1897]) first tried to interpret these terminologies at the end of the nineteenth century are the similar vertical equations linking the corresponding female kin, with a focus on the equation M = MBD, possibly including MZ as well through bifurcate merging. One explanation,

Substitutability of Kin and the Crow-Omaha Problem

suggested by Kohler himself, starts from the unexceptional fact that anywhere in the world, formally speaking, male ego’s mother is his father’s wife and his father’s WBD is his own (ego’s) MBD. Among the Omaha, however, WBD may be one of the father’s options for a second marriage, a rare event even among the Omaha, let alone elsewhere. For Kohler, this provides a sufficient explanation for the specific features of the Omaha terminology. However, a principle of descent is also involved in these equations. Generally speaking, in societies of all types, it is perhaps logical at first sight to conceive of a woman’s same-sex successor as her own daughter. Certainly inheritance may sometimes go matrilineally from mother to daughter even where descent itself (in the sense of the transmission of social status) is patrilineal. For example, in India the stridhana or female ‘portion’, usually consisting of female clothing and jewellery, goes matrilineally from mother to daughter in a society where patrilineal descent and inheritance predominate (see Tambiah 1973). Of course, this rule of matrilineal descent applies in any case in the typical matrilineal society, but Crow terminologies, which express matrilineal descent, do not do so through equations uniting mother and daughter (see next paragraph). In most patrilineal societies, conversely, these two women are not lineally linked, at least not by birth:1 assuming exogamy of patrilines, the daughter will be in a different patriline than her mother. Within the mother’s own natal patriline, therefore, her same-sex successor must be traced through a male, most obviously her brother (ego’s MB), and will therefore be ego’s MBD – hence the reason for the puzzling equation. In short, once one has vertical equations for males, it is only logical to expect an equivalent series for females. Moreover, this set of vertical equations can often be extended further down successive generations, as happens with the male terms, and this can also proceed regularly through male links, thus including MBSD, MBSSD, etc. These latter specifications are not always included in the key textbook definitions of Omaha-type equations, yet they are implied by the logic of the classification and are found in many actual cases too – if not uniformly, then at least sufficiently frequently to invite an explanation such as Kohler’s. Although, as just suggested, Radcliffe-Brown’s theory of the significance of these terminologies has not lasted as well as Lévi-Strauss’s, it is still my starting point here. However, I am not just accepting Radcliffe-Brown’s explanation hook, line and sinker. Quite a lot of



229

230

How Kinship Systems Change

ethnography since his day, and some even before his time, shows that lineal unity is not just a thing in itself but may very well have a function (as it did for Lévi-Strauss, viz. the part it allegedly played in controlling marriage choices), though the function of lineal unity, if any, may vary from case to case. In the present context, I argue that this function can be seen to have two main forms, which can also be connected: 1) Residual interests in, or claims to, property or prestations held but not owned by others are one possible reason for CrowOmaha vertical equations. This will become clearer when I discuss examples below, but the scenario I am referring to runs roughly as follows. The members of descent group or descent line A are related (typically, it seems, by marriage, often in the deep past) to ego’s descent group or descent line B but are and remain separate from it. But further, A has a residual interest in property currently controlled by the members of B or has a claim to prestations that B owes to A, in such a manner that the lineally related members of A are substitutes for one another over time when it comes to these claims being recognized. Moreover, where the terminology is Crow-Omaha in type (and it need not be: see Parkin 2018, reproduced as Ch. 9), the vertical equations make this substitutability clear. This recognition of claims does not necessarily mean that they will be resolved at any particular point in time, and indeed in many cases it seems that they can never be resolved entirely if the relationship is to endure in the long term. It may therefore be the fact of substitutability that is the key to interpreting the vertical terminological equations here, not the interests or claims themselves. Among other things, this would avoid putting forward an argument that is simply a reprise of Radcliffe-Brown’s doctrine of the unity of the lineage: while that unity may exist in and of itself, it may also express the substitutability of lineally related kin for particular, ethnographically varying reasons. Examples of this process are David Labby’s (1976) work on Yapese property and Ross Bowden’s (1983) work on the mortuary prestations of the Kwoma of Papua New Guinea, both discussed at length below, together with other similar cases. 2) The other form, already discussed above, is Kohler’s suggestion that Crow-Omaha terminologies are associated with secondary

Substitutability of Kin and the Crow-Omaha Problem

marriages in which a chain of patrilineally connected women are preferential second spouses for ego, and therefore substitutable for one another. This might, therefore, be considered a special case of 1), though it was dismissed by Radcliffe-Brown, Lévi-Strauss and others (see Barnes 1984: 156) precisely because of its reliance on second marriages that presumably few in the society concerned will ever contract. The rest of this article explores these scenarios further by discussing several pertinent examples. I should emphasize here, however, that I am not claiming to have solved the Crow-Omaha problem but merely attempting to interpret one major theory of the occurrence of these terminologies in a slightly different light. My sources are exclusively published ethnographies and not primary fieldwork of my own.

The General Substitutability Argument (#1 Above) One relevant case that has been discussed extensively, though for other reasons, is social organization among the inhabitants of the island of Yap, a US Trust Territory in Micronesia. Much of this earlier discussion relates to the interpretation of descent on the island, a topic wrapped up with David Schneider’s attempts (1972, 1984) to deconstruct descent entirely for analytical and comparative purposes, even though he originally interpreted the island’s kinship system and practices in conventional genealogical terms (Schneider 1953). This involved him going back on these earlier studies (ibid.) in which he talked freely about both patrilineal and matrilineal descent groups (respectively the tabineau and the genung) being present on the island. In his later book repudiating this position, A Critique of the Study of Kinship (1984), he now decided that the tabineau was in reality just a land-holding corporation, and he sought similarly to dismiss the idea of the genung as a descent group, though it is evident throughout his discussion that he found it no easy task to make these revised arguments. A subsequent study by David Labby (1976) unequivocally supports Schneider’s original view that both groups are descent groups. It is clear, however, when comparing these two works, that the genung is not stressed as a unit by the Yapese; indeed, although exogamous



231

232

How Kinship Systems Change

and therefore governing marriages, it is residentially dispersed (residence being patrilocal). The tabineau, by contrast, is clearly more prominent, being both a residential and a land-holding unit and discernibly patrilineal in that succession to its headship proceeds patrilineally; it is also exogamous (ibid.: 32). In addition, unlike the tabineau, the genung is quasi-secret, not something discussed openly or that one asks questions about. Schneider says (1953: 217) that a mother will inform her children of their membership in it and point out other children who belong to it in the interests of avoiding incest, but anything less like a corporate group would evidently be hard to find. Schneider also says (ibid.: 218) that marriage has little formality, the ceremony often being omitted altogether, and that divorce is easy to enact (a situation often found in societies with matrilineal descent). This, however, is implicitly contradicted by Labby’s detailed description of the very elaborate exchanges of material goods that take place in connection with marriages (1976: 38–44). Status differences play some part in choosing spouses, as women should not marry into families or kin groups of lower status than their own (Labby 1976: 26). Neither Labby nor Schneider mention the existence of positive marriage rules or give any indication of other possible restrictions on affinal alliance; for example, whether there are any bans on two kin groups repeating alliances in the generations following the first marriage. In addition, neither author draws attention to the fact that the Yap kinship terminology is a Crow-type one, despite both of them describing it in some detail; indeed, Labby does not mention the fact at all, and Schneider only does so in passing (1953: 215). In addition, the terminology is used only in reference, and even then there are restrictions on its use as such; it is not used in address, vocative use being by name alone (apparently regardless of age or gender). This may be the reason for both authors minimizing its importance. Nonetheless, taking the two main sources together (ibid.: 219; Labby 1976: Ch. 4), many vertical equations can be identified in the Yap terminology, including those normally definitive of Crow systems. Thus, the term citamangin2 means F, FB, FZS, FZDS, FZDDS, as well as MB, PZH, EF and EFB. The female equivalent is citiningin, FZ, FZD, FZDD, FZDDD, as well as M, MZ, FBW, EM, EMZ.3 Fak links C/BC and MBC in Crow fashion, as well as other specifications not linked lineally. However, other lines are terminologically united.

Substitutability of Kin and the Crow-Omaha Problem

Thus, thang e tu apparently stands for all +3 males,4 plus the line of FFFZS, FFFZDS, FFFZDDS, and thang e thaw for all +3 females, plus the line of FFFZD, FFFZDD, FFFZDDD. Similarly, tutu covers many males in +2, as well as the line of FFZS, FFZDS, FFZDDS, while titaw refers to many +2 females plus the line of FFZD, FFZDD, FFZDDD. Then there is wa’ayengin, a reciprocal term for MB and ZC, but also extending the line of the latter to ZCC and ZCCC. Wolag, given as ‘sibling’, unites MZC and MZCC. Affines can also be involved in this process: thus, le’engin, W, MBW also links MZSW and MZDSW. However, as can be seen, most of these equations involve patrilateral relatives, consonant, it seems, with the obvious importance of ego’s FZ as mafaen, roughly the owner, or joint owner with her brother, of the land on which ego is brought up and will continue to live if male (women marry virilocally). This right of (co-)ownership also extends to FZ’s children and further descendants, and it exists regardless of the fact that neither she nor they live on the land they co-own. This is clearly the reason, Labby notes (ibid.: 54–55), for viewing her whole descent line as a terminological unit (citamangin, citiningin), as well as those of FFZ (tutu, titaw) and FFFZ (thang e tu, thang e thaw) respectively, who are also implicated in co-ownership by virtue of the birth of their ascendants on this land and their departure from it at marriage. Indeed, the three lines headed, respectively, by FZ, FFZ and FFFZ are themselves called mafaen, in the order mafaen ni be’ec, or ‘new mafaen’; mafaen ni le’ or ‘coconut mafaen’, a metaphor for this line’s supposed hardness and durability; and mafaen ni gapalou or ‘black bird mafaen’, a reference to the likelihood of FFFZ no longer being alive and of her spirit having been eaten by a black bird (a starling important in divination). In addition, having enjoyed the fruits of the land for a long time, this line will have been paid off and therefore, of the three lines, will have the least claim on ego’s tabineau. The line headed by FZ, by contrast, which is the closest to ego, has the most claim over the latter by virtue of its residual ownership of the land ego is occupying (ibid.: 54). This also seems to be an explanation for the equations under wa’ayengin, at least in so far as this term unites ZD, ZDD and ZDDD, as they are the future mafaen of this tabineau, though they do not appear to be given a name as a line per se, as they are not mafaen to male ego himself (ibid.: 58). Despite the importance of these relatives as actual or potential co-owners of the tahina’s land, I suggest that as they are resident elsewhere contact with them is infrequent, consonant with their ter-



233

234

How Kinship Systems Change

minological identity as units rather than as individuals. By contrast, matrilateral relatives are not defined vertically as units in the terminology, since they potentially have different origins (there is no suggestion that, for example, male ego’s mother and wife should come from the same genung), and they go to different tahini on marriage; therefore, there is nothing to unite them. North-west of Yap is the island of Truk (Chu(u)k in some later texts), associated especially with fieldwork done by the anthropologist Ward Goodenough. In fact, his first publication on the inhabitants of this island (1947) was a journal article written with, and clearly under the influence of, George Murdock, in which he unambiguously defines the Truk kinship terminology as a Crow one; a later, book-length publication by Goodenough alone (1951) is rather more cautious on this point. Nonetheless, as on Yap, descent per se (i.e. as a marker of identity or as a means to control marriage through exogamy) is fundamentally matrilineal, meaning, of course, that the children of a male ego are in a different matriline than his own. In addition, what the authors call a ‘matrilineal descent line’ is not only a lineage segment but also a property-holding corporation, the most significant property here being land (lineages per se are not property-holders). However, as on Yap, under certain circumstances men can inherit patrilineally from their fathers as öfökür (a general term for −1 descendants). Most such inheritance is of the father’s movables, but if land is involved, the inheritor must recognize the residual rights of those members of his father’s matriline who are the actual owners. And, as these individuals apparently belong to a number of different generations, ‘The result is a kinship system of the Crow type, despite the characteristic Malayo-Polynesian paucity of distinct denotative terms’ (Murdock and Goodenough 1947: 340). Thus, semei is F, FB, FZS, FZDS, or any male of father’s lineage, while inei is FZ, FZD, or any female of father’s lineage, both therefore covering other lineal kin in addition to the already specified genealogical relations. However, the MBC = C equation, which also defines Crow terminologies, does not appear here, except in so far as both of these kin types are subsumed under the general −1 category of öfökür ‘children’, while MB is included under semei (mentioned above), or alternatively under the zero-level term ääi mwään ‘elder opposite-sex sibling’ (Murdock and Goodenough 1947: 340–41). The features described in the last sentence are seen as transitional by Goodenough in his later solo work, except on the island of Pu-

Substitutability of Kin and the Crow-Omaha Problem

luwat, where, at the time he was there, this transition had already taken place (1951: 96; see also below). He explains this transition as follows (ibid.): This redefinition is a logical outgrowth of the organization of lineages and descent lines as corporations whose members are regarded as siblings. It is possible that at one time a person stood in a dual relationship to his mother’s brother’s son just as he does now to his mother’s brother, calling this cousin a sibling when they interacted simply as kinsmen but calling him a child when behaving as a member of a corporation toward one of their jëfëkyr. One might predict that if the same trend continues, reference to mother’s brother as ‘father’ will give way to calling him ‘older brother’ in all contexts, as appears now to be the case on [the Truk island of] Puluwat.

In other words, Goodenough is speculating that Hawaiian equations linking referents in the same generation – namely, F = MB and MBS = B – is giving way to a Crow pattern that makes equations between generations, namely MB = eB and B = S. Another passage, published as an addendum to the earlier joint publication but written by Goodenough alone (1947: 343), provides the following explanation, which, unlike the quotation above, seems to treat the situation purely synchronically rather than as a potential transition supported by precedent: The term ääi mwään . . . refers primarily to an elder sibling of the same sex, and is applied to maternal uncles only in connection with inheritance or other matters when it is important to distinguish between ‘fathers’ . . . of one’s own and one’s father’s lineage.

Although MB ordinarily shares a term with ‘father’, on some occasions, says Goodenough, a distinction should be made between them. Goodenough also gives information on marriage rules and practices (1951: 120, 122–23). Thus, there is no marriage to consanguines, especially within the matrilineage, nor into one’s father’s lineage, and therefore there is no cross-cousin marriage, nor indeed any marriage to first cousins. Another society with matrilineal descent are the Trobriand Islanders, whose terminology, generally reckoned to be of Crow type, was much discussed by Malinowski, as well as Leach (1958) and Lounsbury (1965) in a famous debate. However, it has also been studied since Malinowski by Fathauer (1961) and Weiner (1979), and by



235

236

How Kinship Systems Change

Read and Behrens (1990) with regard to the underlying generative logic of the terminology. The Crow equations Fathauer gives are tabu FMM, FZ, FZD, FZDD, plus a number of other specifications, some male, some in other lines; tama FMMB, FMB, F, FB, FZS, FZDS; and latu MBC, C, ZC ws; these were later confirmed by Weiner (ibid.). Prominent in these discussions has been how to interpret the term tabu. Weiner states that tabu can denote both ancestral kin, those she calls the ‘founders’ of land (including the whole of the +2 genealogical level), and those women in male ego’s father’s clan whom male ego can marry. The latter, inter alia, include the vertically linked specifications FZ, FZD and FZDD – a classic Crow series – as well as FFBDD, another member of the same matriclan (1979: 339–42) and a classificatory FZD. She further states (ibid.: 343) that her informants ‘emphasized the very great importance attached to a marriageable spouse for ego coming from the same clan as ego’s father’ (apparently here both genders of ego are meant). This is apparently connected with mortuary payments ensuring control over land, though Leach (1958) suggested that what accounted for the kin types under this term was, first, the position of many of them as recipients of urigubu payments from WB to ZH and, more generally, wife-givers to wife-takers;5 and secondly, their status as relatives with whom there are potentially hostile relations, often involving joking relations. Weiner tries to reconcile the two meanings of tabu given above, as well as revealing that the earlier assumption of a preference for FZD marriage was denied by her informants. Nonetheless, it appears that the term tabu partially maps out a vertical series of, here, matrilineally related women in different generations whom male ego may marry. At all events, by marrying into their father’s clan, both male and female egos are repeating, though in reverse order, a previous alliance between their and their father’s clan without the affinal alliance system or terminology being at all prescriptive. As is well known in the Trobriands, at least formally the place of the father in procreation is denied, and he is often seen as an outsider, indicating a low degree of emphasis on marriage, coupled with relatively easy divorce. Since much has already been written about the Trobriand case, thus giving it its classic status, I do not wish to add unduly to that corpus here. The upshot appears to be, however, that, roughly as on Yap and Truk, matrilineal descent is opposed by patrilineal descent associated with a rule of patrilocal residence, creating a structural ten-

Substitutability of Kin and the Crow-Omaha Problem

sion with regard to the respective land rights of patrikin and matrikin and of associated gift exchanges. We may conclude that Crow-type equations in the Trobriands thus map out the line(s) of those who are due urigubu payments and are also the preferred source of spouses for male ego and his own line.6 A particularly well-described example of a society with patrilineal descent and an Omaha terminology are the Kwoma, of the Sepik river area in Papua New Guinea, studied by both Williamson (1980) and Bowden (1983), though there are some differences of detail between them.7 The latter gives as key terms magwapa MB, MBS, MBSS, MBSSS; nowkwapa M, MZ, MBD, MBSD, MBSSD (also FBW); wapok MBW, MBSW MBSSW, MBSSSW; yakw FFFZ, FFZ, FZ and their respective husbands; nel WBC, WBSC (as well as BC, BSC); and ruwey ZC, FZC ms (as well as HZC, HFZC ws). Fundamentally, Bowden’s analysis of the Kwoma (1983) links their Omaha-type terminology to how the Kwoma maintain alliances set up by marriage in the following three generations, when repeat marriages between the two groups are banned. While patrilines of wife-givers and wife-takers can be identified here, the social system is predicated on the fact that they must ordinarily be kept distinct from one another.8 There are no positive marriage rules associated with the practice of any form of prescriptive alliance. In Bowden’s words, ‘Kwoma formally prohibit the repetition of marriages, symmetrically or asymmetrically, between affinally linked patrilines (but not clans as wholes) for several generations once a marriage has been contracted’ (ibid.: 748). As a result, alliances are dispersed among patrilines and cannot be focused on specific patrilines, as they are in both the models and many actual instances of prescriptive alliance. Nonetheless, Bowden’s argument is that alliances initially contracted through a marriage are continued into the following generations (up to four) in other ways than through marriage, mainly involving mortuary prestations, and that this is consistent, with some exceptions, with the characteristically Omaha equations in the terminology. Some of these prestations are transferred asymmetrically from wife-takers to wife-givers, mostly as bridewealth, which in the first generation goes from ZH (the groom) to WB and the latter’s same-generation kin in his patriline. Similar asymmetric transfers are repeated in the second generation between the wife-taker’s sons and the same wifegiver’s patriline, represented by ego’s MB. The transfers in these two generations are compulsory, but they may be succeeded by further



237

238

How Kinship Systems Change

discretionary prestations of this kind in the third and fourth generations, after which the initial marriage tie is no longer seen as relevant and the two lines may intermarry again. These later transfers therefore take place between the wife-taker’s SS and SSS respectively and the corresponding survivors of the initial wife-giver’s line. More formally, the relationships typically involved are ZSS to FMB in the third generation and ZSSS to FFMB in the fourth, though whether the senior, wife-giving parties to these relationships will still be alive at this stage is a moot point. Simultaneously with these transfers, food goes in the opposite direction, from wife-givers to wife-takers, and there is also a constant ‘symmetrical exchange of various domestic, social and political services’ (ibid.: 750). However, further marriages in this four-generation period are ruled out – that is, a male ego cannot marry into the patrilines of BWB, MB, FMB or FFMB, nor of WB, nor may he marry wife-taking relatives such as ZHZ, ZD or FZD, as this would involve an illicit reversal of the direction of alliances. In linking the relationships set up by these exchanges to the terminology, which has numerous Omaha-type equations, Bowden has to acknowledge some exceptions. Thus, WB and WF are distinguished here because ego is not obliged to give prestations to WF, even though he must do so to WB. Also, for reasons Bowden admits he cannot account for, WB has a separate term not only from WF but also from WBS and the latter’s descendants, despite the fact that an equation between them would be expected in an Omaha-type terminology, though such an equation does occur in Kwoma Pidgin. However, the equation WBS = WBSS (and implicitly = WBSSS etc.) does obtain here, as they are all WB’s successors in the exchange relationship with the respective wife-taking line and inherit WB’s claims to the wealth payments. Similarly, from the point of view of the second generation of wife-takers, while MF has a separate term from MB, as he (MF) is not involved in these transfers with ego, MB’s lineal descendants are equated with MB for four generations in classic Omaha fashion, such that MB = MBS = MBSS = MBSSS. From the points of view of the third and fourth generations of wife-takers, there are also terms uniting the lineal descendants respectively of FMB and FFMB. However, once this series of generations dies out, mutual terminological recognition of these lines of wife-givers and wife-takers ceases, in line with their ability to renew the initial marriage alliance: at this point, indeed, they become formally unrelated.

