The lowland South America culture area: towards a structural definition¹

Abstract

This article, originally written in 1973 and published here for the first time, argues that, despite their evident cultural variety, the Indigenous societies of lowland South America share an invariant structural characteristic that define them as a ‘culture area’. This characteristic is a two-line relationship terminology articulated by a principle of direct exchange. An analysis of a wide range of societies from the region reveals this invariant characteristic within their terminological differences. The article is preceded by a short introductory note written specifically for this publication.

Key Words:
Lowland South America; Culture Area; Two-line relationship terminology

Resumo

Este artigo, originalmente escrito em 1973 e publicado aqui pela primeira vez, propõe que, a despeito de sua evidente variedade cultural, as sociedades indígenas das terras baixas da América do Sul compartilham uma característica estrutural invariante que as define como “área cultural”. Esta característica é a sua terminologia de relacionamento em duas linhas articulada pela troca direta. Uma análise de terminologias de uma ampla gama de sociedades da região revela tal característica invariante apesar das diferenças terminológicas. O artigo é precedido por uma curta nota introdutória escrita especificamente para a publicação.

Palavras-chave:
Terras Baixas da América do Sul; Área Cultural; Terminologia de relacionamento em duas linhas

Resumen

El presente artículo, escrito originalmente en 1973 y publicado aquí por primera vez, propone que, a pesar de su evidente variedad cultural, las sociedades indígenas de las tierras bajas de América del Sur comparten una característica cultural invariante que las definen como un “área cultural” específica. Esta característica es su terminología de relacionamiento que se articula en dos secciones a partir del intercambio directo. Un análisis de las terminologías propias de una amplia gama de las sociedades de la región revela tal característica invariante a pesar de las diferencias terminológicas. Precede al artículo una breve nota introductoria escrita especialmente para esta publicación.

Palabras clave:
Tierras Bajas de América del Sur; Área Cultural; Terminología de relacionamiento en dos secciones

Foreword

My paper, which the Editors of this journal, have been kind (or foolhardy) enough to publish, first saw the light of day as a paper presented at a symposium at the AAA meetings in New Orleans in December 1973. The symposium was entitled ‘Marriage practices in Lowland South America’ and was organised by the late Kenneth Kensinger. Participants at the symposium were invited to submit versions of their papers for publication and, in due course, i.e., 1984, a volume including some of the papers, but not mine, appeared. I will return to that below but first it should be noted that this was the first of a series of symposia which Kensinger organised in conjunction with others.

The proceedings of some of these symposia were published in a typescript series called ‘Working papers on South American Indians’, of which Kensinger was the series editor and each had a different issue editor. Five volumes of these papers finally appeared between 1979 and 1984, covering an interesting range of topics and all with contributors whose names will be familiar to anyone who works in Lowland South America. Despite that it has been rare to see reference to them in recent works; perhaps the result of the relatively small number produced and their format.

The proceedings of the 1973 meeting were finally published in 1984 as No 14 in the Illinois Studies in Anthropology. The volume did not contain all the papers that were presented at the meeting but did include two that were not. Submitted papers, and I do not know which were not, were sent out for review. As a result of an appeal on the Society of Anthropologists of Lowland South America website for the whereabouts of Kensinger’s papers, especially those relating to the publication of this volume, Norman Whitten got in touch. He very kindly obtained for me from Illinois University Press a letter relating to the review process. The Press had only sent the contributions to one reviewer, F K Lehman, a Burma specialist, at Illinois University. The Press was able to produce a covering letter to what was apparently a seven-page single-spaced report of which it did not have a copy. Nor is this report among his papers deposited at the University of Uppsala which Jan-Ǻke Alvarsson was kind enough to search for me.

The letter does contain mention of a few of the contributions but not mine. The identity of the reviewer, however, is significant as it fits exactly with what I recall happened. Lehmann subscribed to the formal method of kinship analysis propounded at the Yale by Floyd Lounsbury and Harold Scheffler. This would certainly explain why, as I remember it, I should have received a letter from Kensinger proposing that I re-draft my paper using structural semantics. This was so totally against what I was trying to do, I withdrew my contribution and it has remained unpublished until now. It has, however, received a certain accolade as ‘the most quoted unpublished paper on Amazonian kinship’.² Furthermore it contains the form of kinship analysis that has been most widely adopted by Lowland South American anthropologists.

I have left the paper almost exactly as it was presented for publication. I quickly realised that if I tried to modify it the whole argument would quickly unwind. It has to be read as an historical document, nearly half-a-century old, that reflects its age. Since I did not know when I was finalising it for publication which of the papers presented at New Orleans in 1974 were to make it to the published version, I included comment on all of them. This means that reference will be found to the following papers although they did not appear in the published version: Michael Harner on the Jívaro, Napoleon Chagnon and Jacques Lizot on the Yanomam, Janet Siskind on the Cubeo, and general comments on the Northwest Amazon by John Sorenson. Finally, two of the Gê-speaking people, the Apinayé by Roberto Da Matta and the Bororo by Chris Crocker, did not appear in the published volume although they receive comment in my paper. The two contributions that are included in the volume and discussed here but were not presented at the symposium were that on the Kagwahiv (Parintintin) by Waud Kracke and the Canelos Quichua by Norman and Dorothea Whitten.