Substitutability of Kin and the Crow-Omaha Problem

Although at this point the Kwoma also say it is good to reverse the direction of alliance to even up the balance between the two groups, fundamentally the relationship between wife-takers and wife-givers within any period of up to four generations is asymmetric. This is confirmed by Williamson’s slightly earlier work on the Kwoma (1980: 535–44). She starts by stating that the Kwoma prefer to think of their marriage practices as symmetric, despite the fact that there should be no exchange of sisters or of actual FZDs. She also says that the value given to symmetry is expressed in a more specific preference for marriage to classificatory FZD, to the extent that wives are routinely placed in this category regardless of their real relationship to ego. Bowden explicitly states that there is no such marriage preference (1983: 763–64, n. 4), remarking that only four of the hundred marriages for which he collected genealogical data were between male ego and a woman who could be classed as FZD. However, for Williamson this liking for exchange symmetry is also expressed as a preference to take wives from the lineage of FMB, to which actual FZD would also belong in the theoretical model of patrilateral cross-cousin marriage (m.s.), and FMBSD is identified as a legitimate wife, who would be equated with actual FZD in the same model.9 Although Williamson does not say as much, this would also represent male ego repeating the marriage of not his father but his FF. However, her further statement that this woman would be a wife for ego’s son or BS (ibid.: 542) looks odd in these circumstances, since it would involve a marriage between generations. The significance of FMB for Williamson is also that once mortuary payments to him have been completed the initial marriage alliance also comes to an end, meaning that this happens a generation earlier than in Bowden’s account; more particularly, it also suggests that a realliance now becomes possible, hence the suitability of FMBSD as a spouse for male ego. Despite the stress on symmetry in affinal alliance, however, Williamson makes it clear that most exchanges are asymmetric, as are marriage prestations, and that the ideal of symmetry often breaks down in practice, even though one can speak of a rough balance in spouse exchange overall. The discrepancies between Bowden’s and Williamson’s accounts cannot be reconciled further here, but they are probably not enough to affect the overall picture of how Kwoma marriage practices work and the extent to which the terminology can be associated with them.10



239

240

How Kinship Systems Change

The practice of maintaining relationships between two patrilines in the generations after the initial marriage, when no further marriages are possible, is also noted, though in much less detail, by Forge (1971) for another Papua New Guinea people. He states: ‘The Eastern Abelam . . . consider that each marriage sets up a relationship that subsists for three generations between a man, his sons and son’s sons, and the sub-clan that provided his wife, the wife-givers becoming mother’s brother’s sub-clan and father’s mother’s brother’s sub-clan to the two succeeding generations’ (1971: 137). This relationship takes the form of ‘very elaborate and numerous exchanges . . . between the two groups for about a hundred years’, which are ‘unbalanced and dissimilar at all stages, and valuables flow only to, never from, the wife-givers’ sub-clan’ (ibid.). Presumably the Kwoma, like the Abelam, have sufficient genealogical memory to be able to tell who owes what to whom over such a long period and when the exchanges come to an end and marriages can start up again. The Kwoma case also has some resonance for similar practices in parts of eastern Indonesia and Sumatra (discussed in Parkin 2018, reproduced as Ch. 9). Both regions have examples of matrilateral cross-cousin marriage or asymmetric prescriptive alliance (MBD/FZS marriage), there being no Crow-Omaha terminologies at all in these areas as far as I am aware. However, there are societies in these regions that practise a non-prescriptive form of matrilateral marriage, though it might be one or other development of a former prescriptive system and thus represent an evolution away from it. As with matrilateral prescription, strict distinctions are still maintained between superior wife-givers and inferior wife-takers. Again, there are delays in renewing marriages between any two groups for a number of generations, with relationships in this intervening period being pursued in other ways and with kin in succeeding generations tending to substitute for one another when exercising claims to property or prestations. What these kin in different generations mostly consist of, however, are chains of mother’s brothers and their ascendants in different clans or lineages: they are not lineally connected within a clan or lineage. That is, it is the alliance aspect that is the focus rather than the lineal aspect.11 Quite a lot of material relevant to the theme of this chapter comes from Africa. One example is John Beattie’s studies of the Bunyoro of Uganda (1957, 1958, 1960). He clearly states that their terminology is of Omaha type (e.g. 1960: 53–55; 1957: 335), with many cross-

Substitutability of Kin and the Crow-Omaha Problem

generational equations, though he does not offer a very satisfactory explanation for their occurrence. The Bunyoro have non-corporate exogamous patriclans (ibid.: 317) and ban marriage into the clans of mother and MM, as well as one’s own clan (ibid.: 321). Residence is patrilocal (ibid.: 329). There is a slight preference for marriage into FM’s clan, and such marriages are not unusual (ibid.: 339). As just noted, there are many vertical lineal equations in the Bunyoro terminology, which Beattie sets out in a series of diagrams (1957: 331, 334, 337; 1958: 12, 18). Thus, while MF is classed with the grandparents, niynarumi is MB, MBS, MBSS, nyinento MZ, MBD, MBSD. Reciprocally, baihwa is FFZC, FZC, ZC. Other terms with vertical equations are the following: • mwene, MFZS, MZS, MBDS, B, FBS • munyanya, MFZD, MZD, MBDD, Z, FBD • nyinenkuru, FM, FMG, FMBC, FMBSC, MM, MMG, MMBC, MMBSC • mwijukuru, FFZCC, FZCC, ZCC, CC, BCC • muko, FFZH, FZH, ZH, DH, SDH • isezara, WFF, WF, WB, WBS, WBSS, HF • nyinazara, WFM, WM, WBW, WBSW, WBSSW, HM • mukamwana, SW, SSW • baijukuru, FZSC, ZSC Beattie’s reasons for the existence of these vertical equations apply mainly to the lines of MB and FZH and do not go much beyond the obvious statement that they reflect ego’s view of these lines being in some sense discrete units. ‘Thus a man thinks of the children of his father’s sister, just as he thinks of the children of his own sister, primarily and collectively as children whom a woman of his own agnatic group has born to another clan’ (1957: 333). And further, ‘Correspondingly, a Munyoro [singular of Bunyoro] conceives the agnatic group to which he is linked by his mother, a group from membership of which he is excluded, as a unit’ (ibid.: 335), with all the women and men respectively at or below ego’s level having the same terms (as shown above). Another African people with an Omaha terminology are the Haya, who live on the Tanzanian, south-western shore of Lake Victoria (Reining 1972). The Haya have patriclans and patrilineages, but Reining focuses more on a third concept she calls patrilines, or enda



241

242

How Kinship Systems Change

in the Haya language. Most marriages are exogamous with respect to both clan and village; residence is usually patrilocal, but if the bride comes from her mother’s clan, then residence may be in the latter’s village of origin. The rules of marriage seem relatively simple at first sight. Thus, female ego may marry into her mother’s clan, as we have just seen, but not her father’s, while male ego can marry into neither clan of his parents. He can, however, marry into FM’s clan, which normally means marriage to FMBSD (reciprocally FFZSS); that is, repeating a marriage that took place two generations earlier. There is also a sense that in the interim a reciprocal marriage takes place – that is, in the opposite direction: ‘The village patriline gives a daughter in one generation and takes a daughter back in the next’ (ibid.: 103). This suggests that the marriage in the intervening generation is to FZD: although Reining does not say as much, some of the marriages in her diagrams on page 104 are clearly between FZD and MBS. This is reinforced by the additional information that ‘A model consisting of three intermarrying clans satisfies known genealogical data’ (ibid.: 103), this being how the bare model of FZD/MBS marriage is generally diagrammed. Regarding the terminology, Reining distinguishes the vocative or address terminology from the reference one, the former being largely generational in pattern, the latter with at any rate one Omaha term, kusikisa (literally, ‘to install the heir’), linking MB, MBS and MBSS.12 The data Reining provides for the terminology are not sufficiently detailed to tell whether the reciprocal equation, FZC = ZC, also occurs here. In explaining the one equation she does identify, Reining refers back to Lounsbury’s suggestion (1964a) that Crow-Omaha terminologies should be linked to rules of succession and by implication to inheritance. Inheritance here is normally from father to son, occasionally to a SS, rarely to a brother. More particularly here, ‘When a man dies, the mother’s brother of his primary heir oversees the installation of his successors’ (Reining ibid.: 99). In other words, the heir’s MB verifies the succession. However, if he too is deceased, his son and potentially his SS will act instead, an equivalence expressed by the term kusikisa. Yet another case is the Baamba of western Uganda, studied by Edward Winter (1956). These have patrilineal descent, with exogamous lineages and a three-generation ban on repeating previous marriages between lineages. At the time Winter wrote, traditional exchange marriages were being replaced by bridewealth marriages because of

Substitutability of Kin and the Crow-Omaha Problem

missionary opposition to the former and colonial government prohibition of them (ibid.: 21–22, 73, 80ff., 172, 190). The Baamba also have several vertical equations in their kinship terminology, namely nukwayo, MFBS, MB, MBS, MBSS, MBSSS; ama, MFBD, M, MZ, MBD, MBSD, MBSSD, also HM; and mwaliana, FFZC, FZC, ZC, also FBDC, MZDC (ibid.: 256–60). Following Murdock, Winter identifies the terminology as Omaha in type (ibid.: 256) but does not explain why, beyond stating that the patriline of MB and his lineal descendants will appear to ego as a unit, since ego is related to them as a whole. However, MB will not view the line of ego (the ZS of MB) as a unit because MB is actually likely to have several sister’s sons in different lineages, who, by contrast, will all have him as their MB in common and thus gravitate towards him and his line in certain significant contexts of the MB-ZS relationship (ibid.: 186 ff.). The Bangwa of western Cameroon also have some indications of Omaha-type equations in their terminology, though their ethnographer, Brain, is not very clear on the matter and does not give a full or very useable terminology, which he finds ‘very sparse . . . Bangwa categories are not clearly differentiated terminologically . . .’ (1972: 47; see also diagram, ibid.: 48). Thus, atsen’ndia is MB, MBC, MZ, MZC, FB, FBC, ZC, possibly FBSC too, though betat is also indicated as linking FBC and FBSS. The first of these terms is also used for what Brain initially refers to as ‘a shallow clustering of kin, to whom ego is related through his mother’ (ibid.: 53). Etymologically, this term combines atsen, a word for a woman’s lower body, with ndia, the term for house (obviously meant in a sociological sense, as well as referring to the physical structure). From the outset (ibid.: 53 ff.), Brain describes the atsen’ndia as an informal, affective ‘matrigroup’, as opposed to the more formal and structured ‘patrigroup’. Both are small kin groups rather than full lineages, the latter being much more concerned with what we normally understand by descent than the former (Brain sums up Bangwa kinship in general as bilateral, ibid.: 59–60). Brain even suggests that the atsen’ndia should be understood through ‘psychology rather than structural analysis’ (ibid.: 53). It does have a range of descent and inheritance functions (apart from ego’s relationship of mutual support and affection with his MB) in such a manner that suggests that these functions and feelings extend in part to other members of it – one can readily detect Radcliffe-Brown’s influence here – but Brain does not make any explicit link with possible terminological equations of Omaha



243

244

How Kinship Systems Change

type. It also seems that the term atsen’ndia can be used for both a matrilineal grouping involving ties of kinship with and through one’s mother back to an ancestress, and the patrilineage from which one’s mother came upon marriage, making it a matrilateral group of relatives or even affinal in character – that is, alliance-oriented instead of descent-oriented (ibid.: 65). It is in this latter sense that we would expect Omaha equations, if any, to map out a vertical line through the generations, though that is far from clear from the available information. Affinity is certainly important in extending useful ties: marrying an existing affine is disliked as a waste of an opportunity to create a new tie. Marriage to distant patrikin, such as a betat, is therefore preferred. ZHZ marriage and sororal polygyny are also ruled out here, as they are thought to be susceptible to future quarrels among those involved (ibid.: 165). In fact, the importance of affinity is such that Brain finds it easy to distinguish wife-takers from wife-givers in what he calls a relationship of ‘asymmetrical alliance’ (ibid.: 166), the former evidently being patrikin and dependents of the latter, who are generally wealthier (ibid.: 114). He even goes so far as to suggest that the Bangwa are closer in this respect to the Lakher and similar groups in Asia practising matrilateral cross-cousin marriage than to African near-neighbours like the Tiv and the Tallensi (ibid.: 187), though there is no suggestion of a prescriptive system here, nor even of any preference for marriage to cousins. More obviously Omaha in type is the terminology of the Konkomba of northern Ghana (Tait 1961: 133–35). Thus, umwidza is MB, MBS, MBSS, ZS, and nawa is MZ, MBD, MBSD. Other terms with cross-generational equations are tshin, FZH, ZH, WB, WZ; tshuor (ms), WFF, WF, WM, WFFZ, WFZ, WMB; and tshapi (ws), HFZ, HZ. Some of these latter equations also occur in Omaha-type terminologies, such as FZH, ZH and WB, also WF and WFF (Barnes 1984: end chart for male ego). Tait does not explicitly acknowledge the presence of Omaha equations here, but his text does describe circumstances in which whole lineages may be treated as units, regardless of generation. Also, he states that ‘It is clear that kinship terms are extended through the generations only in the lineage of Ego’s mother’ (ibid.: 134). Some of the circumstances where this occurs relate to the role of matrikin in mortuary rituals (ibid.: 134–36), and, in general, matrilateral ties are clearly important (ibid.: 149, 150). The Omaha equations may also go along with marriage restrictions, as in Lévi-Straussian theory. Thus, FM’s lineage or ‘extended house’ is

Substitutability of Kin and the Crow-Omaha Problem

banned as a source of wives for male ego (ibid.: 78, 102), and MM’s lineage might be banned as well (but cf. pp. 78 and 102), as it certainly is for female ego. The ‘extended house’ of FZH is also banned for both male and female egos (ibid.: 103). More generally, repeating alliances between the same kin groups is only possible in the third generation (ibid.: 106). Exchange marriages are allowed (almost certainly within the same generation) but are rare, at a rate of less than 5% of all marriages (ibid.: 95). Also of Ghana, on the central part of the coast, are the Fanti, studied over many years by David Kronenfeld, who has also commented on the skewing of the Crow-Omaha phenomenon more generally. The Fanti are one of the Akan peoples, closely related to the Ashanti, with a population of over a million. They have matrilineages, called ebusua, grouped into a small number of matriclans. Kronenfeld (2009: Ch. 1 [1973]; Ch. 2 [1980]) claims to have uncovered three different terminological structures or ‘subsystems’ (2009: 16) among them.13 One of them need not detain us here, since really it amounts to no more than the ‘extension’, in his words, of terms for focal kin types outwards to remoter relatives of the same and even unrelated alters who can be assimilated to them in terms of age, gender etc. – a familiar enough situation worldwide. More particularly, for our purposes, there is a terminology in day-to-day use that tends to the generational in pattern, plus a Crow terminology that is less used but is seen as more ‘correct’ (ibid.: 17, 49). The Fanti’s Crow terminology has the following minimal diagnostic equations: na FZ, FZD, FZDD, FZDDD (also M, MZ – perhaps influenced by the generational [‘unskewed’ in Kronenfeld’s words] terminology – and other specifications); egya F, FB, FZS, FZDS, FZDDS; and awofasi GC, MBC (both the latter with other specifications as well) (ibid.: 20, Figs. 1–3; also ibid.: 37). Clearly the Crow (‘skewed’) terminology is linked to the assimilation of generations, but the Fanti themselves link one of its key features, the MBC = C equation, to potential matrilineal inheritance, something they talk about and reflect on a lot (ibid.: 33). However, neither the rules nor actual practice conform to a simple model of matrilineal inheritance from, for example, MB to ZS. Although the deceased’s matrilineage takes charge of that portion of the inheritance the deceased himself inherits14 (Kronenfeld adopts a solely male perspective in his description), it generally observes a hierarchy of heirs. At the top of this hierarchy are the younger brother(s) of the deceased, as is common in West Africa, a practice that involves



245

246

How Kinship Systems Change

inheritance within the same generation, not between generations. Inheritance between generations only becomes an issue when the last of the brothers dies, whereupon the inheritance goes to the eldest son of the eldest sister, then to the eldest son of the next eldest sister, and so on down the age hierarchy within the same generation until, on the exhaustion of that generation, it goes one generation lower, to the eldest son of the eldest sister’s eldest daughter, and so on. Despite the adelphic inheritance, generation is seen as fundamental in the sense that a nephew (presumably ZS is meant here) can inherit from an uncle (MB?), but not vice versa. In practice, however, because of the priority given to adelphic inheritance, the small size of movable legacies, which are soon dispersed or used up, and the unlikelihood that the potential residual heirs will live long enough to inherit anything, cross-generational inheritance is rare, only occurring, if at all, late in the inheritor’s life (ibid.: 33–34).15 The link between matrilineal inheritance and the Crow terminology, therefore, seems to reside more in people’s minds and their collective thought than in reality. However, it is reinforced by the theoretical ability of a ZS to inherit the widow(s) and children of his MB, which Kronenfeld calls ‘an extended form of the levirate’ designed to give the survivors ‘a kind of old-age insurance’ (ibid.: 34). This too, though, is rare, with widows generally preferring to remarry or return to their natal homes as dependents of their male kin. There are also some ritual aspects to this relationship between ego and MBW/MBC, but in general the circumstances described above do not seem to be enough to account for the Fanti’s Crow terminology in a sociological sense, being limited to giving the Fanti a cognitive understanding and set of symbolic meanings for the underlying principles of their society.