I

The division of the aboriginal population of South America into culture areas that has become and remained the best known and most influential is that represented by the first four volumes of the Handbook of South American Indians.³ In this work are recognised the Marginal Andean, Tropical Forest, and Circum-Caribbean cultures.

The criteria employed in defining and distinguishing these are various culture traits, although the emphasis placed on a particular element varies from one volume to another. Thus, the Circum-Caribbean culture is identified by the presence of a class system and temple cults, while the Tropical Forest and Marginal cultures are defined by the respective presence and absence of agriculture and the possession of certain material artefacts. That there was something wrong with this approach quickly became apparent and even before the whole handbook was published the editor, J.H. Steward, was admitting that the criteria used and the method in which they were used had given rise to certain classificatory difficulties. Thus, as he wrote: “The Circum-Caribbean peoples would belong with the Tropical Forest if material elements were emphasized and with the Andes if social and ritual elements were given more weight.” (H.S.A.I., Vol. 5: 671). In the summary article in which his criticisms of the earlier volumes appear Steward put forward an alternative classification based primarily on sociopolitical and religious patterns. While this seems a step in the right direction, Steward’s reclassification remains open to the same criticisms as those that can be leveled at the use of mainly cultural criteria.

If one looks at Lowie`s Introduction to Volume 3 of the Handbook, that concerned with the Tropical Forest peoples, the striking thing is that he appears to be paying attention solely to the differences. The impression given is not that we have here a distinct culture are typified by some common characteristic but rather one of endless, almost arbitrary, variation. This situation is little improved by adopting Steward’s criteria of patterns of sociopolitical organization and religious institutions, for the diversity of such phenomena is still too great to form the identifying feature of a Tropical Forest culture.

However, having criticised these approaches for stressing the diversity within the area, it must be admitted that the variation does exist. Indeed a curious, even paradoxical, situation has arisen in as much as the definition the Tropical Forest culture comprises a list of culture elements that tends to emphasize variation, while the blanket term Tropical Forest culture has drawn attention away from what are extraordinarily interesting and often significant changes in the distribution of elements as one moves from tribe to tribe. This does not mean that we are denying the existence of a Tropical Forest culture. Whatever we call it there is something or things that the people who live in the tropical forest region of South America share in common. Those who work with Tropical Forest peoples are fully aware that they are working within a single cultural tradition. They tend to be at home with each other’s ethnographies: they recognise and appreciate the basic backgrounds, and quite often specific details parallel their own experiences and knowledge. They are often familiar not only with the material and institutional features of man’s existence in the area as portrayed in ethnographies from widely dispersed regions of it, but also with how over and over again the same objects and beings - anacondas, jaguars, stones, and rotten trees - are used as symbols to express similar ideas. Accordingly it is not the aim of this article to call in question the existence of this cultural tradition. Indeed there is no obstacle and every advantage in accepting the Tropical Forest culture as a polythetic class. However, there is no need to stop there and what needs to be done is to consider whether there is not after all some features or features that all the peoples that are representative of this cultural tradition share in common. If there is such a feature or features, it seems unlikely that it or they will be found at the phenomenal level of political organization, residential groupings, economic institutions or material culture; in other words, at the level at which variation is greatest. It is necessary to look beyond these superficial aspects to see whether there is not some more fundamental, structural element that is the common characteristic of all Tropical Forest peoples.

II

The aim of this article is tentatively to propose a single characteristic that is to be found among all Tropical Forest peoples. This feature is a two-line relationship terminology that is articulated by the principle of direct exchange. The observable form in which this feature manifests at the level of social organization varies greatly throughout the region. These forms range from societies with unilineal descent rules, exogamous moieties and other social institutions to societies with a cognatic mode of organization and lacking any permanent identifiable groupings. The term dual organization may sometimes be appropriately applied to the former type of society, but a significant distinction must be made between those societies whose relationship terminologies form a consistent fit with the grosser social units and where, accordingly, direct exchange would function in the absence of these grosser units, and those societies in which direct exchange depends on the presence of such grosser social units in order to operate. It will be argued that societies of this latter type fall outside the Tropical Forest culture and the handful of such societies represented in this volume will be dealt with in Section VI

For those societies lacking unilineal descent rules and identifiable social formations, the description dual organization is always inappropriate since the dual nature of such societies exists only as a feature of the inter-relationship of the social categories normally referred to as kinship terms, and does not manifest itself in an institutionalised form except as an occasional occurrence of bilateral cross-cousin marriage. Thus, whatever the mode of social organization, the essential thing is the presence of a relationship terminology that exhibits some quite specific characteristics. These will be discussed in detail below but first there are two further matters that need to be cleared up.