The Secondary Marriage Hypothesis (#2 Above) To understand fully the implications of WBD being a preferred secondary spouse in some societies with at least Omaha-type terminologies (Kohler’s hypothesis, 1975 [1897]), we need to look at marriage practices in such societies in a general sense. A particularly good place to start, given the ample evidence on them (cf. Barnes 1984), are the Omaha themselves. The Omaha are well known for marriage prohibitions that rule out the repetition of a marriage alliance between the same groups

Substitutability of Kin and the Crow-Omaha Problem

in the immediately following generations, viz. for a man, his mother’s clan and the subclans of his FM, FMM, MM and MMM (Barnes 1984: 163). This has the effect of dispersing alliances among several alliance groups over time. Similar though not always identical prohibitions have become a defining feature of Crow-Omaha systems for many authorities, especially followers of Lévi-Strauss (cf. 1966), though they are not exclusive to them – that is, they are not necessarily associated with the vertical equations that characterize Crow-Omaha terminologies by definition. In the actual Omaha case, at least, Barnes (1984) says that these restrictions relate to close lineal kin, not their classificatory equivalents. However, these restrictions do not prevent the repetition of marriages between descent groups in the same generation and may well go along with them. Indeed, in the Omaha case a group of lineage brothers may well marry a group of lineage sisters in another descent group within the same generation. One way of expressing this is to say that such marriages take place between BWZ and ZHB: in both cases, the assumption is that at least one marriage has taken place already in each direction. However, while Barnes (1984) considers marriage practices to be symmetric overall, sister exchange is not permitted and is evidently not practised. Thus, while marriage between BWZ and ZHB (asymmetric) is allowed, marriage between ZHZ and BWB (symmetric) is not (NB: none of these specifications is given against any kin terms in the lists in Barnes, 1984: 132 ff., and Barnes himself does not mention them). However, the actual Omaha terminology does have the equations F = FB = MZH; M = MZ = FBW; and FZ = MBW; though not MB = FZH. These specifications also form part of two classic Omaha-type vertical series respectively. It is tempting to see this as a relic of a prior period of prescription, though there is little evidence of this; but it is also consistent with the situation, just noted, in which descent lines in a single generation intermarry intensively but do not repeat the analogous marriages in later generations, as indicated by the lack of any affinal specifications for these terms. From ego’s point of view, this could still be construed as representing the intermarriage of sibling sets. This would also be the case with cross-cousin marriage, but in that case cross cousins are clearly repeating the marriages of the previous generation and are definable because of this; thus, one would also expect cross-cousin specifications to be linked terminologically with affinal ones. This may not apply neatly to the actual Omaha case, given the ban on



247

248

How Kinship Systems Change

direct exchange mentioned above, but it could apply to other similar cases without this restriction. But that is not all: among the actual Omaha, a man may marry several women from the same patrilineal descent group but from different generations in what by definition will not be his first marriage. This is because they are all defined as wife’s relatives, indicating that he has already married at least once. Thus, WZ, WFZ and WBD are terminologically equated as ihon’ga (reciprocally, ZH, BDH and FZH are equated under the term ishi’e). From the point of view of ego’s children, the former group of specifications translate into MZ, MFZ and MBD, which are accordingly terminologically equated as ihon’, though their reciprocals, ZS, BDS and FZS, are distributed between two different terms. Such marriages are far from universal, but they do appear to reflect known if subsidiary marriage preferences among the Omaha. A low take-up rate is not unusual for such preferences, and even first cross-cousin marriage has repeatedly been shown to be a minority practice, even where it is prescribed. One way of looking at this phenomenon is to return to the practice of BWZ–ZHB marriages noted above. Another way of conceiving these is to say that male ego finds a wife for his brother in the form of his own WZ. In marrying polygynously into his wife’s descent group, however, he provides that wife not for his brother but for himself.16 Kohler showed at the end of the nineteenth century (1975 [1897]) that many of the equations and distinctions in the actual Omaha terminology could be explained in terms of a preference for the polygynous marriages just mentioned, and Barnes subsequently identified yet more such equations (1984: 158–59). As already noted above, authorities coming after Kohler, especially Kroeber, Radcliffe-Brown, Murdock and Lévi-Strauss, rejected this explanation because it depended on preferences for marriages subsequent to ego’s first marriage, marriages they considered less common than the first marriage (though in some cases they may still be practised fairly often, as among the Omaha and Mapuche; see respectively, Barnes 1984: 156 and passim; Faron 1956, 1961). Nonetheless, the structure of the actual Omaha terminology does correspond to the regular operation of these preferences without being defined by it in the same way as a prescriptive terminology would be, as Barnes makes clear. Barnes therefore appears to accept Kohler’s basic premise, though does so rather cautiously (1984: 175, 217).

Substitutability of Kin and the Crow-Omaha Problem

The fact that Barnes (1984: Table 3) found many more equations to add to those Kohler uncovered expressing the secondary marriage preferences discussed above for the Omaha terminology means that the actual Omaha terminology expresses this sociological phenomenon as much as it does patrilineal unity, with which it is more usually associated. However, as Barnes points out (1984: 149, 151), the two phenomena are connected, given that some vertical equations in the actual Omaha terminology could be interpreted as uniting a line of patrilineally linked women with whom male ego might contract a second marriage. Similarly, from the point of view of ego’s children, these women are also equated by the identical circumstance that they are women whom their father might marry secondarily. There also appear to be other examples of terminologies with vertical equations uniting potential marriage partners. Gifford’s early work on the Omaha-type terminology of the Miwok of California has long been seen as pointing in that direction in the context of the rather rare WBD marriage (1916; otherwise MBD marriage occurs here), as does Rivers’ work (1914) on the Banks Islands in Melanesia, where the relevant category is MBW (reciprocally HZS) in a Crow terminology (both cases are summarized by McKinley 1971: 235–36). Not surprisingly, the Osage, closely related to the Omaha, also have evidence of preferential second marriages to WBD and WFZ, and M and MFZ, at least, have the same term (Nett 1952). Finally, the terminology of the Tlingit, a matrilineal people of north-west North America, has the Crow equations FZ = FZD and FB = FZS, which combine the marriage preferences for male nobles and commoner women respectively. In the former case, this is strengthened by the fact that another preference is for marriage to the wife of a deceased MB (which McClennan 1961: 114 calls ‘nepotic widow inheritance’; cf. also the Banks Islands case, just mentioned), who would ideally also be FZ here. This is expressed by a symmetric inter-cognate equation in the +1 level of the terminology despite the distinction between patrilateral and matrilateral cross cousins in ego’s genealogical level (cf. also FF = MMB in +2). One motive here is the desire, in this matrilineal society, for male ego to marry into his father’s line (which would be different from his own, of course), which is achieved through FZD marriage. However, neither preference is stated to be a rule or prescription (De Laguna 1952; McClennan 1961). McClennan (ibid.) adds the information that FZ marriage is quite common, that FB/BD



249

250

How Kinship Systems Change

marriage is also known, and that fraternal polyandry also takes place here. Further, the neighbouring Atna are said to have cross-cousin marriage with a clear preference for FZD over MBD, though whether this amounts to prescription, and whether the Atna are matrilineal too, is not made clear. The Haida of the north-west coast, who are matrilineal, may have a preference for FZD marriage except among chiefs, who inherit their MB’s wives as well as MB’s positions; there is no evidence of FZ marriage (Moore 1963: 302–3). In the Solomon Islands, too, the matrilineal Baniata have a Crow terminology and MBW as a possible marriage partner, alongside third cousins (Scheffler 1972: 350–52). A rather similar situation is also found among the Pende of the Kasai province of the former Belgian Congo, for whom, at least at the time of De Sousberghe’s study (1955), there was a strong preference for FZD marriage so as ‘to see our father’s face again’ (ibid.: 47 ff.) – that is, to return to the father’s clan through marriage. Indeed, the same term is used for both FZD and ‘wife’. Formerly (perhaps as an alternative) there is also said to have been a preference for FZ in marriage for a similar reason. Thus, FZ and FZD can be linked as marriage preferences for a male ego, though they appear to have separate terms among the Pende, and no Crow-type equations are evident in De Sousberghe’s description (he does not give a full terminology). Marriage is devalued in this strongly matrilineal society, as is the case in some other cases of societies with matrilineal descent. Another example is the BaVenda of the Transvaal–Zimbabwe borderlands, studied, apparently in the 1920s, by Hugh Stayt (1931). In the terminology, there are a number of equations between generations equating, for example, +1 affines with +2 consanguines, and affines of ego’s-level with +1 consanguines, but there are also vertical equations of interest in the context of Omaha-type terminologies, though without expressing the classic equations specifically. Thus, the common term in Bantu languages for MB, malume, is here also the term for WB, WBS (which has an alternative descriptive term) and WBSS, the latter three being equivalents of MBS, MBSS, MBSSS when MBD marriage is practised, as is the case here, though non-prescriptively. Other terms of interest are muzadzana (literally ‘little wife’) WyZ, WBD, WBDD; and mukwasha or muduhulu FZH, ZH, DH, GDH, DDH, HZH. There are also equations between +1 parallel kin consistent with, but not diagnostic of, prescriptive marriage, namely FB = MZH and MZ = FBW (with an age distinction).

Substitutability of Kin and the Crow-Omaha Problem

As just noted, the basic form of marriage today is with MBD, ‘practiced wherever possible, and . . . an essential feature of the society’ (ibid.: 175). There are occasional marriages to FZD, but these are considered wrong; nonetheless, the two cross cousins share the same term, muzwala. Of more significance here are marriages to WBD, nowadays rather few but which Stayt speculates may formerly have been more common: ‘Today, if [a man] has a son, this woman must be given to the son [as MBD], and only if he has no son is he allowed to keep her himself [as WBD]’ (ibid.: 177). Evidently, male ego has a claim on the woman for either himself or his son. Stayt puts forward a rather involved explanation for the equation linking MB to WB, WBS and WBSS. It is basically rooted in the fact that if male ego obtains a wife for his son from his (ego’s) WB, a woman who is ego’s son’s MBD, ego is initially regarded as the bride’s legal husband by virtue of having paid the brideprice (lobola) for her. Only later does his son act as the husband. Nonetheless, over time the bride has two husbands in successive generations of the same patriline. Kin-term use therefore becomes skewed because of this feature, with the equating of MB etc. apparently being due to the possibility of repeat marriages between the same two patrilines down the generations (see diagram, ibid.: 177). For example, if male ego’s MB and WB both regard ego as their ZH or DH (themselves having the same term; see above), he will reciprocate by equating them terminologically. There is also the same temporary skewing due to the bride having ego’s father, not ego, as her initial husband. This example may be useful in understanding whether or not Omaha terminologies in particular can be derived from terminologies of asymmetric prescriptive alliance (in light of the frequency of non-prescriptive MBD marriage here), and if so, how. However, one thing Stayt misses is the fact that the lowest marriage on his diagram, between the individuals marked as c and X2, is actually between FMBSD and FFZSS, thus expressing the repetition of a marriage two generations previously but with a gap in between (see above for the Haya; also found elsewhere, as noted where relevant in this chapter). In this case, the marriage in the interim generation is ultimately one between FZS and MBD, not between MBS and FZD, as in the Haya case (q.v.). Stayt claims that WBD marriage is common in other Bantu societies as well. His argument is that male ego has a claim to her if his wife proves unsatisfactory, though actually the claim is to WBW,



251

252

How Kinship Systems Change

whom WB may well be reluctant to give up. This argument is vitiated by the further information that, at least among the BaVenda, WBW is strictly tabooed even when it comes to conversation with her, let alone marriage. More likely, then, the claim really lies with WB, who is expected to provide ego with a replacement wife himself if asked to do so. In any case, we are also told that the BaVenda place a priority on WZ over WBD in these circumstances. Given the tendency to repeat marriage to MBD, she could presumably also be a sibling of WB if ego’s patriline had taken more than one wife from WB’s line. In his work on the Mapuche of southern Chile, Faron (1956, 1961) makes similar observations, as he sees WBD marriage in particular as important in certain, usually unsettled periods of Mapuche history as a way of cementing alliances between kin groups when the main alternative, MBD marriage, was ruled out for some reason or had become less popular. For Faron, it is part of a bundle of practices that includes polygyny more generally and the levirate and sororate as well, an explanation that is potentially relevant to other cases.17 But he is also concerned to try and establish changes in terminological pattern over time and the reasons for them, since the Mapuche have only had an Omaha terminology in more recent periods of their traceable history. It is also clear that some of the differences in terminological patterns reflect regional variation, and some of Faron’s reasoning for terminological change is somewhat speculative. Despite this, very real changes can be discerned in a chronology that, thanks to early Spanish dictionaries, can be traced back to about AD 1600. Thus, at that time, and indeed well into the eighteenth century, the terminology, such as we have a record of it, seems to have been mainly generational, especially in ego’s generation, though with a partly bifurcate merging pattern in +1 and a few +2/−2 equations. By the early twentieth century, the pattern had become Dakota-like (in Faron’s terms) in ego’s level, and indeed in the other two medial levels too. By the middle of the twentieth century, Faron himself was collecting terminologies with some Omaha equations, namely M = MBD and MBW = MM, MFZ, though not MB = MBS or FZC = ZC.18 Unfortunately, what Faron means by ‘Dakota’ is not entirely clear, but it appears to have been the same as Dravidian rather than Iroquois in Lounsbury’s sense of the latter term (1964b). Also, given the almost complete lack of affinal terms in any of Faron’s sources, including the terminologies he collected himself, we are prevented from diagnosing any of these terminologies properly.

Substitutability of Kin and the Crow-Omaha Problem

Another problem is the matching of marriage practices with these various terminologies. Fundamentally, Faron appears to believe in a chronology that starts in recorded history with occasional MBD marriage, which is mentioned in eighteenth-century sources. Then we have a period of WBD marriage as settled affinal alliance groups in the form of patriclans break down in the face of warfare with the Spanish and later the Chilean authorities. The argument here is that WBD marriage was better able to perpetuate alliances between families under these circumstances and that it was part of a package of alliance maintenance combining polygyny more generally with the levirate and sororate. Later, in the more settled period of the twentieth century, patrilines were re-established, and WBD marriage gave way to MBD marriage, which is the preferred mode of affinal alliance today, though still not prescriptive (Faron reports no asymmetric prescriptive terminology). However, the Omaha equations in the later terminology reflect the earlier period of WBD marriage, while the earlier period of MBD marriage allegedly coincided with a longer period when the Mapuche terminology was generational – that is, the chronologies of terminological change and change in marriage patterns and preferences do not coincide.19 Moreover, only WBD is at issue as a marriage preference, not also WZ and WFZ, as for the Omaha. At all events, for Faron the Mapuche are an example of a society where male ego can take WBD as his own wife, possibly at the expense of his own son (for whom she would be MBD), at least in certain periods. And despite the apparent lack of an asymmetric prescriptive terminology for any period, the standard diagram of such a terminology is only one of two in which male ego’s WBD is his son’s MBD, the other being the diagram of symmetric prescriptive alliance (assuming the male perspective is followed). This is not the case with the theoretical diagram of patrilateral cross-cousin marriage (FZD/ MBS), while among the Omaha, MBD, WBD and SW all have different terms and are in different lines in a manner that, expectedly, is not at all prescriptive (Barnes 1984).20

Conclusion The examples discussed in the first ethnographic section above show, with varying degrees of clarity, that Crow-Omaha equations express lineal unity not just because lineal unity itself exists, as Radcliffe-



253

254

How Kinship Systems Change

Brown tended to assume, but because lineages and clans have relationships of affinal alliance with one another. These relationships are often initiated through a marriage that cannot be repeated in the immediately following generations, but they also tend to be based on mutual ritual services or material claims that may continue for several generations or even indefinitely, with successive kin in the same descent line replacing one another over time in exerting these claims. The whole lineage they belong to therefore assumes an identity in its own right, which may or may not also be expressed in the terminology. Therefore, this remains a partial explanation at best for the existence of Crow-Omaha terminologies, not a definitive one, and in all likelihood this is all we shall ever have. Many of the examples in the second ethnographic section above, with their focus on secondary marriages to wife’s kin, while relevant can ultimately be seen as special instances of the substitutability hypothesis. At least that is the case with the Omaha, where a patrilineally linked group of wife’s relatives are suitable partners. That is ruled out in the Mapuche case, since only WBD is at issue. Again, this is only a partial explanation for the marriage patterns in societies with Crow-Omaha terminologies, a theme that deserves a study in its own right. Terminologies like this express lineal unity by definition, but, as I hope to have shown here, it is not enough to leave it at that. Ultimately, however, marriage preferences and lineal unity need to be combined in a single analysis if we are to arrive at a plausible hypothesis for the existence of Crow-Omaha terminologies. The possibility of renewing an alliance after the lapse of two or three generations (perhaps eighty to a hundred years) and the extent to which this happens are also important, as is the implied cyclicity involved in these practices. Indeed, another reason for the unity of some lineages that are marked out by Crow-Omaha equations is presumably the fact that marriages with them are impossible at present but will become possible at some point in the future, even if that future is several decades away. In this way, if in no other, Crow-Omaha terminologies express lineal unity over long time periods. In this respect, they are unlike cross-cousin marriage and the prescriptive terminologies associated with them, where, at least formally, the repetition of an alliance is expected in the following generation, and the return to the starting point represented by an initial marriage accordingly takes place much sooner.

Substitutability of Kin and the Crow-Omaha Problem

Notes * First published in Structure and Dynamics: eJournal of Anthropological and Related Sciences 11(1) (2019): 73–95. Reproduced with permission. 1. A different situation might arise where women become members of their husband’s patrilineages on marriage, which need not cancel their membership of their natal lineage. One possible indicator of such membership is where the woman is buried after her death. 2. I use Labby’s spellings where they differ from Schneider’s. I do not provide a full analysis of this terminology, nor indeed of any other, in this article. 3. The diagnostic terms in Crow-Omaha terminologies do not always restrict themselves to expressing lineal unity and may well have referents in more than one terminological line. See, for a good example, Barnes on the Omaha themselves (1984: Tables 9, 10). 4. There was a vagueness about some of these terms among Labby’s informants, as he indicates on his main chart of the terminology (1976: 65, Fig. 25). 5. A recent study by Mosko (2017: 366) also draws attention to the equivalences between kin involved in urigubu payments, in which they may stand in for one another (especially junior kin for senior kin), though he does not relate that to the Crow shape of the terminology. In any case, the kin term tabu is evidently not relevant here: the kin types involved are linked by real or classificatory father-son ties, not at all like a Crow system, nor a matrilineal one, in so far as a man’s successor within the matriline is his sister’s son, not his own son. 6. Malinowski himself says that strictly speaking urigubu payments are owed by a man to his married sister so as to signal that the latter, like her children, retains rights in her matrilineage’s land despite her moving away from it in a patrilocal marriage; also, when she dies, these payments cease, at least in principle. It is also clear, however, that in effect they also benefit her husband and possibly some of his own kin as well. These payments should also be seen in light of the strict rule of avoidance between brother and sister. These issues are discussed in Malinowski 1929, 1935. 7. In a private conversation to which I was a party back in the 1980s, Rodney Needham, who became notorious for his dismissal of ‘Crow-Omaha’ as a ‘type’ of kinship system (e.g. 1971), suggested that Bowden’s interpretation was as close as we were likely to get to an explanation for these terminological features. It was also crucial in my own epiphany in relation to the theme of this chapter. 8. The actual Omaha system also rules out direct exchange, which formally equates wife-givers and wife-takers. See Barnes (1984). 9. This equivalence does not appear in the Kwoma terminology, in which FZD is ruwey and FMBSD is yey. Marriage to FMBSD nonetheless appears as a possible marriage option in many societies with Crow-Omaha terminologies. 10. The asymmetry of Kwoma practices is anyway a feature of all Omaha-type terminologies. All Crow-Omaha terminologies characteristically distinguish patrilateral and matrilateral cross cousins in making their vertical equations,



255

256

How Kinship Systems Change

11. 12. 13.

14.

15. 16.

17.

18. 19.

20.

and indeed they distinguish groups of wife-takers and wife-givers more generally. However, they do so without MBD marriage (usually banned in itself) or the repetitions of alliance between groups generation after generation that this asymmetric form of marriage prescription entails. Here, too, it was Bowden’s account of the Kwoma that was instrumental in my realizing what was involved in these Indonesian cases. This may therefore be an instance of Kronenfeld’s notion of layering (2009). Kronenfeld also claims to have found a doubling of skewed and unskewed terminologies among the Ashanti, another Akan people (ibid.: 49). He also mentions a fourth pattern among the Fanti, namely their use of some English terms to modify the existing patterns (ibid.: 49, 313 ff.). He links this to legal changes introduced by the modern Ghanaian state favouring some inheritance from father to son even among matrilineal peoples. Some of the property the deceased acquired himself may go out of the lineage to his wives and children (ibid.: 33). Elsewhere, Kronenfeld indicates that such transfers usually take place during the deceased’s lifetime (ibid.: 311). One exception appears to be the ‘stools’ or chiefly positions (ibid.: 312). This, incidentally, recalls a possible explanation for ZD marriage in parts of south India and the Amazon, according to which male ego takes his ZD as a wife not for his son but for himself, ‘anticipating’, as it were, ZD’s marriage. The levirate involves a married woman taking her husband’s brother as a spouse, normally after her husband’s death. The sororate involves a man taking his wife’s sister (possible a ‘lineage’ sister) as a second wife for himself, often polygynously. The second wife could, however, be another relative of the first, like WBD. This seems to be the burden of Faron’s argument here. There are some contradictions between Faron’s different articles: I am relying on the full terminology set out diagrammatically at 1956: 449. In 1967, Needham re-analysed the Mapuche case, concluding that the terminology was not prescriptive and that MBD/FZS marriage was merely preferential. However, I prefer to go back to Faron’s original accounts rather than rely on Needham here, as the latter had his own agenda and only partially succeeded in explaining the historical complexities of the Mapuche case. Gabriela Piña, at the time of writing the original paper in 2018 a Ph.D. student at the London School of Economics who has done fieldwork with the Pehuenche subgroup of the Mapuche since 2011, has informed me that while some MBD marriages take place she has encountered no instances of WBD marriage. I am most grateful to Ms Piña for this information.

References Barnes, R.H. 1976. ‘Dispersed Alliance and the Prohibition of Marriage: Reconsideration of McKinley’s Explanation of Crow-Omaha Terminologies’, Man 11: 384–99.