An aspect of the principle of direct exchange that articulates these terminologies is its prescriptive nature, and it will be as well to make perfectly explicit the way in which this term is being used. Prescription is taken to refer to a formal feature of a set of social categories which we normally refer to as a relationship terminology. Although it may be accompanied by an explicit rule, prescription is an intrinsic feature of certain relationship terminologies and its presence or absence can be discovered by the correct analytical procedure. Because prescription is a formal feature of a relationship terminology it is not possible to ascertain its presence or absence by the observation of the aggregation of individual actions. Prescription as either a rule or an ordering principle can no more be demonstrated to exist by statistical means than any rule or principle. This does not imply that the actions of individuals are unimportant, for clearly they are and have to be taken into account in any field of study. It is simply that if we are to find some feature that all Tropical Forest peoples have in common, it is not at the level of individual behaviour that we must look but at the level of social categories and the principles by which they are ordered. It might be noted here that used in this sense prescriptive and preferential are not exclusive types of society, for preferences may be made within the prescribed category. Thus within the prescribed category there may be certain specific, genealogically defined individuals with whom marriage is preferred because certain advantages accrue to marriage with them. Preferential marriage, as opposed to prescriptive systems, can be expressed statistically. It will be found that many of the societies discussed in Section III have preferences (and prohibitions) about which members of the prescribed category to marry.

The second point is that the term alliance has been avoided. This has been done on purpose since it seems that at least some of the misunderstanding that centres around the study of prescriptive systems results from the use of this term. Alliance in English is a more concrete affair than it need be in French. Alliance in English seems to imply the relationship, often political in tone, between recognisable groups of individuals, however defined. This is fatal for it has had the tendency to make anthropologists think that prescriptive alliance entails relationships between identifiable groups, and furthermore (and worse) that these groups are in some way representative of the relationship categories. In some societies this situation holds good; there are alliances between groups, and even a link between groups and categories. However, in many societies, and among numerous Tropical Forest people, this is not the case, and for this reason the notion of alliance has been excluded as being potentially misleading.

If the feature common to the Tropical Forest culture is a two-line relationship terminology articulated by prescriptive direct exchange, how is its presence to be recognised? In the first place, since the feature is an intrinsic quality of a relationship terminology, it emerges when the correct analytical procedure is applied. A fundamental and essential aspect is that the categories composing the relationship terminology may be coherently and consistently ordered in two lines, on at least two genealogical levels and usually three. For the purposes of this paper these two lines will be referred to as “kin” and “affines’’ although what we call them is relatively unimportant. What is important is to stress that these lines have nothing to do with descent but are a feature of the terminology revealed through analysis. The existence of these two lines is diagnosed by the recognition of certain equations and distinctions that occur in the distribution of genealogical specifications among the terms in a relationship terminology. The correct distribution of the specifications will also identify the direct exchange relationship between the lines and prescriptive nature of the terminology.

The following description outlines the basic equations and distinctions that are to be found in simple forms of two-line terminologies with prescriptive direct exchange. For the present purpose the easiest way to refer to these equations and distinctions is by the genealogical specifications of the particular terms. However, one must be aware of the dangers of this approach and it must be stressed that it is being used as a convenient and idiomatic shorthand that may be totally alien to the people whose categories we are concerned to understand. The numbers in the discussion refer to the positions in Figure 1.

At ego’s own⁴ genealogical level the basic features of the type of terminology under consideration include one term (5) that covers ego’s male siblings and parallel cousins, a term (7) for female siblings and parallel cousins, a term (8) for male cross-cousins, and a term (6) for female cross-cousins.

The prescriptive and direct exchange nature of the terminology becomes apparent when we add affinal specifications. Term 8, that for male cross-cousins, also applies to sister’s husband and wife’s brother, and the term 6, that for female cross-cousins, to wife, potential wife, and brother’s wife.

At the first ascending genealogical level the important equation is that of father with father’s brother (1), and the significant distinction is that of this term from that (4) for mother’s brother. The corresponding terms for women equate mother with mother’s sister (2), and distinguish these from father’s sister (2). The term (4) for mother’s brother covers the affinal specifications father’s sister’s husband and wife’s father, and that (3) for father’s sister the mother’s brother’s wife and wife’s mother.

At the first descending genealogical level the various specifications are usually distributed among four terms in the following way: a term (9) for son, brother’s son, male parallel cousin’s son, and female cross-cousin’s son; a term (11) for daughter, brother’s daughter, male parallel cousin’s daughter, and female cross-cousin’s daughter; a term (12) for sister’s son, female parallel cousin’s son, and male cross-cousin’s son; and a term (10) for sister’s daughter, female parallel cousin’s daughter, and male cross-cousin’s daughter. The affinal aspects equates the sister’s son (12) with the daughter’s husband, and the sister’s daughter (10) with the son’ wife.