Substitutability of Kin and the Crow-Omaha Problem

———. 1984. Two Crows Denies It: A History of Controversy in Omaha Sociology. Lincoln, NE and London: University of Nebraska Press. Beattie, John. 1957. ‘Nyoro Kinship’, Africa 27(4): 317–40. ———. 1958. ‘Nyoro Marriage and Affinity’, Africa 28(1): 1–22. ———. 1960. Bunyoro: An Africa Kingdom. New York: Holt, Rinehart and Winston. Bowden, Ross. 1983. ‘Kwoma Terminology and Marriage Alliance: The “Omaha” Problem Revisited’, Man 18(4): 745–65. Brain, Robert. 1972. Bangwa Kinship and Marriage. Cambridge: Cambridge University Press. De Laguna, F. 1952. ‘Some Dynamic Forces in Tlingit Society’, Southwestern Journal of Anthropology 8(1): 1–12. De Sousberghe, R.P.L. 1955. Structures de parenté et d’alliance d’après les formules Pende (ba-Pende, Congo belge) (Mémoires de l’Académie Royale Sciences Coloniale Belge). Gembloux: Ducolot. Faron, Louis. 1956. ‘Araucanian Patri-organization and the Omaha System’, American Anthropologist 58(3): 435–56. ———. 1961. ‘The Dakota-Omaha Continuum in Mapuche Society’, Journal of the Royal Anthropological Institute 91(1): 11–22. Fathauer, G.H. 1961. ‘Trobriand’, in David M. Schneider and Kathleen Gough (eds), Matrilineal Kinship. Berkeley and Los Angeles: University of California Press, pp. 234–69. Forge, Anthony. 1971. ‘Marriage and Exchange in the Sepik: Comments on Francis Korn’s Analysis of Iatmul Society’, in Rodney Needham (ed.), Rethinking Kinship and Marriage. London: Tavistock Publications, pp. 133–44. Gifford, Edward W. 1916. Miwok Moieties. Berkeley: University of California Press. Goodenough, Ward H. 1951. Property, Kin, and Community on Truk. New Haven: Yale University Publications in Anthropology, no. 46. Kohler, Josef. 1975 [1897]. On the Prehistory of Marriage: Totemism, Group Marriage, Mother Right, trans. R.H. Barnes and Ruth Barnes. Chicago: The University of Chicago Press. Kronenfeld, David. 2009. Fanti Kinship and the Analysis of Kinship Terminologies. Urbana and Chicago: University of Illinois Press. Labby, David. 1976. The Demystification of Yap: Dialectics of Culture on a Micronesian Island. Chicago and London: The University of Chicago Press. Leach, Edmund. 1958. ‘Concerning Trobriand Clans and the Kinship Category “Tabu”’, in J. Goody (ed.), The Developmental Cycle of Domestic Groups. Cambridge: Cambridge University Press, pp. 120–45. Lévi-Strauss, Claude. 1966. ‘The Future of Kinship Studies’, Proceedings of the Royal Anthropological Institute 1965: 13–22. Lounsbury, Floyd G. 1964a. ‘A Formal Account of the Crow- and Omaha-Type Kinship Terminology’, in W. Goodenough (ed.), Explorations in Cultural Anthropology: Essays in Honor of George Peter Murdock. New York: McGraw-Hill, pp. 351–93. ———. 1964b. ‘The Structural Analysis of Kinship Semantics’, in Horace G. Lunt (ed.), Proceedings of the Ninth International Congress of Linguists. The Hague: Mouton, pp. 1073–93.



257

258

How Kinship Systems Change

———. 1965. ‘Another View of the Trobriand Kinship Categories’, in E.A. Hammel (ed.), Formal Semantic Analysis: American Anthropologist Association Special Publication (Vol. 67, no. 5, part 2), pp. 142–85. Malinowski, Bronisław. 1929. The Sexual Life of Savages in Northwestern Melanesia. London: George Routledge. ———. 1935. Coral Gardens and their Magic. London: Allen and Unwin. McClennan, C. 1961. ‘Avoidance between Siblings of the Same Sex in Northwestern North America’, Southwestern Journal of Anthropology 17: 103–23. McKinley, Robert. 1971. ‘A Critique of the Reflectionist Theory of Kinship Terminology: The Crow/Omaha Case’, Man (n.s.) 6(2): 228–47. Moore, Sally Falk. 1963. ‘Oblique and Asymmetrical Cross-Cousin Marriage and Crow-Omaha Terminology’, American Anthropologist 65: 296–312. Mosko, Mark S. 2017. Ways of Baloma: Rethinking Magic and Kinship from the Trobriands. Chicago: Hau Books. Murdock, George P., and Ward H. Goodenough. 1947. ‘Social Organization of Truk’, Southwestern Journal of Anthropology 3(4): 331–43. Needham, Rodney. 1967. ‘Terminology and Alliance II: Mapuche, Conclusions’, Sociologus 17: 39–54. ———. 1971. ‘Remarks on the Analysis of Kinship and Marriage’, in Rodney Needham (ed.), Rethinking Kinship and Marriage. London: Tavistock Publications, pp. 1–34. ———. 1973. ‘Prescription’, Oceania 43: 166–81. Nett, B.R. 1952. ‘Historical Changes in the Osage Kinship System’, Southwestern Journal of Anthropology 8: 164–81. Parkin, Robert. 2018. ‘The Evolution of Kinship Terminologies: Non-prescriptive Forms of Asymmetric Alliance in Indonesia’, Journal of Anthropological Research 74(2): 232–51. Radcliffe-Brown, A.R. 1952. Structure and Function in Primitive Society: Essays and Addresses. London: Cohen & West. Read, Dwight, and Cliff Behrens. 1990. ‘KAES: An Expert System for the Algebraic Analysis of Kinship Terminologies’, Journal of Quantitative Anthropology 2: 353–93. Reining, Paula. 1972. ‘Haya Kinship Terminology: An Explanation and Some Comparisons’, in P. Reining (ed.), Kinship Studies in the Morgan Centennial Year. Washington DC: The Anthropological Society of Washington, pp. 88–112. Rivers, W.H.R. 1914. Kinship and Social Organization. London: Constable. Scheffler, Harold. 1972. ‘Baniata Kinship Classification: The Case for Extensions’, Southwestern Journal of Anthropology 28: 350–81. Schneider, David. 1953. ‘Yap Kinship Terminology and Kin Groups’, American Anthropologist 55: 215–36. ———. 1972. ‘What is Kinship All About?’ In Paula Reining (ed.), Kinship Studies in the Morgan Centennial Year. Washington DC: Anthropological Society of Washington, pp. 32–63. ———. 1984. A Critique of the Study of Kinship. Ann Arbor: The University of Michigan Press. Stayt, H.A. 1931. The BaVenda. London: Humphrey Mitford and Oxford University Press for the International Institute of African Languages and Cultures.

Substitutability of Kin and the Crow-Omaha Problem

Tait, David. 1961. The Konkombra of Northern Ghana. London: Oxford University Press. Tambiah, Stanley J. 1973. ‘Dowry and Bridewealth and the Property Rights of Women in South Asia’, in Jack Goody and Stanley J. Tambiah (eds), Bridewealth and Dowry. Cambridge: Cambridge University Press, pp. 59–166. Tiemann, G. 1970. ‘The Four-Got Rule among the Jat of Haryana in North India’, Anthropos 65: 166–77. Trautmann, Thomas R., and Peter M. Whiteley (eds). 2012. Crow-Omaha: New Light on a Classic Problem of Kinship Analysis. Tucson: The University of Arizona Press. Weiner, Annette. 1979. ‘Trobriand Kinship from Another View: The Reproductive Power of Men and Women’, Man 14: 328–48. Williamson, Margaret H. 1980. ‘Omaha Terminology and Unilateral Marriage on the Sepik’, American Ethnologist 7: 530–48. Winter, Eric H. 1956. Bwamba: A Structural-Functional Analysis of a Patrilineal Society. Cambridge: W. Heffer & Sons.



259

9 

The Evolution of Kinship Terminologies Non-prescriptive Forms of Asymmetric Alliance in Indonesia

Introduction Indonesia’s significance to the anthropology of kinship is partly due to the fact that some of its regions (principally the eastern island chains and certain societies in Sumatra in the west) have proved to be major areas of the world in which marriage between classificatory mother’s brother’s daughter and classificatory father’s sister’s son is practised.* This system has variously been called circulating connubium (by early Dutch anthropologists), generalized exchange (Lévi-Strauss 1949) and asymmetric prescriptive alliance (early Rodney Needham). It was early identified by Edmund Leach among the Kachin in the highlands of Upper Burma (now Myanmar) (1961 [1945]), becoming a textbook example of this form, but was also described by a number of Dutch anthropologists in works going back to the nineteenth century, the best known internationally being F.E. van Wouden (1968 [1935]). Later this topic of research was taken up by a worldwide community, many of them discussed below, but represented in Britain especially by Needham and R.H. Barnes, of the University of Oxford (now deceased and retired respectively). As already noted, one can find plenty of societies with asymmetric systems of affinal alliance underpinned by a prescriptive terminology,

The Evolution of Kinship Terminologies

usually itself also asymmetric in form, in eastern Indonesia and parts of Sumatra; these have been discussed and compared extensively, and I shall define them as prescriptive in what follows. Less well known, and certainly less frequently compared, are societies whose marriage practices often resemble asymmetric prescriptive systems in certain details but that lack a clear rule that classificatory MBD/ FZS marriages are the proper ones to follow and, importantly, that lack a prescriptive terminology; I shall call these non-prescriptive in what follows but also use the abbreviation EAA for them, standing for ‘evolved asymmetric alliance’, to reflect evolutionary possibilities. Formally speaking, indeed, the similarities between the two situations are sufficient to make the marriage practices of the latter group of societies seem as if they have derived from the prescriptive systems of the former group, and the strong version of this argument is indeed that this process of derivation is likely to have been historical and not just formal or logical. One should not, therefore, overemphasize the dichotomy or differences between the two situations: they coexist in the same geographical regions, as well as in an ideological environment that generally prohibits the direct exchange of spouses between affinal alliance groups so that wife-givers are distinguished from wife-takers, and that accords the former a superior status to the latter. Nonetheless, they differ in types of terminology (i.e. prescriptive versus non-prescriptive) and in the extent to which they can be said to have a positive marriage rule of asymmetric type. While there may be uncertainty over which of these two groups particular examples belong in, not least because the terminologies are not always recorded, I argue that there is a distinction to be made here and that an attempt ought to be made to account for it. One aim of this chapter is therefore to start filling the gap in scholarship I have just identified, part of a long-standing interest of mine in the evolution of kinship terminologies and affinal alliance systems as systems of classification. Another aim, however, is to respond to a recent suggestion by Thomas Trautmann (2012) that Crow-Omaha terminologies may have derived from asymmetric prescriptive terminologies on the basis that the latter sometimes have very similar vertical equations (for a more detailed argument, see Ch. 7). However that may be in other parts of the world, it seems less plausible in the case of eastern Indonesia and Sumatra, partly because of the very existence of EAA systems here, which provide an alternative, and partly because these two regions lack known Crow-Omaha terminologies



261

262

How Kinship Systems Change

entirely. I will return to Trautmann’s theory at the end of the chapter after reviewing some relevant ethnographic studies. There is certainly one similarity between EAA systems and CrowOmaha I ought to address here first, however, and that is in respect of the latter’s practices and rules encouraging dispersed alliance, highlighted especially by Lévi-Strauss (1966), though he gives them the term ‘semi-complex’ structures. For example, there is frequently a ban on male ego marrying into his mother’s line in a society that traces descent patrilineally: her marriage to ego’s father constitutes the initial, permitted alliance, which male ego should not repeat, as he should, formally speaking, in prescriptive systems. Very often male ego is banned from marrying into other lines allied to his own in the past as well, such as those of at least some of his grandparents. The practice can be seen as delaying the renewal of alliances till an appropriate number of generations have passed, which can also be interpreted in terms of cousin ranges: for example, no marriages permitted up to third-cousin range. Although in reality alliances with asymmetric prescription are sometimes dispersed as well,1 the formal prescriptive model does not reflect this but instead assumes that all egos repeat the marriages of their parents. Societies with EAA also typically disperse alliances, but they do not have vertical terminological equations of Crow-Omaha type. Also, the prohibitions usually refer to lines of past wife-givers and wife-takers, not the lines of ego’s parents and grandparents, as is usual in Crow-Omaha examples. Recorded examples of EAA are quite varied, but they can all be convincingly linked to asymmetric prescriptive systems, which they resemble much more closely in detail than any other known system of affinal alliance, as the examples discussed below should indicate.2

Evolved Terminologies of Asymmetric Alliance I start by considering affinal alliance arrangements in central Nias, an island some seventy miles west of Sumatra, studied by Andrew Beatty (1990, 1992). While the affinal alliance system here has a definite asymmetric bias, in that patrilines of wife-givers and wife-takers are clearly distinguished and spouses should only go in one direction between them for at least four generations,3 there is no positive marriage rule establishing a prescriptive system. Beatty’s ethnography must be interpreted as ruling out the repetition of marriages in

The Evolution of Kinship Terminologies

later generations for as long as the initial relationship is recognized, though there is no definite number of generations that have to pass before this ceases to apply. In the past, a death penalty was allegedly imposed for violations of this rule, which was still taken extremely seriously by many Niasians at the time of Beatty’s fieldwork. The initial alliance set up by marriage is continued in later generations by means of prestations, mainly of continued bridewealth payments. The different prestations have different names according to their recipients among the wife-givers. First among these recipients are the bride’s direct agnates (F, FB, B) – that is, male ego’s direct wife-givers. Next are her mother’s agnates, her MF and MB. The third group are her MM’s agnates, represented by MMB, and their descendants (obviously with time the main representatives of the remoter groups may have died out). The fourth group is represented by her MMMB and his descendants. A further two such groups (represented formally by MMMMB and MMMMMB respectively, who are most unlikely to be alive themselves at this point) may also be invited to the relevant rituals but would receive little.4 It is also at or about this point that a return marriage may be mooted, despite the ambivalence connected with it, and also that marriages in the same direction between groups again become possible, as their existing relationship as affines ceases to be recognized (though again there is no firm point or generation when this happens). Beatty’s diagram (1990: 456) shows all six patrilines being linked by successive marriages in the second generation of each wife-giving line, but no further marriages are shown as taking place between each respective pair of lines in the following generations. As Beatty remarks, the alliance system links patrilines on the basis of past marriages, not current ones, as would be the case if the arrangements were prescriptive: the representative wife-giver’s wifegiver is MMB (a relative of the +2 generation), not WBWB (a relative of ego’s generation).5 In Nias, the prestations appear to be transferred at one and the same wedding, and not successively over generations. There is also a ‘slight preference’ (Beatty 1990: 467) for marriage into a wife-giving lineage, though as taking a real MBD as a wife would be controversial, it is evident that the practice refers to other lines than that from which one’s mother came (though they may be within the same clan). Indeed, Beatty lists a number of tactics followed by Niasians to get around the dislike of actual MBD marriage, which are mostly found elsewhere in the world as well. The first is marriage to MFBSD,



263

264

How Kinship Systems Change

a classificatory equivalent to MBD by virtue of the parallel cousin/ sibling equivalence represented here by the equation FBS = B. This marriage practice is also found among the Ho of central India (Parkin 1992: 154). The second is marriage to FMBSD, which repeats a marriage between any two lines after the lapse of one generation, not immediately, so that male ego is repeating the marriage of his FF, not his father. This is one of the second-cousin specifications involved in the symmetric prescriptive four-line system of the Aranda, but it is also the main prescribed referent in the Iatmul case as reanalysed by Francis Korn (1973: Ch. 5), not at issue here. The other two options are marriage to FMBDD and MFZSD respectively. As Tjon Sie Fat subsequently showed (1998: 88, Figure 3.9, right-hand diagram), these two options can be combined with other second-cousin specifications to create a formal model in which marriages are only repeated between the same groups in the fourth generation, a model that is relevant to the Santal of central India (Parkin 1992: 160), as well as certain other societies (Tjon Sie Fat 1998: 88). The difference is that, in the Nias case, none of these marriages is shown as being repeated in later generations, so these practices never come to constitute a formal model in the same sense, and they remain non-prescriptive. For the Nias terminology, we have to consult Beatty’s slightly later monograph (1992: Ch. 4). As a whole, the terminology is not at all prescriptive, as there are no cognate-affine equations of prescriptive type, and many terms equate wife-givers and wife-takers in an apparently random fashion. However, one term is significant for the alliance pattern described above, namely sibaya, which equates the male agnates of the three successive wife-giving lines of MB, MMB and MMMB. This is a formal term used especially in contexts of ritual or the formal exchanges described above: the genealogical MB can be isolated from this large group of wife-givers by means of the descriptive phrase makhelo nina (literally ‘mother’s sibling’, so it is also applied to MZ), while, when treated as a genealogical referent lacking any connotation of alliance, he may also be called ama ‘father’, a usage emphasizing his quasi-paternal relationship with ego.6 There is a similar terminological unity of wife-taking lines, expressed by the descriptive phrase ono mbini’ö, not fully glossed here, though ono is C, GC, CC etc. (cf. Indonesian anak, C); thus this term ‘juniorizes’ the wife-takers to which it refers, as does its appearance in certain other descriptive phrases here. Beatty says explicitly that sibaya and ono mbini’ö are reciprocals. Other terms are symmetric, especially la’o

The Evolution of Kinship Terminologies

as (inter alia) WB and ZH, or denote affines in such a way as not to associate them with either wife-giving or wife-taking lines exclusively, like la’o or mbambatö. Although there are some vertical equations, especially between affinal specifications, no terms other than sibaya and ono mbini’ö unite lines of either wife-givers or wife-takers in a consistent fashion. One can nonetheless agree with Beatty when he says that ‘the terminology has a pronounced asymmetric character based on a tracing of the line of maternal affiliation’ (1992: 96), the uniting of successive lines of both wife-givers and wife-takers being reflected in the terminology. Thus Nias is an example of ‘a system of asymmetric alliance even in the absence of prescription’ (Beatty 1992: 97) – that is, despite the lack of any prescriptive cognate-affine equations.7 A similar example, from the western Keo region of Flores, eastern Indonesia, has been described by Gregory Forth (2001). Affinal alliance here is strongly asymmetric – that is, there should be no reversal in the direction of alliances between intermarrying patrilineal ‘houses’ of wife-givers and wife-takers, the former being superior in status to the latter. Although the terminology is non-prescriptive in its lack of cognate-affine equations, it has some inter-cognate equations in ego’s generation that are compatible with prescription, though not diagnostic of it, as they are found in other types of terminology too. However, they indicate a symmetric pattern, the two generations adjacent to ego’s being more asymmetric. There are separate terms for MBD and FZS, and the opposite pair of cross cousins, FZD and MBS, are equated with parallel cousins to indicate their non-marriageability (ibid.: 104). Forth states that marriages between MBD and FZS are accordingly the only ones possible and that they are ‘enjoined’ in what he calls ‘the traditional system’, though it is not clear from his use of ‘traditional’ whether this rule is still in force or is now historical. Three-unit alliance circles do occur, but they are disliked, being seen indigenously, the author states, as a form of direct exchange, which is not held in favour. However, there is a stress on at least one of a group of brothers taking a wife from a known wife-giver. Against this is a desire to disperse alliances and expand one’s range of allies, making alliance repetition in the following generations unlikely, and indeed Keo are also free to marry those with whom they have no known relationship. As in the Nias case, moreover, male ego’s social universe includes a chain of wife-giving groups from several generations. Although there



265

266

How Kinship Systems Change

are apparently no indigenous terms for them, Forth divides them formally into direct wife-givers and what he calls the ‘maternal line of origin’. The former group consists of those wife-giving houses that have given a wife directly to ego’s line – that is, those of WB/WF, MB, FMB, FFMB etc. In the second group are the wife-givers’ wife-givers, namely MB (who may therefore be in both groups, in so far as his line is not also that of WB), MMB, MMMB etc. Moreover, as a man inherits his father’s relationships with the latter’s maternal line of origin, as well as having his own line of origin, his relatives in the wife-giving direction include the houses of FMB, FMMB, FMMMB etc. Among the significant aspects of these links are the funerals of at least the direct wife-givers (WB, MB), at which their descendants are owed mortuary payments from their wife-takers, though indirect wife-givers may also be recipients, and they may substitute for the direct wifegivers (there are different categories of prestation, depending on the nature of the tie with the wife-taker). Although, as already noted, the terminology (Forth 2001: Appendix 1) has both symmetric and asymmetric features, it is the latter that predominate: many referents who would be classed as cross kin in an asymmetric (three-line) scheme, and are so classed here, would be equated with parallel kin (in brackets below) in a symmetric two-line scheme, viz. MBWB (= F), FZHZ (= M), MBWMB (= M), WMBW (= M), WBW = MBSW (= Z), and WBC = MBCC (= C, BC). There are two unambiguously symmetric terms, both of ego’s level, namely éja (WB, ZH, ZHB, BWB, CEF) and ipa (WZ, BW, WBW, ZHZ, CEM). Some other equations would express the circularity of a strictly three-line system, in which case FZHZ = MBW and FZ = MBWBW, WMBW, though as we have seen three-line alliance circles are not liked here. However, yet other equations link alters in different wife-giving lines, namely moi MB (in MF’s line), equated with MBWB and WMB (both in MMB’s line); and mame MBW (in MMB’s line), equated with MBWBW (in MMMB’s line). Given that the present-day terminology is non-prescriptive, therefore, and that marriage seems imperfectly governed at best by a positive marriage rule, the Keo case seems, on balance, to be one of EAA. A further example is the domain of Thie on the island of Roti studied by James Fox (1980: 116–20), who identifies what he calls a ‘line of maternal affiliation’. Here too there is no asymmetric prescriptive terminology, and despite the existence of a verbal preference for the actual MBD in marriage, it is rarely followed except among some

The Evolution of Kinship Terminologies

elite groups as a political strategy. As a consequence, ‘there is no strict entailment to renew a marriage alliance by a further marriage. [. . .] New alliances are formed by the marriages of each generation’ (Fox 1980: 116). The operational alliance units are often patrilineal ‘houses’, but they may be lineages and clans as well, allowing some strategic flexibility in claiming the legitimacy of particular alliances (for example, in some domains on Roti they may take place between lineages of the same patriclan). Bridewealth must be paid to allow male egos to incorporate their children into their own lines, but this is ‘merely an initial contractual payment; a continuing series of obligatory payments maintains the alliance for at least two generations’ (Fox 1980: 117). For male ego, the immediate wife-giver of significance is the toö huk or ‘mother’s brother of origin’ (not necessarily the actual MB), whose rights and obligations in relation to his wife-takers are inherited by his sons and SS. To the wife-taker’s children, the mother’s brother of origin becomes the bai-huk or MMB – that is, a wife-giver one line removed from their own mother’s brother of origin. Turning now to the Tanimbar Islands to the east of Timor, Simone Pauwels and Susan McKinnon worked at opposite ends of this archipelago, on Selaru and Fordata respectively, but they make similar observations concerning the chains of affinal groups or lines connected through the affinal alliance system. Pauwels (1994) treats this theme partly in terms of sibling relations, whereby sisters move on marriage but brothers stay behind to ‘watch over’ the lines into which their sisters have married. It appears that the mythical transition from incest to marriage took place when men began giving their sisters to what became wife-taking groups, at which point society entered a flow of time, as well as of affinal alliances. The arrangement of affinally linked lines closely resembles Beatty’s model for Nias, but Pauwels says that the tracing basically goes through women – that is, M, MM, MMM – until one comes to the line of origin or ‘base of the tree’ (see Pauwels’ diagram, ibid.: 80). However, one can also trace these ties through men – that is, the brothers of the above women – in their role as ‘watchers’; both ways of tracing seem equally proper. Each successive line has its own name, but in the middle more than one line can be called ditlan or ‘middle’, so there is flexibility. The whole series of lines from ego to the ‘base of the tree’ is a lolge or ‘row’, while if the lines of ego and his actual MB are omitted it becomes an areske or ‘ladder’. In the middle of the row, individual names tend to be



267

268

How Kinship Systems Change

forgotten and give way to the names of houses (in the Lévi-Straussian sense). Exchanges of prestations are involved in these relationships, and there is the usual distinction between wife-givers or serimwanire (‘men’s side’) and wife-takers or serimfwetare (‘women’s side’), the former being superior. Each line may be involved in several rows and ladders simultaneously. As for the marriage rule, Pauwels simply says that there is a preference for MBD marriage, though it is also possible to delay renewing a previous alliance for a generation, such that male ego marries FMBSD (cf. the Nias case; it is not very clear which of these situations prevails, if either).8 Pauwels does not give a terminology, so it is uncertain whether it is at all prescriptive or not. The situation on Fordata, described by McKinnon (1991) in much more detail and using arguments that are theoretically out of sympathy with the structuralism of Needham or Lévi-Strauss, seems broadly similar to that described by Pauwels for Selaru. Again, one affinal alliance option is the formation of a chain of allied groups but limited to three such groups through male links (i.e. three in each direction from ego’s point of view, only wife-taking lines being called lolge or ‘rows’ here; the notion of a ‘ladder’ seems to be absent here) and therefore to three generations after the initial alliance, during which period exchanges of prestations are continued. After this point is reached, the initial arrangement comes to an end and either has to be renewed by a further marriage or lapse entirely (in which case the direction of alliance could also be reversed). McKinnon calls this an open system and therefore a hierarchical one, though the classic cycles of asymmetric affinal alliance also exist here in what she calls a closed (i.e. circular) system that also expresses equality. However, where links between lines through females are concerned, the relations between them may continue indefinitely in the form of exchanges of prestations, though without, it seems, any implication of a renewal of the initial marriage alliance. As she sums it up: ‘Thus while the rights and responsibilities of the descendants of [a woman’s] brother extend indefinitely in relation to the descendants of the sister traced through female links, [those] in relation to the descendants of the sister traced through male links may be limited to three generations’ (1991: 202). These arrangements reflect the fact that not everyone marries MBD/ FZS, especially as two brothers (and presumably more than two as well) should not marry into the same wife-giving line, and the terminology she gives, although reserving a special term for MBD ms

The Evolution of Kinship Terminologies

and FZS ws ( fatnima), is not at all prescriptive. However, MBD/FZS marriage is clearly proper, just as FZD/MBS marriage is illicit, as it reverses the proper direction of alliances and cancels out, reverses or otherwise undermines the hierarchical relationship between wifegivers and wife-takers. By contrast to McKinnon, in her discussion of the Mambai of East Timor, Elizabeth Traube (1986) follows a Needhamite line quite strictly. Nonetheless, as on Fordata a short chain of wife-giving lines is recognized, namely the lines of MB, MMB and MMMB. While MBD/FZS marriage is one preference, alliances can also be made with totally new lines. In these cases, in which an unrelated woman is automatically classified as MBD once married, a further and apparently equal preference involves marriage between FMBSD and FFZSS. As on Fordata, prestations are involved in relations between lines, generation after generation, mainly in the form of mortuary payments, as well as bridewealth; and again male ego continues to be responsible for the prestations arising from his own ascendants’ marriages, even if he marries into a wholly different line himself. It is not clear where the cut-off point is regarding such obligations, if indeed there is one. Again here, there is strong condemnation of FZD/MBS marriages that reverse the direction of alliances and confuse the statuses of wife-givers and wife-takers, such marriages being associated with witchcraft and likened to the indiscriminate copulation of animals. Traube does not give a terminology, so it is uncertain whether it is at all prescriptive or not, though her statement about unrelated wives being defined as MBD indicates it may be. Of the three cases just discussed, therefore, only Fordata appears unambiguously to have EAA: the Selaru case may be prescriptive, and for the Mambai there is some uncertainty. The Manggarai in eastern Flores, however, seem more concretely to be an example of EAA. They were studied by James Gordon (1980), who among other things rejects earlier accounts of Manggarai alliance practice as prescriptive, especially Needham’s interpretation (1966) that asymmetric alliance goes along with a symmetric terminology here. Instead, Gordon finds that, while there are some equations that are compatible with prescription, the terminology overall does not specify a category of prescribed spouse, and ego’s level, at least from one point of view (there are alternative terms here), is generational. Moreover, only 15% of marriages, called tungku, are with MBD or a structural equivalent. The remaining 85% of marriages are between



269

270

How Kinship Systems Change

previously unrelated individuals, indicating that there is no regular practice of continuing alliances in later generations through the repetition of marriages. Nonetheless, there are clear indications in the ethnography (Gordon is not very explicit on this point) that such continuity is sought, even though tungku marriages – which would ensure this and, among other things, reduce bridewealth and other obligations – are a distinct minority of marriages. As with previous examples, chains of affinal alliance groups or lines can be identified that are linked not only through marriages but mutual exchange obligations. From ego’s perspective, in one direction these are the anak rona (immediate wife-givers) and anak rona sa’i (wife-givers’ wifegivers), while in the other direction they are the anak wina (wifetakers) and the latter’s anak wina (no Manggarai term given). On East Timor, the Makassae studied by Shepard Forman (1980) distinguish different affinal alliance groups within a society that lacks prescribed marriage with MBD: thus a male ego has to make mortuary payments to the MBs of himself, his father, his wife and his WF (ibid.: 156–57). Another case where asymmetric prescriptive alliance is no longer practised, but there is evidence of different categories of wife-givers, is that of the Lamahalot of East Flores, studied in an unpublished thesis by Dr Sandra Modh (2012).9 While wife-givers in general are simply called ina ama (literally ‘mother-father’), another term, opu laké,10 ‘refers more specifically to wife-givers of the progenitor line (the MB line) and in particular to men who hold the status [of] MB (with preference for the genealogical MB)’ (ibid.: 154). In addition, there is the category of opu (laké) pu’én, ‘root/origin MB’ (ibid.: 157), a reference to the line of the oldest opu laké, rather than to any individual (Sandra Modh, personal communication). The terminology as presented (2012: 162 ff.) appears to be mainly symmetric except for a distinction between FZ and MBW and a separate term for MB. Only one cousin term is given, murén laran, a compound apparently meaning classificatory MBD only. As there is no longer a prescription for MBD/FZS marriage, which apparently used to obtain here according to informants’ statements, it is not surprising that the terminology as recorded lacks clear prescriptive features.11 As we have seen, in a couple of cases, namely Selaru and the Mambai, there is uncertainty whether affinal alliance is prescriptive or not, mainly because the respective authors do not discuss the terminology. Nonetheless, the example of the Ema of Timor, studied by

The Evolution of Kinship Terminologies

Brigitte Clamagirand (1980), indicates that chains of wife-givers and wife-takers may go along with prescription. Indeed, the Ema have a fairly standard prescriptive system, with a rule of MBD marriage, a ban on FZD marriage, distinctions between wife-givers and wifetakers in which the former are superior to the latter, and a preference for closed alliance cycles consisting of four alliance groups. From male ego’s perspective, immediate wife-givers and wife-takers are ai mea, while those involved in subsequent alliances are bae botu. There is also a term for male ego’s wife-givers’ wife-givers, namely uma mane pun, or what Clamagirand calls the ‘base house of the wifegivers’ (ibid.: 142). Again, these lines are linked by exchange obligations invoked in connection with key rituals. However, they do not appear to extend as far back into past generations as some of the cases of EAA I have already discussed. A similar situation obtains in Rindi, a domain in eastern Sumba that was the subject of an earlier study by Forth (1981). Here, although alliance circles are disliked, there is otherwise the whole panoply of features associated with asymmetric prescription, including the word tanggu, ‘rightful share or portion’ (1981: 330), for a man’s right to marry his MBD. Two kin terms, yera and layia, are also used for wife-givers and wife-takers respectively, who together can also be referred to as kalembi. However, the latter term can also denote the more restricted categories of wife-givers’ wife-givers and wife-takers’ wife-takers, while, ‘With regard to immediate wife-givers and wifetakers, the Rindi distinguish affines of long standing from those with whom marriages have only recently been contracted as kalembi ndai and bidi kalembi, “old” and “new” affines’, the cut-off point between them notionally being the third generation (1981: 282–83).12 The extension back in time thus goes a generation deeper than in the Ema case but is still less than for the Niasians or Tanimbar Islands. I turn at this point to western Indonesia to discuss two ethnic groups in northern Sumatra, both belonging to the Batak conglomeration of peoples. The first of these are the Karo Batak, studied by Masri Singarimbun, himself a Karo (1975). One of the peculiarities of this case is the fact that Rodney Needham felt able to contradict certain aspects of Singarimbun’s own interpretation of affinal alliance in this society in an article he (Needham) published three years later (Needham 1978). Singarimbun’s book ends with an appendix by Harold Scheffler, who uses this case to attack Needham’s own position regarding systems of prescriptive alliance, the existence of



271

272

How Kinship Systems Change

which Scheffler was concerned to deny. Singarimbun clearly goes along with this, at least to the extent of denying that affinal alliance takes place here between groups rather than just individuals. However, while mostly accepting the ethnographic facts in Singarimbun’s account (the main divergence here is that Needham raises some queries concerning the terminology), Needham reinterprets some of them to fit his own ideas about affinal alliance. In what follows, I use Needham’s account, as handier and in some respects easier to follow, while paying attention to Singarimbun’s descriptions to ensure ethnographic accuracy where necessary. Affinal alliance among the Karo Batak clearly has some aspects consistent with asymmetric prescription, though it has some divergent features as well. One of a group of brothers should marry a woman classed as MBD, though his choice is restricted by his having to marry the eldest daughter; this may indicate that it is the oldest brother that has to marry in this fashion, but that is not clear. His brothers, by contrast, should marry elsewhere, thus dispersing alliances. This may be a restriction imposed by the wife-givers: stating, in accordance with this dispersal rule, that there should be no marriage to BWZ, Singarimbun also gives as a reason the need to increase one’s count of wife-takers (1975: 154). As on Nias, MFBSD is seen as an equivalent to MBD in marriage, though the conformity of actual marriages with these specifications is less than 5%. The distinction between wife-givers or kalimbubu and wife-takers or anakberu is a characteristic of asymmetric alliance in that the two groups are unequal in status to the advantage of the former. Singarimbun also identifies wife-givers’ wife-givers as puang kalimbubu (apparently traceable through both MMB and WBWB, formally in the same line, though respectively in +2 and ego’s genealogical levels), while wife-takers’ wife-takers are anakberu menteri. However, regarding the former, the situation is a little more complicated. Each marriage establishes a new kalimbubu–anakberu relationship unless it is an MBD marriage or equivalent, in which case male ego’s kalimbubu by birth (line of MF etc.) is also his kalimbubu by marriage (line of WF etc.). Marrying other women from among relatives in the direction of ego’s wife-givers means that these two latter categories overlap but are not entirely the same, potentially giving rise to conflict if, for example, one has a double set of similar exchange obligations to both MB and WF (who would be the same, with just one set of obligations, if MBD marriage were formally obligatory). Marrying an entirely unrelated

The Evolution of Kinship Terminologies

woman, by contrast – which is entirely possible – brings a line of kalimbubu into existence that is both new and separate from all ego’s other (and nearer) wife-giving lines, thus avoiding such conflict. Singarimbun is insistent that all this depends on interpersonal, not intergroup relations – that is, it is egocentric, not sociocentric, unlike asymmetric prescription as conceived by Needham. In his article (1978), Needham reanalysed the kinship terminology, deciding it was basically asymmetric prescriptive, despite the equation ZH = WB in ego’s level and the symmetric pattern of impal, the term for the prescribed spouse. Also, two +1 terms for relatives in the wife-giving direction are of interest in light of the foregoing in that they link chains of wife-givers: thus mama MB links wife-giving males of ego’s generation and above across three such lines, while mami MBW does the same for wife-giving females of ego’s generation and above across as many as four such lines (to WMMBSW). In the wife-taking direction, terms like beréberé and beru also link three lines. The other Batak example is the Toba Batak, examined in an unambiguously Needhamite manner by Kathryn Bovill (1985). A conventional system of asymmetric prescription, wife-givers or hulahula are distinguished from wife-takers or boru, a word meaning both ‘female’ and ‘daughter’, suggesting that boru are both feminized and juniorized in relation to their hulahula. There is little indication of names for remoter lines in either direction, except that the phrase boru ni tulang so siolion, listed as a kin term literally meaning ‘the MBD who must not be married’ (ibid.: 44), also refers to wife-givers’ wife-givers in general, and indeed, as WMMBSD (ibid.: 40, Table 1), to the latter’s wife-givers as well. Thus, some kin terms have lateral extension across the nearest two and occasionally a third and even fourth wife-giving line as well. Other examples are inang WBW but extending to WMBSW and WMBSSSW; tulang MB but also WMMBS; nantulang MBW but also MBWBW, MMBSW, WMMBD, WMBW; and ompung FF, MF but also WMMBW, a specification in the fourth wife-giving line. A similar situation is found on the wife-taking side. Thus pahompu, CC, GCC is also FFZDDH; while among kin terms used by a female ego, there is amang bao, HZH but also HFZDH; amangboru, FZH but also HFFZDH; bere, HZC but also HZDH, HFZDDH, the latter a specification in the fourth wife-taking line; and pahompu, CC, GCC but also FZDDC, HZDC, HFZSDC, HFZDSC, HFZDDC, the latter also a specification in the fourth wife-taking line.



273

274

How Kinship Systems Change

The terminology has some vertical equations, but they all link referents in alternate rather than consecutive generations; for example, tunggane MBS and MBSSS, WB and WBSS.

Discussion These examples are of interest because, apart from the Ema and Rindi cases, they are united by a combination of: 1) systems of asymmetric affinal alliance that retain some features of the asymmetric prescriptive model in, especially, the strict separation of wife-givers and wife-takers; 2) delays in the renewal of alliances for several generations from the point of view of each line, contrary to the model, though not necessarily the practice, of asymmetric prescription; and 3) the perpetuation of the relationships started by any marriage into the following generations through means (especially exchanges of prestations at specific rituals) other than a renewal of the marriage alliance per se. Referring back to Trautman’s hypothesis of a developmental link between asymmetric prescription and Crow-Omaha (cf. Ch. 7), it can be argued that aspects 2) and 3) recall Omaha-type marriage rules and practices, including, it would seem, those of the Omaha themselves (Barnes 1984). Aspect 1), the separation of wife-givers and wifetakers, is also found in the actual Omaha terminology, and among the Omaha themselves direct exchanges of sisters in marriage are disliked and are apparently unknown (ibid.), though alliances between any two lines or groups in the wider sense may be symmetric. However, unlike the Indonesian examples, the Omaha do not have chains of successive wife-giving and wife-taking lines, each of which represent a chronologically prior marriage and concomitant intergroup alliance to that of ego, nor are the diagnostic Crow-Omaha equations apparent anywhere in the Indonesian material. Thus these Indonesian EAA examples are emphatically not ‘Crow-Omaha’ systems, and unlike the latter they convey a much clearer sense of being former systems of prescriptive alliance that have ceased to be prescriptive but have retained certain relevant features; some cases, like the Keo and (especially Toba) Batak, even have terminological equations across lines that fit the principles underlying EAA marriage practices. Most of this chapter has been taken up with describing, in parts of Indonesia, systems and practices of asymmetric alliance that do not

The Evolution of Kinship Terminologies

involve prescription but that might have devolved from ones that do. Moreover, although the actual historical evidence may be thin,13 the remaining structural similarities make it a reasonable hypothesis that these systems and practices represent an evolved or derogated form of asymmetric prescription; in other words, a historical development. Indeed, this is supported by the Ema and Rindi cases, which despite having asymmetric prescription also show some of the features of the evolved forms, though they extend back through fewer generations. And while one might argue that this is what one might expect in such cases, other examples, such as Nias, have developed these features further by maintaining chains of wife-givers and wife-takers going back several generations. Add to this the abandonment of terminological prescription and one ends up with what I have defined as EAA. To repeat, I am not claiming the existence of an absolute difference in all respects between the two situations: indeed, the interest in this topic is precisely the fact that there are overlaps and plenty of variation between them, thus permitting the historical argument to be ventured. Moreover, it is clear that there is no one ‘system’ of EAA in ethnographic reality, any more than there is of asymmetric prescription, despite the ability to model them both, even EAA to an extent, formally. Nonetheless, although these societies and their terminologies lack Crow-Omaha features entirely, and certainly cannot be defined as Crow-Omaha, they do resemble them at least in banning the immediate repetition of marriage alliances between the same groups, in maintaining relations between such groups in the interim in other ways before alliance repetition becomes possible again, and in evidently being developments away from prescription. In fact, there are plenty of examples of this combination of factors around the world without the co-presence of Crow-Omaha terminologies, as in north India (Tiemann 1970; Parry 1979) or rural Greece (Du Boulay 1982); EAA is just another example of this phenomenon. Returning again to Trautmann’s theory, finally, one conclusion of this chapter is that asymmetric prescriptive terminologies do not need to evolve into Crow-Omaha ones because they have their own evolved forms. This is further supported by their being highly localized mainly in two areas of Indonesia, which apparently lack CrowOmaha terminologies. This is not to say that the sorts of evolutionary developments posited by Trautmann are impossible or must be rejected, since systems of asymmetric alliance do occur sporadically



275

276

How Kinship Systems Change

elsewhere in the world, especially in the Americas, and occasionally alongside Crow-Omaha terminologies; better material for his hypothesis might be found, for instance, among Gê-speakers in central Brazil (Maybury-Lewis 1979).14 By the same token, however, such evolutionary developments are neither inevitable nor structurally necessary, as the Indonesian material makes clear.