This is a simplified review (for example, we have chosen to ignore the diagnostically unimportant second ascending and descending genealogical levels) of the most important equations and distinctions characteristic of the type of terminology under consideration. Even so the reader should have no difficulty in appreciating why such a terminology can be described as two-line, direct exchange and prescriptive. There are numerous societies whose terminologies exactly or closely approximate this ideal and simple form, but, not surprisingly, many societies have more complex terminologies. Common variations found within the Tropical Forest area include different terms for older and younger siblings, terms covering specifications from more than one genealogical level, and the absence at some level of the distinction between lines (e.g., a single term for both brother’s and sister’s children). However, variations such as these are ones of degree and not of kind, and none of them threatens the fundamental structure of terminologies.

III

We can now turn to consider actual examples of Tropical Forest peoples and their relationship terminologies. All it is intended to do in this paper is to illustrate the tentative proposal and for this purpose examples have been limited to those peoples who are the subjects of the other papers in this volume. Relationship terminologies are available for all of them except the Wachipaeri. It is not intended to deal with each terminology in detail but rather to review each one, picking out salient features and discussing apparently anomalous characteristics.

1) The Piaroa terminology is an excellent example of the type we are discussing and for that reason it deserves some extended treatment (see Kaplan, 1975). The terminology is as follows:

• cha’o: F, FB, MZH.

• cha’hu: M, MZ, FBW.

• chiminya: MB, FZH, WF.

• chiminyahu: FZ, MBW, WM.

• chú’buo: eB, FBSe, MZSe.

• chihawa: yB, FBSy, MZSy.

• chirekwa: FZD, MBD, D, BW.

• chú’buoa: eZ, FBDe, MZDe.

• chihawahu: yZ, FBDy, MDZy.

• chisapo: FZS, MBS, WB, ZH.

• chitti: S, BS, FBSS, M ZSS, FZDS, MBDS.

• chuhorihu: ZD, FBDD, MZDD, FZSD, MBSD, SW.

• chittihu: D, BD, FBSD, MZSD, FZDD, MBDD.

• chuhori: ZS, FBDS, MZDS, FZSS, MBSS, DH.

Inspection of this set of terms quickly reveals that there are four terms at each genealogical level except ego’s where a distinction is made between elder and younger brothers and sisters. Except for this minor variation which, it has already been noted, is very common in Tropical Forest South America, the distribution of genealogical specifications among the terms is exactly the same as that portrayed in Figure 1. In order to make the comparison easy the terms have been numbered as in Figure 1.

An interesting and unusual feature of the Piaroa relationship terminology is that, with the exception of the cross-cousin terms, the same root occurs in two terms at each genealogical level. A gender suffix is then used to distinguish between the masculine and feminine forms. For example, cha; cha’o, F; cha’hu, M. This usage also serves to accentuate the opposition between kin and affines. At the first ascending genealogical level the same root is shared by mother and father as opposed to wife’s mother and father. At ego’s own level the same root is shared by siblings, and at the first descending genealogical level the same root is shared by son and daughter as opposed to their spouses.

2) The Culina terminology is another example exhibiting all the diagnostic equations and diagnostic equations and distinctions outlined in Section II (see Townsend’s and Adam’s paper). There are two small points that might be made about this terminology. First, while there is a term for elder brothers and male parallel cousins, and another for elder sisters and female parallel cousins, there is only one term for all younger siblings and parallel cousins. This failure to differentiate terminologically between the sex of younger siblings and parallel cousins is very common among Tropical Forest peoples. The second point is that father and mother are distinguished from their same sex siblings in the following way:

• abi: F; abi ohuaha: FB (ohuaha: eB).

• ami: M; ami onihi: MZ (onihi, eZ).

This detail has no consequences for the prescriptive two-line nature of the terminology.

3) The Jívaro are another people who, with one exception, have a perfectly straightforward two-line prescriptive terminology (see Harner, 1973). At ego’s own genealogical level and at the first descending level the standard equations and distinctions are made. The problem lies at the first ascending genealogical level where the term nuku covers the following specifications:

• nuku: M, MZ, FZ, FBW, MBW.

In other words all females of the first ascending genealogical level are terminologically equated. However, Harner notes that there is a term tsatsa that is substituted for nuku if ego marries a daughter of a MBW, a FZ, or a FBW. There is some obscurity here still since it is not clear whether tsatsa applies only to these genealogical specifications when and if such a person becomes ego’s wife’s mother or whether it is the term for wife’s mother generally, a specification that is conspicuous by the absence from the list of reported terms. A possible solution is that nuku means “woman of mother’s genealogical level” and that tsatsa is used to distinguish one particular female status from among all such women, i.e., the wife’s mother. If this surmise is correct, then tsatsa could cover any genealogical specification that the wife’s mother happened to be, although given all the other features of the terminology the FZ and MBW seem more likely candidates for the position of wife’s mother than is the FBW. Indeed it might be noted that there is some support for this suggestion to be found on the kinship chart provided by Harner (1973:102) where both FZ and MBW are shown as being either nuku or tsatsa, but FBW is shown only as nuku.