Notes * Originally published in the Journal of Anthropological Research 74(2) (2018): 232– 51. Reproduced with permission. 1. ‘Dispersed alliance’ has a number of meanings in the anthropology of kin-

ship. First, one of a group of siblings must marry in a prescribed fashion (e.g. to MBD), but other siblings have more freedom to marry elsewhere, and indeed may be obliged to. These are associated with positive marriage rules or prescriptive marriage systems. In another arrangement, marriage is expected to take place with a second cross cousin, as among the Aranda in Australia, meaning that one only repeats a marriage between the same lines every second generation, the system enjoining marriage into a different line in the intervening generation. Permitting marriage only outside certain degrees of relationship, as found traditionally across Europe in various forms, is in effect a third example. Of course, alliances may be dispersed at random, even in a prescriptive system, in that brothers may, in fact, be marrying into quite different lines of affinal alliance groups. It therefore seems most useful to talk of dispersed alliance where it is associated with a rule of some sort, not simply the random distribution of alliances. 2. Note added in 2020: in working out these forms of alliance, I have been inspired by Ross Bowden’s work on the Omaha-type kinship system of the Kwoma of Papua New Guinea (1983). The connection is therefore not terminological but consists in the circumstance that, over periods of several generations, lineally linked kin can substitute for one another in exchange or other intergroup relations. For Bowden, the reason for the vertical, Omaha-type equations in the Kwoma terminology is that they link such kin. In the eastern Indonesian cases, none of the terminologies are Crow-Omaha. 3. Beatty says: ‘A man is prohibited from marrying a woman whose M, MM, MMM, MMMM etc., belonged to his descent group (ideally clan)’ (1990: 455). 4. See the diagram at Beatty 1990: 456. Beatty also says that noble families may recognize even more lines, though he does not say whether the latter too would be invited to the wedding. This has a knock-on effect on the availability of spouses: evidently, it is nobles who are most likely to practise village exogamy in search of them. Partly for this reason, ordinary villagers tend to treat the fourth generation as a suitable cut-off point, rather than having extra lines of wife-givers at their weddings.

The Evolution of Kinship Terminologies

5. There is also a backward chain of women ranging from WM to WMMMMM.

6. 7.

8.

9. 10.

11.

12.

Beatty nonetheless resists seeing Nias society as an example of double descent, or as having matrilineal descent. This chain is ego-focused, not ancestor-focused. This equation between F and MB also occurs on the island of Truk in the northwest Pacific, studied by Ward Goodenough. See Chapter 8. In fact, southern Nias does have an asymmetric prescriptive system in the stricter sense (Beatty 1990: 470 note 14). Unfortunately, we do not have a terminology from there that we could compare with the terminology from central Nias given by Beatty. Basing herself elsewhere (1983: 132) on Father P. Drabbe’s earlier work on the neighbouring island of Yamdena, Pauwels describes a regular system of MBD marriage in which there may be a repetition of the alliance between the generations, but if not the relationship is maintained by continuing the exchange of alliance-type prestations instead; also possible, however, is initiating a marriage alliance with a completely new group or house. Wife-givers are nduan (master), wife-takers uranak (sister’s children), and there is a brief mention of ‘the principal wife-giver of the house, that is, the one who represents the original wife-giver of the house’ (ibid.: 136), possibly hinting at a similar chain of allied groups as on Selaru. One of the anonymous reviewers for this article pointed out that FMBSD is often a candidate for marriage in prescriptive systems, as she also comes from a wife-giving group. I am grateful to Dr Modh for her permission to make use of her thesis here and for her help in answering my queries in relation to it. It seems that opu is a general term for affines in many Lamaholot-speaking areas, though Barnes reported (1977: 146–47) that in Wailolong it had become restricted to wife-takers (opu pain, opu lalant), the term belake having largely replaced opu laké in the meaning of wife-givers. Nonetheless, the latter term was still known in Wailolong, and Barnes glosses it as ‘male opu’, the ‘male’ here probably indicating the higher status of wife-givers rather than being a marker of gender as such; see also Barnes 1974: 269. Dr Modh reports that belaké has been introduced in her village as well, though without replacing the older term (2012: 162). Note added in 2020: Barnes goes into these terms and their meaning in more detail elsewhere (Barnes 1979). Dr Modh also reports (personal communication) that the Lamaholot she worked with were quite different from those studied by Barnes in the 1970s. She found that they also have an affinal alliance system called tiga tungku or ‘three hearths’, possibly expressing the minimal circle of three intermarrying alliance groups associated with asymmetric prescription. Despite their claims to the contrary, Dr Modh suggests further that these Lamaholot might not have had this form of marriage originally but might either have borrowed it from other Lamaholot or simply have had a mixture of practices of their own resulting in a similar pattern of alliances. In fact, it has long been recognized that terminologies of asymmetric prescriptive alliance may have more than the minimum three lines. The classic ‘type case’, the Kachin, have a terminology in five lines – distinguishing



277

278

How Kinship Systems Change

wife-givers’ wife-givers and wife-takers’ wife-takers, as well as ego’s line and those of his or her immediate affines. See Leach 1961 [1945]: 41 (Table 2). 13. Beatty (1992) specifically doubts that the system he describes for Nias (see above) can be explained by historical change. 14. Note added in 2020: see now Chapter 7.

References Barnes, R.H. 1974. Kédang: A Study of the Collective Thought of an Eastern Indonesian People. Oxford: Clarendon Press. ———. 1977. ‘Alliance and Categories in Wailolong, East Flores’, Sociologus 27: 133–57. ———. 1979. ‘Lord, Ancestor and Affine: An Austronesian Relationship Name’, Nusa 7: 18–34. ———. 1984. Two Crows Denies It: A History of Controversy in Omaha Sociology. Lincoln and London: University of Nebraska Press. Beatty, Andrew. 1990. ‘Asymmetric Alliance in Nias, Indonesia’, Man 25: 454–71. ———. 1992. Society and Exchange in Nias. Oxford: Clarendon Press. Bovill, Kathryn J. Brineman. 1985. ‘Toba Batak Relationship Terminology’, Bijdragen tot de Taal-, Land-, en Volkenkunde 141: 36–66. Bowden, Ross. 1983. ‘Kwoma Terminology and Marriage Alliance: The “Omaha” Problem Revisited’, Man 18: 745–65. Clamagirand, Brigitte. 1980. ‘The Social Organization of the Ema of Timor’, in James J. Fox (ed.), The Flow of Life: Essays on Eastern Indonesia. Cambridge, MA and London: Harvard University Press, pp. 134–51. Du Boulay, Juliet. 1982. ‘The Greek Vampire: A Study of Cyclic Symbolism in Marriage and Death’, Man 17: 219–38. Forman, Shepard. 1980. ‘Descent, Alliance and Exchange Ideology among the Makassae of East Timor’, in James J. Fox (ed.), The Flow of Life: Essays on Eastern Indonesia. Cambridge, MA and London: Harvard University Press, pp. 152–77. Forth, Gregory. 1981. Rindi: An Ethnographic Study of a Traditional Domain in Eastern Sumba. The Hague: Martinus Nijhoff. ———. 2001. Dualism and Hierarchy: Processes of Binary Combination in Keo Society. Oxford: Oxford University Press. Fox, James J. 1980. ‘Obligation and Alliance: State Structure and Moiety Organization in Thie, Roti’, in James J. Fox (ed.), The Flow of Life: Essays on Eastern Indonesia. Cambridge, MA and London: Harvard University Press, pp. 98–133. Gordon, John M. 1980. ‘The Marriage Nexus among the Manggarai of West Flores’, in James J. Fox (ed.), The Flow of Life: Essays on Eastern Indonesia. Cambridge, MA and London: Harvard University Press, pp. 48–67. Korn, Francis. 1973. Elementary Structures Reconsidered: Lévi-Strauss on Kinship. London: Tavistock Publications. Leach, Edmund. 1961 [1945]. ‘Jinghpaw Kinship Terminology’, in E.R. Leach, Rethinking Anthropology. London and New York: Athlone Press and Humanities Press, pp. 28–53.

The Evolution of Kinship Terminologies

Lévi-Strauss, Claude. 1949. Les Structures Elémentaires de la Parenté. Paris: Presses Universitaires Françaises. ———. 1966. ‘The Future of Kinship Studies’, Proceedings of the Royal Anthropological Institute 1965: 13–22. McKinnon, Susan. 1991. From a Shattered Sun: Hierarchy, Gender and Alliance in the Tanimbar Islands. Madison: University of Wisconsin Press. Maybury-Lewis, David (ed.). 1979. Dialectical Societies: The Gê and Bororo of Central Brazil. Cambridge, MA and London UK: Harvard University Press. Modh, Sandra. 2012. ‘Lamaholot of East Flores: A Study of a Boundary Community’, DPhil thesis. University of Oxford. Needham, Rodney. 1966. ‘Terminology & Alliance I: Garo, Manggarai’, Sociologus 16: 141–57. ———. 1978. ‘Classification and Alliance among the Karo: An Appreciation’, Bijdragen tot de Taal-, Land-, en Volkenkunde 134: 116–48. Parkin, Robert. 1992. The Munda of Central India: An Account of their Social Organization. New Delhi: Oxford University Press. Parry, Jonathan. 1979. Caste and Kinship in Kangra. London: Routledge & Kegan Paul. Pauwels, Simone. 1983. ‘Some Important Implications of Marriage Alliance: Tanimbar, Indonesia’, in R.H. Barnes, Daniel de Coppet and Robert J. Parkin (eds), Context and Levels: Anthropological Essays on Hierarchy. Oxford: Journal of the Anthropological Society of Oxford, Occasional Papers, no. 4, pp. 131–38. ———. 1994. ‘Sibling Relations and Intemporality: Towards a Definition of the House Eastern Indonesia’, in Leontine E. Visser (ed.), Halmahera and Beyond: Social Science Research in the Moluccas. Leiden: KITLV Press, pp. 79–96. Singarimbun, Misri. 1975. Kinship, Descent and Alliance among the Karo Batak. Berkeley: University of California Press. Tiemann, G. 1970. ‘The Four-Got Rule among the Jat of Haryana in Northern India’, Anthropos 55(1–2): 166–77. Tjon Sie Fat, Franklin. 1998. ‘On the Formal Analysis of “Dravidian”, “Iroquois”, and “Generational” Varieties as Nearly Associative Combinations’, in Maurice Godelier, Thomas T. Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington: Smithsonian Institution Press, pp. 59–93. Traube, Elizabeth. 1986. Cosmology and Social Life: Ritual Exchange among the Mambai of East Timor. Chicago and London: The University of Chicago Press. Trautmann, Thomas R. 2012. ‘Crossness and Crow-Omaha’, in Thomas R. Trautmann and Peter M. Whiteley (eds), Crow-Omaha: New Light on a Classic Problem of Kinship Analysis. Tucson: The University of Arizona Press, pp. 31–50. Wouden, F.E. van. 1968 [1935]. Types of Social Structure in Eastern Indonesia, trans. Rodney Needham. The Hague: Martin Nijhoff.



279



Conclusion

The changes to terminology and marriage practice described in the foregoing chapters have some differences, but the unifying theme is of a change from prescriptive systems involving cross-cousin marriage to various kinds of non-prescriptive systems that either do not insist on cross-cousin marriage or do not allow it at all, nor even, very possibly, marriage to close cousins in general. Chapter 1 described the various trajectories that a move away from an original system of symmetric cross-cousin marriage, whether tetradic or simply two-line prescriptive, might entail, whether to another form of prescriptive system with cross-cousin marriage (asymmetric threeline, symmetric four-line) or to a non-prescriptive pattern like Iroquois, Crow-Omaha or Hawaiian (or generational). Iroquois itself might be a stopover on a path from Dravidian systems of symmetric cross-cousin marriage to Crow-Omaha ones, unless the latter in some cases derive from Kachin-type three-line prescriptive systems, a question dealt with more generally in Chapter 7. While terminologies make equations – that is, they equate more than one kin type under the same term – the distinctions between kin types that different terminologies make can also be seen as informing the diagnosis of their respective types. The equating property is reduced to a minimum in what, following N.J. Allen (e.g. 1989), I have called ‘zero-equation’ terminologies, where kin types that may be equated in other systems have or tend to have their own terms. Once this stage has been reached, a cognatic (English-type) pattern might emerge by reimposing terminological equations but in a different arrangement. The significance of zero-equation terminologies is discussed especially in Chapters 5 on India and Chapter 6 on Europe,

Conclusion

in the latter case drawing partly on written evidence stretching back to antiquity. In fact, Chapter 5 hypothesizes how Indo-European zero-equation terminologies might have developed in north India, while Ch. 6 takes up the story regarding further possible developments among related Indo-European terminologies in Europe. Chapter 2 is also of relevance here, as it represents my initial attempt to link the prescriptive systems of south India diachronically with the non-prescriptive practices and terminologies of north India by using intermediate tribal systems in central India as a waystation; it therefore provides a basis for the more extended treatment of this theme in Chapter 5. Over time, marked by generations, actual marriage patterns tend to result in marriages being dispersed between different kin groups. This might happen with two-line prescriptive alliance (symmetric cross-cousin marriage), but the formal model of this system does not require it, and in the case of the closely related tetradic society it is ruled out entirely, as this formal (and admittedly unattested) model only envisages two vertical moieties of affinal alliance regularly and perpetually intermarrying across the whole society.1 Both formal models envisage all egos marrying as their parents did – that is, into the same kin groups and with the same kin categories, as is entailed by the requirement to marry a cross cousin, but with twoline prescriptive this may not be the case in practice. However, once this repetition of alliances generation after generation is banned, as it is explicitly in most Crow-Omaha examples (Ch. 7), we can no longer talk about prescription. This can also be linked with the practice of encouraging groups of siblings to intermarry amongst themselves (whereby they become siblings-in-law or GEG categories) in such a way that they do not repeat the marriages of the previous generation, nor more than one previous generation in some cases, and the following generation(s) too will have to marry elsewhere (Ch. 3). Both situations have the effect of dispersing alliances among a number of different kin groups, but there are other possibilities as well: in particular, as north Indian practice shows, a Crow-Omaha terminology is not needed for this to take place (Ch. 5). North India is actually predominantly an area of zero-equation terminologies, at least in part, but this cannot be linked specifically to marriage norms, showing that the logical fit between marriage patterns and terminological systems one finds with cross-cousin marriage no longer occurs in



281

282

How Kinship Systems Change

post-prescriptive situations, where the co-occurrence becomes more contingent and even random. The possible reasons for a society moving away from cross-cousin marriage are many and complex; they are discussed at length in Chapter 4. One fundamental consideration here may be demographic: it has often been noted that strict adherence to cross-cousin marriage would not work demographically, as not everyone has a cross cousin, even at second-cousin range or remoter. In these circumstances, if strict cross-cousin marriage is attempted initially, it may quickly become irksome as well as impractical, leading to its modification, perhaps allowing second and remoter cousins, as well as irregular marriages, which are then redefined as if they had conformed to the rules. This, therefore, need not lead to the immediate abandonment of cross-cousin marriage as a formal type of spouse-exchange, but in time the strains may become too great, and societies therefore reach out to other kin groups than those most closely related to them through cross-cousin marriage, hedge their bets with such groups by dispersing alliances among them, and abandon even the pretence that certain marriages conform to the rules of cross-cousin marriage when they do not. For a time, a ghost of the old system may persist with marriages among GEG categories, in which groups of siblings intermarry, possibly quite intensively, as if they were cross cousins but do not form ties of cousinhood through previous generations, as they do not repeat the marriages of those generations (Chapters 2, 3, 5). The much better known restrictions on the repetition of alliances associated with Crow-Omaha terminologies have a similar affect. Their possible affinal alliance correlates are therefore discussed in Chapter 8. In linking vertical Crow-Omaha equations hypothetically with the perpetuation of exchanges and other obligations between different lines of descent (including clans and lineages) and their inheritance generation after generation, it suggests possible reasons for their emergence. The chapter therefore presents a diachronic account in a different sense than the focus on change in marriage practices and terminologies in the chapters of Part I. Here, too, Crow-Omaha terminologies are not essential: Chapter 9 describes a number of societies in Southeast Asia as having similar intergenerational obligations and relations without such terminologies, though these examples all appear to derive from earlier systems of asymmetric cross-cousin marriage or asymmetric prescriptive alliance.

Conclusion

While classifications may seem concrete, natural and inevitable to those they classify, analysis of them shows that this is not always the case and that they do change over time with circumstances. This is obvious in other domains – while the most widely accepted biological classifications may still be Linnaean in their fundamentals in the twenty-first century, their details have both grown and changed enormously in the quarter of a millennium since Linnaeus’s death. Similarly, Berlin and Kay (1969) argued over fifty years ago, not without controversy, that terms for colour change from, typically, an initial triple-category ‘black-white-general colour’ pattern into one with more categories covering fewer and more specific other colours. There may be restrictions on how far this process can proceed in the case of colour, but where biological classification is concerned it may well be endless as more and more species are discovered and as biological organisms themselves evolve new forms. In this regard, kinship classification is closer to colour than to biology; indeed, restrictions on it as a semantic domain are known in the sense that pragmatically there are only a quite limited number of both analytical kin types and, even more so, acknowledged relatives in an indigenous classification.2 There are, of course, oppositional, anti-structuralist views not giving credence to prescriptive alliance and terminologies and/or to genealogical thinking in relation to studies of kinship. An influential figure in rejecting genealogy as a basis for thinking about kinship both analytically and indigenously was David Schneider (e.g. 1984). As far as dismissing the social significance of prescriptive alliance or cross-cousin marriage itself was concerned, another figure took the field, namely Harold Scheffler (e.g. Scheffler and Lounsbury 1971), who dismissed these notions entirely as ‘systems’ of kinship and marriage that provided the social cement of those societies that had them in the way that the structuralists posited. Nonetheless, his views, treated here in Chapter 4, open up a number of issues that structuralist approaches should be aware of and be prepared to answer. Finally in this conclusion, I would like to stress the significance of certain features that have already been mentioned and described more than once in the foregoing pages but that still seem to me to deserve greater attention from kinship specialists than they have received thus far. One is the practice of certain speech communities that appear to have left cross-cousin marriage behind and learned to conceive of



283

284

How Kinship Systems Change

their preferred (NB: not prescribed) marriage option as involving intermarrying groups of siblings-in-law. This goes along with a usually firmer rule that marriages between the same two groups are then banned in succeeding generations, from the number of one upwards. It is not that these features have not been noticed before, as they are clearly evident in the ethnographic data for a number of such communities around the world. Moreover, as I have pointed out repeatedly ever since I experienced this epiphany at some time in the late 1980s, this practice would amount to cross-cousin marriage if links to one’s siblings-in-law were or could be traced via the previous generation. However – and this is where I think a point has been missed that could be developed further – it is also reported in many such cases that, possibly because of the ban on realliances in succeeding generations, alliances within the current one are intensified and replicated as much as possible. What is curious is that, with very few exceptions, there seem to be no kin terms isolating GEG categories in the manner of prescriptive systems identifying the prescribed kin type (cross cousin and occasionally ZD), despite their apparent significance to these speech communities. I am not suggesting that what I am describing amounts to a new and so far undiscovered formal system of alliance, as any society with this practice, given that it cannot be considered prescriptive in any sense, is going to have other marriage options alongside it to choose from.3 As a practice, however, it clearly has significance, varied though it may be, for how many speech communities view such marriage options. The second point I would like to emphasize relates to terminology more than marriage options. It is to stress once again the significance of zero-equation terminologies in facilitating a change from prescriptive or immediately post-prescriptive systems with, perhaps typically, bifurcate merging, generational and/or bifurcate collateral patterns (the latter itself being a form of zero-equation pattern) to a cognatic terminology like the English one. This hypothesis gives zero-equation terminologies a key position in certain evolutionary trajectories, allowing them no longer to be considered as just one pattern, and not a very productive or interesting one, among others. A corollary of this is to ask whether or not cognatic terminologies such as the English one are the end point of such trajectories. This is not to link them with formerly objectionable notions of the superiority of the speech communities that have them in terms of alleged levels of civilizational progress, nor is this a concealed attempt on my

Conclusion

part to revive them. Indeed, they occur in many parts of the world in societies that have little else in common and are very varied in the other features that distinguish them. Instead, it is to recognize that such terminologies have, so to speak, run into an evolutionary dead end. The logical principles underlying the different types of terminology are so few and so predictable that it is hard to see how cognatic terminologies could change into any other type except by putting into reverse what I have repeatedly suggested is a one-way trajectory of change. Perhaps, given the reduction in the size of kinship circles in many societies in the modern world, where people very often do not associate much with their cousins, for example, even if they know they have them, practical kinship terminologies might be reduced still further to cover just the nuclear family, while for certain individuals who lack kin entirely they may become almost entirely redundant. Finally, it is worth saying something briefly about the links between terminology and marriage practice, as well as the relationship between changes within both taken separately. While the practice of cross-cousin marriage in its few variants logically fits actually recorded terminologies, there is no guarantee that the two occur together in the same speech community, and with non-prescriptive systems the logical connection is much reduced or eliminated entirely. There is also no single reason for marriage practices changing, encouraging the view that the reasons for such changes are likely to be purely local. Certainly they seem highly varied, in so far as we can actually recover them. That is not the case with the terminologies, whose structural principles are much more restricted and more predictable. In other words, speech communities are much freer in their marriage practices than in their kinship terminologies, especially once the latter no longer express the former, as in a prescriptive system. That, of course, assumes that actual terminologies are internally consistent. In fact, they are not always, which in itself suggests a terminology in flux and with it the possibility that it is changing. Nevertheless, in terms of global comparisons we seem to have a range of marriage practices each of which, at the level of greatest detail, might ultimately be unique. Alongside them, however, we have a small and limited number of terminological types based on a similarly restricted number of logical principles, limitations on which are known, recognized and predictable, and are themselves logical. The dynamics of classification, therefore, are such as to allow myriad



285

286

How Kinship Systems Change

marriage practices to be associated with, refer to and sometimes be controlled by a restricted number of terminological types.

Notes 1. With tetradic society, there is also a second moiety consisting of two hor-

izontal sets of alternating generations, marriage being prescribed within each set, but this is not immediately relevant here. 2. It is true that, in principle, chains of genealogical symbols could run endlessly, but for real humans in real societies there are practical cognitive limits on the tracing of such relationships. Many ethnographers, for example, have remarked on the shallow genealogical memory of their informants, barely extending back to grandparents in some cases. This would inhibit not just the tracing of vertical links back in time, but also the ability to establish genealogical connections with, for example, remoter cousins. 3. As already noted, such choice is not necessarily denied to the individual in prescriptive systems: it is just that such systems formally recognize only one option and redefine ‘wrong’ marriages in accordance with it. It is this that led Needham (1973) to identify prescription with the terminologies, not marriage practices.

References Allen, Nicholas J. 1989. ‘The Evolution of Kinship Terminologies’, Lingua 77: 173–85. Berlin, Brent, and Paul Kay. 1969. Basic Color Terms: Their Universality and Evolution. Berkeley: University of California Press. Needham, Rodney. 1973. ‘Prescription’, Oceania 43: 166–81. Scheffler, Harold, and Floyd G. Lounsbury. 1971. A Study of Structural Semantics: The Siriono Kinship System. Englewood Cliffs, NJ: Prentice-Hall. Schneider, David. 1984. A Critique of the Study of Kinship. Ann Arbor: The University of Michigan Press.



Appendix Publications on Kinship by Robert Parkin

Books 1992 The Munda of Central India: An Account of their Social Organization. Delhi: Oxford University Press. 1997 Kinship: An Introduction to Basic Concepts. Oxford: Blackwell. 2001 Perilous Transactions and Other Papers in Indian and General Anthropology. Bhubaneswar (India): Sikshasandhan. 2003 (co-edited with Linda Stone) A Reader in Kinship and Social Organization. Malden, MA: Blackwell Publishing, Chs. 5, 7–11, 15 and 16. 2006 Translation of Louis Dumont, Introduction à deux theories d’anthropologie sociale, with an Introduction. Oxford: Berghahn.

Articles 1983 ‘Lévi-Strauss and the Austroasiatics: “Elementary Structures” under the Microscope’, Journal of the Anthropological Society of Oxford 14(1): 79–86. 1985 ‘Munda Kinship Terminologies’, Man 20(4): 705–21. 1986 ‘History and Kinship: Two Recent Books’ [Review Article of Jack Goody, The Development of Family and Marriage in Europe, and Andrejs Plakans, Kinship in the Past: An Anthropology of European Family Life 1500–1900]. Journal of the Anthropological Society of Oxford 17(1): 50–56.

288

Appendix

1986 ‘Prescriptive Alliance in Southeast Asia: The Austroasiatic Evidence’, Sociologus 36(1): 52–64. 1986 ‘Comparative Munda Kinship: A Preliminary Report’, Oxford University Papers on India 1(1): 59–74. 1987 ‘Kin Classification in the Karakorum’, Man 22(1): 157–70. 1987 ‘Tibeto-Burman and Indo-European Loans in Burushaski Kinship Terminology’, Bulletin of the School of Oriental and African Studies 50(2): 325–29. 1987 ‘Studying Kinship [Review Article of Alan Barnard and Anthony Good’, Research Practices in the Study of Kinship]. Journal of the Anthropological Society of Oxford 18(2): 137–44. 1988 ‘Reincarnation and Alternating Generation Equivalence in Middle India’, Journal of Anthropological Research 44(1): 1–20. [This volume, Ch. 10] 1988 ‘Marriage, Behaviour and Generation among the Munda of Eastern India’, Zeitschrift für Ethnologie 113(1): 67–85. 1988 ‘Prescription and Transformation in Mon-Khmer Kinship Terminologies’, Sociologus 38(1): 55–68. 1988 ‘Sur quelques correspondances et emprunts en termes de parenté‚ austroasiatiques’, Bulletin de l’Ecole Française d’ExtrêmeOrient 77: 315–20. 1989 ‘Some Comments on Brahui Kinship Terminology’, IndoIranian Journal 32(1): 37–43. 1990 ‘Terminology and Alliance in India: Tribal Systems and the North-South Problem’, Contributions to Indian Sociology 24(1): 61–76. [This volume, Ch. 2] 1990 ‘Ladders and Circles: Affinal Alliance and the Problem of Hierarchy’, Man 25(3): 472–88. 1990 ‘Descent in Old Cambodia: Deconstructing a Matrilineal Hypothesis’, Zeitschrift für Ethnologie 115: 209–27. 1992 ‘Dispersed Alliance and Terminological Change in South Asia’, Journal of the Anthropological Society of Oxford 23(3): 253–62. 1992 ‘Cognates and Loans among Aslian Kin Terms’, Journal of the Siam Society 80(2): 167–69. 1993 ‘Middle Indian Kinship Systems: A Critique of Georg Pfeffer’s Interpretation’, Anthropos 88(4–6): 323–36. 1993 ‘On the Definition of Prescription: The Problem of Germanic Kinship Terminologies’, Social Anthropology 1(3): 309–25. 1993 ‘The Joking Relationship and Kinship: A Brief Retrospect’, Journal of the Anthropological Society of Oxford 24(3): 251–63.

Appendix

1995 ‘The Contemporary Evolution of Polish Kinship Terminology’, Sociologus 45(2): 140–52. 1995 ‘Transformatzia i metodologia’, in Algebra rodstva: rodstvo, sistemi rodstva, sistemi terminov rodstva. St Petersburg: Russian Academy of Sciences. 1996 ‘Genealogy and Category: An Operational View’, L’Homme 139: 85–106. 1996 ‘On Dumont’s “Affinal Terms”: Comment on Rudner’, Contributions to Indian Sociology 30(2): 289–97. 1996 ‘Scherzbezhiehungen und Verwandtschaft: nachvollziehen einer Abhängigkeit in der Theorie’, in Georg Elwert, Jürgen Jensen and Ivan R. Kortt (eds), Kulturen und Innovationen: Festschrift für Wolfgang Rudolph. Berlin: Duncker & Humblot. 1997 ‘Tree Marriage in India’, in Klaus Seeland (ed.), Nature is Culture: Indigenous Knowledge and Socio-cultural Aspects of Trees and Forests in Non-European Cultures. London: Intermediate Technology Publications. 1997 ‘Introduction [to special issue on “Kinship and Identity”]’, Journal of the Anthropological Society of Oxford 28(3): 241–47. 1997 ‘Caste, Kinship, and Identity in India’, Journal of the Anthropological Society of Oxford 28(3): 339–49. 1998 ‘Dravidian and Iroquois in South Asia’, in Maurice Godelier, Thomas Trautmann and Franklin Tjon Sie Fat (eds), Transformations of Kinship. Washington and London: Smithsonian Institution Press. [This volume, Ch. 5] 2004 ‘New Books on Kinship’, Journal of the Anthropological Society of Oxford 31(2): 185–96. 2005 ‘Tribal Kinship in Central India: A Reply Article’, Anthropos 100: 567–70. 2009 ‘What Shapiro and McKinnon Are All About, and Why Kinship Still Needs Anthropologists’, Social Anthropology 17: 158–70. 2012 ‘Kinship as Classification: Towards a Paradigm of Change’, Journal of the Anthropological Society of Oxford Online 4(2): 183– 211. [This volume, Ch. 1] 2013 ‘Relatedness as Transcendence: On the Renewed Debate over the Meaning of Kinship’, Journal of the Anthropological Society of Oxford Online 5(1): 1–26. 2013 ‘From Tetradic Society to Dispersed Alliance: Notes Arising from a Chapter by N.J. Allen’, JASO-online 5(2): 194–206. [Part reproduced this volume, Ch. 3]



289

290

Appendix

2013 ‘Note on Oblique Exchange in a Matrilineal Society in the Comoro Islands’, JASO-online 5(2): 207–11. 2015 ‘Indo-European Kinship Terminologies in Europe: Trajectories of Change’, JASO-online 7(2): 205–33. [This volume, Ch. 6] 2018 ‘The Evolution of Kinship Terminologies: Non-Prescriptive Forms of Asymmetric Alliance in Indonesia’, Journal of Anthropological Research 74(2): 232–51. [This volume, Ch. 9] 2018 ‘Why Do Societies Abandon Cross-Cousin Marriage?’ in Warren Shapiro (ed.), Focality and Extension in Kinship: Essays in Memory of Harold W. Scheffler. Acton: Australian University Press. [This volume, Ch. 4] 2019 ‘Substitutability of Kin and the Crow-Omaha Problem’, Structure and Dynamics: e-Journal of Anthropological and Related Sciences 11(1): 73–95. [This volume, Ch. 8] Forthcoming. ‘Arranged Marriages: Whose Choice and Why? Reflections on the Principles Underlying Spouse Selection Worldwide’, History and Anthropology.



Glossary

Affinal alliance. Marriage, especially where seen as the basis of an alliance between whole kin groups rather than individuals. Affines. Relatives by marriage. Aranda system. Symmetric form of bilateral cross-cousin marriage or symmetric prescriptive alliance in which egos are modelled as marrying their second, not first cross-cousins. Also ‘four-line symmetric prescriptive alliance’. Bifurcate collateral. Terminological pattern in which terms for lineal kin are separate from those for parallel collaterals, e.g. F ≠ FB. Bifurcate merging. Terminological pattern in which terms for lineal kin are the same as those for parallel collaterals, e.g. F = FB. Clan. Social group united by descent from a real or putative ancestor. May be patrilineal or matrilineal. Larger than a lineage. Classificatory. Used to describe a situation in which the genealogically minimal specification of a term is understood as covering remoter relatives of similar type, usually identified by tracing connections with ego one or more generations further back. For example, the category minimally defined as FB might also include FFBS; that for MBD might include MFBSD and other second cross cousins. Also used to describe merger of lineal and parallel kin, e.g. F = FB. Opposed to ‘descriptive’, q.v. Cognatic. Form or feature of a terminology that does not distinguish side of family (matrilateral from patrilateral relatives) nor different types of cousin but does separate lineal from collateral kin (e.g. F ≠ FB = MB). The full English terminology is an example. Called ‘cognatic’ by Rodney Needham; preferred by him and

292

Glossary

used here in preference to Robert Lowie’s earlier ‘lineal’, with the same meaning. Complex structures. Lévi-Strauss’s term for marriage systems and practices that do not specify a particular category of relative to marry (unlike elementary structures) and that do not prohibit certain kin in marriage by placing their descent groups out of bounds (like semi-complex structures). Cross-cousin marriage, bilateral, matrilateral. Marriage to cross cousins may envisage or allow direct exchange between two spouseexchange groups (bilateral, with marriages between cross cousins who are simultaneously MBC and FZC to each other). Can therefore be considered symmetric. This is also Lévi-Strauss’s ‘restricted exchange’. Alternatively, marriage to cross cousins may require both wife-takers and wife-givers to come from separate groups from ego’s, in which case three spouse-exchange groups are required and exchange is asymmetric. Associated with prescribed MBD/FZS marriage (matrilateral cross-cousin marriage), MBS/ FZD marriage (patrilateral cross-cousin marriage) being banned. Lévi-Strauss’s generalized exchange. This and Lévi-Strauss’s restricted exchange compose his ‘elementary structures’. In both matrilateral and patrilateral cross-cousin marriage, it is the female spouse that is being referred to (i.e. MBD as a matrilateral relative, FZD as a patrilateral relative). Patrilateral cross-cousin marriage can be modelled theoretically, but it does not appear to exist anywhere in the world in reality, unlike the matrilateral form. Crow-Omaha. Terminologies that make characteristic equations between generations in descent lines (not normally ego’s) without being prescriptive. Crow terminologies do this matrilineally, Omaha ones patrilineally. Associated with Claude Lévi-Strauss’s notion of semi-complex structures characterized by restrictions on ego marrying into the descent lines or descent groups of near relatives like ego’s parents. Descent, line of. Descent is the sociological feature whereby kin of different generations are linked to one another. May or may not involve actual named social groups, but lines of descent between individuals in different generations can still be traced. Descriptive. Used to describe terminologies that distinguish lineal kin from parallel kin, e.g. F ≠ FB, and where the term for the parallel relative (here FB) is not extended to cover analogous rel-

Glossary

atives traced through generations further back (here FFBS). Opposed to classificatory, q.v. Dravidian. See two-line prescriptive terminologies/bilateral crosscousin marriage. Elementary structures. See cross-cousin marriage. Evolved asymmetric alliance. My term for non-prescriptive but still quite regular patterns of affinal alliance among certain societies in Southeast Asia that possibly evolved from asymmetric prescriptive alliance. Explained in full in Chapter 9 of this collection. GEG marriage. Marriage to a sibling’s spouse’s sibling (GEG in standard notation), or more broadly between groups of siblings. Seen here as an early move away from prescription, which also involves intermarriage between sibling groups but defines them as cross cousins. GEG marriages typically also involve banning repetitions of the same marriages in the following generations. This means that GEG categories cannot also be defined as cross-cousin ones, as there are no traceable links through previous generations of the sort that inform relations between cousins. Genealogical level. Units of differentiation into generations reckoned genealogically. Generation. Similar to genealogical level, but the stress may be on age, not genealogy, in differentiating generations. Generational. Terminologies that merge all consanguines in a genealogical level into the same category/term. Among other things, this collapses the cross-parallel distinctions in a terminology. Also known as Hawaiian. Hawaiian. See generational. Iroquois. A non-prescriptive terminological type seemingly derived from Dravidian or two-line prescriptive. Unlike Dravidian, Iroquois terminologies only recognize the distinction between cross and parallel for the closest kin and do not extend it throughout the terminology. Thus the relative-sex distinctions that underpin the cross/parallel distinction throughout Dravidian are based on absolute sex differences for remoter kin in Iroquois. Iroquois-type societies typically ban first cousins in marriage but may allow remoter ones. Kachin. See three-line prescriptive alliance/matrilateral cross-cousin marriage.



293

294

Glossary

Kariera. Variant of two-line prescriptive/bilateral cross-cousin marriage in which society is divided into four sections created by the intersection of two vertical moieties and two horizontal sets of alternating generations, such that ego marries into the opposite moiety and the same generation set. Kin category. Indigenous semantic unit in a kin terminology; the emic aspect. Kin type. Analytical unit used in discussions and analyses of kinship; the etic aspect. Lineage. Smaller version of a clan, within which all connections are generally traceable. Lineal. Robert Lowie’s earlier term for terminologies that I, following Needham, call cognatic (q.v.). For Needham and I, a lineal terminology is one that sorts categories into descent lines. This includes prescriptive and Crow-Omaha terminologies, as well as Iroquois ones. Negative marriage rules. Rules that lay down categories of relative one should not marry, for example, because of incest rules. Positive marriage rules. Rules that lay down categories of relative one is expected to marry, typically some sort of cross cousin. Prescriptive alliance, terminologies. Asymmetric (see matrilateral cross-cousin marriage), symmetric (see bilateral cross-cousin marriage), four-line (see Aranda), three-line (see matrilateral crosscousin marriage), two-line (see bilateral cross-cousin marriage). Claude Lévi-Strauss’s elementary systems. See also Introduction. Semi-complex structures. See Crow-Omaha. Sudanese. See zero equation. Tetradic kinship/society. Model of kinship based on the Kariera system (q.v.) extrapolated further back in evolutionary time in order to distil its fundamental principles. The four sections of the Kariera are seen in this model as both fundamental sociological units and terminological categories such that only four kin terms are needed. Not actually attested; developed by and identified with N.J. Allen. Zero equation. Terminologies in which or in parts of which every kin type has its own term and does not share a term with – that is, is not equated terminologically with – any other kin type. Murdock’s ‘Sudanese’ type.