4) Although there is considerable variation in detail and organization from one Yanomam⁵ subgroup to another, it now seems perfectly clear that all the subgroups have prescriptive two-line relationship terminologies. As far as the terminologies themselves are concerned there is differing emphasis from one subgroup to the next in the use of relative age terms, but a second and more interesting feature is the use of a term to cover specifications from more than once genealogical level, something we have not met in the examples already considered although quite common in the area. In this case it is the first and second ascending levels that are concerned and the exact way in which the specifications are distributed shows some intriguing regional variation.

In the extreme northeast among the Sanuma, the following distribution is found:

• haya: F, FB, FF, MF.

• shoaya: MB, FZH, WF.

In the extreme southeast among the Shamatari the terms are distributed thus:

• hayä: F, FB.

• shoaya: MB, FZH, WF, FF, MF.

In keeping with this Lizot (1971: p. 30) reports from a central region that there is some confusion about whether the males of the second ascending genealogical level should be referred to as hayä or shoaya.

The situation with regard to the terms for the women of these two genealogical levels has not been so clearly reported by the ethnographers but there is evidence of a similar shift. Thus Chagnon reports for the Shamatari the following (1974: pp. 22-3):

• nayä: M, MZ.

• yayä: FZ, MBW, WM, FM, MM.

While Liozt from the central region records the following (1971: pp. 28-30)

• naya: M, MZ, MM, FM.

• amiwa: eZ, FBD, MZD, MM, FM.

• yayä: FZ, MBW, WM.

While this variation is of great potential interest in the study of Yanomam society, its immediate value is in demonstrating how considerable change in the distribution of genealogical specifications among terms can occur without the essential two-line and prescriptive nature of a terminology being disrupted.

Before leaving the Yanomam, there is a point from Chagnon’s contribution to this volume that is worth touching on briefly. He refers to “types of prescriptive marriages” that occur among the Yanomamo. His use of the word “prescriptive” differs from mine. In my terms the Yanomamo have a single prescribed category, and what Chagnon is referring to are the various preferences, for social, economic or political reasons, that are made within that category. Furthermore, while in my terms it makes sense to talk about the strength of a preference, it makes none to refer to the strength of a prescription.

5) The Xingu peoples are represented in this volume by two different, although both Carib-speaking groups, the Kalapalo and the Kuikuru. There has been some discussion about the Kalapalo and Kuikuru terminologies, neither of which at first sight seems to distinguish four categories at ego’s genealogical level. Dole (1969) claims that the cross-cousin terms among the Kuikuru disappeared as a result of demographic and other changes. Basso (1970) queries this and demonstrates that for the Kalapalo the distinction between siblings and parallel cousins on the one hand and cross-cousins on the other is a matter of context and whether the contrast between marriageable and non-marriageable is relevant.

From the point of view of this article’s argument the disagreement between Baso and Dole is unimportant. In her contribution to this volume Basso discusses Kalapalo affinity and its terminology in operation. Not surprisingly this makes the system seem far more complex, but it does not negate the two-line prescriptive nature of the terminology⁶. Dole’s argument, for its part, does not deny but assumes the earlier presence in the terminology of the distinction between siblings and parallel cousins, and cross-cousins. An assumption supported by a Kuikuru terminology collected by Oberg (1953) that exhibits most of the important equations and distinctions that characterise a two-line terminology.

There seems no reason to linger on these Xingu cases, and it seems as though the situation can be summed up thus: the two-line nature of these terminologies is not in dispute at the first ascending and first descending genealogical levels, but there is some obscurity at ego’s own level where the kin/affine distinction either has terminologically died out or is only used in certain contexts. The weight of evidence, however, undoubtedly indicates that the Kalapalo and the Kuikuru have now or had in the recent past a two-line prescriptive terminology⁷.

6) The next area we may take is that of the Northwest Amazon, and three contributors refer to peoples of this region: Jackson to the Bara, Siskind to the Cubeo, and Sorensen provides some general but useful comments.

The peoples of the area are marked by a greater elaboration of social organization than the cases already examined. Even so it can readily be shown that the terminologies of these people are of the two-line prescriptive type. The particular features of the social organization in the region are a rule of patrilineal descent, the presence of patrilineal descent groups, and in most cases, although not in that of the Cubeo, there is a rule of language exogamy. For present purposes one of the most interesting things is how the patrilineality is reflected in the relationship terminology, especially at ego’s level.

If we take the Bará first, we find that at the first ascending genealogical level the terminology exhibits all the conventional equations and distinctions to be expected in a two-line terminology. Minor variations are a specific term for father which is distinct from that for father’s brother and mother’s sister’s husband, and one for mother which is distinct from that for mother’s sister and father’s brother’s wife. At the first descending genealogical level there exist the normal divisions of son, daughter, sister`s son, and sister`s daughter.

The important and interesting variations in the terminology occur at ego’s own level. There are terms for brother and sister that are applied to the male and female patrilateral parallel cousins respectively but not to the matrilateral parallel cousins unless these last also happen to be patrilateral, i.e., bilateral, parallel cousins. The matrilateral parallel cousin terms are formed from the term for mother and that for son and daughter. This usage can be related to the rule of patrilineal descent and to the fact that if ego’s mother’s sister is not married to a man of ego’s patrilineal group then ego’s mother’s sister’s child belongs to a patrilineal group that is neither ego’s nor his mother’s.

A similar situation holds good in the case of cross-cousins. The bilateral and patrilateral cross-cousins are known by one of two terms depending on sex, and the matrilateral cross-cousins by one of two terms depending on sex. Although both categories of cross-cousin are marriageable the preference is for the bilateral/patrilateral cross-cousin because such individuals give rise to a direct exchange, either immediate or delayed,⁸

To clarify the matter, here are the Bará sibling and cousin terms (Jackson 1972: 174):

• bái: B, FBS, FBS/MZS.

• bayó: Z, FBD, FBD/MZD.

• pahkó/mahkû: MZS (literally pahkó, M; and mahkû, S).

• pahkó-mahkó: MZD (literally pahkó, and mahkó, D)

• mehkó-mahkú: FZS, FZS/MBS (literally mehkó, FZ; and mahkû, S).

• mehkó-mahkó: FZD, FZD/MBD (literally mehkó, FZ; and mahkó, D).

• mehkú-mahkú: MBS (literally mehkú, MB; and mahkú, S).

• mehkú-mahkó: MDB (literally mehkú, MB; and mahkó, D)

The Cubeo provide some interesting contrasts to the Bará for while their relationship terminologies are similar in so far as the distribution of genealogical specifications among terms is concerned, there are a number of organizational differences. Firstly, as mentioned above, the Cubeo are unusual in the region because of their preference for language endogamy. Second, unlike the Bará, the Cubeo regard as marriageable a matrilateral parallel cousin who is not also a patrilateral parallel cousins (Goldman, 1963: 126). Third, within their bilateral prescription the Cubeo express a matrilateral rather than a patrilateral preference. Goldman writes (ibis: 137): “there is a strong preference for taking a wife from the same sib from which one’s mother has come’, although, curiously enough, they see such unions as forms of direct exchange. Sorensen in his contribution similarly notes a preference for the matrilateral cross-cousin and gives as the reason (which cannot be applicable in the case of the endogamous Cubeo) the fact that a man is more likely to speak the language of his mother’s natal group than that of his father’s sister’s husband’s.

These differences in social practice have been mentioned in order, once again, to stress that the same underlying structure can give rise to great variation at the level of behavior.

7) The Cashinahua’s terminology is clearly a two-line prescriptive one but it contains one feature which, although quite common in the area, has not been present in any of the terminologies so far considered. The feature is that of alternation so that the terms that occur at ego’s genealogical level recur at the second ascending and descending levels. For example:

• betsa: FF, B, FBS, SS.

• chai: MF, MBS/FZS, DS.

Alternation in some form or another has been reported in recent years from the Panare (Dumont, 1971: 87), the Cuiva (Arcand, 1972), and the Yanomamo (Chagnon, 1973: 221).

The Cashinahua also possess two forms of social classification that deserve some consideration. First, they have named exogamous patrimoieties for both men and women. Second they have named sections. Dealing only with the men, the system works thus: the men’s exogamous moieties are called inubakebu and duabakebu. All men of inubakebu belong to either awabakebu or kababakebu, and those of duabakebu to either yawabakebu or dunubakebu. While recruitment to the exogamous moieties is through patrilinity, membership of the section within the moiety results from alternation. Thus a man will belong to the same named section as his father’s father and son’s son, while his father and son will belong to the other section of the moiety. Marriage rules link sections of the different moieties. Thus awabake men marry yawabake women and vice-versa. Full details are given in Kensinger’s contribution so there is no need to repeat them other than to say that this is the only reported case from South America of a so-called Kariera system.

The Cashinahua case has some importance for the present argument. It was suggested that the universal feature of the Tropical Forest culture was the presence of a specific type of relationship terminology. It was further claimed that those societies which lacked such a terminology but were characterised by a dual organization and operated a direct exchange system by means of grosser social units would not be classed as Tropical Forest. The Cashinahua are a useful example of the opposite case. Although they possess two forms of dual organization, sections and moieties, their relationship terminology fits consistently with both these forms. Thus, an inubakebu man who belongs to the awabakebu section will call all men of that section betsa. Indeed the remarkable thing about the Cashinahua system is the degree of redundancy built into it.