Index

abbreviations, kinship, 6–7, 35 Abelam, 82, 240 Aberle, David, 28, 206–13, 222n8 Achuar, 85 Ackermann, Charles, 198, 217, 221n4 Acoma, 222n11 affinity, Ch. 4, and passim Aguarana, 85 Akan, 245, 256n13 Akwe, 200 Albania, 189n17 Albanian, 163 Algonkian (also Algonquian), 113, 206, 212, 221 Ali, Tahir, 135–37, 155–6n3 Allen, N.J., 3, 12, 20n3, 21n15, 28–9, 36–7, 47, 48, 73, 97, 105, 125n12, 145, 162, 187n1, 191n33, 213–14, 221n2, 222n10, 280 alliance, dispersed, 108–9, 276n1, 281, and Chs. 7–9 passim evolved asymmetric, 221, 293, and Ch. 9 passim prescriptive, 7–10, and passim alternate generation equations, 30, 31, 36, 63–4, 75, 217–18, 221n2, 222n11, 286n1 Ambryn, 90 analysis, componential, 34 formal (also semantic), 34, 96, 124n6, 134 Anglo-Saxon, 183, 191n29

Anindilyakwa, 216 Anlo Ewe, 91 Anthropocene, 20n5 anthropology, biological, 3 cultural, 3 evolutionary, 3 physical, 3 social, 3 Apache, 89, 206 Apachean, 206–7 Arabic, 42, 44, 145, 160, 162 Aranda (also Arunta), 38, 39, 47, 83–4, 176n1, 218, 264 Ardener, Edwin, 72–3, 121, 127n33, 162 Ashanti, 245, 256n13 Athapaskan, 113, 207–13, 222n7, 222n8 Atna, 250 Austronesian, 29 Baamba, 242–43 Bahnar, 156n5 Baltic (languages), 14, 161, 163–67, 186 Banaro, 90 Bangwa, 243–44 Baniata, 250 Banks Islands, 249 Bantu (languages), 29 Barnard, Alan, 54, 107 Barnes, Robert, 29, 44, 124n5, 126n20, 198–200, 202–6,

296

Index

217–19, 221, 227–28, 246–49, 255n3, 260, 277n10, 277n11 Barth, Fredrik, 5 Baruya, 91n3 Basso, Ellen B., 31–2, 33, 48, 83, 106, 114, 125n16, 126n22 Batak, 271–74 Karo, 271–73 Toba, 273–74 BaVenda, 250–52 Beattie, John, 240–41 Beatty, Andrew, 262–65, 267, 276n3, 276n4, 277n5, 277n7, 278n13 Beck, Brenda, 31 behaviour, 102, 105, 121 Bemba, 91 Bengali, 63, 122 Berlin, Brent, 29–30, 74, 120–21, 161–63, 283 Bhuiya, 11, 77, 156n6 Bhumij, 141 Bielorussian, 172–73, 188n4, 189n11 bifurcate collateral (terminologies), 30, 42–3, 45, 114, 160, 165, 284 bifurcate generational (terminologies), 40, 114, 125n16 bifurcate merging (terminologies), 30, 42–3, 45, 160, 177, 284 Birhor, 69 Bloomfield, Leonard, 34 Blust, Robert, 29, 33 Boas, Franz, 3, 4, 28 Bovill, Kathryn, 273–74 Bowden, Ross, 230, 237–39, 240, 255n9, 256n11, 276n2 Brahmans, Nambudiri, 126n24 Brahui, 77 Brain, Robert, 243–44 Breton, 185 bridewealth (also brideprice), 63, 87, 237, 242, 263, 267, 269, 270 Brown, Cecil, 120 Bruner, Edward, 127n34 Bulgarian, 169–70, 185–86, 188n4, 190n18 Bunyoro, 240–41

Burushaski (also Burusho), 13, 15, 134–37, 147–48, 160, 161, and Ch. 5 passim Byansi, 41, 45 Canadian (branch of Athapaskan), 209–10 Carneiro, Robert, 3 Carrier, 212 caste, 5, 11, 48, 62, 72, 78, 82, 116, 119–20 Catalan, 161, 163, 177–78 categories, kinship, 6, 35, 294 category, Ch. 4 passim Celtic, 14, 15, 159, 184–85 Chagnon, Napoleon, 112 Cheyenne, 40, 139–41, 153 Chiricahua, 209 Chinese (kinship), 32, 122 Chipewyan, 209, 211 Chomsky, Noam, 34 Chuuk, 234–35. See also Truk Clark-Decès, Isabelle, 118–19 classes, marriage, 37, 89, 109 classification, dual symbolic, 103 colour terms, 6, 29–30, 54, 74, 120–21, 283 communities, speech, 5–7, 33, 55, 115, 122, 207, 283–85 concubinage, 115 consanguinity, Ch. 4 passim Cook, E.A., 86–7 cores, sibling, 113 cousins, cross, passim Cree, 84, 110 Crow-Omaha, Chs. 7–9 and passim Cyrillic, 188n6 Czech, 14, 173–74, 185, 188n4 Dakota, 219–20, 223n15, 252 Dani, 215 Danish, 42, 163, 179–80 Daribi, 87, 92n6 demography, 9 descent, 35–6, 48, 96, 102, 117, 125n9, 207, 228–29, 231–32, 234, 236, 242, 292



297

Index

diagrams, 16–20 Dinka, 5 Dole, Gertrude, 20n1, 28, 31–3, 40, 48, 106, 114, 125n16, 126n21, 126n22, 191n33 dowry, 117–19, 127n29 Drabbe, Father P., 277n8 Dravidian, passim Dumont, Louis, 9, 21n9, 71–2, 78, 97, 99, 101, 115, 118, 126n24, 140 Dutch, 14, 163, 179, 183–85 Dyen, Isidore, 28, 206–13, 222n8 Eggan, Fred, 28, 88 Ehret, Christopher, 29, 42–4 Elkin, A.P., 215 Elmendorf, William W., 28, 33, 114 el Guindi, Fadwa, 2 Ema, 270–71, 274–75 endogamy, 48, 85, 112–14, 126n20 English, 1, 6,14, 15, 31, 35, 43, 47, 162–63, 179, 180, 182–83, 185, 187n1, 189n13, 280, 284 Ensor, Bradley, 5 equations, classificatory, 30–1 Eskimo, 1, 20n1, 43, 46, 47, 187, 191n33 Evans-Pritchard, E.E., 5 Euripides, 174 evolution, evolutionism, 2–6, 10, 15, 27–30, 55, 97, 197–98, 261, 284–85, Ch. 9 passim evolutionists, lexical, 120, 161–62. See also universalists, lexical exchange, exclusive straight sister, 89 generalized, 292 restricted, 292 exogamy, 40, 48, 63, 85, 90, 112, 113, 126, 135, 173, 228–29, 231–32, 234, 241–42, 276n11 extensions, extensionism, 99, 100, 125n9 family, nuclear, 98, 99, 117, 173, 185, 187, 285 Fanti, 245–46

Faron, Louis, 219–20, 252–53, 256n17, 256n18, 256n19 Fêng, Han-yi, 122 Fiji, 114 filiation, 102, 125n9 focality, 99–100 Foi, 90 Fordata, 267–69 Forge, Anthony, 240 Fortes, Meyer, 54, 228 Forth, Gregory, 29, 32, 124n5, 265–66, 271 Fox, James, 32–3, 266–67 French, 161, 163, 166, 171, 174, 177–80, 182–84, 186, 190n26 Friedrich, Paul, 170–73 Fuller, Chris, 116–17 functionalism, 3, 47 Gamble, Clive, 54 Garo, 126n23 Gaulish, 184 Gê, 220, 223n15, 276 genealogy, 9, Ch. 4 passim generation, 7, 293 generational (terminologies), 36, 41–2, 67–8, 140–41, 146, 148, 242, 245, 252, 269, 280, 284, 293. See also Hawaiian German, 14, 42, 163, 166, 168, 172, 178–80, 182–85, 191n32 Germanic, 14, 159, 167, 179–86 Gifford, Edward W., 249 Gnau, 82–4 Gonja, 91 Good, Anthony, 21n11, 31, 53, 118, 125n14, 127n28 Goody, Jack, 47, 175 Goodenough, Warren, 34, 89, 234–35 Gordon, James, 269–70 Graham, Penelope, 125n5 Greek, 14, 159, 161, 174–77, 179, 186, 188n6 Guermonprez, Jean-François, 32–3 Gujarati, 77, 122 Gumuz, 83 Gurage, 41, 88

298

Index

Hage, Per, 29, 55 Haida, 250 Hallowell, A. Irving, 28, 84 Hamid, S. Shahid, 135 Hammel, Eugene, 172–73, 189n16, 189–90n17, 190n24 Hano, 88 Hawaiian (terminologies), 30, 36, 40–3, 45–8, 113, 114, 146, 163, 176, 207, 209–10, 235, 280, 293. See also generational Haya, 241–42, 251 Henderson, Richard N., 222n11 Henley, Paul, 92n9, 111 Heredotos, 175 Héritier, Françoise, 218–19 Herzfeld, Michael, 176 Hickerson, H., 113 Hicks, David, 29, 107, 124n5 Hindi, 42, 122, 160–61, 163–65, 171, 185–86; also Ch. 5 passim history, 3, 4, 27 speculative, 3 Ho, 141, 264 Hocart, Arthur Maurice, 54, 174–75 Homer, 174–76 Homo sapiens, 3–4, 20n5 Hopi, 88, 211 humans, cognitively modern, 4, 20n5 Hungarian, 42 hunter-gatherers, 47–8, 112 Iatmul, 82, 197, 217–18, 264 Ibo, Onitsha, 160, 222n11 Icelandic, 163, 181–82, 191n30, 191n31 incest, 6, 41, 45–6, 64, 67, 73, 90, 109, 198, 6, 41, 45–6, 64, 67, 73, 80, 109, 198, 232, 267, 294 Indo-European, 13–14, 44, 63, 77, 126n20, 127n34, 152–54, 156n6, 281, and Ch. 6 passim Inuit, 47, 191n33. See also Eskimo Irish, 184–85 Iroquois, Chs, 5, 7, and passim Islam, 44

Italian, 43, 163, 177, 190n27 Ives, John W., 113, 126n20 Jat, 138 Juang, 10–12, 21n12, 91n1, 138–39, 141, Ch. 2 passim Just, Roger, 176, 189n16, 190n24 Kachin, 1, 15, 37, 47, 103, 197, 200–2, 222n12, 260, 277–8n12, 293 Kalapalo, see Xingu Carib Kammu (also Khmu), 107 Kandoshi, 84–6, 92n5 kanya dan, 116–17 Kapadia, Karen, 117–19, 127n27 Kariera, 28, 30, 38, 73, 105, 221n2, 294 Kaska, 209–12 Kato, 210, 212 Kay, Paul, 29–30, 74, 120–21, 161–63, 283 Kayapó, 90 Kédang, 21n7 Keo, 265–66, 274 Ketchika River (dialect), 210, 212 Kham Magar, 126n23 Khambu, 213–14 Kharia, 69, 77 Khasi, 69, 156n5 kin, substitutability of, 15, 16, Chs. 8, 9 kinship, Amazonian, 92n9, 111 North Indian, 116–18, 228, and Chs. 2, 5 passim Kinship Circle, 2 Kohler, Josef, 228–31, 246, 248–49 Konkani, 77, 153 Konkomba, 244–45 Konso, 91 Koraput, 11, 75, 152 Korku, 69 Korn, Francis, 217–18, 222n12 Korwa, 69, 138 Kpelle, 91 Kroeber, Alfred, 8, 34, 50, 248



299

Index

Kronenfeld, David B., 48–9, 214, 222n11, 245–46. 256n12, 256n13, 256n14 Kryukov, Michael, 32, 44–6, 127– 8n35, 145, 170, 185n5, 190n18 Kuikuru, 114 Kuma, 84 Kumaon, 148–50, 152–53 Kwoma, 230, 237–40, 255n9, 255n10, 256n11, 276n2 Labby, David, 230–33, 255n2, 255n4 Lakher, 244 Lamahalot, 110, 126n20, 270, 277n10, 277n11 Lamet, 125n10, 202 Lassik, 210, 212 Latin, 14, 41–2, 154–55, 159, 164, 177–79, 183, 184, 185–86, 188n9 Latvian, 163–67, 186, 188n4, 189n12 Lavrovski, P.A., 170, 190n18 Lawa (also Luaq), 107 Layton, Robert, 54, 107, 126n19 Leach, Edmund, 99, 200, 235–36, 260 learning (kin-term use), 100 Lembata, 21n7 levels, genealogical, 7, 20n1, 293 levirate, 189n17, 246, 252–53, 256n17 Lévi-Strauss, Claude, 45–7, 50, 81, 88–9, 98–9, 108–9, 125n11, 197, 218, 228–31, 244, 247–48, 260, 262, 268, 292 life forms, 6, 74, 120–21, 162 line of maternal affiliation, 266 line of origin, maternal, 266 Lithuanian, 14, 163–67, 186, 188n4, 188n9, 189n11, 189n14 loans (between languages), 186 Lomblen, 21n7 Lounsbury, Floyd, 9, 34, 133, 160, 203, 206–7, 235, 242, and Ch. 4 passim Lowie, Robert, 1, 8, 22n16, 35, 43, 187–8n1, 292, 294

Makassae, 270 Malay, 41 Maler, 21n13 Malinowski, Bronisław, 3, 4, 235, 255n7 Malpahariya, 21n13, 76–7, 125n15, 151, 152 Malto, 21n13, 76–7 Mambai, 269–70 Mamboru, 125n5 Manga, 86–7 Manggarai, 269–70 Mapuche, 82, 219–20, 248, 252–54, 256n19, 256n20 Marathi, 73, 77, 153 Maravar, 21n11 Marett, Robert L., 3 Marra, 216 marriage rules, negative, 9, 294. See also incest positive, 9, 64, 70, 92n9, 135, 154, 232, 237, 272n1, 294 marriage, passim companionate, 116–17 FBC (father’s brother’s child), 43–4 GEG (sibling’s spouse’s sibling), 10–13, 15, 51, 108, 247–48, 281–82, 284, 293, and Chs. 2, 3, 5 passim oblique, 43. See also ZD marriage patrilateral parallel cousin, 135 ZD (sister’s daughter), 118–19, 127n28, 256n16. See also marriage, oblique Mauss, Marcel, 28, 102 Maybury-Lewis, David, 8, 9, 220, 222–3n15 Mbae, 107 McConvell, Patrick, 38, 215–16, 222n11 McDougal, Charles, 11–12, and Ch. 2 passim McKinley, Robert, 41, 198, 227 McKinnon, Susan, 267–69 Melpa, 90–1 memory, genealogical, 67, 83–6, 286n2

300

Index

Merton, Robert Mescalero, 209 metaphor, 100 Miao, 126n17 Mickey, Barbara H., 222n11 Middle East, 44 missions, missionaries, 84–5, 110–11 Miwok, 249 Mkako, 88, 91n3 Mnong Gar, 202, 221n5 Modh, Sandra, 270, 277n9, 277n10, 277n11 Moi, 91n3 moieties, 12, 36, 37, 63, 109, 215, 221n2, 281, 286n1 Mon-Khmer, 77, 156n5 Moore, Melinda, 127n26 Morgan, Lewis Henry, 3, 10, 27–8, 42, 45, 50, 52, 54, 104, 122 morphology, social, Ch. 4 passim Mosko, Mark S., 255n5 Moso, 127n25 mother’s brother of origin, 267 Muller, Jean-Claude, 89 Munda, 11, 32, 39, 40, 62, 68–9, 75, 77, 81–4, 91n1, 91n5, 108, 119–20, and Chs. 2, 5 passim Mundumugor, 84 Murdock, George, 21n15, 42–3, 46, 145, 160, 191n33, 234, 243, 248 Muria, 125n17 Muyu, 88 myth, 32 Nabesna, 212 NaDene, 211 Nakane, Chie, 116 Narasimhan, Haripriya, 116–17 Navaho (also Navajo), 206–11 Nayar, 77, 115–16 Needham, Rodney, passim Ngawbe, 84 Nias, 222n13, 262–65, 267–68, 271, 275, 277n5, 277n7, 278n13 Norwegian, 161, 163, 179–81, 183, 188n4

Nostratic, 29 Nuer, 5 Numan, 113 Nungyuburu, 216 Ojibwa, 84, 110, 113 Old Church Slavonic, 186 Old Frisian, 182 Old Norse, 179–80, 190n28, 191n29 Old Prussian, 163–64 Opler, Morris, 208–9, 212 Oraon, 69 Oriya, 63 Osage, 249 Oxford, 3, 28 Pacific (branch of Athapaskan), 209–10 Pahari, 77 Pathan, 5 Pauwels, Simone, 267–68, 277n8 Pende, 250 Pfeffer, Georg, 155–6n3 Piña, Gabrielle, 256n20 Pleistocene, 20n5 Polish, 42–3, 46, 52, 156n7, 161, 164, 167–74, 185–86, 189n11, 189n16, 190n18 polysemy, 98, 100 Portuguese, 163, 177–78 prescription, 7–10, 102, 105, 123, 138, 154, 197, 247, 250, 265, 269, 275, 281, 286n3, 293 Pukupuka, 89 Puluwat, 234–35 Punjabi, 44 Purum, 74, 125n10, 126n23, 202 Radcliffe, Brown, A.R., 3, 4, 47, 102, 215, 228–31, 241, 248, 253–34 Read, Dwight, 2, 34 Reining, Priscilla, 241–42 relativism, cultural, 4, 47 replacement, lexical, 14, 33, 77–8, 153 Rindi, 271, 274–75



301

Index

Rivers, W.H.R., 28, 52–3 Rivière, Peter, 50, 124n5 Romance (languages), 14, 15, 159, 163, 174, 176–79, 185, 186 Romanian, 43, 174, 177, 188n4, 190n27 rule(s), equivalence, 99 extension, 99 four-got, 51, 70, 82 sapinda, 51, 82 Russian, 164, 169, 170–72, 185–86, 188n4, 190n18 Sahlins, Marshall, 3, 114 Saki, 212 Salish, 114 Samo, 39, 41, 83, 197, 218–19 San Carlos, 207, 209 Sanskrit, 154–55, 174 Santal, 68, 75, 138, 141, 264 Sapir, Edward, 208–9, 212 Scandinavian (languages), 14, 42, 154–55, 156n7167, 179–82, 185–86 Scheffler, Harold, 9, 13, 34, 90, 134, 139, 213, 271–72, 283, and Ch. 4 passim Schneider, David, 2, 47, 124n2, 231–32, 255n2, 283 sections, marriage, 37, 49, 102, 107 Sekani, 210, 212 Selaru, 267–70, 277n8 Seneca-Iroquois, 104 sense generalization, 100 sense specialization, 100 Serbo-Croat, 167–69, 170–71, 185–86, 188n4, 190n18 Service, Elman, 3, 28 Shapiro, Warren, 126n21 Shavante, 90, 220, 222–3n15 Sherente, 220, 222–3n15 Sherpa, 41, 45, 213–14 Shina, 77, 153 Shoshone, 88 Siane, 214 Singarimbun, Misri, 271–73

Sinhalese, 33, 35, 77, 110, 127n34, 153 Sinkyone, 212 Sirionó, 34, 37, 99, 103, 124n6, 126n21, 202 skewing, 200, 245, 251, 256n13 Slave (dialect), 210 Slavonic (also Slavic), 14, 42, 159, 161, 164, 167–74, 186, 188n6, 188n9, 188n11 Smith, James, 113 society, tetradic, 28–9, 30, 36–7, 49, 53, 73–4, 107, 221n11, 280, 281, 286n1, 294 sororate, 189n17, 252–53, 256n17 soul substance, recirculation of, 64 South Tutchone, 209, 211 Spanish, 163, 177–78 Spoehr, Alexander, 28 Stayt, H.A., 250–52 Steward, Julian, 3 structuralism, 2, 28, 34, 47, 138, 283, Ch. 4 passim structures, complex, 45–6, 50, 107, 197–98, 292 elementary, 45, 197–98, 292–94. See also prescription semi-complex, 45, 51, 81, 88, 92n9, 107, 197–98, 218, 262, 292, 294. See also Crow-Omaha Sudanese, 21n15, 92–3, 145, 160, 294. See also terminologies, zero-equation Swedish, 163, 179–80, 190n28 Szemerényi, Oswald, 174–75, 191n29 Tahltan 209–11 Tait, David, 244–45 Tallensi, 244 Tamil, 31, 35, 97, 101, 118, 141 Taylor, Anne-Christine, 85, 112 terminologies, cognatic, passim lineal, 1, 20n1, 22n16, 35, 43, 114, 126n21, 146, 187n1, 292, 294 prescriptive, passim zero-equation, passim

302

Index

terms, deletion of, 33, 38, 43, 46–7, 74, 173, 186 terms, kin, 6, 35, also passim classificatory, 7, 100, 108, 138, 146–48, 154 Tevar, 118 theories, mid-range, 47 theory, overlay, 214, 222n11, 256n12 tetradic, 3, 12, 30, 54, 75, 105, 107 Thie, 266–67 Tibeto-Burman, 69, 153–54 time, 20n11 Tiv, 244 Tiwi, 110 Tlingit, 211, 249–50 Tonga, 91 Traube, Elizabeth, 269 Trautmann, Thomas, 44, 101, 104, 115, 122, 124n6, 141, 199–206, 221, 261, 274–76 tribes, 5, 10, 48, 156n6, 281, Ch. 2 Trio, 50, 124n5 Trobriand Islands, 235–37 Truk, 89, 234–35, 277n6. See also Chuuk Tsimshian, 37 Tsumje, 213–14 Tupian, 40 Turanian, 41 Turkish, 176, 189n16, 190n24 Txicao, 37 types, kin, 6, 35, 294 Tzotzil, 88 Umeda, 84, 91n3 Ungarinyin, 197, 215–17, 222n11

universalists, lexical, 120, 161–62, 173 Upper Liard, 210–11 Urarina, 112 Urdu, 44 van Wouden, F.E., 260 Vattima, 116–17 Vellan Chattiar, 118 Viveiro de Castro, Eduardo, 83–4 Wailaki, 210, 212 Walker, Ian, 92n6 Warao, 105, 112, 120 Weiner, Annette, 235–36 Welsh, 162, 185 Wheeler, Charles, 29 White, Leslie, 3, 28, 50 Wikmunkan, 35, 50, 52 Williamson, Margaret H., 237, 239 Winter, Eric H., 242–43 witchcraft, 83, 269 Witowski, Stanley, 29–31, 120–21, 162–63 Wrigley Slavey, 113 Xingu Carib, 32, 48, 83, 106, 114 Yafar, 84 Yamdena, 277n8 Yanomamo, 112 Yap, 230–34 zadruga, 168, 173, 187, 189n16, 189–90n17