8) The Parintintin case is also instructive. The terminology recorded by Kracke is unquestionably of the two-line prescriptive type. Furthermore the terminology fits with the exogamous moieties. However, the situation is confused in the way described by Kracke, and as his title indicates the moieties seem to have lost their function. Although it can only be speculative there is perhaps a lesson to be derived from this example. The official ideology is one of moiety exogamy, but in practice marriages are frequently not consistent with the ideology. Rather, it appears, most marriages occur between “patrilines” and involve direct exchange, and one would guess that at this level the terminology is fully operational. In other words, while the grosser social units are falling into decay for some reason (demographic decline, introduction into the system of a third element), the more flexible relationship terminology continues to operate perfectly satisfactorily. I would like to see this as evidence of my claim that the relationship terminologies, or better the principles at articulate them, are fundamental to the Tropical Forest societies, while the organizational forms are accessories that can be dispensed with.

This concludes the brief survey of those peoples who form the subjects of other chapters in this collection and for whom we have found clear evidence of the presence of a two-line prescriptive terminology. It might further be noted that by other criteria, cultural or geographical, all these peoples would be unquestionably classed as belonging to the Tropical Forest culture.

IV

The four remaining cases in this volume are important since they help us to examine the other side of the argument. While it is claimed that the Tropical Forest peoples are characterised by the possession of a relationship terminology structured in a specific way, it is also useful to consider what happens as one moves away from the region. The peoples to be considered now all inhabit regions peripheral to the tropical forest, and in every case the two-line terminology, so readily identifiable in the examples dealt with, is missing. Interestingly enough, however, while in some of these cases the emphasis on duality is even more marked than in the cases considered, nowhere is it in the same way intrinsic to the relationship terminology.

9) The Canelos Quichua, located between the lowlands and the highlands, share in the culture from both areas. Given these peoples linguistic affiliation it is not perhaps surprising to find that they do not have a two-line prescriptive terminology although it is interesting to note the degree to which the terminology does exhibit the characteristics of such a terminology. Whitten plays with the idea of identifying the terminology as one of restricted exchange, but rightly refrains from doing so. The reason he offers, however, is not in itself adequate as we have noted examples in which siblings and parallel cousins are terminologically equated with cross-cousins, and in which certain members of the prescribed category are prohibited. A more basic reason for not accepting the Canelos’ terminology as a prescriptive one is the equation of the parents with the parents-in-law.

• yaya: F, FB, WF, HF.

• mama: M, MZ, WM, HM.

Furthermore there are terms for mother’s brother (jachi) and father’s sister (miquia), but they lack essential affinal denotations.⁹

10) The last three cases, the Apinayè, Bororo, and Canella, are all Gê-speaking peoples, a group that in many social and cultural aspects are clearly differentiated from the Tropical Forest peoples. None of the examples of the Gê in this volume has a two-line prescriptive terminology although they are societies marked by various forms of dual organization. Thus Christopher Crocker writes of the Bororo that “the only prescriptive rule is moiety exogamy” (1969: 238), but this prescription is not reflected in the relationship terminology. Unlike the case with the Parintintin the Bororo have nothing to fall back on and the disappearance of their moiety system would entail the loss of their prescriptive system.

The Apinayé and Canella cases are equally clear cut. Although both societies contain forms of dual organization these are not underpinned, even where they are concerned with marriage, by a two-line prescriptive terminology. The terminologies of these two societies do contain many equations and distinctions typical of two-line terminologies but only sporadically which is not enough since it is the systematic nature of such equations and distinctions that is essential.

The discovery outside the tropical forest of a two-line prescriptive relationship terminology would not invalidate the argument in this paper which is that the people belonging to the Tropical Forest culture share at least one common feature. Under the diversity of social forms there is one common structure. It would be an interesting exercise to examine whether this particular structure disappears as one reaches the margins of the tropical forest. If this is the case the structure in question is not simply a general feature of the Tropical Forest culture but also something that distinguished them from their non-Tropical Forest neighbours. For the time being, however, there is enough to do check whether, as is claimed here, the two-line prescriptive terminology is the common factor that gives unity beneath the Tropical Forest’s cultural variety.

Referências bibliográficas

• ARCAND, Bernard. 1972. “A contribution to Cuiva ethnohraphy”. Tese de Doutorado, University of Cambridge, Cambridge.
• BASSO, Ellen. 1970. “Xingu Carib kinship terminology and marriage”. Southwestern Journal of Anthropology, 26:402-416.
• BENNET, Wendell & BIRD, Junius. 1949. Andean Culture History Garden City, NY: American Museum of Natural History.
• CHAGNON, Napoleon. 1974. “Studying the Yanomamö”. New York: Holt, Rinehart & Winston.
• COOPER, John M. 1942. “Areal and temporal aspects of aboriginal South American culture”. Primitive Man, 14:1-38.
• CROCKER, Jon Christopher. 1969. “Men’s house associates among the Eastern Bororo”. Southwestern Journal of Anthropology , 25:236-260.
• DOLE, Gertrude. 1969. “Generation kinship nomenclature as an adaptation to endogamy”. Southwestern Journal of Anthropology , 25:105-123.
• DUMONT, Jean-Pierre. 1971. “Compte rendu de mission chez les indiens Panará”. L”Homme, 11:83-88.
• GALVÃO, Eduardo. 1953. “Cultura e sistema de parentesco das tribos do Alto Rio Xingu”. Boletim do Museu Nacional, n. 14, Rio de Janeiro.
• GOLDMAN, Irving. 1963. The Cubeo: Indians of the Northwest Amazon Urbana: University of Illiniois Press.
• HARNER, Michael. 1973. The Jívaro: People of the Sacred Waterfalls New York: Hale.
• JACKSON, Jean. 1972. Marriage and Linguistic Identity among the Bará Indians of the Vaupés, Colombia Tese de Doutorado, Universidade de Stanford.
• KAPLAN, Joanna Overing. 1975. The Piaroa: A People of the Orinoco Basin Oxford: The Clarendon Press.
• LIZOT, Jacques. 1971. “Remarques sur le vocabulaire de parenté Yanomami”. L’Homme, 11:25-38.
• MURDOCK, George Peter. 1951. “South American cuture areas”. Southwestern Journal of Anthropology , 7:415-436.
• OBERG, Kalevo. 1953. Indian Tribes of Northern Mato Grosso Washington: Institute of Social Anthropology.
• RIVIÈRE, Peter. 1969a. Marriage Among the Trio Oxford: The Clarendon Press .
• RIVIÈRE, Peter.1969b. “Myth and material culture: Some symbolic interrelations”. In: Victor Turner (org.), Forms of Symbolic Action Seattle: University of Washington Press.
• RIVIÈRE, Peter.1974. “Some problems in the comparative study of Carib societies”. Atas do XL Congresso Internacional de Americanistas 1972, v. 2:639-643.
• STEWARD, Julian (org.). 1946-1950. Handbook of South American Indians 7 vols. Bureau of American Ethnology.
• STEWARD, Julian & FARON, Louis. 1959. Native Peoples of South America New York: McGraw-Hill.
• STOUT, Dietrich. 1938. “Culture types and culture areas in South America”. Paper of the Michigan Academy of Science, Arts, and Letters, v. 23:73-86.
• WISSLER, Clark. 1917. The American Indian: An Introduction to the Anthropology of the New World New York: Oxford University Press.

Notes

• 1
  Since this paper was first presented in 1973 I have changed my mind on certain points. This is the result of new ethnographic data and of discussion with other people, in particular those who have contributed to this volume. I am grateful to many of them who replied to my letters with information, advice and criticism. The one major change involves my view concerning the position of the Gê-speaking peoples. I had originally argued that they should be included with the Tropical Forest culture; I now argue that they should not.
• 2
  This remark was made by Paul Henley. However, while we both remember his making it neither of us can find it! I am further grateful to Paul for still having a copy of this paper. I found I had not, and I asked him on the hope that he still had his papers and notes assembled while preparing his ‘South Indian models in the Amazonia Lowlands’, Manchester Papers in Social Anthropology 1 (1996).
• 3
  There have been numerous attempts, using an assortment of criteria, at the classification of South American cultures (e.g., Wisslet, 1917; Stout, 1938; Cooper, 1942; Bennett & Bird, 1949; Murdock, 1951; and Steward & Faron, 1959). None of them has survived and most of them were soon forgotten. It might also be noted that there is no apparent correlation between culture and language (for a demonstration of this in one small area, see Rivière, 1969b), although I have made a tentative effort to isolate a Carib culture (Rivière, 1974).
• 4
  Unless otherwise stated a male ego’s point of view is adopted throughout this paper.
• 5
  The term Yanomam is used here to refer to the whole group of linguistically related Indians who live in Brazil and Venezuela, near the headwaters of the Orinoco.
• 6
  For an example of the contrast between the relative simplicity of a terminology’s formal structure and its complicated social usage, see Rivère, 1969a: Chapters IV & V
• 7
  On this point evidence from the Mehinacu, another Xingu people who were originally to be included in this volume, is of interest. The Mehinacu terminology as reported by Gavlão (1953), although incomplete, has most of the features that Dole records for the Kuikuru. However, Gregor who has worked more recently among the Mehinacu sides with Basso and has written to say that while Galvão’s terminology is correct, in practice the situation is more complex. “There are a number of terms in everyday use that cut right across the system… There is, for example, a referential term used for cross-cousins that is very similar in use to the cross-cousin term that Basso describes for the Kalapalo.”.
• 8
  Jackson has written to say that although the Barà do express a patrilateral preference they also have an alternative model of society in which the matrilateral groups are preferred.
• 9
  The Canelos Quichua system as represented by Whitten in his diagram (p. 18 of typescript) is one of patrilateral indirect exchange, with each ego marrying a FZD. As Whitten notes, such systems readily collapse in bilateral, or direct exchange, systems.

Publication Dates

• Publication in this collection
  10 Oct 2022
• Date of issue
  2022

History

• Received
  02 Aug 2022
• Accepted
  10 Aug 2022